Commit Graph

  • 70d2acb579 * Fix edge features Matthew Honnibal 2015-07-09 12:15:01 +0200
  • 8a7bbd5850 * Announce v0.88 Matthew Honnibal 2015-07-09 12:12:29 +0200
  • 703ca40420 * Inc version Matthew Honnibal 2015-07-08 20:07:23 +0200
  • adb868bdad * Add warning for models not found in parser Matthew Honnibal 2015-07-08 20:04:55 +0200
  • 05b28ec9eb * Add warning for models not found in parser Matthew Honnibal 2015-07-08 20:02:13 +0200
  • ef700401a6 * Add warning for models not found in parser Matthew Honnibal 2015-07-08 20:00:46 +0200
  • 6218d8b389 * Add warning for models not found in parser Matthew Honnibal 2015-07-08 19:59:16 +0200
  • f6a6c39ce8 * Add warning for models not found in parser Matthew Honnibal 2015-07-08 19:52:30 +0200
  • 78db7e32f7 * Remove has_sense method from Lexeme declaration Matthew Honnibal 2015-07-08 19:41:20 +0200
  • 6ddb2f5e45 * Restore merge_mwe in English class Matthew Honnibal 2015-07-08 19:35:30 +0200
  • 6859f6adac * Restore merge_mwe in English class Matthew Honnibal 2015-07-08 19:34:55 +0200
  • 3c270fc8ff * Remove has_sense method from Lexeme Matthew Honnibal 2015-07-08 19:28:29 +0200
  • b64c843861 * Remove senses attr Matthew Honnibal 2015-07-08 19:26:24 +0200
  • 1d3a592edf * Remove the senses attr from LexemeC, to keep data compatibility Matthew Honnibal 2015-07-08 19:24:44 +0200
  • 0ceb1f71c2 * Update parse features Matthew Honnibal 2015-07-08 19:11:36 +0200
  • 2e51b5027a * Alias Doc to Tokens, for backwards compatibility Matthew Honnibal 2015-07-08 18:59:35 +0200
  • 462301d9e6 * Fix reference to Tokens in documentation Matthew Honnibal 2015-07-08 18:58:25 +0200
  • e3c53f5ecd * Fix mention of Tokens in docstring Matthew Honnibal 2015-07-08 18:56:27 +0200
  • 783867c1ad * Update quickstart.rst for Tokens --> Doc rename Matthew Honnibal 2015-07-08 18:54:08 +0200
  • bb522496dd * Rename Tokens to Doc Matthew Honnibal 2015-07-08 18:53:00 +0200
  • d0fc7f5ba9 * Relabel docs sections Matthew Honnibal 2015-07-08 18:23:49 +0200
  • ec398ef1d0 * Relabel docs sections Matthew Honnibal 2015-07-08 18:20:00 +0200
  • 38f6e92ffb * Add docs for vocab and string store Matthew Honnibal 2015-07-08 18:00:33 +0200
  • 2f7110e852 * Add using/ docs. Matthew Honnibal 2015-07-08 17:59:07 +0200
  • 4b07c17d6f * More work on reorganized docs. Getting close to useable Matthew Honnibal 2015-07-08 17:58:49 +0200
  • 2566c16c7e * Remove obsolete docs/guide dir Matthew Honnibal 2015-07-08 17:51:55 +0200
  • 2702105183 * Improve index.html table Matthew Honnibal 2015-07-08 17:11:03 +0200
  • 6a7a059660 * Improve index.html table Matthew Honnibal 2015-07-08 17:09:26 +0200
  • a32c6ff930 * Add links in reference.rst Matthew Honnibal 2015-07-08 16:51:55 +0200
  • 9fa46743c0 * Add reference/index.rst Matthew Honnibal 2015-07-08 15:33:47 +0200
  • 1a95b490a8 * Add loading.rst reference Matthew Honnibal 2015-07-08 15:13:47 +0200
  • 79abe2860a * Add processing.rst reference docs Matthew Honnibal 2015-07-08 15:11:09 +0200
  • b24e8be2b9 * Whitespace in docstring Matthew Honnibal 2015-07-08 12:37:03 +0200
  • abc43b852d * Add pos_tags attr to Vocab. Matthew Honnibal 2015-07-08 12:36:38 +0200
  • 935bcdf3e5 * Remove redundant tag_names argument to Tokenizer Matthew Honnibal 2015-07-08 12:36:04 +0200
  • ff885e8511 * Add ParserFactory convenience function Matthew Honnibal 2015-07-08 12:35:46 +0200
  • 4e4fac452b * Refactor __init__ for simplicity. Allow parse=True, tag=True etc flags to be passed at top-level. Do not lazy-load parser. Matthew Honnibal 2015-07-08 12:35:29 +0200
  • 4d24d513ad * Add fab docs command Matthew Honnibal 2015-07-08 12:34:35 +0200
  • b42db257b7 * Work on API reference docs Matthew Honnibal 2015-07-08 12:34:23 +0200
  • 99e84488da * Add draft doc describing annotation standards Matthew Honnibal 2015-07-08 10:27:35 +0200
  • 68eff957a5 * Work on API docs Matthew Honnibal 2015-07-07 21:35:22 +0200
  • 1d2deb4616 * Work on refactoring default arguments to English.__init__ Matthew Honnibal 2015-07-07 15:53:25 +0200
  • 2d0e99a096 * Pass pos_tags into Tokenizer.from_dir Matthew Honnibal 2015-07-07 14:23:08 +0200
  • 6788c86b2f * Begin refactor Matthew Honnibal 2015-07-07 14:00:07 +0200
  • be5affe390 * Fix import of sense tagger sense_unsupervised Matthew Honnibal 2015-07-06 09:33:58 +0200
  • a916f6a109 * Compile spacy.wsd module Matthew Honnibal 2015-07-06 09:33:41 +0200
  • 5ec2ce4dcb * Fix spacy.wsd module Matthew Honnibal 2015-07-06 09:33:26 +0200
  • eb3057d806 * Add updated unsupervised_train script, from the wsd directory Matthew Honnibal 2015-07-06 09:33:00 +0200
  • 1d21eebda4 Update gitignore for new wsd module Matthew Honnibal 2015-07-06 09:32:10 +0200
  • 300eb44848 * Add corpus.py, with DocsDB class Matthew Honnibal 2015-07-06 09:31:40 +0200
  • 2e4cfe5255 * Add script to train the dictionary-supervised supersense tagger Matthew Honnibal 2015-07-06 09:06:22 +0200
  • 88a4e53fcb * Begin refactoring sense tagger Matthew Honnibal 2015-07-06 09:01:21 +0200
  • 2133c2d299 * Don't expect WSD in gold tuples Matthew Honnibal 2015-07-06 08:45:05 +0200
  • 0be251776e * Supply templates as an argument to the parser Config object Matthew Honnibal 2015-07-06 08:44:39 +0200
  • 316a0772b2 * Remove WSD from gold.pyx Matthew Honnibal 2015-07-06 08:43:59 +0200
  • b61b495024 * Start adding parse features to sense_tagger Matthew Honnibal 2015-07-06 08:43:24 +0200
  • cb628ba352 * Add document features to sense_tagger. Matthew Honnibal 2015-07-05 21:05:38 +0200
  • 8f0fe1a4ea * Note broken sense data in prepare_treebank Matthew Honnibal 2015-07-05 21:04:57 +0200
  • 96442d9c3e * Put supersenses.json in the wordnet directory, not in a wsd directory Matthew Honnibal 2015-07-05 21:03:59 +0200
  • 3eff39ff63 * Prevent supersenses from being assigned to CONJ, DET, NUM and PRON words. Matthew Honnibal 2015-07-05 14:20:07 +0200
  • 9534d336ed * Ensure word senses are loaded, even if not in probabilities file Matthew Honnibal 2015-07-05 11:31:07 +0200
  • 149a901ea7 * Don't use POS tags in supersense dict Matthew Honnibal 2015-07-05 10:50:02 +0200
  • 4e0cd8def8 * Remove score_senses method from Scorer Matthew Honnibal 2015-07-05 09:15:17 +0200
  • 211058f7a6 * Load adverb senses Matthew Honnibal 2015-07-05 09:13:22 +0200
  • 427ea16b27 * Use tagdict in sense_tagger Matthew Honnibal 2015-07-05 09:12:53 +0200
  • 5e0545be5c * Fix 32bit/64bit int problem when setting flags Matthew Honnibal 2015-07-05 09:11:55 +0200
  • 4c6533a019 * Write a supersenses.json fil into a wsd directory in init_model Matthew Honnibal 2015-07-04 17:24:32 +0200
  • 00c9acbf42 * Add hacky distribution over supersenses, using a half-assed thing like a stick-breaking process Matthew Honnibal 2015-07-04 16:45:04 +0200
  • 153758bf65 * Hack on index.rst Matthew Honnibal 2015-07-04 12:26:45 +0200
  • 893b5fd42c * Hack on sense tagger Matthew Honnibal 2015-07-04 12:26:16 +0200
  • 389dcd3fb2 * Fix setting of supersense bits in lexeme.pyx Matthew Honnibal 2015-07-04 12:25:21 +0200
  • 948ea9333a * Fix alignment of supersenses in init_model Matthew Honnibal 2015-07-04 12:24:40 +0200
  • fb68df91b8 * Work on sense tagger Matthew Honnibal 2015-07-03 15:25:41 +0200
  • 2fbcdd0ea8 * Refactor sense tagger to get rid of intermediary layers Matthew Honnibal 2015-07-03 13:31:11 +0200
  • 6735439abf * Fix the way supersenses are loaded from the json file Matthew Honnibal 2015-07-03 13:29:22 +0200
  • ff1f9fe246 * Fix init_model to read supersenses from wordnet, not pre-computed supersenses file Matthew Honnibal 2015-07-03 13:28:39 +0200
  • b977d60bf4 * Hack in WSD scoring Matthew Honnibal 2015-07-03 09:25:52 +0200
  • 68f174b235 * Remove adjectives from supersense list. This seems to be associated with current memory errors Matthew Honnibal 2015-07-03 09:24:45 +0200
  • 12dd4f745a * Add validation for argmaxing in _ml.pyx Matthew Honnibal 2015-07-03 09:18:33 +0200
  • 5d933eec8e * Use the gold sense labels for training Matthew Honnibal 2015-07-03 05:45:42 +0200
  • 4a60b68a24 * Add encode_sense_strs function Matthew Honnibal 2015-07-03 05:45:16 +0200
  • 1be5ab200f * Add some of the sensetagger changes Matthew Honnibal 2015-07-03 05:18:15 +0200
  • b7e9c1da85 * Begin writing score_senses method Matthew Honnibal 2015-07-03 05:10:52 +0200
  • 8464378a85 * Initialize Lexeme.senses to zero Matthew Honnibal 2015-07-03 05:03:16 +0200
  • e99e15574e * Add sense and sense_ properties to Token objects Matthew Honnibal 2015-07-03 04:59:20 +0200
  • 8f068dc6fe * Set scores to 0 before prediction Matthew Honnibal 2015-07-03 04:55:30 +0200
  • 2be517ba6d * Read in gold wsd data, as supersenses Matthew Honnibal 2015-07-03 04:47:23 +0200
  • c60cc22390 * Ignore adjective supersenses Matthew Honnibal 2015-07-03 04:46:11 +0200
  • dbcef2b76e * Read in new WSD gold data Matthew Honnibal 2015-07-03 04:43:23 +0200
  • 333e414e9f * Hack prepare_treebank script to load wordnet supersenses Matthew Honnibal 2015-07-02 08:31:12 +0200
  • 05146a4578 * Add script to read wordnet data for supersense stuff Matthew Honnibal 2015-07-02 08:30:43 +0200
  • 2256ba7590 * Integrate sense tagger module Matthew Honnibal 2015-07-02 00:54:46 +0200
  • 9c74f82d20 * Add rough sense tagger Matthew Honnibal 2015-07-02 00:54:26 +0200
  • 4e830b9d41 * Add N_SENSES in senses.pxd Matthew Honnibal 2015-07-02 00:54:06 +0200
  • 041908a272 * Merge neuralnet branch into sense-tagger Matthew Honnibal 2015-07-01 22:38:22 +0200
  • 3992724685 * Compile sense_tagger Matthew Honnibal 2015-07-01 22:37:31 +0200
  • 52fd80c6c6 * Add experimental supersense features for parsing, based on lookup into wordnet. Matthew Honnibal 2015-07-01 20:12:44 +0200
  • e6d828a9af * Set up an array POS_SENSES that denotes the set of valid senses for each POS tag. This way, we can do bitwise & between a lexeme's senses and the ones available for its POS tag, to get the allowable senses for the token. Matthew Honnibal 2015-07-01 20:12:13 +0200
  • 2b8459d9a8 * Add senses flag to Lexeme Matthew Honnibal 2015-07-01 20:10:41 +0200
  • e23d1582a2 * Add supersense data to Lexeme objects. Add simple has_sense method to check the flag. Matthew Honnibal 2015-07-01 18:50:37 +0200