Ines Montani
30765674d0
Merge branch 'master' into develop
2021-01-30 12:20:28 +11:00
Pamphile ROY
41ee75ac6d
Remove --no-cache-dir when downloading models
...
When `--no-cache-dir` is present, it prevents caching to properly function.
If the user still wants to do this, there is the possibility to pass options with `user_pip_args`.
But you should not enforce options like these. In my case this is preventing some docker build (using buildkit caching) to have proper caching of models.
2021-01-29 15:37:44 +01:00
Ines Montani
bbf080dfe5
Merge pull request #6645 from bittlingmayer/patch-3
2021-01-30 01:26:28 +11:00
Adriane Boyd
bced6309e5
Add full exceptions with spaces
2021-01-29 14:27:22 +01:00
Ines Montani
7886d59c56
Add check for remove_listener method
2021-01-29 23:47:30 +11:00
Ines Montani
7694f76dd1
Update warning and mention replace_listeners
2021-01-29 23:46:01 +11:00
Ines Montani
94232aea08
Improve E889
2021-01-29 23:39:23 +11:00
Ines Montani
924396c20c
Merge branch 'feature/replace-listeners' of https://github.com/explosion/spaCy into feature/replace-listeners
2021-01-29 21:43:10 +11:00
Ines Montani
2102082478
Make Tok2Vec.remove_listener return bool
...
Whether listener was removed
2021-01-29 21:41:38 +11:00
Ines Montani
e766e8c56d
Apply suggestions from code review
...
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
2021-01-29 21:41:17 +11:00
Ines Montani
bc089b693c
Update tests
2021-01-29 19:38:09 +11:00
Ines Montani
325f47500d
Move replacement logic to Language.from_config
2021-01-29 19:37:04 +11:00
Ines Montani
0f3e3eedc2
Add Tok2vec.remove_listener
2021-01-29 19:36:38 +11:00
Ines Montani
99842387cb
Remove default value
2021-01-29 18:45:37 +11:00
Ines Montani
44b5542d14
Change method order
2021-01-29 18:42:41 +11:00
Ines Montani
8c15d1daec
Update and validate config first and exit early if paths don't exist
2021-01-29 18:24:47 +11:00
Ines Montani
bbb94b37c6
Update error handling and docstring
2021-01-29 16:27:49 +11:00
Ines Montani
01ecfbcc45
Merge branch 'develop' into feature/replace-listeners
2021-01-29 15:57:32 +11:00
Ines Montani
911dfcccfc
Add option to replace listeners for sourced components
2021-01-29 15:57:04 +11:00
Adriane Boyd
fcce3600ed
Forbid OP matching 2+ tokens in DependencyMatcher ( #6824 )
...
Instead of silently using only the first token in each matched span:
* Forbid `OP: ?/*/+` through `DependencyMatcher` validation
* As a fail-safe, add warning if a token match that's not exactly one
token long is found by a token pattern.
2021-01-29 08:52:01 +08:00
Sofie Van Landeghem
24a697abb8
avoid empty aliases and improve UX and docs ( #6840 )
2021-01-29 08:51:40 +08:00
Sofie Van Landeghem
837a4f53c2
Error handling in nlp.pipe ( #6817 )
...
* add error handler for pipe methods
* add unit tests
* remove pipe method that are the same as their base class
* have Language keep track of a default error handler
* cleanup
* formatting
* small refactor
* add documentation
2021-01-29 08:51:21 +08:00
Ines Montani
cc18f3f23c
Improve Example error handling for NER data ( #6835 )
...
* Improve Example error handling for NER data
* Fix conditional
2021-01-28 13:11:20 +11:00
Ines Montani
78d6ff4dd4
Update quickstart recommendations
2021-01-28 11:14:49 +11:00
Ines Montani
ec5f55aa5b
Update config generation defaults and transformers ( #6832 )
2021-01-27 23:56:33 +11:00
Adriane Boyd
4096a79de7
Add alignment mode error and fix Doc.char_span docs ( #6820 )
...
* Raise an error on an unrecognized alignment mode rather than
defaulting to `strict`
* Fix the `Doc.char_span` API doc alignment mode details
2021-01-27 23:40:42 +11:00
Sofie Van Landeghem
6b68ad027b
Fix beam NER resizing ( #6834 )
...
* move label check to sub methods
* add tests
2021-01-27 23:39:14 +11:00
Ines Montani
5ed51c9dd2
Merge pull request #6828 from explosion/master-tmp
2021-01-27 23:05:46 +11:00
Adriane Boyd
d17afb4826
Add Spanish rule-based lemmatizer ( #6833 )
...
* Initial Spanish lemmatizer
* Handle merged verb+pron(s) multi-word tokens
* Use VERB for AUX rule lookup
* Add morph to lemma cache key
* Fix aux lookups, minor refactoring
* Improve verb+pron handling
* Move verb+pron handling into its own method
* Check for exceptions (primarily for se)
* Collect pronouns in the same (not reversed) order
* Only add modified possible lemmas
2021-01-27 19:21:35 +08:00
Ines Montani
615dba9d99
Fix tokenizer exceptions
2021-01-27 22:11:42 +11:00
Ines Montani
abb24fdc0f
Merge pull request #6827 from explosion/feature/add-labels-implicitly
2021-01-27 21:34:58 +11:00
Ines Montani
80ba9eaf7d
Fix test
2021-01-27 21:29:02 +11:00
Ines Montani
e3f8be9a94
Update language data
2021-01-27 13:29:22 +11:00
Ines Montani
230e651ad6
Merge branch 'develop' into master-tmp
2021-01-27 13:26:29 +11:00
Matthew Honnibal
05050210f3
Dont add labels implicitly for parser
2021-01-27 13:04:47 +11:00
Matthew Honnibal
1d20e21f3e
Add labels implicitly for parser and ner
2021-01-27 12:54:47 +11:00
Matthew Honnibal
68b1c2984d
Test labels are added implicitly
2021-01-27 12:52:29 +11:00
Ines Montani
fabd3a3394
Tidy up code comments [ci skip]
2021-01-27 12:40:03 +11:00
Dhruv Naik
e7db07a0b9
Fix Span.char_span bug ( #6816 )
...
* Create dhruvrnaik.md
* add test for issue #6815
* bugfix for issue #6815
* update dhruvrnaik.md
* add span.vector test for #6815
2021-01-26 15:50:37 +08:00
Matthew Honnibal
e8674c5c42
Set version to v3.0.0rc5
2021-01-26 14:55:41 +11:00
Adriane Boyd
71a6350744
Implement overwrite param for all custom lemmatizers ( #6794 )
2021-01-26 14:53:43 +11:00
Adriane Boyd
2263bc7b28
Update develop from master for v3.0.0rc5 ( #6811 )
...
* Fix `spacy.util.minibatch` when the size iterator is finished (#6745 )
* Skip 0-length matches (#6759 )
Add hack to prevent matcher from returning 0-length matches.
* support IS_SENT_START in PhraseMatcher (#6771 )
* support IS_SENT_START in PhraseMatcher
* add unit test and friendlier error
* use IDS.get instead
* ensure span.text works for an empty span (#6772 )
* Remove unicode_literals
Co-authored-by: Santiago Castro <bryant@montevideo.com.uy>
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
2021-01-26 14:52:45 +11:00
Ines Montani
c0926c9088
WIP: Various small training changes ( #6818 )
...
* Allow output_path to be None during training
* Fix cat scoring (?)
* Improve error message for weighted None score
* Improve messages
So we can call this in other places etc.
* FIx output path check
* Use latest wasabi
* Revert "Improve error message for weighted None score"
This reverts commit 7059926763
2021-01-26 14:51:52 +11:00
Matthew Honnibal
f049df1715
Revert "Set annotations in update" ( #6810 )
...
* Revert "Set annotations in update (#6767 )"
This reverts commit e680efc7cc
2021-01-25 22:18:45 +08:00
Matthew Honnibal
42b117e561
Fix Doc.copy bugs ( #6809 )
...
* Dont let the Doc own LexemeC, to fix Doc.copy
* Copy doc.spans
* Copy doc.spans
2021-01-25 21:40:18 +08:00
Adriane Boyd
0f2de39efb
Fix types for exclude args in info CLI ( #6808 )
2021-01-25 20:00:22 +08:00
muratjumashev
2b19ebad59
Remove Kyrgyz chars fr. char_classes since Tatar ones already cover
2021-01-25 00:46:45 +06:00
muratjumashev
87168eb81f
Add tests
2021-01-24 20:56:16 +06:00
muratjumashev
53abf759ad
Fix punctuation
2021-01-24 20:54:22 +06:00
Matthew Honnibal
ffc371350a
Avoid assuming encode.get_dim('nO') is set in tok2vec ( #6800 )
2021-01-24 14:37:33 +11:00
