Commit Graph

14763 Commits

Author SHA1 Message Date
Paul O'Leary McCann
0f01f46e02
Update Cython string types (#9143)
* Replace all basestring references with unicode

`basestring` was a compatability type introduced by Cython to make
dealing with utf-8 strings in Python2 easier. In Python3 it is
equivalent to the unicode (or str) type.

I replaced all references to basestring with unicode, since that was
used elsewhere, but we could also just replace them with str, which
shoudl also be equivalent.

All tests pass locally.

* Replace all references to unicode type with str

Since we only support python3 this is simpler.

* Remove all references to unicode type

This removes all references to the unicode type across the codebase and
replaces them with `str`, which makes it more drastic than the prior
commits. In order to make this work importing `unicode_literals` had to
be removed, and one explicit unicode literal also had to be removed (it
is unclear why this is necessary in Cython with language level 3, but
without doing it there were errors about implicit conversion).

When `unicode` is used as a type in comments it was also edited to be
`str`.

Additionally `coding: utf8` headers were removed from a few files.
2021-09-13 17:02:17 +02:00
Adriane Boyd
c5de9b463a
Update custom tokenizer APIs and pickling (#8972)
* Fix incorrect pickling of Japanese and Korean pipelines, which led to
the entire pipeline being reset if pickled

* Enable pickling of Vietnamese tokenizer

* Update tokenizer APIs for Chinese, Japanese, Korean, Thai, and
Vietnamese so that only the `Vocab` is required for initialization
2021-08-19 14:37:47 +02:00
Adriane Boyd
b278f31ee6
Document scorers in registry and components from #8766 (#8929)
* Document scorers in registry and components from #8766

* Update spacy/pipeline/lemmatizer.py

Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>

* Update website/docs/api/dependencyparser.md

Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>

* Reformat

Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
2021-08-12 12:50:03 +02:00
Edward
944ad6b1d4
Add new parameter for saving every n epoch in pretraining (#8912)
* Add parameter for saving every n epoch

* Add new parameter in schemas

* Add new parameter in default_config

* Adjust schemas

* format code
2021-08-12 11:14:48 +02:00
Adriane Boyd
f99d6d5e39
Refactor scoring methods to use registered functions (#8766)
* Add scorer option to components

Add an optional `scorer` parameter to all pipeline components. If a
scoring function is provided, it overrides the default scoring method
for that component.

* Add registered scorers for all components

* Add `scorers` registry
* Move all scoring methods outside of components as independent
  functions and register
* Use the registered scoring methods as defaults in configs and inits

Additional:

* The scoring methods no longer have access to the full component, so
  use settings from `cfg` as default scorer options to handle settings
  such as `labels`, `threshold`, and `positive_label`
* The `attribute_ruler` scoring method no longer has access to the
  patterns, so all scoring methods are called
* Bug fix: `spancat` scoring method is updated to set `allow_overlap` to
  score overlapping spans correctly

* Update Russian lemmatizer to use direct score method

* Check type of cfg in Pipe.score

* Fix check

* Update spacy/pipeline/sentencizer.pyx

Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>

* Remove validate_examples from scoring functions

* Use Pipe.labels instead of Pipe.cfg["labels"]

Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
2021-08-10 15:13:39 +02:00
fgaim
ee011ca963
Update Tigrinya ትግርኛ language support (#8900)
* Add missing punctuation for Tigrinya and Amharic

* Fix numeral and ordinal numbers for Tigrinya

 - Amharic was used in many cases
 - Also fixed some typos

* Update Tigrinya stop-words

* Contributor agreement for fgaim

* Fix typo in "ti" lang test

* Remove multi-word entries from numbers and ordinals
2021-08-10 13:55:08 +02:00
Dimitar Ganev
733ffe439d
Improve the stop words and the tokenizer exceptions in Bulgarian language. (#8862)
* Add more stop words and Improve the readability

* Add and categorize the tokenizer exceptions for `bg` lang

* Create syrull.md

* Add references for the additional stop words and tokenizer exc abbrs
2021-08-10 13:44:23 +02:00
Adriane Boyd
415dee587c
Merge pull request #8911 from adrianeboyd/chore/update-develop-from-master-v3.1-1
Update develop from master
2021-08-09 15:41:36 +02:00
Adriane Boyd
a79888ed67 Merge remote-tracking branch 'upstream/master' into chore/update-develop-from-master-v3.1-1 2021-08-09 13:13:13 +02:00
Paul O'Leary McCann
cac298471f
Fix #8902 (bad link in docs)
typo fix
2021-08-08 22:04:00 +09:00
Eduard Zorita
439f30faad
Add stub files for main cython classes (#8427)
* Add stub files for main API classes

* Add contributor agreement for ezorita

* Update types for ndarray and hash()

* Fix __getitem__ and __iter__

* Add attributes of Doc and Token classes

* Overload type hints for Span.__getitem__

* Fix type hint overload for Span.__getitem__

Co-authored-by: Luca Dorigo <dorigoluca@gmail.com>
2021-08-07 12:30:03 +02:00
github-actions[bot]
56d4d87aeb
Auto-format code with black (#8895)
Co-authored-by: explosion-bot <explosion-bot@users.noreply.github.com>
2021-08-06 13:38:06 +02:00
Kabir Khan
1dfffe5fb4
No output info message in train (#8885)
* Add info message that no output directory was provided in train

* Update train.py

* Fix logging
2021-08-05 09:21:22 +02:00
Adriane Boyd
fa2e7a4bbf
Fix spancat tests on GPU (#8872)
* Fix spancat tests on GPU

* Fix more spancat tests
2021-08-04 14:29:43 +02:00
Paul O'Leary McCann
77d698dcae
Fix check for RIGHT_ATTRS in dep matcher (#8807)
* Fix check for RIGHT_ATTRs in dep matcher

If a non-anchor node does not have RIGHT_ATTRS, the dep matcher throws
an E100, which says that non-anchor nodes must have LEFT_ID, REL_OP, and
RIGHT_ID. It specifically does not say RIGHT_ATTRS is required.

A blank RIGHT_ATTRS is also valid, and patterns with one will be
excepted. While not normal, sometimes a REL_OP is enough to specify a
non-anchor node - maybe you just want the head of another node
unconditionally, for example.

This change just sets RIGHT_ATTRS to {} if not present. Alternatively
changing E100 to state RIGHT_ATTRS is required could also be reasonable.

* Fix test

This test was written on the assumption that if `RIGHT_ATTRS` isn't
present an error will be raised. Since the proposed changes make it so
an error won't be raised this is no longer necessary.

* Revert test, update error message

Error message now lists missing keys, and RIGHT_ATTRS is required.

* Use list of required keys in error message

Also removes unused key param arg.
2021-08-04 09:20:41 +02:00
Adriane Boyd
941a591f3c
Pass excludes when serializing vocab (#8824)
* Pass excludes when serializing vocab

Additional minor bug fix:

* Deserialize vocab in `EntityLinker.from_disk`

* Add test for excluding strings on load

* Fix formatting
2021-08-03 14:42:44 +02:00
Adriane Boyd
175847f92c
Support list values and INTERSECTS in Matcher (#8784)
* Support list values and IS_INTERSECT in Matcher

* Support list values as token attributes for set operators, not just as
pattern values.

* Add `IS_INTERSECT` operator.

* Fix incorrect `ISSUBSET` and `ISSUPERSET` in schema and docs.

* Rename IS_INTERSECT to INTERSECTS
2021-08-02 19:39:26 +02:00
Adriane Boyd
fbbbda1954
Fix start/end chars for empty and out-of-bounds spans (#8816) 2021-08-02 19:07:19 +02:00
Adriane Boyd
9ad3b8cf8d
Only add sourced vectors hashes to meta if necessary (#8830) 2021-08-02 18:22:35 +02:00
Nick Sorros
0485cdefcc
Add logger debug for project push and pull (#8860)
* Add logger debug for project push and pull

* Sign contributor agreement
2021-08-02 18:13:53 +02:00
themrmax
de076194c4
Make ConsoleLogger flush after each logging line (#8810)
This is necessary to avoid "logging blackouts" when running training on Kubernetes pods
2021-08-02 14:33:38 +02:00
Ines Montani
30f20496d5
Merge pull request #8840 from polm/docs/evaluate-speed [ci skip] 2021-07-30 09:10:15 +10:00
Ines Montani
65d163fab5
Adjust formatting [ci skip] 2021-07-30 09:10:04 +10:00
Ines Montani
3a701d3645
Merge pull request #8841 from adrianeboyd/docs/ent-id-sep [ci skip]
Fix formatting of ent_id_sep in EntityRuler API docs
2021-07-30 09:09:25 +10:00
Ines Montani
f08be084fb
Merge pull request #8844 from thomashacker/bugfix/fix-doc-transformer-typo [ci skip]
Fix typo in Tok2VecTransformer example config
2021-07-30 09:08:59 +10:00
thomashacker
02258916c8 Fix example config typo for transformer architecture 2021-07-29 11:19:40 +02:00
Adriane Boyd
15b12f3e35 Fix formatting of ent_id_sep in EntityRuler API docs 2021-07-29 10:10:12 +02:00
Paul O'Leary McCann
a60cb13910 Update speed entry in metrics table 2021-07-29 16:35:19 +09:00
Paul O'Leary McCann
e125313a50 Revert "Add note about SPEED in output"
This reverts commit c92d268176.
2021-07-29 16:34:08 +09:00
Ines Montani
0a1e299d30
Merge pull request #8814 from polm/docs/migrate-lexeme-tables [ci skip] 2021-07-29 17:18:02 +10:00
Paul O'Leary McCann
c92d268176 Add note about SPEED in output
In #8823 it was pointed out that the `SPEED` value wasn't documented
anywhere.
2021-07-29 15:03:07 +09:00
Paul O'Leary McCann
8867e60fbb
Update website/docs/usage/v3.md
Co-authored-by: Ines Montani <ines@ines.io>
2021-07-29 14:56:56 +09:00
Adriane Boyd
8547514aa4
Remove labels from textcat component config example (#8815) 2021-07-27 13:14:38 +02:00
Paul O'Leary McCann
76ac95923a Add note to migration guide about lexeme tables (fix #7290)
This just adds the resolution from #6388 to the docs.
2021-07-27 19:19:25 +09:00
Paul O'Leary McCann
67ecdcc3ac
Update subset/superset docs (#8795)
* Update subset/superset docs

* Update website/docs/usage/rule-based-matching.md

Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>

Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
2021-07-27 12:08:46 +02:00
Adriane Boyd
81d3a1edb1
Use tokenizer URL_MATCH pattern in LIKE_URL (#8765) 2021-07-27 12:07:01 +02:00
Adriane Boyd
4f28190afe
Merge pull request #8813 from adrianeboyd/chore/develop-v3.2
Update develop for v3.2
2021-07-27 11:26:18 +02:00
Ines Montani
7f21c7dfa2
Merge pull request #8794 from explosion/autoblack
Auto-format code with black
2021-07-27 12:17:15 +10:00
Ines Montani
34c401f04f
Merge pull request #8801 from polm/fix/respect-no-skip (fixes #8796)
Respect the no_skip value
2021-07-27 12:16:47 +10:00
Ines Montani
134cb06af3
Merge pull request #8808 from kevinlu1248/master [ci skip]
Changed a CLI command in data-formats.md due to erroneous information
2021-07-27 12:15:16 +10:00
Ines Montani
9bf0d6f2fd
Merge pull request #8806 from Ledenel/master [ci skip]
fix typo
2021-07-27 12:14:22 +10:00
Kevin Lu
4a8e9e4e4e
Update data-formats.md 2021-07-25 22:58:53 -07:00
Ledenel
413f745c68 fix broken example in spaCy universe Chatterbot 2021-07-25 15:53:32 +00:00
Paul O'Leary McCann
284b530c63 Respect the no_skip value
Seems like the logic for this was just left out. See #8796.
2021-07-24 15:31:17 +09:00
explosion-bot
a58ab6ea22 Auto-format code with black 2021-07-23 08:04:09 +00:00
Adriane Boyd
6bbc2b1956
Reload train corpus in debug data after initialize (#8776) 2021-07-21 22:38:40 +02:00
Adriane Boyd
d48c01a6f7
Remove extraneous grc test file (#8768) 2021-07-20 15:51:15 +02:00
Sofie Van Landeghem
ffaead8fe0
bump to 3.1.1 2021-07-19 14:48:27 +02:00
Sofie Van Landeghem
83e27d262e
negative tag annotation (#8731)
* unit test to unlearn tag via negative annotation

* bump thinc to 8.0.8
2021-07-19 14:39:11 +02:00
Adriane Boyd
0e4b96c97e
Update lexeme ranks for loaded vectors (#8640)
Update the ranks for any lexemes that have been added to the vocab
before the vectors are added to the model.
2021-07-19 18:25:54 +10:00