6f54e59fe7
Fix util.filter_spans() to prefer first span in overlapping sam… ( #4414 )
...
* Update util.filter_spans() to prefer earlier spans
* Add filter_spans test for first same-length span
* Update entity relation example to refer to util.filter_spans()
2019-10-10 17:00:03 +02:00
da6e0de34f
fix attrs field in the matcher ( #4423 )
...
* raise specific error when removing a matcher rule that doesn't exist
* rephrasing
* ensure attrs is NULL when nr_attr == 0 + several fixes to prevent OOB
2019-10-10 15:20:59 +02:00
5efae495f1
Error when removing a matcher rule that doesn't exist ( #4420 )
...
* raise specific error when removing a matcher rule that doesn't exist
* rephrasing
2019-10-10 14:01:53 +02:00
fa95c030a5
Unify matcher get_ent_id and get_pattern_key ( #4415 )
...
This is basically stabbing blindly at the ghost match problem, but it at
least seems like there was a bug previously here --- so this should
hopefully be an improvement, even if it doesn't fix the ghost match
problem.
2019-10-09 15:26:31 +02:00
77643de2ca
Downgrade importlib_metadata requirement
2019-10-08 23:43:24 +02:00
5cbe21700b
Only show label scheme if not empty [ci skip]
2019-10-08 15:52:59 +02:00
8f76d6c9ef
Update transformer model details [ci skip]
2019-10-08 15:39:38 +02:00
dd30d3ec99
Add setuptools as runtime dependency
2019-10-08 12:46:59 +02:00
c4f95c1569
Update formatting and docstrings [ci skip]
2019-10-08 12:25:23 +02:00
ddd6fda59c
Add registry for model creation functions ('architectures') ( #4395 )
...
* Add architecture registry
* Add test for arch registry
* Add error for model architectures
2019-10-08 12:21:03 +02:00
650cbfe82d
multiprocessing pipe ( #1303 ) ( #4371 )
...
* refactor: separate formatting docs and golds in Language.update
* fix return typo
* add pipe test
* unpickleable object cannot be assigned to p.map
* passed test pipe
* passed test!
* pipe terminate
* try pipe
* passed test
* fix ch
* add comments
* fix len(texts)
* add comment
* add comment
* fix: multiprocessing of pipe is not supported in 2
* test: use assert_docs_equal
* fix: is_python3 -> is_python2
* fix: change _pipe arg to use functools.partial
* test: add vector modification test
* test: add sample ner_pipe and user_data pipe
* add warnings test
* test: fix user warnings
* test: fix warnings capture
* fix: remove islice import
* test: remove warnings test
* test: add stream test
* test: rename
* fix: multiproc stream
* fix: stream pipe
* add comment
* mp.Pipe seems to be able to use with relative small data
* test: skip stream test in python2
* sort imports
* test: add reason to skiptest
* fix: use pipe for docs communucation
* add comments
* add comment
2019-10-08 12:20:55 +02:00
14841d0aa6
Fix PhraseMatcher callback and add tests ( #4399 )
...
* Fix callback lookup in PhraseMatcher (string key rather than hash key)
* Add callback tests for Matcher and PhraseMatcher
2019-10-08 12:07:02 +02:00
fd4a5341b0
Fix ner_jsonl2json converter ( fix #4389 ) ( #4394 )
2019-10-08 00:52:45 +02:00
29f9fec267
Improve spacy pretrain ( #4393 )
...
* Support bilstm_depth arg in spacy pretrain
* Add option to ignore zero vectors in get_cossim_loss
* Use cosine loss in Cloze multitask
2019-10-07 23:34:58 +02:00
9cd6ca3e4d
Improve usage of pkg_resources and handling of entry points ( #4387 )
...
* Only import pkg_resources where it's needed
Apparently it's really slow
* Use importlib_metadata for entry points
* Revert "Only import pkg_resources where it's needed"
This reverts commit 5ed8c03afa8b30b579579f071f5c4002e12a17ec
2019-10-07 17:22:09 +02:00
d53a8d9313
Consider batch_size when sorting similar vectors ( #4388 )
2019-10-07 13:38:35 +02:00
a3509f67d4
Extend unicode character block for Sinhala ( #4378 )
...
* Extend unicode character block for Sinhala
* Add sentencizer tests for more languages
2019-10-07 13:17:03 +02:00
573e543e4a
Alphanumeric -> alphabetic [ci skip]
...
see ines/spacy-course#38
2019-10-06 13:30:01 +02:00
cbc2cee2c8
Improve URL_PATTERN and handling in tokenizer ( #4374 )
...
* Move prefix and suffix detection for URL_PATTERN
Move prefix and suffix detection for `URL_PATTERN` into the tokenizer.
Remove associated lookahead and lookbehind from `URL_PATTERN`.
Fix tokenization for Hungarian given new modified handling of prefixes
and suffixes.
* Match a wider range of URI schemes
2019-10-05 13:00:09 +02:00
e65dffd80b
Clarify serialization of extension attributes ( closes #4377 ) [ci skip]
2019-10-05 11:58:00 +02:00
fec9433044
Make PhraseMatcher.vocab consistent with Matcher.vocab ( closes #4373 )
2019-10-04 12:18:41 +02:00
e7ddc6f662
Add conda install for lookups [ci skip]
2019-10-03 17:52:53 +02:00
37ef874d8b
Set version to v2.2.1
2019-10-03 14:50:39 +02:00
4e7259c6cf
Bugfix initializing DocBin with attributes ( #4368 )
...
* docbin init fix + documentation fix + unit tests
* newline
* try with zlib instead of gzip (python 2 incompatibilities)
2019-10-03 14:48:45 +02:00
ce1d441de5
Add docs for Vectors.most_similar [ci skip]
2019-10-03 14:29:47 +02:00
1db79a33cb
most_similar() return the k most similar vectors ( #4364 )
...
* most_similar return n-most similar vectors
* updated most_similar comment
* add bintay contributor agreement
* sign bintay contributor agreement
* fix most_similar documentation typo
* fixed error in prune_vectors
* updated prune_vectors test
2019-10-03 14:09:44 +02:00
4159936720
Update README.md [ci skip]
2019-10-02 19:15:22 +02:00
e4782feae9
Update README.md [ci skip]
2019-10-02 18:49:55 +02:00
80cf385f65
Update v2-2.md [ci skip]
2019-10-02 16:58:21 +02:00
f8e606c303
Update README.md [ci skip]
2019-10-02 16:47:10 +02:00
12a941d841
Update binder version [ci skip]
2019-10-02 16:47:01 +02:00
2eb31012e7
Set version to v2.2.0
2019-10-02 14:40:06 +02:00
796072e560
Set version to v2.2.0.dev19
2019-10-02 12:51:29 +02:00
9d3ce7cba2
Ensure training doesn't crash with empty batches ( #4360 )
...
* unit test for previously resolved unflatten issue
* prevent batch of empty docs to cause problems
2019-10-02 12:50:47 +02:00
52b5912dbf
Tidy up [ci skip]
2019-10-02 12:05:59 +02:00
d82241218a
Make the default NER labels less model-specific [ci skip] ( #4361 )
2019-10-02 12:05:17 +02:00
dda86118bd
Update Ukrainian lemmatizer with new lookups ( #4359 )
...
* Update Ukrainian lemmatizer with new lookups
* Add missing import
Co-authored-by: Ines Montani <ines@ines.io>
2019-10-02 12:04:06 +02:00
b6670bf0c2
Use consistent spelling
2019-10-02 10:37:39 +02:00
208629615d
Auto-format
2019-10-02 10:37:04 +02:00
867e93aae2
Add Streamlit example [ci skip]
2019-10-02 01:21:20 +02:00
38b6e69389
Merge branch 'master' of https://github.com/explosion/spaCy
2019-10-01 22:28:25 +02:00
d4b63bb6dd
Set version to v2.2.0
2019-10-01 22:28:13 +02:00
9885b5ae68
Update spacy_lookups_data version [ci skip]
2019-10-01 22:21:21 +02:00
475e3188ce
Add docs on filtering overlapping spans for merging ( resolves #4352 ) [ci skip]
2019-10-01 21:59:50 +02:00
667f294627
Merge branch 'master' of https://github.com/explosion/spaCy
2019-10-01 21:37:25 +02:00
0dd127bb00
Update v2-2.md [ci skip]
2019-10-01 21:37:06 +02:00
64a9577d43
Set version to v2.2.0.dev17
2019-10-01 21:36:59 +02:00
cf65a80f36
Refactor lemmatizer and data table integration ( #4353 )
...
* Move test
* Allow default in Lookups.get_table
* Start with blank tables in Lookups.from_bytes
* Refactor lemmatizer to hold instance of Lookups
* Get lookups table within the lemmatization methods to make sure it references the correct table (even if the table was replaced or modified, e.g. when loading a model from disk)
* Deprecate other arguments on Lemmatizer.__init__ and expect Lookups for consistency
* Remove old and unsupported Lemmatizer.load classmethod
* Refactor language-specific lemmatizers to inherit as much as possible from base class and override only what they need
* Update tests and docs
* Fix more tests
* Fix lemmatizer
* Upgrade pytest to try and fix weird CI errors
* Try pytest 4.6.5
2019-10-01 21:36:03 +02:00
3297a19545
Warn in Tagger.begin_training if no lemma tables are available ( #4351 )
2019-10-01 15:13:55 +02:00
bc7e7db208
Fix wording [ci skip]
2019-10-01 14:20:44 +02:00
adrianeboyd