Morph MORPH
SIGN IN SIGN UP
microsoft / semantic-kernel UNCLAIMED

Integrate cutting-edge LLM technology quickly and easily into your apps

0 0 11 C#
Normal View History Raw
Initial commit
2023-03-16 19:54:34 -07:00
# Copyright (c) Microsoft. All rights reserved.
Python: Uses Python standard logging (#4115) ### Motivation and Context This replaces passing Logger instances as arguments around with the preferred Python way of having each module define a module-wide `logger` instance at the top of the module. Fixes #3711 ### Description Removes any `log` argument and argument documentation. Replaces usage of those `log` arguments with the module-wide `logger` instance. ### Contribution Checklist - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Eduard van Valkenburg <eavanvalkenburg@users.noreply.github.com>
2023-12-16 11:41:41 -07:00
import logging
Python: Feature new memory stores and collections (#7614) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds new vector store and vector store record collections as well as implementations for: - Azure AI Search - Redis - Qdrant - Volatile (in-memory) It also adds samples, tests, and unit-tests for these. Next it adds the vector store record fields, definition and decorator, with tests and samples. All marked experimental. Existing Redis, Azure AI Search, Qdrant and Volatile will be marked as deprecated in the future, once the new collections are feature complete. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-06 22:30:26 +02:00
from collections.abc import AsyncGenerator, AsyncIterable, Callable
Python: Introducing SK as a MCP Server (#11362) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds the ability to stand up your Kernel as a MCP Server, it returns the Server object, which you can then run as either a Stdio or Sse server. Includes a sample to demonstrate. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Nico Möller <nkm.moeller@gmail.com>
2025-04-11 07:22:07 +02:00
from contextlib import AbstractAsyncContextManager
Python: Add ability to clone a kernel (#11680) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Currently, we don't have a dedicated method to create a clone of a kernel instance. Using `model_copy(deep=True)` will result in an error because some service clients aren't serializable. Being able to clone a kernel will be critical in many scenarios where developers don't want to mutate the original one that is attached to an agent or some other objects. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> 1. Create a clone method in the `Kernel` class that will return a new instance of the kernel with the same configuration, including the list of filters, plugins and the AI selector. It will also include the same list of services references. 2. Add unit tests and integration tests to ensure the new method works as expected. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-23 17:26:51 -07:00
from copy import copy, deepcopy
Python: mypy coverage enhancement (#6250) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> extend mypy coverage for prompt_templates, template_engine and some other smaller folders. still todo: connectors, memory and planner folders. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-31 18:37:19 +02:00
from typing import TYPE_CHECKING, Any, Literal, TypeVar
Python: Add kernel plugin collection and remove old plugin collection classes (#4764) ### Motivation and Context This PR solves two works items: #4560 and #4667. The Python SDK had different ways of handling plugins and collections. The work in the PR helps align the SDK with dotnet nomenclature and structure. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description A `KernelPluginCollection` is introduced with an underlying `DefaultKernelPlugin` and a `KernelPlugin` base class. Other code is refactored to use this new `KernelPluginCollection` model which should simplify how plugins are handled. Unit tests are added to test the new functionality. Existing unit tests and integration tests, along with kernel examples and notebooks were updated based on these changes. Unlike before, we're requiring plugins to be defined with a name, which removes the GLOBAL_PLUGIN constant name of "__GLOBAL_PLUGIN__". In some senses, while creating a plugin or a function, a name may not be provided and so a random name in the form of a plugin (p_<random_ascii_chars>) or function (f_<random_ascii_chars>) is used. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-01 13:57:00 -05:00
Python: Add embedding and text completion samples (#10797) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Closing https://github.com/microsoft/semantic-kernel/issues/9963 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> 1. Add text embedding and text completion samples. 2. Move the embedding generator base to a `semantic_kernel.connectors.ai`. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-03-05 11:39:07 -08:00
from semantic_kernel.connectors.ai.embedding_generator_base import EmbeddingGeneratorBase
Python: Unsafe input handling (#6003) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Implements dealing with unsafe content, by doing HTML parsing on variables and function results. Closes: #5889 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds parameter `allow_dangerously_set_content` to: - InputVariable - PromptTemplateConfig - PromptTemplateBase The behavior is that if the flag is set to True on the template itself (KernelPromptTemplate, Jinja2PromptTemplate or HandlebarsPromptTemplate) the behavior is the same, no encoding is done on inputs. Otherwise: - variables are encoded by default, this can be switched off using the InputVariables class for that variable. - function output is encoded by default, this can be switched off using the flag in the PromptTemplateConfig, this is not yet possible to do on a per function basis. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-16 15:02:54 +02:00
from semantic_kernel.const import METADATA_EXCEPTION_KEY
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
from semantic_kernel.contents.chat_history import ChatHistory
from semantic_kernel.contents.function_call_content import FunctionCallContent
from semantic_kernel.contents.function_result_content import FunctionResultContent
Python: Yield FunctionResultContent in streaming chat completion path. Update tests. (#9974) ### Motivation and Context Currently, if using SK's Python streaming chat completion path, we only yield the following content types: `StreamingChatMessageContent` and `FunctionCallContent`. We are not yielding `FunctionResultContent` which is valuable for some use cases. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR updates the code to yield `FunctionResultContent` if it exists in the streaming chat completion path. When we merge the function call result content together into a `StreamingChatMessageContent` type, we check if that message has items (which are of type `FunctionResultContent`) and if so, we yield them. The filter path still works because once they're yielded, we break out of the function calling loop. We need to include the `ai_model_id` if it exists for the current PromptExecutionSettings because if performing a `reduce` operation to add two streaming chat message chunks together, the `StreamingChatMessageContent` that has the function results will break if the `ai_model_id` is not set (the error is thrown in the `__add__` override for the `StreamingChatMessageContent`. Some unit tests that cover function calling were also updated -- during the test, the test JSON function args were breaking in the `json.loads` call because they contained single quotes and not double. We're now sanitizing the args, just in case, so we don't break there. This PR fixes: - #9408 - #9968 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-12-18 10:09:25 +09:00
from semantic_kernel.contents.streaming_chat_message_content import StreamingChatMessageContent
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
from semantic_kernel.contents.streaming_content_mixin import StreamingContentMixin
Python: rebuilt exceptions structure; pythonic version (#5231) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> The existing Exceptions structure was very much inspired by dotnet, this is now replaced with a pythonic implementations. This means all Exceptions derive from KernelException and then specialise for different purposes. Closes #2194 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added folder for all exception types, all can be imported through `from semantic_kernel.exceptions import ...` no need for a user to know which file the relevant one is in, but keeps things tidy for developers. Removed old KernelException, added back subtypes for the errorcodes. Went through everything to make sure the `raise ... from exc` pattern is used as much as possible as that returns a better stacktrace. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-02-28 19:53:53 +01:00
from semantic_kernel.exceptions import (
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
FunctionCallInvalidArgumentsException,
FunctionExecutionException,
Python: rebuilt exceptions structure; pythonic version (#5231) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> The existing Exceptions structure was very much inspired by dotnet, this is now replaced with a pythonic implementations. This means all Exceptions derive from KernelException and then specialise for different purposes. Closes #2194 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added folder for all exception types, all can be imported through `from semantic_kernel.exceptions import ...` no need for a user to know which file the relevant one is in, but keeps things tidy for developers. Removed old KernelException, added back subtypes for the errorcodes. Went through everything to make sure the `raise ... from exc` pattern is used as much as possible as that returns a better stacktrace. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-02-28 19:53:53 +01:00
KernelFunctionNotFoundError,
KernelInvokeException,
Python: implement filters (#5681) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This pull request includes significant changes across multiple files, mainly related to the addition of hooks and the modification of function invocations in the `semantic_kernel` module. The changes also include the addition of a new sample and a YAML file, and modifications to the `__init__.py` files. Removals: * [`python/semantic_kernel/events`](diffhunk://#diff-ebda9504832b19ab83239a92c9a6d5f8c744deff9fef86071c13956ec92bb010L1-L11): Removed the previously used events. New Exceptions: * [`python/semantic_kernel/exceptions/kernel_exceptions.py`](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48): Added new exception classes `OperationCancelledException` and `HookInvalidSignatureError`. [[1]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48) [[2]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR57-R58) Fixes: #3038 Fixes: #6276 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-05-17 23:05:56 +02:00
OperationCancelledException,
Python: rebuilt exceptions structure; pythonic version (#5231) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> The existing Exceptions structure was very much inspired by dotnet, this is now replaced with a pythonic implementations. This means all Exceptions derive from KernelException and then specialise for different purposes. Closes #2194 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added folder for all exception types, all can be imported through `from semantic_kernel.exceptions import ...` no need for a user to know which file the relevant one is in, but keeps things tidy for developers. Removed old KernelException, added back subtypes for the errorcodes. Went through everything to make sure the `raise ... from exc` pattern is used as much as possible as that returns a better stacktrace. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-02-28 19:53:53 +01:00
TemplateSyntaxError,
)
Python: Feature new memory stores and collections (#7614) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds new vector store and vector store record collections as well as implementations for: - Azure AI Search - Redis - Qdrant - Volatile (in-memory) It also adds samples, tests, and unit-tests for these. Next it adds the vector store record fields, definition and decorator, with tests and samples. All marked experimental. Existing Redis, Azure AI Search, Qdrant and Volatile will be marked as deprecated in the future, once the new collections are feature complete. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-06 22:30:26 +02:00
from semantic_kernel.exceptions.kernel_exceptions import KernelServiceNotFoundError
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
from semantic_kernel.filters.auto_function_invocation.auto_function_invocation_context import (
AutoFunctionInvocationContext,
)
from semantic_kernel.filters.filter_types import FilterTypes
from semantic_kernel.filters.kernel_filters_extension import (
KernelFilterExtension,
_rebuild_auto_function_invocation_context,
)
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
from semantic_kernel.functions.function_result import FunctionResult
from semantic_kernel.functions.kernel_arguments import KernelArguments
Python: split kernel into kernel extensions for relevant pieces (#6361) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> In order to keep better track of related pieces within the Kernel, including easier unit testing, this PR splits the kernel into pieces that are self-contained, but abstract. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> New: - KernelServicesExtension - KernelFunctionExtension - KernelReliabilityExtension Changed: - Kernel, now imports the new ones. - KernelFiltersExtension, moved to filters folder, in line with the rest. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-22 17:03:34 +02:00
from semantic_kernel.functions.kernel_function_extension import KernelFunctionExtension
Python: implement filters (#5681) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This pull request includes significant changes across multiple files, mainly related to the addition of hooks and the modification of function invocations in the `semantic_kernel` module. The changes also include the addition of a new sample and a YAML file, and modifications to the `__init__.py` files. Removals: * [`python/semantic_kernel/events`](diffhunk://#diff-ebda9504832b19ab83239a92c9a6d5f8c744deff9fef86071c13956ec92bb010L1-L11): Removed the previously used events. New Exceptions: * [`python/semantic_kernel/exceptions/kernel_exceptions.py`](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48): Added new exception classes `OperationCancelledException` and `HookInvalidSignatureError`. [[1]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48) [[2]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR57-R58) Fixes: #3038 Fixes: #6276 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-05-17 23:05:56 +02:00
from semantic_kernel.functions.kernel_function_from_prompt import KernelFunctionFromPrompt
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
from semantic_kernel.functions.kernel_plugin import KernelPlugin
Python: Introducing SK as a MCP Server (#11362) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds the ability to stand up your Kernel as a MCP Server, it returns the Server object, which you can then run as either a Stdio or Sse server. Includes a sample to demonstrate. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Nico Möller <nkm.moeller@gmail.com>
2025-04-11 07:22:07 +02:00
from semantic_kernel.kernel_types import AI_SERVICE_CLIENT_TYPE, OneOrMany, OptionalOneOrMany
Python: split kernel into kernel extensions for relevant pieces (#6361) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> In order to keep better track of related pieces within the Kernel, including easier unit testing, this PR splits the kernel into pieces that are self-contained, but abstract. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> New: - KernelServicesExtension - KernelFunctionExtension - KernelReliabilityExtension Changed: - Kernel, now imports the new ones. - KernelFiltersExtension, moved to filters folder, in line with the rest. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-22 17:03:34 +02:00
from semantic_kernel.prompt_template.const import KERNEL_TEMPLATE_FORMAT_NAME
Python: add support for prompts and sample (#11508) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Allow a list of prompts to be added and exposed with the MCP server setup. This uses the built-in rendering types (all of them) to be used and exposed as prompts. Closes: #11143 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-15 10:25:35 +02:00
from semantic_kernel.prompt_template.prompt_template_base import PromptTemplateBase
Python: .Net: Updated encoding logic in prompt templates (#12983) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Resolves: https://github.com/microsoft/semantic-kernel/issues/11821 Today, the encoding of template arguments is performed only if argument type is `string`. In case of custom type, anonymous type or collection - the encoding is not performed. This PR contains changes to throw an exception in case if encoding is enabled but complex type is used. In case of complex type, the encoding should be performed manually according to business logic and automatic encoding should be explicitly disabled. This enforces stricter, but more secure template rendering rules. **Note**: this is a breaking change for customers who use Handlebars or Liquid template with complex type arguments. Code changes are required when initializing template arguments: ```diff var arguments = new KernelArguments() { { "customer", new { - firstName = userInput.FirstName, - lastName = userInput.LastName, + firstName = HttpUtility.HtmlEncode(userInput.FirstName), + lastName = HttpUtility.HtmlEncode(userInput.LastName), } } }; var templateFactory = new LiquidPromptTemplateFactory(); var promptTemplateConfig = new PromptTemplateConfig() { TemplateFormat = "liquid" + InputVariables = new() + { + // We set AllowDangerouslySetContent to 'true' because each property of this argument is encoded manually. + new() { Name = "customer", AllowDangerouslySetContent = true }, + } }; var promptTemplate = templateFactory.Create(promptTemplateConfig); var renderedPrompt = await promptTemplate.RenderAsync(kernel, arguments); ``` ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-08-26 22:09:00 -07:00
from semantic_kernel.prompt_template.prompt_template_config import PromptTemplateConfig
Python: split kernel into kernel extensions for relevant pieces (#6361) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> In order to keep better track of related pieces within the Kernel, including easier unit testing, this PR splits the kernel into pieces that are self-contained, but abstract. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> New: - KernelServicesExtension - KernelFunctionExtension - KernelReliabilityExtension Changed: - Kernel, now imports the new ones. - KernelFiltersExtension, moved to filters folder, in line with the rest. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-22 17:03:34 +02:00
from semantic_kernel.reliability.kernel_reliability_extension import KernelReliabilityExtension
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
from semantic_kernel.services.ai_service_selector import AIServiceSelector
Python: mypy coverage enhancement (#6250) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> extend mypy coverage for prompt_templates, template_engine and some other smaller folders. still todo: connectors, memory and planner folders. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-31 18:37:19 +02:00
from semantic_kernel.services.kernel_services_extension import KernelServicesExtension
Python: Introducing SK as a MCP Server (#11362) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds the ability to stand up your Kernel as a MCP Server, it returns the Server object, which you can then run as either a Stdio or Sse server. Includes a sample to demonstrate. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Nico Möller <nkm.moeller@gmail.com>
2025-04-11 07:22:07 +02:00
from semantic_kernel.utils.feature_stage_decorator import experimental
Python: slight improvements in kernel and function result content, sample of local function calling with Nexus Raven (#6982) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This introduces a sample that uses function calling with a local model, Nexus Raven V2, in this case using the API, but the same model can be loaded using the Hugging Face Connection, the class here is highly adapted and specific to Nexus including parsing the returned function call value, into FunctionCallContent and then running that. It then returns one or more ChatMessageContents with the function call content and function result content which can then be shared with a OpenAI call. Error handling improvements: * [`python/semantic_kernel/contents/chat_history.py`](diffhunk://#diff-b57a30f97fec1a91bbf6f7d15584473a8f0cf409797bdcd90d8c80bedc670892L266-R267): Improved error logging when parsing prompts as XML fails. * [`python/semantic_kernel/contents/function_result_content.py`](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10): Removed field validator for `result` and allow result to be any value, instead of casting everything to string. [[1]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10) [[2]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L50-R51) [[3]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L63-L69) [[4]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87R92-R115) Minor changes: * [`python/semantic_kernel/core_plugins/math_plugin.py`](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38): Removed some docstrings for clarity. [[1]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38) [[2]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L55-R47) * [`python/semantic_kernel/functions/kernel_function_from_prompt.py`](diffhunk://#diff-eabd53da2ec241cce6f92de85283065e7b478d62cefb52817c5b29ff07e28189L173-R174): Changed the order of condition checks for better readability. * [`python/semantic_kernel/kernel.py`](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10): Added functionality to update kernel arguments in `invoke_stream` method and made `function_name` and `plugin_name` optional in `invoke_prompt` method. [[1]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10) [[2]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR29) [[3]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR119-R120) [[4]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL195-R202) [[5]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL209-R216) Closes #6804 Closes #6944 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-28 22:17:05 +02:00
from semantic_kernel.utils.naming import generate_random_ascii_name
Initial commit
2023-03-16 19:54:34 -07:00
Python: mypy clears kernel.py (#5689) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Fixed all mypy errors for kernel.py In preparation of extending that to all files in semantic_kernel Also got the pre-commit hook working on a commit, with mypy checks fixes #5630 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Moved pre-commit-config to base folder, can now be extended to have specific hooks for other languages fixed things in kernel.py to satisfy mypy used assert for some things, so that they can be skipped when running in -o mode. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-28 16:10:28 +01:00
if TYPE_CHECKING:
Python: Introducing SK as a MCP Server (#11362) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds the ability to stand up your Kernel as a MCP Server, it returns the Server object, which you can then run as either a Stdio or Sse server. Includes a sample to demonstrate. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Nico Möller <nkm.moeller@gmail.com>
2025-04-11 07:22:07 +02:00
from mcp.server.lowlevel.server import LifespanResultT, Server
Python: Fix KernelArgument bug in invoke_prompt functions (#8414) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> In `invoke_prompt(...)` and `invoke_prompt_async(...)`, the `arguments` parameter is tested to see if it's empty before proceeding. The type of the `arguments` parameter is `KernelArgument`, which derives from a `dict` type, with an additional field for execution settings. When `arguments` contains only the execution settings, it will be considered as an empty dictionary. This will result in the `arguments` parameter being overridden completely, invalidating all execution settings passed to the APIs. To reproduce (assuming we have a plugin/function that will return the weather): ``` result = await kernel.invoke_prompt( "What is the weather like in my location?", arguments=KernelArguments( settings={ service_id: PromptExecutionSettings( service_id=service_id, function_choice_behavior=FunctionChoiceBehavior.Auto(), ), } ), ) ``` In the example above, the settings will be overridden to None. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> 1. Fix the bug. 2. Fix mypy in the pre-commit hook in order to commit. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-29 08:40:53 -07:00
from semantic_kernel.connectors.ai.function_choice_behavior import FunctionChoiceBehavior
Python: Feature new memory stores and collections (#7614) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds new vector store and vector store record collections as well as implementations for: - Azure AI Search - Redis - Qdrant - Volatile (in-memory) It also adds samples, tests, and unit-tests for these. Next it adds the vector store record fields, definition and decorator, with tests and samples. All marked experimental. Existing Redis, Azure AI Search, Qdrant and Volatile will be marked as deprecated in the future, once the new collections are feature complete. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-06 22:30:26 +02:00
from semantic_kernel.connectors.ai.prompt_execution_settings import PromptExecutionSettings
Python: Updated plugins (#5827) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> The loading of plugins and functions into Kernel was not fully consistent and clear, this is now simplified. Adding a plugin can now be done in two ways: - multiple: `kernel.add_plugins[plugin1, plugin2]` - single plugin: `kernel.add_plugin(plugin)` - if the plugin here is a class, with methods that have the kernel_function decorator then that is parsed into a plugin Adding a function can now be done in three ways: - multiple: `kernel.add_functions([func1, func2]` or `kernel.add_function({'func1': func1})` - single: `kernel.add_function('plugin_name', func1)` - from a prompt: `kernel.add_function(function_name='func1', prompt='test prompt', ...)` In other words, all the different methods that were available have been simplified, these methods also return the created plugin (or updated plugin for the `add_functions` method. The `add_function` (singular) method has a parameter `return_plugin` to control whether you get the created or updated plugin, instead of the created function. **An important callout:** One big internal change that this introduces is that a function that gets added to a plugin (new or existing) is automatically copied, the metadata is deep-copied and the plugin_name in the metadata is updated, so this sequence is valid now: ```python @kernel_function(name='test') def f(...): etc func = KernelFunctionFromMethod(method=f) func2 = kernel.add_function('plugin', func) assert func != func2 ``` Also closes: #5855 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Removed KernelPluginCollection Updated KernelPlugin added from_... method to load from different sources Updated KernelFunctionFromPrompt added from_yaml and from_directory methods. Added and updated tests to 99% coverage of KernelPlugin and KernelFunctionFromPrompt ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-04-15 15:51:59 +02:00
from semantic_kernel.functions.kernel_function import KernelFunction
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
Python: mypy coverage enhancement (#6250) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> extend mypy coverage for prompt_templates, template_engine and some other smaller folders. still todo: connectors, memory and planner folders. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-31 18:37:19 +02:00
T = TypeVar("T")
Python: Feature new memory stores and collections (#7614) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds new vector store and vector store record collections as well as implementations for: - Azure AI Search - Redis - Qdrant - Volatile (in-memory) It also adds samples, tests, and unit-tests for these. Next it adds the vector store record fields, definition and decorator, with tests and samples. All marked experimental. Existing Redis, Azure AI Search, Qdrant and Volatile will be marked as deprecated in the future, once the new collections are feature complete. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-06 22:30:26 +02:00
TDataModel = TypeVar("TDataModel")
Python: Unsafe input handling (#6003) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Implements dealing with unsafe content, by doing HTML parsing on variables and function results. Closes: #5889 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds parameter `allow_dangerously_set_content` to: - InputVariable - PromptTemplateConfig - PromptTemplateBase The behavior is that if the flag is set to True on the template itself (KernelPromptTemplate, Jinja2PromptTemplate or HandlebarsPromptTemplate) the behavior is the same, no encoding is done on inputs. Otherwise: - variables are encoded by default, this can be switched off using the InputVariables class for that variable. - function output is encoded by default, this can be switched off using the flag in the PromptTemplateConfig, this is not yet possible to do on a per function basis. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-16 15:02:54 +02:00
Python: Uses Python standard logging (#4115) ### Motivation and Context This replaces passing Logger instances as arguments around with the preferred Python way of having each module define a module-wide `logger` instance at the top of the module. Fixes #3711 ### Description Removes any `log` argument and argument documentation. Replaces usage of those `log` arguments with the module-wide `logger` instance. ### Contribution Checklist - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Eduard van Valkenburg <eavanvalkenburg@users.noreply.github.com>
2023-12-16 11:41:41 -07:00
logger: logging.Logger = logging.getLogger(__name__)
Python: fixes and release (#855) * Fix lint errors * Bump version number * Add retry loop to integration tests * Enable all integration tests: some tests will fail if OpenAI/Azure are throttling.
2023-05-08 15:52:48 -07:00
Python: split kernel into kernel extensions for relevant pieces (#6361) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> In order to keep better track of related pieces within the Kernel, including easier unit testing, this PR splits the kernel into pieces that are self-contained, but abstract. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> New: - KernelServicesExtension - KernelFunctionExtension - KernelReliabilityExtension Changed: - Kernel, now imports the new ones. - KernelFiltersExtension, moved to filters folder, in line with the rest. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-22 17:03:34 +02:00
class Kernel(KernelFilterExtension, KernelFunctionExtension, KernelServicesExtension, KernelReliabilityExtension):
Python: removed D101 and fixes for it (#7853) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Added comments about each of the rules to pyproject.toml Removed D101 from ignore list and fixed that, now docstrings at the class level are required. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-05 18:33:49 +02:00
"""The Kernel of Semantic Kernel.
Python: new pre-commit actions and pre-commit as GHA (#6376) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> - Added static code checking to pre-commit (check-ast and nbqa-check-ast) - Added Bandit security checking to pre-commit - Added pre-commit step to python-lint workflow, if it works, can delete seperate mypy, ruff and black tests ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-28 17:28:04 +02:00
Python: removed D101 and fixes for it (#7853) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Added comments about each of the rules to pyproject.toml Removed D101 from ignore list and fixed that, now docstrings at the class level are required. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-05 18:33:49 +02:00
This is the main entry point for Semantic Kernel. It provides the ability to run
functions and manage filters, plugins, and AI services.
Python: Make the Kernel class Pydantic. (#4850) ### Motivation and Context Completes work item: #4846. The Kernel class was not following the Pydantic convention. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR updates the Kernel class to use the KernelBaseModel and Pydantic fields. All unit and integration tests passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com> Co-authored-by: Eduard van Valkenburg <eavanvalkenburg@users.noreply.github.com>
2024-02-02 07:46:45 -05:00
Attributes:
Python: removed D101 and fixes for it (#7853) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Added comments about each of the rules to pyproject.toml Removed D101 from ignore list and fixed that, now docstrings at the class level are required. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-05 18:33:49 +02:00
function_invocation_filters: Filters applied during function invocation, from KernelFilterExtension.
prompt_rendering_filters: Filters applied during prompt rendering, from KernelFilterExtension.
auto_function_invocation_filters: Filters applied during auto function invocation, from KernelFilterExtension.
plugins: A dict with the plugins registered with the Kernel, from KernelFunctionExtension.
services: A dict with the services registered with the Kernel, from KernelServicesExtension.
ai_service_selector: The AI service selector to be used by the kernel, from KernelServicesExtension.
Python: Make the Kernel class Pydantic. (#4850) ### Motivation and Context Completes work item: #4846. The Kernel class was not following the Pydantic convention. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR updates the Kernel class to use the KernelBaseModel and Pydantic fields. All unit and integration tests passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com> Co-authored-by: Eduard van Valkenburg <eavanvalkenburg@users.noreply.github.com>
2024-02-02 07:46:45 -05:00
"""
Initial commit
2023-03-16 19:54:34 -07:00
def __init__(
self,
Python: Updated plugins (#5827) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> The loading of plugins and functions into Kernel was not fully consistent and clear, this is now simplified. Adding a plugin can now be done in two ways: - multiple: `kernel.add_plugins[plugin1, plugin2]` - single plugin: `kernel.add_plugin(plugin)` - if the plugin here is a class, with methods that have the kernel_function decorator then that is parsed into a plugin Adding a function can now be done in three ways: - multiple: `kernel.add_functions([func1, func2]` or `kernel.add_function({'func1': func1})` - single: `kernel.add_function('plugin_name', func1)` - from a prompt: `kernel.add_function(function_name='func1', prompt='test prompt', ...)` In other words, all the different methods that were available have been simplified, these methods also return the created plugin (or updated plugin for the `add_functions` method. The `add_function` (singular) method has a parameter `return_plugin` to control whether you get the created or updated plugin, instead of the created function. **An important callout:** One big internal change that this introduces is that a function that gets added to a plugin (new or existing) is automatically copied, the metadata is deep-copied and the plugin_name in the metadata is updated, so this sequence is valid now: ```python @kernel_function(name='test') def f(...): etc func = KernelFunctionFromMethod(method=f) func2 = kernel.add_function('plugin', func) assert func != func2 ``` Also closes: #5855 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Removed KernelPluginCollection Updated KernelPlugin added from_... method to load from different sources Updated KernelFunctionFromPrompt added from_yaml and from_directory methods. Added and updated tests to 99% coverage of KernelPlugin and KernelFunctionFromPrompt ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-04-15 15:51:59 +02:00
plugins: KernelPlugin | dict[str, KernelPlugin] | list[KernelPlugin] | None = None,
Python: implement filters (#5681) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This pull request includes significant changes across multiple files, mainly related to the addition of hooks and the modification of function invocations in the `semantic_kernel` module. The changes also include the addition of a new sample and a YAML file, and modifications to the `__init__.py` files. Removals: * [`python/semantic_kernel/events`](diffhunk://#diff-ebda9504832b19ab83239a92c9a6d5f8c744deff9fef86071c13956ec92bb010L1-L11): Removed the previously used events. New Exceptions: * [`python/semantic_kernel/exceptions/kernel_exceptions.py`](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48): Added new exception classes `OperationCancelledException` and `HookInvalidSignatureError`. [[1]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48) [[2]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR57-R58) Fixes: #3038 Fixes: #6276 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-05-17 23:05:56 +02:00
services: (
AI_SERVICE_CLIENT_TYPE | list[AI_SERVICE_CLIENT_TYPE] | dict[str, AI_SERVICE_CLIENT_TYPE] | None
) = None,
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
ai_service_selector: AIServiceSelector | None = None,
Python: Make the Kernel class Pydantic. (#4850) ### Motivation and Context Completes work item: #4846. The Kernel class was not following the Pydantic convention. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR updates the Kernel class to use the KernelBaseModel and Pydantic fields. All unit and integration tests passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com> Co-authored-by: Eduard van Valkenburg <eavanvalkenburg@users.noreply.github.com>
2024-02-02 07:46:45 -05:00
**kwargs: Any,
Initial commit
2023-03-16 19:54:34 -07:00
) -> None:
Python: new pre-commit actions and pre-commit as GHA (#6376) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> - Added static code checking to pre-commit (check-ast and nbqa-check-ast) - Added Bandit security checking to pre-commit - Added pre-commit step to python-lint workflow, if it works, can delete seperate mypy, ruff and black tests ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-28 17:28:04 +02:00
"""Initialize a new instance of the Kernel class.
Python: sync C# for Pre/Post RunAsync Handlers (#3063) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> #3038 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [ ] The code builds clean without any errors or warnings - [ ] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [ ] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Roger Barreto <19890735+RogerBarreto@users.noreply.github.com> Co-authored-by: Dmytro Struk <13853051+dmytrostruk@users.noreply.github.com>
2023-11-14 06:45:04 +09:00
Python: Make the Kernel class Pydantic. (#4850) ### Motivation and Context Completes work item: #4846. The Kernel class was not following the Pydantic convention. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR updates the Kernel class to use the KernelBaseModel and Pydantic fields. All unit and integration tests passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com> Co-authored-by: Eduard van Valkenburg <eavanvalkenburg@users.noreply.github.com>
2024-02-02 07:46:45 -05:00
Args:
Python: removed D101 and fixes for it (#7853) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Added comments about each of the rules to pyproject.toml Removed D101 from ignore list and fixed that, now docstrings at the class level are required. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-05 18:33:49 +02:00
plugins: The plugins to be used by the kernel, will be rewritten to a dict with plugin name as key
services: The services to be used by the kernel, will be rewritten to a dict with service_id as key
ai_service_selector: The AI service selector to be used by the kernel,
Python: Deprecated retry_mechanism (#9965) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> In Python, we have the `retry_mechanism` property in the kernel but it's never used. Address: https://github.com/microsoft/semantic-kernel/issues/9015 ### Description 1. Deprecate the retry_mechanism field. 2. Add a sample showing how to perform retries with the kernel using filter. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-12-13 14:22:58 -08:00
default is based on order of execution settings.
**kwargs: Additional fields to be passed to the Kernel model, these are limited to filters.
Python: Add kernel plugin collection and remove old plugin collection classes (#4764) ### Motivation and Context This PR solves two works items: #4560 and #4667. The Python SDK had different ways of handling plugins and collections. The work in the PR helps align the SDK with dotnet nomenclature and structure. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description A `KernelPluginCollection` is introduced with an underlying `DefaultKernelPlugin` and a `KernelPlugin` base class. Other code is refactored to use this new `KernelPluginCollection` model which should simplify how plugins are handled. Unit tests are added to test the new functionality. Existing unit tests and integration tests, along with kernel examples and notebooks were updated based on these changes. Unlike before, we're requiring plugins to be defined with a name, which removes the GLOBAL_PLUGIN constant name of "__GLOBAL_PLUGIN__". In some senses, while creating a plugin or a function, a name may not be provided and so a random name in the form of a plugin (p_<random_ascii_chars>) or function (f_<random_ascii_chars>) is used. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-01 13:57:00 -05:00
"""
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
args = {
"services": services,
Python: Updated plugins (#5827) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> The loading of plugins and functions into Kernel was not fully consistent and clear, this is now simplified. Adding a plugin can now be done in two ways: - multiple: `kernel.add_plugins[plugin1, plugin2]` - single plugin: `kernel.add_plugin(plugin)` - if the plugin here is a class, with methods that have the kernel_function decorator then that is parsed into a plugin Adding a function can now be done in three ways: - multiple: `kernel.add_functions([func1, func2]` or `kernel.add_function({'func1': func1})` - single: `kernel.add_function('plugin_name', func1)` - from a prompt: `kernel.add_function(function_name='func1', prompt='test prompt', ...)` In other words, all the different methods that were available have been simplified, these methods also return the created plugin (or updated plugin for the `add_functions` method. The `add_function` (singular) method has a parameter `return_plugin` to control whether you get the created or updated plugin, instead of the created function. **An important callout:** One big internal change that this introduces is that a function that gets added to a plugin (new or existing) is automatically copied, the metadata is deep-copied and the plugin_name in the metadata is updated, so this sequence is valid now: ```python @kernel_function(name='test') def f(...): etc func = KernelFunctionFromMethod(method=f) func2 = kernel.add_function('plugin', func) assert func != func2 ``` Also closes: #5855 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Removed KernelPluginCollection Updated KernelPlugin added from_... method to load from different sources Updated KernelFunctionFromPrompt added from_yaml and from_directory methods. Added and updated tests to 99% coverage of KernelPlugin and KernelFunctionFromPrompt ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-04-15 15:51:59 +02:00
"plugins": plugins,
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
**kwargs,
}
if ai_service_selector:
args["ai_service_selector"] = ai_service_selector
super().__init__(**args)
async def invoke_stream(
Initial commit
2023-03-16 19:54:34 -07:00
self,
Python: upgraded all files to 310 plus format and removed from future (#6353) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Now that we only support 310 and up, we could unify all files, to use the new style typing and annotations. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-21 18:33:08 +02:00
function: "KernelFunction | None" = None,
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
arguments: KernelArguments | None = None,
function_name: str | None = None,
plugin_name: str | None = None,
Python: Replace mutable default arguments and prevent unintended side effects (#11849) ### Motivation and Context Some functions and methods were using mutable objects (typically `dict`) as default argument values. In Python, default arguments are evaluated only once at the time of function definition. This can lead to unexpected behavior when the default value is modified within the function, as subsequent calls to the function may reuse the modified object. As a result, calling such functions or methods multiple times can lead to unintended behavior. Reference: [8.7. Function definitions (Official Python Documentation)](https://docs.python.org/3.12/reference/compound_stmts.html#function-definitions) > Default parameter values are evaluated from left to right when the function definition is executed. This means that the expression is evaluated once, when the function is defined, and that the same “pre-computed” value is used for each call. This is especially important to understand when a default parameter value is a mutable object, such as a list or a dictionary: if the function modifies the object (e.g. by appending an item to a list), the default parameter value is in effect modified. This is generally not what was intended. A way around this is to use None as the default, and explicitly test for it in the body of the function, e.g.: ### Description - Replaced mutable types (e.g., `dict`, `list`) used as default argument values with `None`, and properly initialized them within the function body. - Reviewed and updated the parts of the code that performed destructive modifications on the arguments, to avoid unintentionally modifying variables from the caller’s scope. - ~Added the `flake8-bugbear` rule in Ruff to detect this kind of issue.~ - ~For now, errors in `tests` and `samples` directories are ignored due to the large number of violations.~ ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [ ] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-05-07 04:03:05 +09:00
metadata: dict[str, Any] | None = None,
Python: implement filters (#5681) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This pull request includes significant changes across multiple files, mainly related to the addition of hooks and the modification of function invocations in the `semantic_kernel` module. The changes also include the addition of a new sample and a YAML file, and modifications to the `__init__.py` files. Removals: * [`python/semantic_kernel/events`](diffhunk://#diff-ebda9504832b19ab83239a92c9a6d5f8c744deff9fef86071c13956ec92bb010L1-L11): Removed the previously used events. New Exceptions: * [`python/semantic_kernel/exceptions/kernel_exceptions.py`](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48): Added new exception classes `OperationCancelledException` and `HookInvalidSignatureError`. [[1]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48) [[2]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR57-R58) Fixes: #3038 Fixes: #6276 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-05-17 23:05:56 +02:00
return_function_results: bool = False,
Python: fix for streaming openai responses, and first parts of fixes for Chat With Your Data (#5387) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [ ] The code builds clean without any errors or warnings - [ ] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [ ] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-03-08 22:10:43 +01:00
**kwargs: Any,
Python: mypy coverage for functions (#5883) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> extend mypy typing coverage to all classes in 'functions' folder. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-04-16 16:21:29 +02:00
) -> AsyncGenerator[list["StreamingContentMixin"] | FunctionResult | list[FunctionResult], Any]:
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
"""Execute one or more stream functions.
Initial commit
2023-03-16 19:54:34 -07:00
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
This will execute the functions in the order they are provided, if a list of functions is provided.
When multiple functions are provided only the last one is streamed, the rest is executed as a pipeline.
Python: Add kernel plugin collection and remove old plugin collection classes (#4764) ### Motivation and Context This PR solves two works items: #4560 and #4667. The Python SDK had different ways of handling plugins and collections. The work in the PR helps align the SDK with dotnet nomenclature and structure. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description A `KernelPluginCollection` is introduced with an underlying `DefaultKernelPlugin` and a `KernelPlugin` base class. Other code is refactored to use this new `KernelPluginCollection` model which should simplify how plugins are handled. Unit tests are added to test the new functionality. Existing unit tests and integration tests, along with kernel examples and notebooks were updated based on these changes. Unlike before, we're requiring plugins to be defined with a name, which removes the GLOBAL_PLUGIN constant name of "__GLOBAL_PLUGIN__". In some senses, while creating a plugin or a function, a name may not be provided and so a random name in the form of a plugin (p_<random_ascii_chars>) or function (f_<random_ascii_chars>) is used. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-01 13:57:00 -05:00
Python: new pre-commit actions and pre-commit as GHA (#6376) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> - Added static code checking to pre-commit (check-ast and nbqa-check-ast) - Added Bandit security checking to pre-commit - Added pre-commit step to python-lint workflow, if it works, can delete seperate mypy, ruff and black tests ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-28 17:28:04 +02:00
Args:
function (KernelFunction): The function to execute,
this value has precedence when supplying both this and using function_name and plugin_name,
if this is none, function_name and plugin_name are used and cannot be None.
arguments (KernelArguments | None): The arguments to pass to the function(s), optional
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
function_name (str | None): The name of the function to execute
plugin_name (str | None): The name of the plugin to execute
Python: implement filters (#5681) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This pull request includes significant changes across multiple files, mainly related to the addition of hooks and the modification of function invocations in the `semantic_kernel` module. The changes also include the addition of a new sample and a YAML file, and modifications to the `__init__.py` files. Removals: * [`python/semantic_kernel/events`](diffhunk://#diff-ebda9504832b19ab83239a92c9a6d5f8c744deff9fef86071c13956ec92bb010L1-L11): Removed the previously used events. New Exceptions: * [`python/semantic_kernel/exceptions/kernel_exceptions.py`](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48): Added new exception classes `OperationCancelledException` and `HookInvalidSignatureError`. [[1]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48) [[2]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR57-R58) Fixes: #3038 Fixes: #6276 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-05-17 23:05:56 +02:00
metadata (dict[str, Any]): The metadata to pass to the function(s)
return_function_results (bool): If True, the function results are yielded as a list[FunctionResult]
in addition to the streaming content, otherwise only the streaming content is yielded.
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
kwargs (dict[str, Any]): arguments that can be used instead of supplying KernelArguments
Python: Add kernel plugin collection and remove old plugin collection classes (#4764) ### Motivation and Context This PR solves two works items: #4560 and #4667. The Python SDK had different ways of handling plugins and collections. The work in the PR helps align the SDK with dotnet nomenclature and structure. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description A `KernelPluginCollection` is introduced with an underlying `DefaultKernelPlugin` and a `KernelPlugin` base class. Other code is refactored to use this new `KernelPluginCollection` model which should simplify how plugins are handled. Unit tests are added to test the new functionality. Existing unit tests and integration tests, along with kernel examples and notebooks were updated based on these changes. Unlike before, we're requiring plugins to be defined with a name, which removes the GLOBAL_PLUGIN constant name of "__GLOBAL_PLUGIN__". In some senses, while creating a plugin or a function, a name may not be provided and so a random name in the form of a plugin (p_<random_ascii_chars>) or function (f_<random_ascii_chars>) is used. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-01 13:57:00 -05:00
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
Yields:
Python: Enhance Chat Message Content creation and parsing (#5520) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Introduces a polymorphic approach to creating and rendering Chat Message Contents and it's subclasses, leverages Pydantic Rootmodel for that. Fixes #5496 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds ChatMessageContentBase that can create all types of CMC's, based on a type field (str) Redoes some of the rendering and parsing also retooled the hierarchy structure for Contents, including introducing a StreamingContentMixin instead of the StreamingKernelContent (which was never used), now all content classes (TextContent, CMC, etc) determine the fields used, while the Mixin adds one additional field and two additional methods for streaming messages, especially __add__ everything else runs the same way. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-28 13:45:56 +01:00
StreamingContentMixin: The content of the stream of the last function provided.
Python: Add kernel plugin collection and remove old plugin collection classes (#4764) ### Motivation and Context This PR solves two works items: #4560 and #4667. The Python SDK had different ways of handling plugins and collections. The work in the PR helps align the SDK with dotnet nomenclature and structure. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description A `KernelPluginCollection` is introduced with an underlying `DefaultKernelPlugin` and a `KernelPlugin` base class. Other code is refactored to use this new `KernelPluginCollection` model which should simplify how plugins are handled. Unit tests are added to test the new functionality. Existing unit tests and integration tests, along with kernel examples and notebooks were updated based on these changes. Unlike before, we're requiring plugins to be defined with a name, which removes the GLOBAL_PLUGIN constant name of "__GLOBAL_PLUGIN__". In some senses, while creating a plugin or a function, a name may not be provided and so a random name in the form of a plugin (p_<random_ascii_chars>) or function (f_<random_ascii_chars>) is used. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-01 13:57:00 -05:00
"""
Python: split kernelfunction into fromprompt and frommethod classes (#5159) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Closes #4633 and brings this part of the python sdk in sync with dotnet, also makes for a much more readable set of classes. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added: - KernelFunctionFromPrompt class - KernelFunctionFromMethod class - tests for both. Changed: - removed parts of KernelFunction that are specific to one of the two subtypes - removed expose attribute of KernelParameterMetadata as that is no longer needed ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-02-26 21:20:25 +01:00
if arguments is None:
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
arguments = KernelArguments(**kwargs)
Python: slight improvements in kernel and function result content, sample of local function calling with Nexus Raven (#6982) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This introduces a sample that uses function calling with a local model, Nexus Raven V2, in this case using the API, but the same model can be loaded using the Hugging Face Connection, the class here is highly adapted and specific to Nexus including parsing the returned function call value, into FunctionCallContent and then running that. It then returns one or more ChatMessageContents with the function call content and function result content which can then be shared with a OpenAI call. Error handling improvements: * [`python/semantic_kernel/contents/chat_history.py`](diffhunk://#diff-b57a30f97fec1a91bbf6f7d15584473a8f0cf409797bdcd90d8c80bedc670892L266-R267): Improved error logging when parsing prompts as XML fails. * [`python/semantic_kernel/contents/function_result_content.py`](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10): Removed field validator for `result` and allow result to be any value, instead of casting everything to string. [[1]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10) [[2]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L50-R51) [[3]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L63-L69) [[4]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87R92-R115) Minor changes: * [`python/semantic_kernel/core_plugins/math_plugin.py`](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38): Removed some docstrings for clarity. [[1]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38) [[2]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L55-R47) * [`python/semantic_kernel/functions/kernel_function_from_prompt.py`](diffhunk://#diff-eabd53da2ec241cce6f92de85283065e7b478d62cefb52817c5b29ff07e28189L173-R174): Changed the order of condition checks for better readability. * [`python/semantic_kernel/kernel.py`](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10): Added functionality to update kernel arguments in `invoke_stream` method and made `function_name` and `plugin_name` optional in `invoke_prompt` method. [[1]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10) [[2]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR29) [[3]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR119-R120) [[4]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL195-R202) [[5]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL209-R216) Closes #6804 Closes #6944 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-28 22:17:05 +02:00
else:
arguments.update(kwargs)
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
if not function:
Python: Update kernel tests (#5341) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Adding a whole bunch of tests to better cover Kernel.py. Also reviewed and updated the conftest and tweaked the scope, now runs faster! Skipped efforts for handlers since work is being done there. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-07 08:38:04 +01:00
if not function_name or not plugin_name:
raise KernelFunctionNotFoundError("No function(s) or function- and plugin-name provided")
Python: Updated plugins (#5827) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> The loading of plugins and functions into Kernel was not fully consistent and clear, this is now simplified. Adding a plugin can now be done in two ways: - multiple: `kernel.add_plugins[plugin1, plugin2]` - single plugin: `kernel.add_plugin(plugin)` - if the plugin here is a class, with methods that have the kernel_function decorator then that is parsed into a plugin Adding a function can now be done in three ways: - multiple: `kernel.add_functions([func1, func2]` or `kernel.add_function({'func1': func1})` - single: `kernel.add_function('plugin_name', func1)` - from a prompt: `kernel.add_function(function_name='func1', prompt='test prompt', ...)` In other words, all the different methods that were available have been simplified, these methods also return the created plugin (or updated plugin for the `add_functions` method. The `add_function` (singular) method has a parameter `return_plugin` to control whether you get the created or updated plugin, instead of the created function. **An important callout:** One big internal change that this introduces is that a function that gets added to a plugin (new or existing) is automatically copied, the metadata is deep-copied and the plugin_name in the metadata is updated, so this sequence is valid now: ```python @kernel_function(name='test') def f(...): etc func = KernelFunctionFromMethod(method=f) func2 = kernel.add_function('plugin', func) assert func != func2 ``` Also closes: #5855 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Removed KernelPluginCollection Updated KernelPlugin added from_... method to load from different sources Updated KernelFunctionFromPrompt added from_yaml and from_directory methods. Added and updated tests to 99% coverage of KernelPlugin and KernelFunctionFromPrompt ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-04-15 15:51:59 +02:00
function = self.get_function(plugin_name, function_name)
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
function_result: list[list["StreamingContentMixin"] | Any] = []
async for stream_message in function.invoke_stream(self, arguments):
if isinstance(stream_message, FunctionResult) and (
Python: Unsafe input handling (#6003) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Implements dealing with unsafe content, by doing HTML parsing on variables and function results. Closes: #5889 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds parameter `allow_dangerously_set_content` to: - InputVariable - PromptTemplateConfig - PromptTemplateBase The behavior is that if the flag is set to True on the template itself (KernelPromptTemplate, Jinja2PromptTemplate or HandlebarsPromptTemplate) the behavior is the same, no encoding is done on inputs. Otherwise: - variables are encoded by default, this can be switched off using the InputVariables class for that variable. - function output is encoded by default, this can be switched off using the flag in the PromptTemplateConfig, this is not yet possible to do on a per function basis. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-16 15:02:54 +02:00
exception := stream_message.metadata.get(METADATA_EXCEPTION_KEY, None)
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
):
raise KernelInvokeException(
f"Error occurred while invoking function: '{function.fully_qualified_name}'"
) from exception
function_result.append(stream_message)
yield stream_message
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
if return_function_results:
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
output_function_result: list["StreamingContentMixin"] = []
for result in function_result:
for choice in result:
if not isinstance(choice, StreamingContentMixin):
continue
if len(output_function_result) <= choice.choice_index:
output_function_result.append(copy(choice))
else:
output_function_result[choice.choice_index] += choice
yield FunctionResult(function=function.metadata, value=output_function_result)
Python: temporal kernel run stream async (#1691) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Temporal solution for #1606 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [ ] The code builds clean without any errors or warnings - [ ] The PR follows SK Contribution Guidelines (https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) - [ ] The code follows the .NET coding conventions (https://learn.microsoft.com/dotnet/csharp/fundamentals/coding-style/coding-conventions) verified with `dotnet format` - [ ] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Shawn Callegari <36091529+shawncal@users.noreply.github.com> Co-authored-by: Abby Harrison <54643756+awharrison-28@users.noreply.github.com> Co-authored-by: Abby Harrison <abby.harrison@microsoft.com>
2023-07-18 03:05:39 +09:00
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
async def invoke(
Python: linter fix (#578) ### Motivation and Context lint fix for pre commit checks Co-authored-by: Abby Harrison <54643756+awharrison-28@users.noreply.github.com>
2023-04-22 04:47:37 +09:00
self,
Python: upgraded all files to 310 plus format and removed from future (#6353) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Now that we only support 310 and up, we could unify all files, to use the new style typing and annotations. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-21 18:33:08 +02:00
function: "KernelFunction | None" = None,
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
arguments: KernelArguments | None = None,
function_name: str | None = None,
plugin_name: str | None = None,
Python: Replace mutable default arguments and prevent unintended side effects (#11849) ### Motivation and Context Some functions and methods were using mutable objects (typically `dict`) as default argument values. In Python, default arguments are evaluated only once at the time of function definition. This can lead to unexpected behavior when the default value is modified within the function, as subsequent calls to the function may reuse the modified object. As a result, calling such functions or methods multiple times can lead to unintended behavior. Reference: [8.7. Function definitions (Official Python Documentation)](https://docs.python.org/3.12/reference/compound_stmts.html#function-definitions) > Default parameter values are evaluated from left to right when the function definition is executed. This means that the expression is evaluated once, when the function is defined, and that the same “pre-computed” value is used for each call. This is especially important to understand when a default parameter value is a mutable object, such as a list or a dictionary: if the function modifies the object (e.g. by appending an item to a list), the default parameter value is in effect modified. This is generally not what was intended. A way around this is to use None as the default, and explicitly test for it in the body of the function, e.g.: ### Description - Replaced mutable types (e.g., `dict`, `list`) used as default argument values with `None`, and properly initialized them within the function body. - Reviewed and updated the parts of the code that performed destructive modifications on the arguments, to avoid unintentionally modifying variables from the caller’s scope. - ~Added the `flake8-bugbear` rule in Ruff to detect this kind of issue.~ - ~For now, errors in `tests` and `samples` directories are ignored due to the large number of violations.~ ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [ ] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-05-07 04:03:05 +09:00
metadata: dict[str, Any] | None = None,
Python: fix for streaming openai responses, and first parts of fixes for Chat With Your Data (#5387) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [ ] The code builds clean without any errors or warnings - [ ] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [ ] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-03-08 22:10:43 +01:00
**kwargs: Any,
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
) -> FunctionResult | None:
Python: further improvements to checks and docs in dev_setup (#6445) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Added a section on documentation to the DEV_SETUP Added a check for the copyright row to pre-commit Closes #5629 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-31 15:06:41 +02:00
"""Execute a function and return the FunctionResult.
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
Python: new pre-commit actions and pre-commit as GHA (#6376) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> - Added static code checking to pre-commit (check-ast and nbqa-check-ast) - Added Bandit security checking to pre-commit - Added pre-commit step to python-lint workflow, if it works, can delete seperate mypy, ruff and black tests ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-28 17:28:04 +02:00
Args:
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
function (KernelFunction): The function or functions to execute,
Python: further improvements to checks and docs in dev_setup (#6445) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Added a section on documentation to the DEV_SETUP Added a check for the copyright row to pre-commit Closes #5629 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-31 15:06:41 +02:00
this value has precedence when supplying both this and using function_name and plugin_name,
if this is none, function_name and plugin_name are used and cannot be None.
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
arguments (KernelArguments): The arguments to pass to the function(s), optional
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
function_name (str | None): The name of the function to execute
plugin_name (str | None): The name of the plugin to execute
Python: implement filters (#5681) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This pull request includes significant changes across multiple files, mainly related to the addition of hooks and the modification of function invocations in the `semantic_kernel` module. The changes also include the addition of a new sample and a YAML file, and modifications to the `__init__.py` files. Removals: * [`python/semantic_kernel/events`](diffhunk://#diff-ebda9504832b19ab83239a92c9a6d5f8c744deff9fef86071c13956ec92bb010L1-L11): Removed the previously used events. New Exceptions: * [`python/semantic_kernel/exceptions/kernel_exceptions.py`](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48): Added new exception classes `OperationCancelledException` and `HookInvalidSignatureError`. [[1]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48) [[2]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR57-R58) Fixes: #3038 Fixes: #6276 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-05-17 23:05:56 +02:00
metadata (dict[str, Any]): The metadata to pass to the function(s)
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
kwargs (dict[str, Any]): arguments that can be used instead of supplying KernelArguments
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
Python: further improvements to checks and docs in dev_setup (#6445) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Added a section on documentation to the DEV_SETUP Added a check for the copyright row to pre-commit Closes #5629 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-31 15:06:41 +02:00
Raises:
KernelInvokeException: If an error occurs during function invocation
Python: linter fix (#578) ### Motivation and Context lint fix for pre commit checks Co-authored-by: Abby Harrison <54643756+awharrison-28@users.noreply.github.com>
2023-04-22 04:47:37 +09:00
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
"""
Python: split kernelfunction into fromprompt and frommethod classes (#5159) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Closes #4633 and brings this part of the python sdk in sync with dotnet, also makes for a much more readable set of classes. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added: - KernelFunctionFromPrompt class - KernelFunctionFromMethod class - tests for both. Changed: - removed parts of KernelFunction that are specific to one of the two subtypes - removed expose attribute of KernelParameterMetadata as that is no longer needed ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-02-26 21:20:25 +01:00
if arguments is None:
Python: major features for the Kernel Arguments, Function Result and new Prompt Templating (#5077) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Major work to replace the context with Kernel Arguments on the front-end and Function Results on the back. Updated the function decorator to a new approach, in line with dotnet. Revamps the way function are called, allowing native functions to be completely ignorant of SK, other then the decorator. This also moves things into the new folder structure in sync with dotnet. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds: - KernelArguments, a dict-like class that replaces the variables in KernelContext. Closes #4565 - FunctionResult, a class to hold the results of function executions, includes the function, the value and metadata, as well as two convenient function to get the value out of it (str and get_inner_content) the first is generic, the second specifically for KernelContent responses (from AI services). - AI Service Selector class, has a default, which is based on the order in the arguments followed by the order in the functions, can be overridden to implement your own strategy. Closes #4631 - Introduces ChatHistory and refactors the PromptTemplateConfig. Closes #4856, #4630 - Improves parsing of templates, will now all validate during creation and throw an error then, instead of some that do not check for validaty until used. - Introduces named_args block and thereby the ability to have multiple arguments for a function call in a template. Closes #5003 Updates: - kernel_function decorators, the parameter decorator was removed and instead we now use Annotated to add a description of a field and we get the type and required from the function definition. - core plugins, use the new approach to kernel_function decorators. - planners, template engines have all been updated to use the kernel and kernelarguments instead of Context. - Events have been updated, now use kernelarguments and function_result - Tokenizers support for named_args and improvements on parsing and checking. - Kernel examples and notebooks to use the latest code. - All unit and integration tests. There is more code coverage now than before. Removed: - kernelContext - kernel_function_parameter_decorator - delegate handling code for native functions - file_io_plugin and tests - SemanticFunctionConfig - ChatPromptTemplate ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com>
2024-02-24 21:05:40 +01:00
arguments = KernelArguments(**kwargs)
Python: update ToolCallBehavior and rename to FunctionCallBehavior to match dotnet and extended capabilities (#5919) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Updated ToolCallBehavior to match the approach in dotnet Closes #5727 Closes #5447 Closes #5414 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Extends ToolCallBehavior class with fields: - enable_kernel_functions - max_use_attempts - allow_any_request_kernel_function Added subclasses of TCB: - KernelFunctions, the configure_options set tools to all functions in the kernel and tool_choice to auto - EnabledFunctions, created with a filter dict, sets tool_choice to auto and tools to the filtered list - RequiredFunction, created with a function fully qualified name (plugin-function), sets tool_choice to that name and adds the definition of just that tool to tools. Methods: - configure_options(kernel, exection_settings) - This sets the execution settings depending on the field in toolcallbehavior - Does nothing in the default ToolCallBehavior class, so there you have to manually set tools and tool_choice ClassMethods: - AutoInvokeKernelFunctions, returns KernelFunctions class with max_auto_invoke_attempts set to 5 (default) - EnabelKernelFunctions, return KernelFunctions but with max_auto_invoke_attempts set to 0, disabling auto invoke, but it might return toolcalls from the model - EnableFunctions, takes the filter and a auto_invoke params and returns a EnabledFunctions class, if auto_invoke == True then it will auto invoke, otherwise it wont. - RequiredFunctions, returns RequiredFunction class with either max_invoke_attempts 0 or 1 depending on auto_invoke param Changed OpenAIChatPromptExecutionSettings to have a field called tool_call_behavior instead of max_invoke_attempts and auto_invoke_tool_calls fields. Some changes in openai_chat_completion_base to handle this. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <evmattso@microsoft.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-05-07 20:33:16 +02:00
else:
arguments.update(kwargs)
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
if not function:
Python: Update kernel tests (#5341) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Adding a whole bunch of tests to better cover Kernel.py. Also reviewed and updated the conftest and tweaked the scope, now runs faster! Skipped efforts for handlers since work is being done there. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-07 08:38:04 +01:00
if not function_name or not plugin_name:
Python: implement filters (#5681) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This pull request includes significant changes across multiple files, mainly related to the addition of hooks and the modification of function invocations in the `semantic_kernel` module. The changes also include the addition of a new sample and a YAML file, and modifications to the `__init__.py` files. Removals: * [`python/semantic_kernel/events`](diffhunk://#diff-ebda9504832b19ab83239a92c9a6d5f8c744deff9fef86071c13956ec92bb010L1-L11): Removed the previously used events. New Exceptions: * [`python/semantic_kernel/exceptions/kernel_exceptions.py`](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48): Added new exception classes `OperationCancelledException` and `HookInvalidSignatureError`. [[1]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48) [[2]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR57-R58) Fixes: #3038 Fixes: #6276 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-05-17 23:05:56 +02:00
raise KernelFunctionNotFoundError("No function, or function name and plugin name provided")
Python: Updated plugins (#5827) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> The loading of plugins and functions into Kernel was not fully consistent and clear, this is now simplified. Adding a plugin can now be done in two ways: - multiple: `kernel.add_plugins[plugin1, plugin2]` - single plugin: `kernel.add_plugin(plugin)` - if the plugin here is a class, with methods that have the kernel_function decorator then that is parsed into a plugin Adding a function can now be done in three ways: - multiple: `kernel.add_functions([func1, func2]` or `kernel.add_function({'func1': func1})` - single: `kernel.add_function('plugin_name', func1)` - from a prompt: `kernel.add_function(function_name='func1', prompt='test prompt', ...)` In other words, all the different methods that were available have been simplified, these methods also return the created plugin (or updated plugin for the `add_functions` method. The `add_function` (singular) method has a parameter `return_plugin` to control whether you get the created or updated plugin, instead of the created function. **An important callout:** One big internal change that this introduces is that a function that gets added to a plugin (new or existing) is automatically copied, the metadata is deep-copied and the plugin_name in the metadata is updated, so this sequence is valid now: ```python @kernel_function(name='test') def f(...): etc func = KernelFunctionFromMethod(method=f) func2 = kernel.add_function('plugin', func) assert func != func2 ``` Also closes: #5855 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Removed KernelPluginCollection Updated KernelPlugin added from_... method to load from different sources Updated KernelFunctionFromPrompt added from_yaml and from_directory methods. Added and updated tests to 99% coverage of KernelPlugin and KernelFunctionFromPrompt ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-04-15 15:51:59 +02:00
function = self.get_function(plugin_name, function_name)
Python: implement filters (#5681) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This pull request includes significant changes across multiple files, mainly related to the addition of hooks and the modification of function invocations in the `semantic_kernel` module. The changes also include the addition of a new sample and a YAML file, and modifications to the `__init__.py` files. Removals: * [`python/semantic_kernel/events`](diffhunk://#diff-ebda9504832b19ab83239a92c9a6d5f8c744deff9fef86071c13956ec92bb010L1-L11): Removed the previously used events. New Exceptions: * [`python/semantic_kernel/exceptions/kernel_exceptions.py`](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48): Added new exception classes `OperationCancelledException` and `HookInvalidSignatureError`. [[1]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48) [[2]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR57-R58) Fixes: #3038 Fixes: #6276 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-05-17 23:05:56 +02:00
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
try:
Python: implement filters (#5681) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This pull request includes significant changes across multiple files, mainly related to the addition of hooks and the modification of function invocations in the `semantic_kernel` module. The changes also include the addition of a new sample and a YAML file, and modifications to the `__init__.py` files. Removals: * [`python/semantic_kernel/events`](diffhunk://#diff-ebda9504832b19ab83239a92c9a6d5f8c744deff9fef86071c13956ec92bb010L1-L11): Removed the previously used events. New Exceptions: * [`python/semantic_kernel/exceptions/kernel_exceptions.py`](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48): Added new exception classes `OperationCancelledException` and `HookInvalidSignatureError`. [[1]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48) [[2]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR57-R58) Fixes: #3038 Fixes: #6276 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-05-17 23:05:56 +02:00
return await function.invoke(kernel=self, arguments=arguments, metadata=metadata)
except OperationCancelledException as exc:
logger.info(f"Operation cancelled during function invocation. Message: {exc}")
return None
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
except Exception as exc:
logger.error(
"Something went wrong in function invocation. During function invocation:"
Python: further improvements to checks and docs in dev_setup (#6445) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Added a section on documentation to the DEV_SETUP Added a check for the copyright row to pre-commit Closes #5629 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-31 15:06:41 +02:00
f" '{function.fully_qualified_name}'. Error description: '{exc!s}'"
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
)
raise KernelInvokeException(
f"Error occurred while invoking function: '{function.fully_qualified_name}'"
Python: implement filters (#5681) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This pull request includes significant changes across multiple files, mainly related to the addition of hooks and the modification of function invocations in the `semantic_kernel` module. The changes also include the addition of a new sample and a YAML file, and modifications to the `__init__.py` files. Removals: * [`python/semantic_kernel/events`](diffhunk://#diff-ebda9504832b19ab83239a92c9a6d5f8c744deff9fef86071c13956ec92bb010L1-L11): Removed the previously used events. New Exceptions: * [`python/semantic_kernel/exceptions/kernel_exceptions.py`](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48): Added new exception classes `OperationCancelledException` and `HookInvalidSignatureError`. [[1]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR41-R48) [[2]](diffhunk://#diff-450aaa5595a8b22cd6ee212eb79b7d6b0d4e9c1072063ef32018a3e7d3fdf21dR57-R58) Fixes: #3038 Fixes: #6276 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-05-17 23:05:56 +02:00
) from exc
Initial commit
2023-03-16 19:54:34 -07:00
Python: Remove the memory tied to the kernel. Update kernel examples and notebooks. (#5143) ### Motivation and Context As we move toward v1, it is necessary to remove the single kernel memory attribute so that memory connectors can be handled via plugins. This change also refactors the `import_plugin` method name to be similar with dotnet as `import_plugin_from_object`, and introduces a simple `kernel.invoke_prompt` method. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description Updates are: - The memory connector is now passed to the `SemanticTextMemory` class, along with the embedding generator. This allows the user to add memory to multiple plugins/services as needed. Closes #4632 - Kernel Examples and Notebooks were updated to reflect these changes. Now using Service Enum for all notebooks. Closes #4557 - There were planners using memory related to the kernel, to store the functions, and many of these planners are non-existent in dotnet's v1 SDK. Python has pending work to update the planner/remove deprecated planners before getting to v1. - A Kernel method called `invoke_prompt` was introduced to invoke a simple prompt, and return the `FunctionResult` when complete. - All unit/integration tests are passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-02-26 12:50:23 -05:00
async def invoke_prompt(
self,
prompt: str,
Python: slight improvements in kernel and function result content, sample of local function calling with Nexus Raven (#6982) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This introduces a sample that uses function calling with a local model, Nexus Raven V2, in this case using the API, but the same model can be loaded using the Hugging Face Connection, the class here is highly adapted and specific to Nexus including parsing the returned function call value, into FunctionCallContent and then running that. It then returns one or more ChatMessageContents with the function call content and function result content which can then be shared with a OpenAI call. Error handling improvements: * [`python/semantic_kernel/contents/chat_history.py`](diffhunk://#diff-b57a30f97fec1a91bbf6f7d15584473a8f0cf409797bdcd90d8c80bedc670892L266-R267): Improved error logging when parsing prompts as XML fails. * [`python/semantic_kernel/contents/function_result_content.py`](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10): Removed field validator for `result` and allow result to be any value, instead of casting everything to string. [[1]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10) [[2]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L50-R51) [[3]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L63-L69) [[4]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87R92-R115) Minor changes: * [`python/semantic_kernel/core_plugins/math_plugin.py`](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38): Removed some docstrings for clarity. [[1]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38) [[2]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L55-R47) * [`python/semantic_kernel/functions/kernel_function_from_prompt.py`](diffhunk://#diff-eabd53da2ec241cce6f92de85283065e7b478d62cefb52817c5b29ff07e28189L173-R174): Changed the order of condition checks for better readability. * [`python/semantic_kernel/kernel.py`](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10): Added functionality to update kernel arguments in `invoke_stream` method and made `function_name` and `plugin_name` optional in `invoke_prompt` method. [[1]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10) [[2]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR29) [[3]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR119-R120) [[4]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL195-R202) [[5]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL209-R216) Closes #6804 Closes #6944 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-28 22:17:05 +02:00
function_name: str | None = None,
plugin_name: str | None = None,
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
arguments: KernelArguments | None = None,
Python: Handlebars template support (#5466) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Introducing Handlebars templating to the Python Semantic Kernel! Closes: #5446 Prework for: #4641 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds HandlebarsPromptTemplate as a subclass of PromptTemplateBase Adds second literal to the PromptTemplateConfig definition Adds helpers for Handlebars (system_helpers) and `create_helper_from_function` function Moves fully_qualified_name into KernelFunctionMetadata instead of in different extensions. Defaults are still 'semantic-kernel' templates, and without making explicit changes that still is the norm. Adds a azure_chat_gpt_handlebars.py sample ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-14 15:00:05 +01:00
template_format: Literal[
Python: support Jinja2 templating (#5483) ### Motivation and Context Jinja2 is a heavily supported Python templating language and we're introducing it to SK Python. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR brings in the ability to specify Jinja2 templates for use within SK Python: - Closes #5343 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-03-15 09:23:25 -04:00
"semantic-kernel",
"handlebars",
"jinja2",
Python: Handlebars template support (#5466) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Introducing Handlebars templating to the Python Semantic Kernel! Closes: #5446 Prework for: #4641 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds HandlebarsPromptTemplate as a subclass of PromptTemplateBase Adds second literal to the PromptTemplateConfig definition Adds helpers for Handlebars (system_helpers) and `create_helper_from_function` function Moves fully_qualified_name into KernelFunctionMetadata instead of in different extensions. Defaults are still 'semantic-kernel' templates, and without making explicit changes that still is the norm. Adds a azure_chat_gpt_handlebars.py sample ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-14 15:00:05 +01:00
] = KERNEL_TEMPLATE_FORMAT_NAME,
Python: .Net: Updated encoding logic in prompt templates (#12983) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Resolves: https://github.com/microsoft/semantic-kernel/issues/11821 Today, the encoding of template arguments is performed only if argument type is `string`. In case of custom type, anonymous type or collection - the encoding is not performed. This PR contains changes to throw an exception in case if encoding is enabled but complex type is used. In case of complex type, the encoding should be performed manually according to business logic and automatic encoding should be explicitly disabled. This enforces stricter, but more secure template rendering rules. **Note**: this is a breaking change for customers who use Handlebars or Liquid template with complex type arguments. Code changes are required when initializing template arguments: ```diff var arguments = new KernelArguments() { { "customer", new { - firstName = userInput.FirstName, - lastName = userInput.LastName, + firstName = HttpUtility.HtmlEncode(userInput.FirstName), + lastName = HttpUtility.HtmlEncode(userInput.LastName), } } }; var templateFactory = new LiquidPromptTemplateFactory(); var promptTemplateConfig = new PromptTemplateConfig() { TemplateFormat = "liquid" + InputVariables = new() + { + // We set AllowDangerouslySetContent to 'true' because each property of this argument is encoded manually. + new() { Name = "customer", AllowDangerouslySetContent = true }, + } }; var promptTemplate = templateFactory.Create(promptTemplateConfig); var renderedPrompt = await promptTemplate.RenderAsync(kernel, arguments); ``` ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-08-26 22:09:00 -07:00
prompt_template_config: PromptTemplateConfig | None = None,
Python: fix for streaming openai responses, and first parts of fixes for Chat With Your Data (#5387) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [ ] The code builds clean without any errors or warnings - [ ] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [ ] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-03-08 22:10:43 +01:00
**kwargs: Any,
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
) -> FunctionResult | None:
Python: new pre-commit actions and pre-commit as GHA (#6376) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> - Added static code checking to pre-commit (check-ast and nbqa-check-ast) - Added Bandit security checking to pre-commit - Added pre-commit step to python-lint workflow, if it works, can delete seperate mypy, ruff and black tests ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-28 17:28:04 +02:00
"""Invoke a function from the provided prompt.
Python: Remove the memory tied to the kernel. Update kernel examples and notebooks. (#5143) ### Motivation and Context As we move toward v1, it is necessary to remove the single kernel memory attribute so that memory connectors can be handled via plugins. This change also refactors the `import_plugin` method name to be similar with dotnet as `import_plugin_from_object`, and introduces a simple `kernel.invoke_prompt` method. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description Updates are: - The memory connector is now passed to the `SemanticTextMemory` class, along with the embedding generator. This allows the user to add memory to multiple plugins/services as needed. Closes #4632 - Kernel Examples and Notebooks were updated to reflect these changes. Now using Service Enum for all notebooks. Closes #4557 - There were planners using memory related to the kernel, to store the functions, and many of these planners are non-existent in dotnet's v1 SDK. Python has pending work to update the planner/remove deprecated planners before getting to v1. - A Kernel method called `invoke_prompt` was introduced to invoke a simple prompt, and return the `FunctionResult` when complete. - All unit/integration tests are passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-02-26 12:50:23 -05:00
Args:
prompt (str): The prompt to use
Python: slight improvements in kernel and function result content, sample of local function calling with Nexus Raven (#6982) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This introduces a sample that uses function calling with a local model, Nexus Raven V2, in this case using the API, but the same model can be loaded using the Hugging Face Connection, the class here is highly adapted and specific to Nexus including parsing the returned function call value, into FunctionCallContent and then running that. It then returns one or more ChatMessageContents with the function call content and function result content which can then be shared with a OpenAI call. Error handling improvements: * [`python/semantic_kernel/contents/chat_history.py`](diffhunk://#diff-b57a30f97fec1a91bbf6f7d15584473a8f0cf409797bdcd90d8c80bedc670892L266-R267): Improved error logging when parsing prompts as XML fails. * [`python/semantic_kernel/contents/function_result_content.py`](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10): Removed field validator for `result` and allow result to be any value, instead of casting everything to string. [[1]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10) [[2]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L50-R51) [[3]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L63-L69) [[4]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87R92-R115) Minor changes: * [`python/semantic_kernel/core_plugins/math_plugin.py`](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38): Removed some docstrings for clarity. [[1]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38) [[2]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L55-R47) * [`python/semantic_kernel/functions/kernel_function_from_prompt.py`](diffhunk://#diff-eabd53da2ec241cce6f92de85283065e7b478d62cefb52817c5b29ff07e28189L173-R174): Changed the order of condition checks for better readability. * [`python/semantic_kernel/kernel.py`](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10): Added functionality to update kernel arguments in `invoke_stream` method and made `function_name` and `plugin_name` optional in `invoke_prompt` method. [[1]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10) [[2]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR29) [[3]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR119-R120) [[4]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL195-R202) [[5]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL209-R216) Closes #6804 Closes #6944 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-28 22:17:05 +02:00
function_name (str): The name of the function, optional
plugin_name (str): The name of the plugin, optional
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
arguments (KernelArguments | None): The arguments to pass to the function(s), optional
template_format (str | None): The format of the prompt template
Python: .Net: Updated encoding logic in prompt templates (#12983) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Resolves: https://github.com/microsoft/semantic-kernel/issues/11821 Today, the encoding of template arguments is performed only if argument type is `string`. In case of custom type, anonymous type or collection - the encoding is not performed. This PR contains changes to throw an exception in case if encoding is enabled but complex type is used. In case of complex type, the encoding should be performed manually according to business logic and automatic encoding should be explicitly disabled. This enforces stricter, but more secure template rendering rules. **Note**: this is a breaking change for customers who use Handlebars or Liquid template with complex type arguments. Code changes are required when initializing template arguments: ```diff var arguments = new KernelArguments() { { "customer", new { - firstName = userInput.FirstName, - lastName = userInput.LastName, + firstName = HttpUtility.HtmlEncode(userInput.FirstName), + lastName = HttpUtility.HtmlEncode(userInput.LastName), } } }; var templateFactory = new LiquidPromptTemplateFactory(); var promptTemplateConfig = new PromptTemplateConfig() { TemplateFormat = "liquid" + InputVariables = new() + { + // We set AllowDangerouslySetContent to 'true' because each property of this argument is encoded manually. + new() { Name = "customer", AllowDangerouslySetContent = true }, + } }; var promptTemplate = templateFactory.Create(promptTemplateConfig); var renderedPrompt = await promptTemplate.RenderAsync(kernel, arguments); ``` ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-08-26 22:09:00 -07:00
prompt_template_config (PromptTemplateConfig | None): The prompt template configuration
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
kwargs (dict[str, Any]): arguments that can be used instead of supplying KernelArguments
Python: Remove the memory tied to the kernel. Update kernel examples and notebooks. (#5143) ### Motivation and Context As we move toward v1, it is necessary to remove the single kernel memory attribute so that memory connectors can be handled via plugins. This change also refactors the `import_plugin` method name to be similar with dotnet as `import_plugin_from_object`, and introduces a simple `kernel.invoke_prompt` method. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description Updates are: - The memory connector is now passed to the `SemanticTextMemory` class, along with the embedding generator. This allows the user to add memory to multiple plugins/services as needed. Closes #4632 - Kernel Examples and Notebooks were updated to reflect these changes. Now using Service Enum for all notebooks. Closes #4557 - There were planners using memory related to the kernel, to store the functions, and many of these planners are non-existent in dotnet's v1 SDK. Python has pending work to update the planner/remove deprecated planners before getting to v1. - A Kernel method called `invoke_prompt` was introduced to invoke a simple prompt, and return the `FunctionResult` when complete. - All unit/integration tests are passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-02-26 12:50:23 -05:00
Returns:
Python: Enhanced pre commit and tasks (#5512) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> To further make the SK python robust and make it easier to get high-quality PRs, some work on the pre-commit-config.yaml and adding a mypy settings ini. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added two steps to the pre-commit-config, the first for pypy, the second for tests with coverage. Over time we want to mandate these checks to run against a PR before it goes to github, that will also reduce the number of ruff and black fix commits, and non-passing unit tests. Added a mypy.ini, currently all but the root folder is excluded, so that we can gradually introduce mypy coverage, did the first pieces in kernel.py, including switch to new style annotations (using from `__future__ import annotations`) in Kernel. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-19 21:20:09 +01:00
FunctionResult | list[FunctionResult] | None: The result of the function(s)
Python: Remove the memory tied to the kernel. Update kernel examples and notebooks. (#5143) ### Motivation and Context As we move toward v1, it is necessary to remove the single kernel memory attribute so that memory connectors can be handled via plugins. This change also refactors the `import_plugin` method name to be similar with dotnet as `import_plugin_from_object`, and introduces a simple `kernel.invoke_prompt` method. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description Updates are: - The memory connector is now passed to the `SemanticTextMemory` class, along with the embedding generator. This allows the user to add memory to multiple plugins/services as needed. Closes #4632 - Kernel Examples and Notebooks were updated to reflect these changes. Now using Service Enum for all notebooks. Closes #4557 - There were planners using memory related to the kernel, to store the functions, and many of these planners are non-existent in dotnet's v1 SDK. Python has pending work to update the planner/remove deprecated planners before getting to v1. - A Kernel method called `invoke_prompt` was introduced to invoke a simple prompt, and return the `FunctionResult` when complete. - All unit/integration tests are passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-02-26 12:50:23 -05:00
"""
Python: Fix KernelArgument bug in invoke_prompt functions (#8414) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> In `invoke_prompt(...)` and `invoke_prompt_async(...)`, the `arguments` parameter is tested to see if it's empty before proceeding. The type of the `arguments` parameter is `KernelArgument`, which derives from a `dict` type, with an additional field for execution settings. When `arguments` contains only the execution settings, it will be considered as an empty dictionary. This will result in the `arguments` parameter being overridden completely, invalidating all execution settings passed to the APIs. To reproduce (assuming we have a plugin/function that will return the weather): ``` result = await kernel.invoke_prompt( "What is the weather like in my location?", arguments=KernelArguments( settings={ service_id: PromptExecutionSettings( service_id=service_id, function_choice_behavior=FunctionChoiceBehavior.Auto(), ), } ), ) ``` In the example above, the settings will be overridden to None. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> 1. Fix the bug. 2. Fix mypy in the pre-commit hook in order to commit. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-29 08:40:53 -07:00
if arguments is None:
Python: Remove the memory tied to the kernel. Update kernel examples and notebooks. (#5143) ### Motivation and Context As we move toward v1, it is necessary to remove the single kernel memory attribute so that memory connectors can be handled via plugins. This change also refactors the `import_plugin` method name to be similar with dotnet as `import_plugin_from_object`, and introduces a simple `kernel.invoke_prompt` method. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description Updates are: - The memory connector is now passed to the `SemanticTextMemory` class, along with the embedding generator. This allows the user to add memory to multiple plugins/services as needed. Closes #4632 - Kernel Examples and Notebooks were updated to reflect these changes. Now using Service Enum for all notebooks. Closes #4557 - There were planners using memory related to the kernel, to store the functions, and many of these planners are non-existent in dotnet's v1 SDK. Python has pending work to update the planner/remove deprecated planners before getting to v1. - A Kernel method called `invoke_prompt` was introduced to invoke a simple prompt, and return the `FunctionResult` when complete. - All unit/integration tests are passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-02-26 12:50:23 -05:00
arguments = KernelArguments(**kwargs)
if not prompt:
Python: rebuilt exceptions structure; pythonic version (#5231) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> The existing Exceptions structure was very much inspired by dotnet, this is now replaced with a pythonic implementations. This means all Exceptions derive from KernelException and then specialise for different purposes. Closes #2194 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Added folder for all exception types, all can be imported through `from semantic_kernel.exceptions import ...` no need for a user to know which file the relevant one is in, but keeps things tidy for developers. Removed old KernelException, added back subtypes for the errorcodes. Went through everything to make sure the `raise ... from exc` pattern is used as much as possible as that returns a better stacktrace. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-02-28 19:53:53 +01:00
raise TemplateSyntaxError("The prompt is either null or empty.")
Python: Update kernel tests (#5341) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Adding a whole bunch of tests to better cover Kernel.py. Also reviewed and updated the conftest and tweaked the scope, now runs faster! Skipped efforts for handlers since work is being done there. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-03-07 08:38:04 +01:00
function = KernelFunctionFromPrompt(
Python: slight improvements in kernel and function result content, sample of local function calling with Nexus Raven (#6982) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This introduces a sample that uses function calling with a local model, Nexus Raven V2, in this case using the API, but the same model can be loaded using the Hugging Face Connection, the class here is highly adapted and specific to Nexus including parsing the returned function call value, into FunctionCallContent and then running that. It then returns one or more ChatMessageContents with the function call content and function result content which can then be shared with a OpenAI call. Error handling improvements: * [`python/semantic_kernel/contents/chat_history.py`](diffhunk://#diff-b57a30f97fec1a91bbf6f7d15584473a8f0cf409797bdcd90d8c80bedc670892L266-R267): Improved error logging when parsing prompts as XML fails. * [`python/semantic_kernel/contents/function_result_content.py`](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10): Removed field validator for `result` and allow result to be any value, instead of casting everything to string. [[1]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10) [[2]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L50-R51) [[3]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L63-L69) [[4]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87R92-R115) Minor changes: * [`python/semantic_kernel/core_plugins/math_plugin.py`](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38): Removed some docstrings for clarity. [[1]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38) [[2]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L55-R47) * [`python/semantic_kernel/functions/kernel_function_from_prompt.py`](diffhunk://#diff-eabd53da2ec241cce6f92de85283065e7b478d62cefb52817c5b29ff07e28189L173-R174): Changed the order of condition checks for better readability. * [`python/semantic_kernel/kernel.py`](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10): Added functionality to update kernel arguments in `invoke_stream` method and made `function_name` and `plugin_name` optional in `invoke_prompt` method. [[1]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10) [[2]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR29) [[3]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR119-R120) [[4]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL195-R202) [[5]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL209-R216) Closes #6804 Closes #6944 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-28 22:17:05 +02:00
function_name=function_name or generate_random_ascii_name(),
Python: Add function name and plugin name to invoke prompt. Add unit tests. (#5318) ### Motivation and Context The kernel.invoke_prompt was missing the function name and plugin name args and was failing. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR: - Adds the required arguments. Closes #5313 - Adds unit tests. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-03-06 03:38:47 -05:00
plugin_name=plugin_name,
Python: Remove the memory tied to the kernel. Update kernel examples and notebooks. (#5143) ### Motivation and Context As we move toward v1, it is necessary to remove the single kernel memory attribute so that memory connectors can be handled via plugins. This change also refactors the `import_plugin` method name to be similar with dotnet as `import_plugin_from_object`, and introduces a simple `kernel.invoke_prompt` method. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description Updates are: - The memory connector is now passed to the `SemanticTextMemory` class, along with the embedding generator. This allows the user to add memory to multiple plugins/services as needed. Closes #4632 - Kernel Examples and Notebooks were updated to reflect these changes. Now using Service Enum for all notebooks. Closes #4557 - There were planners using memory related to the kernel, to store the functions, and many of these planners are non-existent in dotnet's v1 SDK. Python has pending work to update the planner/remove deprecated planners before getting to v1. - A Kernel method called `invoke_prompt` was introduced to invoke a simple prompt, and return the `FunctionResult` when complete. - All unit/integration tests are passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-02-26 12:50:23 -05:00
prompt=prompt,
template_format=template_format,
Python: .Net: Updated encoding logic in prompt templates (#12983) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Resolves: https://github.com/microsoft/semantic-kernel/issues/11821 Today, the encoding of template arguments is performed only if argument type is `string`. In case of custom type, anonymous type or collection - the encoding is not performed. This PR contains changes to throw an exception in case if encoding is enabled but complex type is used. In case of complex type, the encoding should be performed manually according to business logic and automatic encoding should be explicitly disabled. This enforces stricter, but more secure template rendering rules. **Note**: this is a breaking change for customers who use Handlebars or Liquid template with complex type arguments. Code changes are required when initializing template arguments: ```diff var arguments = new KernelArguments() { { "customer", new { - firstName = userInput.FirstName, - lastName = userInput.LastName, + firstName = HttpUtility.HtmlEncode(userInput.FirstName), + lastName = HttpUtility.HtmlEncode(userInput.LastName), } } }; var templateFactory = new LiquidPromptTemplateFactory(); var promptTemplateConfig = new PromptTemplateConfig() { TemplateFormat = "liquid" + InputVariables = new() + { + // We set AllowDangerouslySetContent to 'true' because each property of this argument is encoded manually. + new() { Name = "customer", AllowDangerouslySetContent = true }, + } }; var promptTemplate = templateFactory.Create(promptTemplateConfig); var renderedPrompt = await promptTemplate.RenderAsync(kernel, arguments); ``` ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-08-26 22:09:00 -07:00
prompt_template_config=prompt_template_config,
Python: Remove the memory tied to the kernel. Update kernel examples and notebooks. (#5143) ### Motivation and Context As we move toward v1, it is necessary to remove the single kernel memory attribute so that memory connectors can be handled via plugins. This change also refactors the `import_plugin` method name to be similar with dotnet as `import_plugin_from_object`, and introduces a simple `kernel.invoke_prompt` method. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description Updates are: - The memory connector is now passed to the `SemanticTextMemory` class, along with the embedding generator. This allows the user to add memory to multiple plugins/services as needed. Closes #4632 - Kernel Examples and Notebooks were updated to reflect these changes. Now using Service Enum for all notebooks. Closes #4557 - There were planners using memory related to the kernel, to store the functions, and many of these planners are non-existent in dotnet's v1 SDK. Python has pending work to update the planner/remove deprecated planners before getting to v1. - A Kernel method called `invoke_prompt` was introduced to invoke a simple prompt, and return the `FunctionResult` when complete. - All unit/integration tests are passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-02-26 12:50:23 -05:00
)
Python: removed pipeline (#5779) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Remove the ability to run multiple functions in go in kernel.invoke. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2024-04-04 17:54:19 +02:00
return await self.invoke(function=function, arguments=arguments)
Python: Remove the memory tied to the kernel. Update kernel examples and notebooks. (#5143) ### Motivation and Context As we move toward v1, it is necessary to remove the single kernel memory attribute so that memory connectors can be handled via plugins. This change also refactors the `import_plugin` method name to be similar with dotnet as `import_plugin_from_object`, and introduces a simple `kernel.invoke_prompt` method. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description Updates are: - The memory connector is now passed to the `SemanticTextMemory` class, along with the embedding generator. This allows the user to add memory to multiple plugins/services as needed. Closes #4632 - Kernel Examples and Notebooks were updated to reflect these changes. Now using Service Enum for all notebooks. Closes #4557 - There were planners using memory related to the kernel, to store the functions, and many of these planners are non-existent in dotnet's v1 SDK. Python has pending work to update the planner/remove deprecated planners before getting to v1. - A Kernel method called `invoke_prompt` was introduced to invoke a simple prompt, and return the `FunctionResult` when complete. - All unit/integration tests are passing. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2024-02-26 12:50:23 -05:00
Python: add new samples and fix streaming tool call FunctionCallContent formation (#5877) ### Motivation and Context We're working towards creating a core set of syntax examples (there are four total). The core examples will be available in all 3 SK languages. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces two (of the four) new kernel syntax examples, which will align with the new kernel examples coming soon for both dotnet and Java. #5784 - Introduce a custom weather plugin that in conjunction with the core TimePlugin, make use of auto function calling. - Introduce a kernel syntax example that shows how to integrate with the Microsoft Graph API to create a "restaurant booking." Note: this doesn't actually place a real reservation, but shows how to interact with msgraph. - Also fixes an issue where the streaming tool call argument formation was broken. Closes #6106 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-05-02 18:15:07 -04:00
async def invoke_prompt_stream(
self,
prompt: str,
Python: slight improvements in kernel and function result content, sample of local function calling with Nexus Raven (#6982) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This introduces a sample that uses function calling with a local model, Nexus Raven V2, in this case using the API, but the same model can be loaded using the Hugging Face Connection, the class here is highly adapted and specific to Nexus including parsing the returned function call value, into FunctionCallContent and then running that. It then returns one or more ChatMessageContents with the function call content and function result content which can then be shared with a OpenAI call. Error handling improvements: * [`python/semantic_kernel/contents/chat_history.py`](diffhunk://#diff-b57a30f97fec1a91bbf6f7d15584473a8f0cf409797bdcd90d8c80bedc670892L266-R267): Improved error logging when parsing prompts as XML fails. * [`python/semantic_kernel/contents/function_result_content.py`](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10): Removed field validator for `result` and allow result to be any value, instead of casting everything to string. [[1]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10) [[2]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L50-R51) [[3]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L63-L69) [[4]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87R92-R115) Minor changes: * [`python/semantic_kernel/core_plugins/math_plugin.py`](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38): Removed some docstrings for clarity. [[1]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38) [[2]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L55-R47) * [`python/semantic_kernel/functions/kernel_function_from_prompt.py`](diffhunk://#diff-eabd53da2ec241cce6f92de85283065e7b478d62cefb52817c5b29ff07e28189L173-R174): Changed the order of condition checks for better readability. * [`python/semantic_kernel/kernel.py`](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10): Added functionality to update kernel arguments in `invoke_stream` method and made `function_name` and `plugin_name` optional in `invoke_prompt` method. [[1]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10) [[2]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR29) [[3]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR119-R120) [[4]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL195-R202) [[5]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL209-R216) Closes #6804 Closes #6944 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-28 22:17:05 +02:00
function_name: str | None = None,
plugin_name: str | None = None,
Python: add new samples and fix streaming tool call FunctionCallContent formation (#5877) ### Motivation and Context We're working towards creating a core set of syntax examples (there are four total). The core examples will be available in all 3 SK languages. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces two (of the four) new kernel syntax examples, which will align with the new kernel examples coming soon for both dotnet and Java. #5784 - Introduce a custom weather plugin that in conjunction with the core TimePlugin, make use of auto function calling. - Introduce a kernel syntax example that shows how to integrate with the Microsoft Graph API to create a "restaurant booking." Note: this doesn't actually place a real reservation, but shows how to interact with msgraph. - Also fixes an issue where the streaming tool call argument formation was broken. Closes #6106 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-05-02 18:15:07 -04:00
arguments: KernelArguments | None = None,
template_format: Literal[
"semantic-kernel",
"handlebars",
"jinja2",
] = KERNEL_TEMPLATE_FORMAT_NAME,
return_function_results: bool | None = False,
Python: .Net: Updated encoding logic in prompt templates (#12983) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Resolves: https://github.com/microsoft/semantic-kernel/issues/11821 Today, the encoding of template arguments is performed only if argument type is `string`. In case of custom type, anonymous type or collection - the encoding is not performed. This PR contains changes to throw an exception in case if encoding is enabled but complex type is used. In case of complex type, the encoding should be performed manually according to business logic and automatic encoding should be explicitly disabled. This enforces stricter, but more secure template rendering rules. **Note**: this is a breaking change for customers who use Handlebars or Liquid template with complex type arguments. Code changes are required when initializing template arguments: ```diff var arguments = new KernelArguments() { { "customer", new { - firstName = userInput.FirstName, - lastName = userInput.LastName, + firstName = HttpUtility.HtmlEncode(userInput.FirstName), + lastName = HttpUtility.HtmlEncode(userInput.LastName), } } }; var templateFactory = new LiquidPromptTemplateFactory(); var promptTemplateConfig = new PromptTemplateConfig() { TemplateFormat = "liquid" + InputVariables = new() + { + // We set AllowDangerouslySetContent to 'true' because each property of this argument is encoded manually. + new() { Name = "customer", AllowDangerouslySetContent = true }, + } }; var promptTemplate = templateFactory.Create(promptTemplateConfig); var renderedPrompt = await promptTemplate.RenderAsync(kernel, arguments); ``` ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-08-26 22:09:00 -07:00
prompt_template_config: PromptTemplateConfig | None = None,
Python: add new samples and fix streaming tool call FunctionCallContent formation (#5877) ### Motivation and Context We're working towards creating a core set of syntax examples (there are four total). The core examples will be available in all 3 SK languages. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces two (of the four) new kernel syntax examples, which will align with the new kernel examples coming soon for both dotnet and Java. #5784 - Introduce a custom weather plugin that in conjunction with the core TimePlugin, make use of auto function calling. - Introduce a kernel syntax example that shows how to integrate with the Microsoft Graph API to create a "restaurant booking." Note: this doesn't actually place a real reservation, but shows how to interact with msgraph. - Also fixes an issue where the streaming tool call argument formation was broken. Closes #6106 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-05-02 18:15:07 -04:00
**kwargs: Any,
) -> AsyncIterable[list["StreamingContentMixin"] | FunctionResult | list[FunctionResult]]:
Python: new pre-commit actions and pre-commit as GHA (#6376) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> - Added static code checking to pre-commit (check-ast and nbqa-check-ast) - Added Bandit security checking to pre-commit - Added pre-commit step to python-lint workflow, if it works, can delete seperate mypy, ruff and black tests ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-28 17:28:04 +02:00
"""Invoke a function from the provided prompt and stream the results.
Python: add new samples and fix streaming tool call FunctionCallContent formation (#5877) ### Motivation and Context We're working towards creating a core set of syntax examples (there are four total). The core examples will be available in all 3 SK languages. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces two (of the four) new kernel syntax examples, which will align with the new kernel examples coming soon for both dotnet and Java. #5784 - Introduce a custom weather plugin that in conjunction with the core TimePlugin, make use of auto function calling. - Introduce a kernel syntax example that shows how to integrate with the Microsoft Graph API to create a "restaurant booking." Note: this doesn't actually place a real reservation, but shows how to interact with msgraph. - Also fixes an issue where the streaming tool call argument formation was broken. Closes #6106 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-05-02 18:15:07 -04:00
Args:
prompt (str): The prompt to use
Python: slight improvements in kernel and function result content, sample of local function calling with Nexus Raven (#6982) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This introduces a sample that uses function calling with a local model, Nexus Raven V2, in this case using the API, but the same model can be loaded using the Hugging Face Connection, the class here is highly adapted and specific to Nexus including parsing the returned function call value, into FunctionCallContent and then running that. It then returns one or more ChatMessageContents with the function call content and function result content which can then be shared with a OpenAI call. Error handling improvements: * [`python/semantic_kernel/contents/chat_history.py`](diffhunk://#diff-b57a30f97fec1a91bbf6f7d15584473a8f0cf409797bdcd90d8c80bedc670892L266-R267): Improved error logging when parsing prompts as XML fails. * [`python/semantic_kernel/contents/function_result_content.py`](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10): Removed field validator for `result` and allow result to be any value, instead of casting everything to string. [[1]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10) [[2]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L50-R51) [[3]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L63-L69) [[4]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87R92-R115) Minor changes: * [`python/semantic_kernel/core_plugins/math_plugin.py`](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38): Removed some docstrings for clarity. [[1]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38) [[2]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L55-R47) * [`python/semantic_kernel/functions/kernel_function_from_prompt.py`](diffhunk://#diff-eabd53da2ec241cce6f92de85283065e7b478d62cefb52817c5b29ff07e28189L173-R174): Changed the order of condition checks for better readability. * [`python/semantic_kernel/kernel.py`](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10): Added functionality to update kernel arguments in `invoke_stream` method and made `function_name` and `plugin_name` optional in `invoke_prompt` method. [[1]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10) [[2]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR29) [[3]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR119-R120) [[4]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL195-R202) [[5]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL209-R216) Closes #6804 Closes #6944 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-28 22:17:05 +02:00
function_name (str): The name of the function, optional
plugin_name (str): The name of the plugin, optional
Python: add new samples and fix streaming tool call FunctionCallContent formation (#5877) ### Motivation and Context We're working towards creating a core set of syntax examples (there are four total). The core examples will be available in all 3 SK languages. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces two (of the four) new kernel syntax examples, which will align with the new kernel examples coming soon for both dotnet and Java. #5784 - Introduce a custom weather plugin that in conjunction with the core TimePlugin, make use of auto function calling. - Introduce a kernel syntax example that shows how to integrate with the Microsoft Graph API to create a "restaurant booking." Note: this doesn't actually place a real reservation, but shows how to interact with msgraph. - Also fixes an issue where the streaming tool call argument formation was broken. Closes #6106 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-05-02 18:15:07 -04:00
arguments (KernelArguments | None): The arguments to pass to the function(s), optional
template_format (str | None): The format of the prompt template
Python: new pre-commit actions and pre-commit as GHA (#6376) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> - Added static code checking to pre-commit (check-ast and nbqa-check-ast) - Added Bandit security checking to pre-commit - Added pre-commit step to python-lint workflow, if it works, can delete seperate mypy, ruff and black tests ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-28 17:28:04 +02:00
return_function_results (bool): If True, the function results are yielded as a list[FunctionResult]
Python: .Net: Updated encoding logic in prompt templates (#12983) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Resolves: https://github.com/microsoft/semantic-kernel/issues/11821 Today, the encoding of template arguments is performed only if argument type is `string`. In case of custom type, anonymous type or collection - the encoding is not performed. This PR contains changes to throw an exception in case if encoding is enabled but complex type is used. In case of complex type, the encoding should be performed manually according to business logic and automatic encoding should be explicitly disabled. This enforces stricter, but more secure template rendering rules. **Note**: this is a breaking change for customers who use Handlebars or Liquid template with complex type arguments. Code changes are required when initializing template arguments: ```diff var arguments = new KernelArguments() { { "customer", new { - firstName = userInput.FirstName, - lastName = userInput.LastName, + firstName = HttpUtility.HtmlEncode(userInput.FirstName), + lastName = HttpUtility.HtmlEncode(userInput.LastName), } } }; var templateFactory = new LiquidPromptTemplateFactory(); var promptTemplateConfig = new PromptTemplateConfig() { TemplateFormat = "liquid" + InputVariables = new() + { + // We set AllowDangerouslySetContent to 'true' because each property of this argument is encoded manually. + new() { Name = "customer", AllowDangerouslySetContent = true }, + } }; var promptTemplate = templateFactory.Create(promptTemplateConfig); var renderedPrompt = await promptTemplate.RenderAsync(kernel, arguments); ``` ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-08-26 22:09:00 -07:00
prompt_template_config (PromptTemplateConfig | None): The prompt template configuration
Python: add new samples and fix streaming tool call FunctionCallContent formation (#5877) ### Motivation and Context We're working towards creating a core set of syntax examples (there are four total). The core examples will be available in all 3 SK languages. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces two (of the four) new kernel syntax examples, which will align with the new kernel examples coming soon for both dotnet and Java. #5784 - Introduce a custom weather plugin that in conjunction with the core TimePlugin, make use of auto function calling. - Introduce a kernel syntax example that shows how to integrate with the Microsoft Graph API to create a "restaurant booking." Note: this doesn't actually place a real reservation, but shows how to interact with msgraph. - Also fixes an issue where the streaming tool call argument formation was broken. Closes #6106 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-05-02 18:15:07 -04:00
kwargs (dict[str, Any]): arguments that can be used instead of supplying KernelArguments
Returns:
AsyncIterable[StreamingContentMixin]: The content of the stream of the last function provided.
"""
Python: Fix KernelArgument bug in invoke_prompt functions (#8414) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> In `invoke_prompt(...)` and `invoke_prompt_async(...)`, the `arguments` parameter is tested to see if it's empty before proceeding. The type of the `arguments` parameter is `KernelArgument`, which derives from a `dict` type, with an additional field for execution settings. When `arguments` contains only the execution settings, it will be considered as an empty dictionary. This will result in the `arguments` parameter being overridden completely, invalidating all execution settings passed to the APIs. To reproduce (assuming we have a plugin/function that will return the weather): ``` result = await kernel.invoke_prompt( "What is the weather like in my location?", arguments=KernelArguments( settings={ service_id: PromptExecutionSettings( service_id=service_id, function_choice_behavior=FunctionChoiceBehavior.Auto(), ), } ), ) ``` In the example above, the settings will be overridden to None. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> 1. Fix the bug. 2. Fix mypy in the pre-commit hook in order to commit. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-29 08:40:53 -07:00
if arguments is None:
Python: add new samples and fix streaming tool call FunctionCallContent formation (#5877) ### Motivation and Context We're working towards creating a core set of syntax examples (there are four total). The core examples will be available in all 3 SK languages. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces two (of the four) new kernel syntax examples, which will align with the new kernel examples coming soon for both dotnet and Java. #5784 - Introduce a custom weather plugin that in conjunction with the core TimePlugin, make use of auto function calling. - Introduce a kernel syntax example that shows how to integrate with the Microsoft Graph API to create a "restaurant booking." Note: this doesn't actually place a real reservation, but shows how to interact with msgraph. - Also fixes an issue where the streaming tool call argument formation was broken. Closes #6106 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-05-02 18:15:07 -04:00
arguments = KernelArguments(**kwargs)
if not prompt:
raise TemplateSyntaxError("The prompt is either null or empty.")
from semantic_kernel.functions.kernel_function_from_prompt import KernelFunctionFromPrompt
function = KernelFunctionFromPrompt(
Python: slight improvements in kernel and function result content, sample of local function calling with Nexus Raven (#6982) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This introduces a sample that uses function calling with a local model, Nexus Raven V2, in this case using the API, but the same model can be loaded using the Hugging Face Connection, the class here is highly adapted and specific to Nexus including parsing the returned function call value, into FunctionCallContent and then running that. It then returns one or more ChatMessageContents with the function call content and function result content which can then be shared with a OpenAI call. Error handling improvements: * [`python/semantic_kernel/contents/chat_history.py`](diffhunk://#diff-b57a30f97fec1a91bbf6f7d15584473a8f0cf409797bdcd90d8c80bedc670892L266-R267): Improved error logging when parsing prompts as XML fails. * [`python/semantic_kernel/contents/function_result_content.py`](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10): Removed field validator for `result` and allow result to be any value, instead of casting everything to string. [[1]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10) [[2]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L50-R51) [[3]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L63-L69) [[4]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87R92-R115) Minor changes: * [`python/semantic_kernel/core_plugins/math_plugin.py`](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38): Removed some docstrings for clarity. [[1]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38) [[2]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L55-R47) * [`python/semantic_kernel/functions/kernel_function_from_prompt.py`](diffhunk://#diff-eabd53da2ec241cce6f92de85283065e7b478d62cefb52817c5b29ff07e28189L173-R174): Changed the order of condition checks for better readability. * [`python/semantic_kernel/kernel.py`](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10): Added functionality to update kernel arguments in `invoke_stream` method and made `function_name` and `plugin_name` optional in `invoke_prompt` method. [[1]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10) [[2]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR29) [[3]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR119-R120) [[4]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL195-R202) [[5]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL209-R216) Closes #6804 Closes #6944 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-28 22:17:05 +02:00
function_name=function_name or generate_random_ascii_name(),
Python: add new samples and fix streaming tool call FunctionCallContent formation (#5877) ### Motivation and Context We're working towards creating a core set of syntax examples (there are four total). The core examples will be available in all 3 SK languages. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces two (of the four) new kernel syntax examples, which will align with the new kernel examples coming soon for both dotnet and Java. #5784 - Introduce a custom weather plugin that in conjunction with the core TimePlugin, make use of auto function calling. - Introduce a kernel syntax example that shows how to integrate with the Microsoft Graph API to create a "restaurant booking." Note: this doesn't actually place a real reservation, but shows how to interact with msgraph. - Also fixes an issue where the streaming tool call argument formation was broken. Closes #6106 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-05-02 18:15:07 -04:00
plugin_name=plugin_name,
prompt=prompt,
template_format=template_format,
Python: .Net: Updated encoding logic in prompt templates (#12983) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Resolves: https://github.com/microsoft/semantic-kernel/issues/11821 Today, the encoding of template arguments is performed only if argument type is `string`. In case of custom type, anonymous type or collection - the encoding is not performed. This PR contains changes to throw an exception in case if encoding is enabled but complex type is used. In case of complex type, the encoding should be performed manually according to business logic and automatic encoding should be explicitly disabled. This enforces stricter, but more secure template rendering rules. **Note**: this is a breaking change for customers who use Handlebars or Liquid template with complex type arguments. Code changes are required when initializing template arguments: ```diff var arguments = new KernelArguments() { { "customer", new { - firstName = userInput.FirstName, - lastName = userInput.LastName, + firstName = HttpUtility.HtmlEncode(userInput.FirstName), + lastName = HttpUtility.HtmlEncode(userInput.LastName), } } }; var templateFactory = new LiquidPromptTemplateFactory(); var promptTemplateConfig = new PromptTemplateConfig() { TemplateFormat = "liquid" + InputVariables = new() + { + // We set AllowDangerouslySetContent to 'true' because each property of this argument is encoded manually. + new() { Name = "customer", AllowDangerouslySetContent = true }, + } }; var promptTemplate = templateFactory.Create(promptTemplateConfig); var renderedPrompt = await promptTemplate.RenderAsync(kernel, arguments); ``` ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-08-26 22:09:00 -07:00
prompt_template_config=prompt_template_config,
Python: add new samples and fix streaming tool call FunctionCallContent formation (#5877) ### Motivation and Context We're working towards creating a core set of syntax examples (there are four total). The core examples will be available in all 3 SK languages. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces two (of the four) new kernel syntax examples, which will align with the new kernel examples coming soon for both dotnet and Java. #5784 - Introduce a custom weather plugin that in conjunction with the core TimePlugin, make use of auto function calling. - Introduce a kernel syntax example that shows how to integrate with the Microsoft Graph API to create a "restaurant booking." Note: this doesn't actually place a real reservation, but shows how to interact with msgraph. - Also fixes an issue where the streaming tool call argument formation was broken. Closes #6106 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-05-02 18:15:07 -04:00
)
function_result: list[list["StreamingContentMixin"] | Any] = []
async for stream_message in self.invoke_stream(function=function, arguments=arguments):
if isinstance(stream_message, FunctionResult) and (
Python: Unsafe input handling (#6003) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Implements dealing with unsafe content, by doing HTML parsing on variables and function results. Closes: #5889 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds parameter `allow_dangerously_set_content` to: - InputVariable - PromptTemplateConfig - PromptTemplateBase The behavior is that if the flag is set to True on the template itself (KernelPromptTemplate, Jinja2PromptTemplate or HandlebarsPromptTemplate) the behavior is the same, no encoding is done on inputs. Otherwise: - variables are encoded by default, this can be switched off using the InputVariables class for that variable. - function output is encoded by default, this can be switched off using the flag in the PromptTemplateConfig, this is not yet possible to do on a per function basis. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-05-16 15:02:54 +02:00
exception := stream_message.metadata.get(METADATA_EXCEPTION_KEY, None)
Python: add new samples and fix streaming tool call FunctionCallContent formation (#5877) ### Motivation and Context We're working towards creating a core set of syntax examples (there are four total). The core examples will be available in all 3 SK languages. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces two (of the four) new kernel syntax examples, which will align with the new kernel examples coming soon for both dotnet and Java. #5784 - Introduce a custom weather plugin that in conjunction with the core TimePlugin, make use of auto function calling. - Introduce a kernel syntax example that shows how to integrate with the Microsoft Graph API to create a "restaurant booking." Note: this doesn't actually place a real reservation, but shows how to interact with msgraph. - Also fixes an issue where the streaming tool call argument formation was broken. Closes #6106 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-05-02 18:15:07 -04:00
):
raise KernelInvokeException(
f"Error occurred while invoking function: '{function.fully_qualified_name}'"
) from exception
function_result.append(stream_message)
yield stream_message
if return_function_results:
output_function_result: list["StreamingContentMixin"] = []
for result in function_result:
for choice in result:
if not isinstance(choice, StreamingContentMixin):
continue
if len(output_function_result) <= choice.choice_index:
output_function_result.append(copy(choice))
else:
output_function_result[choice.choice_index] += choice
yield FunctionResult(function=function.metadata, value=output_function_result)
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
async def invoke_function_call(
self,
function_call: FunctionCallContent,
chat_history: ChatHistory,
Python: filter improvements (#10588) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Fixes some inconsistency in the filter setup between dotnet and python Adds a sample for how to use a filter to retry a prompt with a different model. Adds a sample for semantic caching Closes: #10572 #5924 #10595 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2025-02-19 01:40:51 +01:00
*,
Python: slight improvements in kernel and function result content, sample of local function calling with Nexus Raven (#6982) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This introduces a sample that uses function calling with a local model, Nexus Raven V2, in this case using the API, but the same model can be loaded using the Hugging Face Connection, the class here is highly adapted and specific to Nexus including parsing the returned function call value, into FunctionCallContent and then running that. It then returns one or more ChatMessageContents with the function call content and function result content which can then be shared with a OpenAI call. Error handling improvements: * [`python/semantic_kernel/contents/chat_history.py`](diffhunk://#diff-b57a30f97fec1a91bbf6f7d15584473a8f0cf409797bdcd90d8c80bedc670892L266-R267): Improved error logging when parsing prompts as XML fails. * [`python/semantic_kernel/contents/function_result_content.py`](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10): Removed field validator for `result` and allow result to be any value, instead of casting everything to string. [[1]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10) [[2]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L50-R51) [[3]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L63-L69) [[4]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87R92-R115) Minor changes: * [`python/semantic_kernel/core_plugins/math_plugin.py`](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38): Removed some docstrings for clarity. [[1]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38) [[2]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L55-R47) * [`python/semantic_kernel/functions/kernel_function_from_prompt.py`](diffhunk://#diff-eabd53da2ec241cce6f92de85283065e7b478d62cefb52817c5b29ff07e28189L173-R174): Changed the order of condition checks for better readability. * [`python/semantic_kernel/kernel.py`](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10): Added functionality to update kernel arguments in `invoke_stream` method and made `function_name` and `plugin_name` optional in `invoke_prompt` method. [[1]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10) [[2]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR29) [[3]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR119-R120) [[4]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL195-R202) [[5]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL209-R216) Closes #6804 Closes #6944 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-28 22:17:05 +02:00
arguments: "KernelArguments | None" = None,
Python: filter improvements (#10588) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Fixes some inconsistency in the filter setup between dotnet and python Adds a sample for how to use a filter to retry a prompt with a different model. Adds a sample for semantic caching Closes: #10572 #5924 #10595 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2025-02-19 01:40:51 +01:00
execution_settings: "PromptExecutionSettings | None" = None,
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
function_call_count: int | None = None,
request_index: int | None = None,
Python: filter improvements (#10588) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Fixes some inconsistency in the filter setup between dotnet and python Adds a sample for how to use a filter to retry a prompt with a different model. Adds a sample for semantic caching Closes: #10572 #5924 #10595 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2025-02-19 01:40:51 +01:00
is_streaming: bool = False,
Python: Introducing Realtime Clients for OpenAI and Azure OpenAI (#10127) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Implements the OpenAI Realtime API with Semantic Kernel ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Adds a RealtimeClientBase Adds RealtimeEvents Adds OpenAI and Azure Realtime clients Closes: #10108 ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2025-03-04 08:44:10 +01:00
function_behavior: "FunctionChoiceBehavior | None" = None,
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
) -> "AutoFunctionInvocationContext | None":
"""Processes the provided FunctionCallContent and updates the chat history."""
try:
if function_call.name is None:
raise FunctionExecutionException("The function name is required.")
if function_behavior is not None and function_behavior.filters:
allowed_functions = [
func.fully_qualified_name for func in self.get_list_of_function_metadata(function_behavior.filters)
]
if function_call.name not in allowed_functions:
raise FunctionExecutionException(
f"Only functions: {allowed_functions} are allowed, {function_call.name} is not allowed."
)
function_to_call = self.get_function(function_call.plugin_name, function_call.function_name)
except Exception as exc:
logger.exception(f"The function `{function_call.name}` is not part of the provided functions: {exc}.")
frc = FunctionResultContent.from_function_call_content_and_result(
function_call_content=function_call,
result=(
f"The tool call with name `{function_call.name}` is not part of the provided tools, "
"please try again with a supplied tool call name and make sure to validate the name."
),
)
chat_history.add_message(message=frc.to_chat_message_content())
return None
Python: Validate for missing or unexpected params from model (#12071) ### Motivation and Context I've experienced times where even though we're sending the required params to the model, for example `input_text`, the model is sending back the tool call argument as `text`. Telling the model that we received an unexpected param and/or are missing one helps it re-send the correct one for function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description Improve the function calling path. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2025-05-19 14:13:27 +09:00
args_cloned = copy(arguments) if arguments else KernelArguments()
try:
parsed_args = function_call.to_kernel_arguments()
# Check for missing or unexpected parameters
required_param_names = {
param.name for param in function_to_call.parameters if param.name is not None and param.is_required
}
received_param_names = set(parsed_args or {})
missing_params = required_param_names - received_param_names
unexpected_params = received_param_names - {param.name for param in function_to_call.parameters}
if missing_params or unexpected_params:
msg_parts = []
if missing_params:
msg_parts.append(f"Missing required argument(s): {sorted(missing_params)}.")
if unexpected_params:
msg_parts.append(f"Received unexpected argument(s): {sorted(unexpected_params)}.")
msg = " ".join(msg_parts) + " Please revise the arguments to match the function signature."
logger.info(msg)
frc = FunctionResultContent.from_function_call_content_and_result(
function_call_content=function_call,
result=msg,
)
chat_history.add_message(message=frc.to_chat_message_content())
return None
if parsed_args:
args_cloned.update(parsed_args)
except (FunctionCallInvalidArgumentsException, TypeError) as exc:
logger.info(f"Received invalid arguments for function {function_call.name}: {exc}. Trying tool call again.")
frc = FunctionResultContent.from_function_call_content_and_result(
function_call_content=function_call,
result="The tool call arguments are malformed. Arguments must be in JSON format. Please try again.",
)
chat_history.add_message(message=frc.to_chat_message_content())
return None
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
num_required_func_params = len([param for param in function_to_call.parameters if param.is_required])
if parsed_args is None or len(parsed_args) < num_required_func_params:
msg = (
f"There are `{num_required_func_params}` tool call arguments required and "
f"only `{len(parsed_args) if parsed_args is not None else 0}` received. The required arguments are: "
f"{[param.name for param in function_to_call.parameters if param.is_required]}. "
"Please provide the required arguments and try again."
)
logger.info(msg)
frc = FunctionResultContent.from_function_call_content_and_result(
function_call_content=function_call,
result=msg,
)
chat_history.add_message(message=frc.to_chat_message_content())
return None
logger.info(f"Calling {function_call.name} function with args: {function_call.arguments}")
_rebuild_auto_function_invocation_context()
invocation_context = AutoFunctionInvocationContext(
function=function_to_call,
kernel=self,
arguments=args_cloned,
Python: filter improvements (#10588) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Fixes some inconsistency in the filter setup between dotnet and python Adds a sample for how to use a filter to retry a prompt with a different model. Adds a sample for semantic caching Closes: #10572 #5924 #10595 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2025-02-19 01:40:51 +01:00
is_streaming=is_streaming,
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
chat_history=chat_history,
Python: Update kernel function span (#12285) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Closing #12262 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> This PR updates tracing support for function execution across both .NET and Python according to the latest OTel GenAI semantic conventions: https://opentelemetry.io/docs/specs/semconv/gen-ai/gen-ai-spans/#execute-tool-span ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-05-27 16:41:44 -07:00
function_call_content=function_call,
Python: filter improvements (#10588) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Fixes some inconsistency in the filter setup between dotnet and python Adds a sample for how to use a filter to retry a prompt with a different model. Adds a sample for semantic caching Closes: #10572 #5924 #10595 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com>
2025-02-19 01:40:51 +01:00
execution_settings=execution_settings,
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
function_result=FunctionResult(function=function_to_call.metadata, value=None),
Python: slight improvements in kernel and function result content, sample of local function calling with Nexus Raven (#6982) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This introduces a sample that uses function calling with a local model, Nexus Raven V2, in this case using the API, but the same model can be loaded using the Hugging Face Connection, the class here is highly adapted and specific to Nexus including parsing the returned function call value, into FunctionCallContent and then running that. It then returns one or more ChatMessageContents with the function call content and function result content which can then be shared with a OpenAI call. Error handling improvements: * [`python/semantic_kernel/contents/chat_history.py`](diffhunk://#diff-b57a30f97fec1a91bbf6f7d15584473a8f0cf409797bdcd90d8c80bedc670892L266-R267): Improved error logging when parsing prompts as XML fails. * [`python/semantic_kernel/contents/function_result_content.py`](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10): Removed field validator for `result` and allow result to be any value, instead of casting everything to string. [[1]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L7-R10) [[2]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L50-R51) [[3]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87L63-L69) [[4]](diffhunk://#diff-3b69e6aedf4c29038cdff0ac99602b70583ab6c85f7bd777b08ad96b7fa48d87R92-R115) Minor changes: * [`python/semantic_kernel/core_plugins/math_plugin.py`](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38): Removed some docstrings for clarity. [[1]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L32-R38) [[2]](diffhunk://#diff-632444474bd3ea991b4409dfd0ab194e0789c73617bef4a3dcae4db5c4b0b6c0L55-R47) * [`python/semantic_kernel/functions/kernel_function_from_prompt.py`](diffhunk://#diff-eabd53da2ec241cce6f92de85283065e7b478d62cefb52817c5b29ff07e28189L173-R174): Changed the order of condition checks for better readability. * [`python/semantic_kernel/kernel.py`](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10): Added functionality to update kernel arguments in `invoke_stream` method and made `function_name` and `plugin_name` optional in `invoke_prompt` method. [[1]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR9-R10) [[2]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR29) [[3]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aR119-R120) [[4]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL195-R202) [[5]](diffhunk://#diff-3db6fe98818eda0b824788901c718d35eac8b13d0ff428d3dfd417866d922d3aL209-R216) Closes #6804 Closes #6944 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-28 22:17:05 +02:00
function_count=function_call_count or 0,
request_sequence_index=request_index or 0,
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
)
if function_call.index is not None:
invocation_context.function_sequence_index = function_call.index
stack = self.construct_call_stack(
filter_type=FilterTypes.AUTO_FUNCTION_INVOCATION,
inner_function=self._inner_auto_function_invoke_handler,
)
await stack(invocation_context)
Python: Deep-copy `function_result.value` in `invoke_function_call` to prevent state mutation (#11826) ### Motivation and Context The existing implementation of `kernel.invoke_function_call(...)` stores mutable tool return values by reference. Subsequent mutations to the same in-memory state bleed into earlier chat messages, which corrupts conversation history and breaks reproducibility. By deep-copying the function result immediately after execution, we can guarantee an immutable snapshot of each tool's output at the time of invocation, which preserves previous tool invocations and removes unwanted side effects. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description After `await stack(invocation_context)` in `kernel.invoke_function_call(...)`, we now have: ```python if invocation_context.function_result and invocation_context.function_result.value is not None: invocation_context.function_result.value = copy.deepcopy( invocation_context.function_result.value ) ``` This has no change to our public API or user-facing behavior; internal state handling is now robust against undesired mutations. - Adds a unit test to exercise the new behavior and assert on the proper mutations/states. <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2025-05-01 09:19:34 +09:00
# Snapshot the tool's return value so later mutations don't leak back
if invocation_context.function_result and invocation_context.function_result.value is not None:
invocation_context.function_result.value = deepcopy(invocation_context.function_result.value)
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
frc = FunctionResultContent.from_function_call_content_and_result(
Python: Introduce the OpenAI Responses Agent (#11240) ### Motivation and Context OpenAI recently released their Responses API, which is their newest core API and an agentic API primitive, combining the simplicity of Chat Completions with the ability to do more agentic tasks. Azure OpenAI also recently released the Responses API, with close-to feature parity with OpenAI's SDK. It gives us enough to introduce an `AzureResponsesAgent` and an `OpenAIResponsesAgent` to Semantic Kernel. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR introduces: - The `AzureResponsesAgent` and `OpenAIResponsesAgent` - Allows use of the web search tool with `OpenAI` (`Azure OpenAI` currently doesn't support this). - Allows use of the file search tool with both `OpenAI` and `Azure OpenAI`. - Provides getting started samples for both agent types. **TODO:** - Add unit test coverage for the `ResponsesAgentThreadActions` class. - Add support for the computer user agent tool. - Improve typing in the `ResponsesAgentThreadActions` class. - Closes #11026 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2025-04-03 08:18:57 +09:00
function_call_content=function_call,
result=invocation_context.function_result,
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
)
Python: Yield FunctionResultContent in streaming chat completion path. Update tests. (#9974) ### Motivation and Context Currently, if using SK's Python streaming chat completion path, we only yield the following content types: `StreamingChatMessageContent` and `FunctionCallContent`. We are not yielding `FunctionResultContent` which is valuable for some use cases. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR updates the code to yield `FunctionResultContent` if it exists in the streaming chat completion path. When we merge the function call result content together into a `StreamingChatMessageContent` type, we check if that message has items (which are of type `FunctionResultContent`) and if so, we yield them. The filter path still works because once they're yielded, we break out of the function calling loop. We need to include the `ai_model_id` if it exists for the current PromptExecutionSettings because if performing a `reduce` operation to add two streaming chat message chunks together, the `StreamingChatMessageContent` that has the function results will break if the `ai_model_id` is not set (the error is thrown in the `__add__` override for the `StreamingChatMessageContent`. Some unit tests that cover function calling were also updated -- during the test, the test JSON function args were breaking in the `json.loads` call because they contained single quotes and not double. We're now sanitizing the args, just in case, so we don't break there. This PR fixes: - #9408 - #9968 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-12-18 10:09:25 +09:00
is_streaming = any(isinstance(message, StreamingChatMessageContent) for message in chat_history.messages)
message = frc.to_streaming_chat_message_content() if is_streaming else frc.to_chat_message_content()
chat_history.add_message(message=message)
Python: Update auto invoke function filter handling for proper functionality for OpenAI models. (#8071) ### Motivation and Context Our auto function invoke filter does not have the correct behavior. Currently, it will return the straight chat completion results which contain the function call content, instead of the function result content. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description This PR fixes this by adding the function result content to the chat history, and if the auto function invoke filter has the context as terminated then it will return the last item of the chat history (function result content) otherwise it will be the completions (in the event that the function isn't run for some reason). - It also fixes the `auto_function_invoke_filters` sample to show the FunctionResultContent if it exists. - Aligns to the same behavior as in dotnet. - Fixes #8020 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2024-08-12 11:37:08 -04:00
return invocation_context if invocation_context.terminate else None
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
async def _inner_auto_function_invoke_handler(self, context: AutoFunctionInvocationContext):
"""Inner auto function invocation handler."""
try:
Python: Update kernel function span (#12285) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Closing #12262 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> This PR updates tracing support for function execution across both .NET and Python according to the latest OTel GenAI semantic conventions: https://opentelemetry.io/docs/specs/semconv/gen-ai/gen-ai-spans/#execute-tool-span ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone :smile:
2025-05-27 16:41:44 -07:00
result = await context.function.invoke(
context.kernel,
context.arguments,
metadata=context.function_call_content.metadata | context.function_call_content.to_dict()
if context.function_call_content
else {},
)
Python: Introduce the new function calling abstraction, FunctionChoiceBehavior (#6910) ### Motivation and Context The current `FunctionCallBehavior` has allowed us to utilize auto function calling for OpenAI type models. As we proceed to support more AI connectors, that differ from OpenAI models, we need to be able to handle functions for all models that support function calling. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description In this PR: - We introduce a new function calling abstraction called `FunctionChoiceBehavior` which has three types: Auto, Required, and None. - One is able to configure the `function_choice_behavior` along with fully qualified function names (e.g., plugin1.function1), `maximum_auto_invoke_attempts` or `auto_invoke_kernel_functions` in yaml and JSON prompts. We have a new concept example showing how to do this for for yaml and JSON prompts. - If the fully qualified names are specified in the config file, they take precedence over the filters, if specified at a later point. - To make sure this isn't a breaking change, we still handle the previous `FunctionCallBehavior`, but so we can make decisions on the new `FunctionChoiceBehavior` we map the `FunctionCallBehavior` to `FunctionChoiceBehavior`. Each time `FunctionCallBehavior` is updated, `FunctionChoiceBehavior` will be updated, too. - The `_process_tool_call()` method in the `open_ai_chat_completion_base` needs to maintain the argument name as `function_call_behavior` as we cannot introduce a breaking change. - All concept samples have been converted to use `FunctionChoiceBehavior`. - New unit tests have been added for `FunctionChoiceBehavior` and we still currently support `FunctionCallBehavior` tests to make sure we haven't broken the backwards compatibility. - Added `deprecated` typing decorators to classes/methods to alert users that it would be best to transition to `FunctionChoiceBehavior` even though `FunctionCallBehavior` is still supported. - The `FunctionCallingStepwisePlanner` was updated to use the new `FunctionChoiceBehavior`. - Closes #6496, #6458 - Addresses #6626 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-06-27 16:34:59 -04:00
if result:
context.function_result = result
except Exception as exc:
logger.exception(f"Error invoking function {context.function.fully_qualified_name}: {exc}.")
value = f"An error occurred while invoking the function {context.function.fully_qualified_name}: {exc}"
if context.function_result is not None:
context.function_result.value = value
else:
context.function_result = FunctionResult(function=context.function.metadata, value=value)
return
Python: Feature new memory stores and collections (#7614) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds new vector store and vector store record collections as well as implementations for: - Azure AI Search - Redis - Qdrant - Volatile (in-memory) It also adds samples, tests, and unit-tests for these. Next it adds the vector store record fields, definition and decorator, with tests and samples. All marked experimental. Existing Redis, Azure AI Search, Qdrant and Volatile will be marked as deprecated in the future, once the new collections are feature complete. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2024-08-06 22:30:26 +02:00
async def add_embedding_to_object(
self,
inputs: OneOrMany[TDataModel],
field_to_embed: str,
field_to_store: str,
execution_settings: dict[str, "PromptExecutionSettings"],
container_mode: bool = False,
cast_function: Callable[[list[float]], Any] | None = None,
**kwargs: Any,
):
"""Gather all fields to embed, batch the embedding generation and store."""
contents: list[Any] = []
dict_like = (getter := getattr(inputs, "get", False)) and callable(getter)
list_of_dicts: bool = False
if container_mode:
contents = inputs[field_to_embed].tolist() # type: ignore
elif isinstance(inputs, list):
list_of_dicts = (getter := getattr(inputs[0], "get", False)) and callable(getter)
for record in inputs:
if list_of_dicts:
contents.append(record.get(field_to_embed)) # type: ignore
else:
contents.append(getattr(record, field_to_embed))
else:
if dict_like:
contents.append(inputs.get(field_to_embed)) # type: ignore
else:
contents.append(getattr(inputs, field_to_embed))
vectors = None
service: EmbeddingGeneratorBase | None = None
for service_id, settings in execution_settings.items():
service = self.get_service(service_id, type=EmbeddingGeneratorBase) # type: ignore
if service:
vectors = await service.generate_raw_embeddings(texts=contents, settings=settings, **kwargs) # type: ignore
break
if not service:
raise KernelServiceNotFoundError("No service found to generate embeddings.")
if vectors is None:
raise KernelInvokeException("No vectors were generated.")
if cast_function:
vectors = [cast_function(vector) for vector in vectors]
if container_mode:
inputs[field_to_store] = vectors # type: ignore
return
if isinstance(inputs, list):
for record, vector in zip(inputs, vectors):
if list_of_dicts:
record[field_to_store] = vector # type: ignore
else:
setattr(record, field_to_store, vector)
return
if dict_like:
inputs[field_to_store] = vectors[0] # type: ignore
return
setattr(inputs, field_to_store, vectors[0])
Python: Add ability to clone a kernel (#11680) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Currently, we don't have a dedicated method to create a clone of a kernel instance. Using `model_copy(deep=True)` will result in an error because some service clients aren't serializable. Being able to clone a kernel will be critical in many scenarios where developers don't want to mutate the original one that is attached to an agent or some other objects. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> 1. Create a clone method in the `Kernel` class that will return a new instance of the kernel with the same configuration, including the list of filters, plugins and the AI selector. It will also include the same list of services references. 2. Add unit tests and integration tests to ensure the new method works as expected. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-23 17:26:51 -07:00
def clone(self) -> "Kernel":
"""Clone the kernel instance to create a new one that may be mutated without affecting the current instance.
The current instance is not mutated by this operation.
Note: The same service clients are used in the new instance, so if you mutate the service clients
in the new instance, the original instance will be affected as well.
New lists of plugins and filters are created. It will not affect the original lists when the new instance
is mutated. A new `ai_service_selector` is created. It will not affect the original instance when the new
instance is mutated.
Python: Safely clone plugins (#13109) ### Motivation and Context Currently, we attempt to deep copy plugins, no matter what the plugin instance holds. This will cause issues when plugins contain unpickleable items like an async generator. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description - Closes #12819 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2025-09-12 09:25:07 +09:00
Important: Plugins are cloned without deep-copying their underlying callable methods. This avoids attempting
to pickle/clone unpickleable objects (e.g., async generators), which can be present when plugins wrap async
context managers such as MCP client sessions. Function metadata is deep-copied while callables are shared.
Python: Add ability to clone a kernel (#11680) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Currently, we don't have a dedicated method to create a clone of a kernel instance. Using `model_copy(deep=True)` will result in an error because some service clients aren't serializable. Being able to clone a kernel will be critical in many scenarios where developers don't want to mutate the original one that is attached to an agent or some other objects. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> 1. Create a clone method in the `Kernel` class that will return a new instance of the kernel with the same configuration, including the list of filters, plugins and the AI selector. It will also include the same list of services references. 2. Add unit tests and integration tests to ensure the new method works as expected. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-23 17:26:51 -07:00
"""
Python: Safely clone plugins (#13109) ### Motivation and Context Currently, we attempt to deep copy plugins, no matter what the plugin instance holds. This will cause issues when plugins contain unpickleable items like an async generator. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description - Closes #12819 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2025-09-12 09:25:07 +09:00
# Safely clone plugins by copying function metadata while retaining callable references.
# This avoids deepcopying bound methods that may reference unpickleable async components.
new_plugins: dict[str, KernelPlugin] = {}
for plugin_name, plugin in self.plugins.items():
cloned_plugin = KernelPlugin(name=plugin.name, description=plugin.description)
# Using KernelPlugin.add will copy functions via KernelFunction.function_copy(),
# which deep-copies metadata but keeps callables shallow.
cloned_plugin.add(plugin.functions)
new_plugins[plugin_name] = cloned_plugin
Python: Add ability to clone a kernel (#11680) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Currently, we don't have a dedicated method to create a clone of a kernel instance. Using `model_copy(deep=True)` will result in an error because some service clients aren't serializable. Being able to clone a kernel will be critical in many scenarios where developers don't want to mutate the original one that is attached to an agent or some other objects. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> 1. Create a clone method in the `Kernel` class that will return a new instance of the kernel with the same configuration, including the list of filters, plugins and the AI selector. It will also include the same list of services references. 2. Add unit tests and integration tests to ensure the new method works as expected. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-23 17:26:51 -07:00
return Kernel(
Python: Safely clone plugins (#13109) ### Motivation and Context Currently, we attempt to deep copy plugins, no matter what the plugin instance holds. This will cause issues when plugins contain unpickleable items like an async generator. <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> ### Description - Closes #12819 <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [X] The code builds clean without any errors or warnings - [X] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [X] All unit tests pass, and I have added new tests where possible - [X] I didn't break anyone :smile:
2025-09-12 09:25:07 +09:00
plugins=new_plugins,
Python: Add ability to clone a kernel (#11680) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Currently, we don't have a dedicated method to create a clone of a kernel instance. Using `model_copy(deep=True)` will result in an error because some service clients aren't serializable. Being able to clone a kernel will be critical in many scenarios where developers don't want to mutate the original one that is attached to an agent or some other objects. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> 1. Create a clone method in the `Kernel` class that will return a new instance of the kernel with the same configuration, including the list of filters, plugins and the AI selector. It will also include the same list of services references. 2. Add unit tests and integration tests to ensure the new method works as expected. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-23 17:26:51 -07:00
# Shallow copy of the services, as they are not serializable
services={k: v for k, v in self.services.items()},
ai_service_selector=deepcopy(self.ai_service_selector),
function_invocation_filters=deepcopy(self.function_invocation_filters),
prompt_rendering_filters=deepcopy(self.prompt_rendering_filters),
auto_function_invocation_filters=deepcopy(self.auto_function_invocation_filters),
)
Python: Introducing SK as a MCP Server (#11362) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds the ability to stand up your Kernel as a MCP Server, it returns the Server object, which you can then run as either a Stdio or Sse server. Includes a sample to demonstrate. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Nico Möller <nkm.moeller@gmail.com>
2025-04-11 07:22:07 +02:00
@experimental
def as_mcp_server(
self,
Python: add support for prompts and sample (#11508) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Allow a list of prompts to be added and exposed with the MCP server setup. This uses the built-in rendering types (all of them) to be used and exposed as prompts. Closes: #11143 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-15 10:25:35 +02:00
prompts: list[PromptTemplateBase] | None = None,
Python: Introducing SK as a MCP Server (#11362) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds the ability to stand up your Kernel as a MCP Server, it returns the Server object, which you can then run as either a Stdio or Sse server. Includes a sample to demonstrate. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Nico Möller <nkm.moeller@gmail.com>
2025-04-11 07:22:07 +02:00
server_name: str = "Semantic Kernel MCP Server",
version: str | None = None,
instructions: str | None = None,
lifespan: Callable[["Server[LifespanResultT]"], AbstractAsyncContextManager["LifespanResultT"]] | None = None,
excluded_functions: OptionalOneOrMany[str] = None,
**kwargs: Any,
) -> "Server":
"""Create a MCP server from this kernel.
This function automatically creates a MCP server from a kernel instance, it uses the provided arguments to
configure the server and expose functions as tools and prompts, see the mcp documentation for more details.
Python: add support for prompts and sample (#11508) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Allow a list of prompts to be added and exposed with the MCP server setup. This uses the built-in rendering types (all of them) to be used and exposed as prompts. Closes: #11143 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-15 10:25:35 +02:00
By default, all functions are exposed as Tools, you can control this by
using use the `excluded_functions` argument.
These need to be set to the function name, without the plugin_name.
Python: Introducing SK as a MCP Server (#11362) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds the ability to stand up your Kernel as a MCP Server, it returns the Server object, which you can then run as either a Stdio or Sse server. Includes a sample to demonstrate. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Nico Möller <nkm.moeller@gmail.com>
2025-04-11 07:22:07 +02:00
Args:
kernel: The kernel instance to use.
Python: add support for prompts and sample (#11508) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Allow a list of prompts to be added and exposed with the MCP server setup. This uses the built-in rendering types (all of them) to be used and exposed as prompts. Closes: #11143 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-15 10:25:35 +02:00
prompts: A list of prompt templates to expose as prompts.
Python: Introducing SK as a MCP Server (#11362) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds the ability to stand up your Kernel as a MCP Server, it returns the Server object, which you can then run as either a Stdio or Sse server. Includes a sample to demonstrate. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Nico Möller <nkm.moeller@gmail.com>
2025-04-11 07:22:07 +02:00
server_name: The name of the server.
version: The version of the server.
instructions: The instructions to use for the server.
lifespan: The lifespan of the server.
Python: add support for prompts and sample (#11508) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Allow a list of prompts to be added and exposed with the MCP server setup. This uses the built-in rendering types (all of them) to be used and exposed as prompts. Closes: #11143 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-15 10:25:35 +02:00
excluded_functions: The list of function names to exclude from the server.
Python: Introducing SK as a MCP Server (#11362) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds the ability to stand up your Kernel as a MCP Server, it returns the Server object, which you can then run as either a Stdio or Sse server. Includes a sample to demonstrate. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Nico Möller <nkm.moeller@gmail.com>
2025-04-11 07:22:07 +02:00
if None, no functions will be excluded.
kwargs: Any extra arguments to pass to the server creation.
Returns:
The MCP server instance, it is a instance of
mcp.server.lowlevel.Server
"""
from semantic_kernel.connectors.mcp import create_mcp_server_from_kernel
return create_mcp_server_from_kernel(
Python: add support for prompts and sample (#11508) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> Allow a list of prompts to be added and exposed with the MCP server setup. This uses the built-in rendering types (all of them) to be used and exposed as prompts. Closes: #11143 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile:
2025-04-15 10:25:35 +02:00
kernel=self,
prompts=prompts,
Python: Introducing SK as a MCP Server (#11362) ### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR adds the ability to stand up your Kernel as a MCP Server, it returns the Server object, which you can then run as either a Stdio or Sse server. Includes a sample to demonstrate. ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [x] I didn't break anyone :smile: --------- Co-authored-by: Nico Möller <nkm.moeller@gmail.com>
2025-04-11 07:22:07 +02:00
server_name=server_name,
version=version,
instructions=instructions,
lifespan=lifespan,
excluded_functions=excluded_functions,
**kwargs,
)
Reference in New Issue Copy Permalink