Cogcomp Nlp Versions Save

CogComp's Natural Language Processing Libraries and Demos: Modules include lemmatizer, ner, pos, prep-srl, quantifier, question type, relation-extraction, similarity, temporal normalizer, tokenizer, transliteration, verb-sense, and more.

4.0.13

5 years ago

NER now supports use of multiple models within same VM context. Improved documentation for configuration in pipeline and in core-utilities. Various minor fixes to improve performance (core-utilities, Chunker). Minor improvements to StatefulTokenizer.

4.0.12

5 years ago

Changes:

  • Added the ability to use JSON Serialized Format with NerTagger #676
  • bugs with multiple NERAnnotators per process space #675
  • Revert "Incremental Training" #672

4.0.10

5 years ago
  1. fix the timex bug when may and sat appear as verbs instead of timexe #663
  2. Incremental Training #667
  3. Added get and post functions for adding Views to JsonStr serialized TA #671

4.0.9

5 years ago

CoreUtils:

  • BasicTextAnnotationBuilder.java now accepts list of list of tokens #670
  • fixed a bug in json serializer. Also updated to explicitly store and… #662
  • fix a TextAnnotation builder bug on Windows #639
  • deleted duplicated DBHelper.java #632

NER:

  • NER training #666
  • NER Model Loading #654

CorpusReaders:

  • Add MascXCESReaderTest corpus to resources #650
  • Ontonotes 5 readers #627
  • TACReader #615

Chunker:

  • Chunker training data fix #627

Similarity:

  • Allow user to provide types for one or both names in NESim.compare() #625

Tokenizer:

  • Fix common dates like "10/14/2016" are not parsed to a single token #654
  • Add option to split on multiple newlines, capture emails as single token. #647

Temporal normalizer:

  • Bug fixes #646
  • Fix temporal component cannot capture obvious timexes #636
  • Populate timex normalization type to the TIMEX View in TextAnnotation #630

4.0.2

6 years ago

Summary of the changes:

  • Question Type Classifier #601
  • Bug fixes for dataless classifier #601
  • Pipeline-Client: now a separate module #601

4.0.1

6 years ago
  • Fix cogcomp website links in the readmes #585
  • Dropping some old dependencies in transliterator #584
  • Dataless Classifier #544

4.0.0

6 years ago
  • Fixes in the readmes #585
  • Clean up old dependencies in transliterator #585
  • Double to Float when loading models #583
  • An ACE reader with TrueCaser #581

3.1.35

6 years ago
  • Added relation extraction #572
  • Added transliteration models #577
  • Extend the list of languages and add ISO 639-3 standard 3-digit ids, with @mayhewsw 's suggestion. #576
  • AnnotatorService can receive parameters now parameters. #576
  • Ignore a few external test, to make CIs faster #577

3.1.34

6 years ago
  • Adding Transliteration #563
  • Limit testing logs #564
  • Propbank readers for Ontonotes 5 #569
  • Improvements to MD and NER #570
  • support for initializing MD with local model #571

3.1.22

6 years ago

Updated version of the MapDB used throughout caching