2023-11-13 20:11:29 +01:00
# Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
# SPDX-License-Identifier: Apache-2.0
"""
Purpose
Shows how to use the AWS SDK for Python (Boto3) with Amazon Bedrock to manage
Bedrock models.
"""
import logging
import boto3
from botocore . exceptions import ClientError
logger = logging . getLogger ( __name__ )
# snippet-start:[python.example_code.bedrock.BedrockWrapper.class]
# snippet-start:[python.example_code.bedrock.BedrockWrapper.decl]
class BedrockWrapper :
""" Encapsulates Amazon Bedrock foundation model actions. """
def __init__ ( self , bedrock_client ) :
"""
:param bedrock_client: A Boto3 Amazon Bedrock client, which is a low-level client that
represents Amazon Bedrock and describes the API operations for
creating and managing Bedrock models.
"""
self . bedrock_client = bedrock_client
# snippet-end:[python.example_code.bedrock.BedrockWrapper.decl]
2024-02-21 14:12:39 +01:00
# snippet-start:[python.example_code.bedrock.GetFoundationModel]
def get_foundation_model ( self , model_identifier ) :
"""
Get details about an Amazon Bedrock foundation model.
:return: The foundation model ' s details.
"""
try :
return self . bedrock_client . get_foundation_model (
modelIdentifier = model_identifier
2024-03-05 00:59:53 +01:00
) [ " modelDetails " ]
2024-02-21 14:12:39 +01:00
except ClientError :
2024-03-05 00:59:53 +01:00
logger . error (
f " Couldn ' t get foundation models details for { model_identifier } "
)
2024-02-21 14:12:39 +01:00
raise
# snippet-end:[python.example_code.bedrock.GetFoundationModel]
2023-11-13 20:11:29 +01:00
# snippet-start:[python.example_code.bedrock.ListFoundationModels]
def list_foundation_models ( self ) :
"""
List the available Amazon Bedrock foundation models.
:return: The list of available bedrock foundation models.
"""
try :
response = self . bedrock_client . list_foundation_models ( )
models = response [ " modelSummaries " ]
logger . info ( " Got %s foundation models. " , len ( models ) )
2024-02-21 14:12:39 +01:00
return models
2023-11-13 20:11:29 +01:00
except ClientError :
logger . error ( " Couldn ' t list foundation models. " )
raise
# snippet-end:[python.example_code.bedrock.ListFoundationModels]
# snippet-end:[python.example_code.bedrock.BedrockWrapper.class]
def usage_demo ( ) :
"""
Shows how to list the available foundation models.
This demonstration gets the list of available foundation models and
prints their respective summaries.
"""
logging . basicConfig ( level = logging . INFO )
print ( " - " * 88 )
print ( " Welcome to the Amazon Bedrock demo. " )
print ( " - " * 88 )
bedrock_client = boto3 . client ( service_name = " bedrock " , region_name = " us-east-1 " )
wrapper = BedrockWrapper ( bedrock_client )
2024-02-21 14:12:39 +01:00
print ( " Listing the available foundation models. " )
2023-11-13 20:11:29 +01:00
try :
for model in wrapper . list_foundation_models ( ) :
2024-02-21 14:12:39 +01:00
print_model_details ( model )
except ClientError :
logger . exception ( " Couldn ' t list foundation models. " )
raise
print ( " Getting the details of an individual foundation model. " )
model_id = " amazon.titan-embed-text-v1 "
try :
print_model_details ( wrapper . get_foundation_model ( model_id ) )
2023-11-13 20:11:29 +01:00
except ClientError :
2024-02-21 14:12:39 +01:00
logger . exception ( f " Couldn ' t get foundation model { model_id } . " )
2023-11-13 20:11:29 +01:00
raise
2024-02-21 14:12:39 +01:00
def print_model_details ( model ) :
print ( " \n " + " = " * 42 )
print ( f ' Model: { model [ " modelId " ] } ' )
print ( " - " * 42 )
print ( f ' Name: { model [ " modelName " ] } ' )
print ( f ' Provider: { model [ " providerName " ] } ' )
print ( f ' Model ARN: { model [ " modelArn " ] } ' )
print ( f ' Lifecycle status: { model [ " modelLifecycle " ] [ " status " ] } ' )
print ( f ' Input modalities: { model [ " inputModalities " ] } ' )
print ( f ' Output modalities: { model [ " outputModalities " ] } ' )
print ( f ' Supported customizations: { model [ " customizationsSupported " ] } ' )
print ( f ' Supported inference types: { model [ " inferenceTypesSupported " ] } ' )
if " responseStreamingSupported " in model :
print ( f ' Response streaming supported: { model [ " responseStreamingSupported " ] } ' )
print ( " = " * 42 )
2023-11-13 20:11:29 +01:00
if __name__ == " __main__ " :
usage_demo ( )
