Rooba
/
spaCy
mirror of https://github.com/explosion/spaCy.git
1
1
Fork 0
Code Issues Packages Projects Releases Wiki Activity
5,885 Commits 70 Branches 177 Tags 298 MiB
Commit Graph

180 Commits

Author SHA1 Message Date
Matthew Honnibal 3e688e6d4b Fix issue #514 -- serializer fails when new entity type has been added. The fix here is quite ugly. It's best to add the entities ASAP after loading the NLP pipeline, to mitigate the brittleness. 2016-10-23 17:45:44 +02:00
Matthew Honnibal f62088d646 Fix compile error 2016-10-23 14:50:50 +02:00
Matthew Honnibal a0a4ada42a Fix calculation of L2-norm for Lexeme 2016-10-23 14:44:45 +02:00
Matthew Honnibal 7ab03050d4 Add resize_vectors method to Vocab 2016-10-21 01:44:50 +02:00
Matthew Honnibal f5fe4f595b Fix json loading, for Python 3. 2016-10-20 21:23:26 +02:00
Matthew Honnibal 5ec32f5d97 Fix loading of GloVe vectors, to address Issue #541 2016-10-20 18:27:48 +02:00
Matthew Honnibal d10c17f2a4 Fix Issue #536: oov_prob was 0 for OOV words. 2016-10-19 23:38:47 +02:00
Matthew Honnibal 2bbb050500 Fix default of serializer_freqs 2016-10-18 19:55:41 +02:00
Matthew Honnibal 2cc515b2ed Add add_flag method to Vocab, re Issue #504. 2016-10-14 12:15:38 +02:00
Matthew Honnibal ea23b64cc8 Refactor training, with new spacy.train module. Defaults still a little awkward. 2016-10-09 12:24:24 +02:00
Matthew Honnibal ca32a1ab01 Revert "Work on Issue #285: intern strings into document-specific pools, to address streaming data memory growth. StringStore.__getitem__ now raises KeyError when it can't find the string. Use StringStore.intern() to get the old behaviour. Still need to hunt down all uses of StringStore.__getitem__ in library and do testing, but logic looks good." 
This reverts commit 8423e8627f.
 2016-09-30 20:20:22 +02:00
Matthew Honnibal 1f1cd5013f Revert "Changes to vocab for new stringstore scheme" 
This reverts commit a51149a717.
 2016-09-30 20:10:30 +02:00
Matthew Honnibal a51149a717 Changes to vocab for new stringstore scheme 2016-09-30 20:01:19 +02:00
Matthew Honnibal 8423e8627f Work on Issue #285: intern strings into document-specific pools, to address streaming data memory growth. StringStore.__getitem__ now raises KeyError when it can't find the string. Use StringStore.intern() to get the old behaviour. Still need to hunt down all uses of StringStore.__getitem__ in library and do testing, but logic looks good. 2016-09-30 10:14:47 +02:00
Matthew Honnibal 95aaea0d3f Refactor so that the tokenizer data is read from Python data, rather than from disk 2016-09-25 14:49:53 +02:00
Matthew Honnibal df88690177 Fix encoding of path variable 2016-09-24 21:13:15 +02:00
Matthew Honnibal af847e07fc Fix usage of pathlib for Python3 -- turning paths to strings. 2016-09-24 21:05:27 +02:00
Matthew Honnibal 453683aaf0 Fix spacy/vocab.pyx 2016-09-24 20:50:31 +02:00
Matthew Honnibal fd65cf6cbb Finish refactoring data loading 2016-09-24 20:26:17 +02:00
Matthew Honnibal 83e364188c Mostly finished loading refactoring. Design is in place, but doesn't work yet. 2016-09-24 15:42:01 +02:00
Matthew Honnibal 872695759d Merge pull request #306 from wbwseeker/german_noun_chunks 
add German noun chunk functionality
 2016-04-08 00:54:24 +10:00
Henning Peters b8f63071eb add lang registration facility 2016-03-25 18:54:45 +01:00
Wolfgang Seeker 5e2e8e951a add baseclass DocIterator for iterators over documents 
add classes for English and German noun chunks

the respective iterators are set for the document when created by the parser
as they depend on the annotation scheme of the parsing model
 2016-03-16 15:53:35 +01:00
Wolfgang Seeker 03fb498dbe introduce lang field for LexemeC to hold language id 
put noun_chunk logic into iterators.py for each language separately
 2016-03-10 13:01:34 +01:00
Matthew Honnibal 963fe5258e * Add missing __contains__ method to vocab 2016-03-08 15:49:10 +00:00
Matthew Honnibal 478aa21cb0 * Remove broken __reduce__ method on vocab 2016-03-08 15:48:21 +00:00
Henning Peters 931c07a609 initial proposal for separate vector package 2016-03-04 11:09:06 +01:00
Matthew Honnibal a95974ad3f * Fix oov probability 2016-02-06 15:13:55 +01:00
Matthew Honnibal dcb401f3e1 * Remove broken Vocab pickling 2016-02-06 14:08:47 +01:00
Matthew Honnibal 63e3d4e27f * Add comment on Vocab.__reduce__ 2016-01-19 20:11:25 +01:00
Henning Peters 235f094534 untangle data_path/via 2016-01-16 12:23:45 +01:00
Henning Peters 846fa49b2a distinct load() and from_package() methods 2016-01-16 10:00:57 +01:00
Henning Peters 788f734513 refactored data_dir->via, add zip_safe, add spacy.load() 2016-01-15 18:01:02 +01:00
Henning Peters bc229790ac integrate with sputnik 2016-01-13 19:46:17 +01:00
Matthew Honnibal eaf2ad59f1 * Fix use of mock Package object 2015-12-31 04:13:15 +01:00
Matthew Honnibal aec130af56 Use util.Package class for io 
Previous Sputnik integration caused API change: Vocab, Tagger, etc
were loaded via a from_package classmethod, that required a
sputnik.Package instance. This forced users to first create a
sputnik.Sputnik() instance, in order to acquire a Package via
sp.pool().

Instead I've created a small file-system shim, util.Package, which
allows classes to have a .load() classmethod, that accepts either
util.Package objects, or strings. We can later gut the internals
of this and make it a proxy for Sputnik if we need more functionality
that should live in the Sputnik library.

Sputnik is now only used to download and install the data, in
spacy.en.download
 2015-12-29 18:00:48 +01:00
Matthew Honnibal 0e2498da00 * Replace from_package with load() classmethod in Vocab 2015-12-29 16:56:51 +01:00
Henning Peters 8359bd4d93 strip data/ from package, friendlier Language invocation, make data_dir backward/forward-compatible 2015-12-18 09:52:55 +01:00
Henning Peters 9027cef3bc access model via sputnik 2015-12-07 06:01:28 +01:00
Matthew Honnibal 6ed3aedf79 * Merge vocab changes 2015-11-06 00:48:08 +11:00
Matthew Honnibal 1e99fcd413 * Rename .repvec to .vector in C API 2015-11-03 23:47:59 +11:00
Matthew Honnibal 5887506f5d * Don't expect lexemes.bin in Vocab 2015-11-03 13:23:39 +11:00
Matthew Honnibal f11030aadc * Remove out-dated TODO comment 2015-10-26 12:33:38 +11:00
Matthew Honnibal a371a1071d * Save and load word vectors during pickling, re Issue #125 2015-10-26 12:33:04 +11:00
Matthew Honnibal 314090cc78 * Set vectors length when unpickling vocab, re Issue #125 2015-10-26 12:05:08 +11:00
Matthew Honnibal 2348a08481 * Load/dump strings with a json file, instead of the hacky strings file we were using. 2015-10-22 21:13:03 +11:00
Matthew Honnibal 7a15d1b60c * Add Python 2/3 compatibility fix for copy_reg 2015-10-13 20:04:40 +11:00
Matthew Honnibal 20fd36a0f7 * Very scrappy, likely buggy first-cut pickle implementation, to work on Issue #125: allow pickle for Apache Spark. The current implementation sends stuff to temp files, and does almost nothing to ensure all modifiable state is actually preserved. The Language() instance is a deep tree of extension objects, and if pickling during training, some of the C-data state is hard to preserve. 2015-10-13 13:44:41 +11:00
Matthew Honnibal f8de403483 * Work on pickling Vocab instances. The current implementation is not correct, but it may serve to see whether this approach is workable. Pickling is necessary to address Issue #125 2015-10-13 13:44:41 +11:00
Matthew Honnibal 85e7944572 * Start trying to pickle Vocab 2015-10-13 13:44:41 +11:00
Matthew Honnibal 41012907a8 * Fix variable name 2015-10-13 13:44:40 +11:00
Matthew Honnibal 37b909b6b6 * Use the symbols file in vocab instead of the symbols subfiles like attrs.pxd 2015-10-13 13:44:40 +11:00
Matthew Honnibal d70e8cac2c * Fix empty values in attributes and parts of speech, so symbols align correctly with the StringStore 2015-10-13 13:44:40 +11:00
Matthew Honnibal a29c8ee23d * Add symbols to the vocab before reading the strings, so that they line up correctly 2015-10-13 13:44:39 +11:00
Matthew Honnibal 85ce36ab11 * Refactor symbols, so that frequency rank can be derived from the orth id of a word. 2015-10-13 13:44:39 +11:00
Matthew Honnibal 83dccf0fd7 * Use io module insteads of deprecated codecs module 2015-10-10 14:13:01 +11:00
Matthew Honnibal 3d9f41c2c9 * Add LookupError for better error reporting in Vocab 2015-10-06 10:34:59 +11:00
alvations 8caedba42a caught more codecs.open -> io.open 2015-09-30 20:20:09 +02:00
Matthew Honnibal abf0d930af * Fix API for loading word vectors from a file. 2015-09-23 23:51:08 +10:00
Matthew Honnibal f7283a5067 * Fix vectors bugs for OOV words 2015-09-22 02:10:25 +02:00
Matthew Honnibal ac459278d1 * Fix vector length error reporting, and ensure vec_len is returned 2015-09-21 18:08:32 +10:00
Matthew Honnibal ba4e563701 * Ensure vectors are same length, and return vector length in load_vectors_bz2 2015-09-21 18:03:08 +10:00
Matthew Honnibal d6945bf880 * Add way to load vectors from bz2 file to vocab 2015-09-17 12:58:23 +10:00
Matthew Honnibal 3d87519f64 * Remove vectors argument from Vocab object 2015-09-15 14:47:14 +10:00
Matthew Honnibal 27f988b167 * Remove the vectors option to Vocab, preferring to either load vectors from disk, or set them on the Lexeme objects. 2015-09-15 14:41:48 +10:00
Matthew Honnibal e9c59693ea * Remove assertion from vocab.pyx 2015-09-13 10:30:08 +10:00
Matthew Honnibal e1dfaeed8a * Check serializer freqs exist before loading 2015-09-12 23:49:38 +02:00
Matthew Honnibal a412c66c8c * Check serializer freqs exist before loading 2015-09-12 23:40:01 +02:00
Matthew Honnibal e285ca7d6c * Load serializer freqs in vocab 2015-09-10 15:22:48 +02:00
Matthew Honnibal 094440f9f5 Merge branch 'develop' of ssh://github.com/honnibal/spaCy into develop 2015-09-10 14:51:17 +02:00
Matthew Honnibal 90da3a695d * Load lemmatizer from disk in Vocab.from_dir 2015-09-10 14:49:10 +02:00
Matthew Honnibal f634191e27 * Fix vocab read/write 2015-09-10 14:44:38 +02:00
Matthew Honnibal a7f4b26c8c * Tmp 2015-09-09 14:33:26 +02:00
Matthew Honnibal d6561988cf * Fix lexemes.bin 2015-09-09 11:49:51 +02:00
Matthew Honnibal c301bebd33 Merge branch 'master' of https://github.com/honnibal/spaCy into develop 2015-09-09 10:55:39 +02:00
Matthew Honnibal 623329b19a Merge branch 'master' of ssh://github.com/honnibal/spaCy into develop 2015-09-08 14:27:01 +02:00
Matthew Honnibal 62a01dd41d * Fix issue #92: lexemes.bin read error on 32-bit platforms. 2015-09-08 14:23:58 +02:00
Matthew Honnibal f6ec5bf1b0 * Use empty tag map in vocab if none supplied 2015-09-06 20:19:27 +02:00
Matthew Honnibal 534e3dda3c * More work on language independent parsing 2015-08-28 03:44:54 +02:00
Matthew Honnibal c2307fa9ee * More work on language-generic parsing 2015-08-28 02:02:33 +02:00
Matthew Honnibal 1302d35dff * Rework interfaces in vocab 2015-08-26 19:21:46 +02:00
Matthew Honnibal 6f1743692a * Work on language-independent refactoring 2015-08-23 20:49:18 +02:00
Matthew Honnibal cad0cca4e3 * Tmp 2015-08-22 22:04:34 +02:00
Matthew Honnibal 3d43f49f69 * Revert prev change 2015-07-27 10:58:15 +02:00
Matthew Honnibal 6b586cdad4 * Change lexemes.bin format. Add a header specifying size of LexemeC and number of lexemes, and don't have the redundant orth information. 2015-07-27 08:31:51 +02:00
Matthew Honnibal 8e4c69ee8c * Add is_oov property, and fix up handling of attributes 2015-07-27 01:50:06 +02:00
Matthew Honnibal fc268f03eb * Assert against null pointer exceptions in vocab 2015-07-27 01:00:10 +02:00
Matthew Honnibal 0f093fdb30 * Fix get_by_orth for py3 2015-07-26 19:26:41 +02:00
Matthew Honnibal ceeda5a739 * Fix get_by_orth for py3 2015-07-26 18:39:27 +02:00
Matthew Honnibal 6bb96c122d * Host IS_ flags in attrs.pxd, and add properties for them on Token and Lexeme objects 2015-07-26 16:37:16 +02:00
Matthew Honnibal 7eb2446082 * Return empty lexeme on empty string 2015-07-26 00:18:30 +02:00
Matthew Honnibal fd525f0675 * Pass OOV probability around 2015-07-25 23:29:51 +02:00
Matthew Honnibal 22028602a9 * Add unicode_literals declaration in vocab.pyx 2015-07-23 13:24:20 +02:00
Matthew Honnibal a7c4d72e83 * Add serializer property to Vocab, and lazy-load it. Add get_by_orth method. 2015-07-23 01:18:19 +02:00
Matthew Honnibal 109106a949 * Replace UniStr, using unicode objects instead 2015-07-22 04:52:05 +02:00
Matthew Honnibal 1f7170e0e1 * Reinstate the fixed vocabulary --- words are only added to the lexicon in init_model, after that we create LexemeC structs with the Pool given to us. 2015-07-20 01:37:34 +02:00
Matthew Honnibal 317cbbc015 * Serialization round trip now working with decent API, but with rough spots in the organisation and requiring vocabulary to be fixed ahead of time. 2015-07-19 15:18:17 +02:00
Matthew Honnibal 82d84b0f2b * Index lexemes by orth, instead of a lexemes vector. Breaks the mechanism for deciding not to own LexemeC structs during parsing. Need to reinstate this. 2015-07-18 22:42:15 +02:00
Matthew Honnibal c2c83120d4 * Remove codec property from Vocab 2015-07-17 16:40:11 +02:00
Matthew Honnibal db9dfd2e23 * Major refactor of serialization. Nearly complete now. 2015-07-17 01:27:54 +02:00