Commit Graph

  • 5e748dcfd7 Clarify a comment regarding the SD version of the multiple NMLs all becoming nn relations. Add the same relation to the UD version of the converter John Bauer 2023-08-08 12:51:04 -07:00
  • 8e7ba0ee74 dependency outputs have blank space at the end John Bauer 2023-08-02 13:57:11 -07:00
  • eef354ab04 fix failing test - i believe this is now testing the expected behavior John Bauer 2023-08-02 13:16:27 -07:00
  • 61ef545efa Separate off the NML part of the nn relationship so that we can do it without the comma restriction. This especially finds lists of NML as nn modifiers of a noun John Bauer 2023-08-02 12:00:58 -07:00
  • 133ddd8965 Document use of num for list item. Christopher Manning 2023-08-01 08:55:10 -07:00
  • 6e145278f8 Fix 'up to 1,700 kilograms' in SD and UD, even with RP error Christopher Manning 2023-07-31 15:31:02 -07:00
  • a000fe349f Javadoc and Java 8 fixes Christopher Manning 2023-07-31 15:30:09 -07:00
  • 86b6a03cd6 Move oldest Klein parseCommandLineArguments out of CoreNLP Christopher Manning 2023-07-28 17:01:34 -07:00
  • 61e4ae1ced Mindless javadoc improvement Christopher Manning 2023-07-28 16:40:56 -07:00
  • 7f0d8a41e2 Copyedit, update CoreNLP version Christopher Manning 2023-07-28 16:40:30 -07:00
  • d8f5559a0b Remove json-simple.jar from the main CoreNLP John Bauer 2023-07-17 08:10:44 -07:00
  • 334687d894 Need to remove the simple-json use from SceneGraph (not the /image version) as well John Bauer 2023-07-17 11:09:00 -07:00
  • cc7983ec26 Fix optional CoordinationPatterns - both conjunctions and disjunctions would infinite loop, as the matcher would accept a failed match even after a previous success (meaning it would always be willing to accept the same failed match again). https://github.com/stanfordnlp/CoreNLP/issues/1375 John Bauer 2023-07-17 01:20:24 -07:00
  • 50f0985647 This method needs to be overridden by any useful subclass, so make it abstract John Bauer 2023-07-17 00:49:55 -07:00
  • 8659653dc8 ... also fix the optional output for disjunctions John Bauer 2023-07-17 00:10:38 -07:00
  • a9965b2bbc Fix the output of optional CoordinationPatterns in Tregex John Bauer 2023-07-17 00:07:07 -07:00
  • 9c67d39d8a Add a test that patterns are not both negated and optional. The logic is hard to define (perhaps someone will eventually come up with a use case) John Bauer 2023-07-17 00:00:54 -07:00
  • 357b1bb222 Overhaul the usage of json in SceneGraph so that we only need one json library in the CoreNLP distro - remove simple-json John Bauer 2023-07-16 23:09:20 -07:00
  • 7f13d34bd5 Add a couple error checks when parsing a json SceneGraphImage. Also, add equals() methods to objects for which it is obvious what to do. Use these to make a rudimentary test of reading & writing json John Bauer 2023-07-16 18:00:04 -07:00
  • c8ca737b80 Whitespace (again) John Bauer 2023-07-16 12:32:53 -07:00
  • 7725e39236 Add a brief test that the json output is as expected for a single scene in SceneGraph John Bauer 2023-07-16 11:51:01 -07:00
  • 9bb91d815e Whitespace John Bauer 2023-07-16 10:45:15 -07:00
  • 8b40947914 Add an endpoint for 'scenegraph' to the server. John Bauer 2023-07-09 19:18:12 -07:00
  • 267041ea7c Remove extra non-matching right parentheses TAN Long 2023-07-11 19:02:04 +08:00
  • ad4556d8c1 Converter can compensate for an error in the parser for vmod (SD) and acl (UD) by looking for NML where there should be an NP. #1363 John Bauer 2023-07-08 17:05:05 -04:00
  • 5745de5b43 Update lists, LST, to be num (SD) or nummod (UD) of the governor. Addresses another portion of #1363 John Bauer 2023-07-08 15:52:30 -04:00
  • 9a86ece4dd Add a fake XSL node when converting constituency trees to dependencies for SD. This makes it easy to treat 'up to' as an MWE. #1363 John Bauer 2023-07-08 06:14:56 -07:00
  • 8c46648e45 Partial fix for '(QP up to ...)' as reported in #1363 John Bauer 2023-07-07 21:13:22 -07:00
  • 5c57fde5a8 Explain wtf is happening if a tree that shouldn't cause a problem does cause a problem John Bauer 2023-07-07 16:35:10 -07:00
  • 553a3f9f89 Get rid of an import .* to make there be fewer collisions between imports John Bauer 2023-07-07 16:24:56 -07:00
  • a606afa9e2 Include {} as punct words, as some parsers don't produce LCB/RCB John Bauer 2023-07-07 06:53:40 -07:00
  • 56cd6bb3e7 Add a note on why the old INTJ rule was no longer used. Addresses #1370 John Bauer 2023-07-07 05:29:32 -07:00
  • 5e0e44ad33 Fix up some usages of Timing in the LexicalizedParser John Bauer 2023-06-11 22:59:30 -07:00
  • b2e67b0004 Rather than gather all the parse trees, then score them all, score them all while processings them. Saves a significant amount of memory. This is especially relevant in the PCFG version (hence not having noticed in two years, since no one retrains that any more these days) John Bauer 2023-06-11 20:15:48 -07:00
  • f6366737db Add an ability to print out the feature index as text rather than a serialized object John Bauer 2023-06-07 23:16:38 -07:00
  • 2dd2014bd8 Add a small bit about the English tagger to the FAQ John Bauer 2023-05-16 23:26:08 -07:00
  • c8772b740d Update xom to 1.3.9, as it no longer has potentially unwanted transitive dependencies John Bauer 2023-05-16 22:23:05 -07:00
  • b577597578 Small javadoc improvement (only) Christopher Manning 2023-04-15 21:07:28 -07:00
  • f1b929e47a Add an mwt misc field to the wire format, including the protobuf annotation serializer John Bauer 2023-03-27 02:00:27 -07:00
  • 9805fb9edd Switch Ssurgeon MergeNodes to use multiple -node entries instead of a custom syntax. This will make it significantly easier to eventually support more than 2 nodes to be merged John Bauer 2023-04-11 00:23:13 -07:00
  • e52fd28b7e Remove test of the lemma being wiped out by setWord John Bauer 2023-04-05 14:46:30 -07:00
  • 61bb448051 Add a test that only setting the lemma still does what we want John Bauer 2023-04-05 12:43:27 -07:00
  • 2f4b60fbd2 Temporarily at least, MergeNodes compensates for setWord doing weird things to the lemma John Bauer 2023-04-05 12:40:53 -07:00
  • 429f61aafd Add a DeleteLeaf operation to Ssurgeon. Will delete a leaf (node with no children) and rearrange all the indices appropriately. John Bauer 2023-04-05 09:39:14 -07:00
  • 203eb065cb Allow dashes as the word in an expression with indices in SemanticGraph.valueOf John Bauer 2023-04-05 09:32:42 -07:00
  • 429688d092 Add comments on a missing feature John Bauer 2023-04-05 08:49:21 -07:00
  • e5dc7d976a Make a variable final for readability John Bauer 2023-04-05 08:44:33 -07:00
  • 2151b1e92a Fix a broken doc in Ssurgeon John Bauer 2023-04-04 11:57:12 -07:00
  • fa2d495632 Some documentation on why we used a specific protocol for Ssurgeon John Bauer 2023-03-29 21:40:45 -07:00
  • d8d9d9fdde Making this an IdentityHashSet hilariously makes the graph serialization stop working for sentences longer than N words. The problem is that the identity of an integer isn't cached beyond some predefined limit, 128 for a standard Integer John Bauer 2023-03-28 10:04:31 -07:00
  • 117615cecb Whitespace John Bauer 2023-03-27 13:44:35 -07:00
  • 785620d76f Oops, the default toString() being used in the Ssurgeon CombineMWT operation was capturing the tag as well John Bauer 2023-03-26 23:17:39 -07:00
  • 010a955f6f Add some more doc to the new Ssurgeon CombineMWT operation John Bauer 2023-03-26 22:07:48 -07:00
  • e5cad737df Add an edit line to mark MWT on a set of nodes to Ssurgeon John Bauer 2023-03-26 22:04:27 -07:00
  • 4fadb511aa Set up potentially multiple -node arguments in the Ssurgeon ArgsBox, so that some editlines can use multiple node names John Bauer 2023-03-26 21:05:52 -07:00
  • 0660fa9d09 Add a two-word MergeNodes operation. John Bauer 2023-03-25 08:13:49 -07:00
  • db0bd459e7 Tiny bit of clarification about a test John Bauer 2023-03-25 10:28:00 -07:00
  • b8e6555c0f Update links to account for 4.5.4 John Bauer 2023-03-24 10:45:20 -07:00
  • 1398932428 Update links for 4.5.4 instead of 4.5.3 v4.5.4 John Bauer 2023-03-15 17:44:27 -07:00
  • 4226765c52 rstrip the release descriptions in README John Bauer 2023-03-15 17:43:14 -07:00
  • 9f3c3c761d Update various notes for a small modification to ssurgeon and a forgotten CoreNLP CLI John Bauer 2023-03-15 17:41:48 -07:00
  • a654241034 Get 'isChanged' directly from the SsurgeonPattern, rather than trying to compare the graphs, which is problematic in the case of EditNode or something else that changes information which isn't compared by SemanticGraph.equals() John Bauer 2023-03-13 18:31:05 -07:00
  • b8bcecd2b5 Update a comment with an explanation of how we handle something John Bauer 2023-03-13 18:10:35 -07:00
  • 81141f690c Add an operation, ReattachNamedEdge, which reattaches a named edge to a new gov and/or dep John Bauer 2023-03-13 13:18:01 -04:00
  • c1ed2deb2a Add a bit of doc on a possible fix needed John Bauer 2023-03-13 13:17:33 -04:00
  • 183e0eb53e Fix a couple comments John Bauer 2023-03-13 13:07:57 -04:00
  • 37a71fb465 Remove an unused import John Bauer 2023-03-13 12:49:18 -04:00
  • 820144fbec Bomb-proof the RelabelNamedEdge operation John Bauer 2023-03-13 12:39:35 -04:00
  • 9ea4f394d2 Keep features in the order they are in when we receive them in a proto v4.5.3 John Bauer 2023-03-11 10:14:48 -05:00
  • 901b0f0def Update 4.5.2 links -> 4.5.3 for direct download. Maven links are caught up to 4.5.2 at least... will need to push to Maven soon if everything checks out John Bauer 2023-03-11 09:15:16 -05:00
  • c4fc4a5eec Update 4.5.2 links to 4.5.3, include some 'what happened' notes John Bauer 2023-03-10 21:17:19 -08:00
  • 81a6a4bbd6 Update 4.5.2 -> 4.5.3, with collinizer fixes and an ssurgeon interface John Bauer 2023-03-10 21:00:14 -08:00
  • 26ab9c010c Add a test of setting the MWT attributes via ssurgeon John Bauer 2023-03-10 23:42:00 -05:00
  • b36d9eb2fe Add MWT fields to AnnotationLookup, which necessitates parsing Boolean in CoreLabel John Bauer 2023-03-10 23:40:35 -05:00
  • b265c16f82 Oops, only added CoNLLUFeats to the string/string constructor, not class/string John Bauer 2023-03-10 22:59:19 -05:00
  • 0f901cff7c Use a LinkedHashMap to keep the feature annotations in the order they are created John Bauer 2023-03-10 14:09:38 -08:00
  • eb4d58d792 encoding problem? John Bauer 2023-03-10 15:48:46 -05:00
  • ef9baf0bef Update doc - spaces are allowed if properly quoted John Bauer 2023-03-10 15:41:49 -05:00
  • 2efab6d69b This part of the test was a duplicate of an earlier test John Bauer 2023-03-10 14:37:31 -05:00
  • 947ab2c4cf Add a test of quoting attributes within the edit-list, which it turns out you can already do John Bauer 2023-03-10 14:36:55 -05:00
  • dc2847c9e7 Pass arguments to an EditLine via both the text of the edit-line node and its attributes. Allows for whitespace in the xml John Bauer 2023-03-10 08:43:11 -05:00
  • 0dc6e105bd Support one specific case of Map<String, String> in AnnotationLookup, the morphological features. Adds that as a feature to ssurgeon John Bauer 2023-03-10 00:02:08 -05:00
  • cb6a2f6ec0 Add a large chunk of documentation to Ssurgeon John Bauer 2023-03-09 10:28:18 -08:00
  • 5250f9faf9 Pass a language to ProcessSsurgeonRequest John Bauer 2023-03-06 16:46:44 -08:00
  • db74e51dfb Add a Language field to the Ssurgeon. This will allow the tool to make edges with the proper dependency scheme. Also, use the version of GrammaticalRelation.valueOf which creates an edge if it doesn't currently exist. This will allow for unknown dependency schemes to work for the most part (although perhaps some bug will crop up anyway) John Bauer 2023-03-06 15:30:51 -08:00
  • ec8576b8a0 Add a method to make a case-insensitive enum search for Language John Bauer 2023-03-06 14:08:12 -08:00
  • fde788f0bc Update named edges in the SemgrexMatcher when rearranging nodes in the SemanticGraph. Includes a test of that functionality John Bauer 2023-03-06 09:30:54 -08:00
  • 90886db324 Wrap all of Ssurgeon's editline processing in a try/catch so that it can report the error line no matter where the error occurs (eg, if it bubbles up from someplace that doesn't know about the editline) John Bauer 2023-03-06 08:37:07 -08:00
  • 71ad47e748 Push the Ssurgeon exception handling into the individual operation. Check for unreadable attributes at the same time as checking for illegal attributes - this will make EditNode also support checking those attributes John Bauer 2023-03-06 08:24:07 -08:00
  • 5a0c2d5e55 Update the SemgrexMatcher's nodes. Still should do edges John Bauer 2023-03-05 23:27:31 -08:00
  • 0240694fc3 Error check that no one tries to update idx, sentIdx, or DocId John Bauer 2023-03-05 23:13:25 -08:00
  • b2840067f4 Add an EditNode which reuses the attribute parsing from AddDep to update the values on a node John Bauer 2023-03-05 23:06:14 -08:00
  • 7e9e95ff02 Update documentation now that the position is correctly handled in AddDep John Bauer 2023-03-05 14:22:54 -08:00
  • 2cdd6e5ba1 Add the ability to control position with + or - relative to a node John Bauer 2023-03-05 14:14:06 -08:00
  • ec517e1a59 Add documentation for before / after operations John Bauer 2023-03-05 14:08:55 -08:00
  • 9140384200 Add a -position argument for AddDep. Currently works only for - (start of sentence) and + (end of sentence, also the default). Need to process -word and +word as well. Will also want to update the SemgrexMatcher so that potentially further operations are possible. John Bauer 2023-03-05 13:39:38 -08:00
  • 15e6186b1f Add utilities to check the min & max index of a SemanticGraph John Bauer 2023-03-05 11:14:23 -08:00
  • 1c193132ba Check that a couple annotation types we won't be able to translate cause exceptions John Bauer 2023-03-04 20:07:56 -08:00
  • 01a61cc933 Switch AddDep to use AnnotationLookup to allow for generic keys. Will need to turn generic keys which can't be translated into SsurgeonParseException, and will also want to add a position to the AddDep John Bauer 2023-03-04 18:41:26 -08:00
  • 4d4056c988 Add a RelabelNamedEdge. John Bauer 2023-03-01 18:13:52 -08:00