Commit Graph

15831 Commits

Author SHA1 Message Date
Paul O'Leary McCann
f8d05e5aa1 make spacy-legacy version explicit in error message 2023-01-23 18:25:33 +09:00
Paul O'Leary McCann
b536ed5978 Remove azure build workarounds 2023-01-23 18:24:52 +09:00
Paul O'Leary McCann
97b31bcf0b Raise spacy legacy version in setup.cfg 2023-01-23 18:24:29 +09:00
Paul O'Leary McCann
d08961b94f Fix requirements.txt, increasing spacy-legacy version 2023-01-23 18:23:05 +09:00
Paul O'Leary McCann
98d82608cf Revert "TODO REVERT Mark dependency check as xfail"
This reverts commit d4863ec156.
2023-01-23 18:22:14 +09:00
Paul O'Leary McCann
d773ef0952 Clean up debugging output
The manual install step with the URL fragment seems to have caused
issues on Windows due to the = in the URL being misinterpreted. On the
other hand, removing it seems to mean the git version of spacy-legacy
isn't actually installed.

This PR removes the URL fragment but keeps the direct command-line
install. Additionally, since it looks like this job is configured to use
the default shell (and not bash), it removes a comment that upsets the
Windows cmd shell.
2022-12-29 17:29:44 +09:00
Paul O'Leary McCann
6801e80225 Revert "Remove debugging output / install step"
This reverts commit 923ea7448b.
2022-12-29 17:27:55 +09:00
Paul O'Leary McCann
923ea7448b Remove debugging output / install step 2022-12-29 16:54:28 +09:00
Paul O'Leary McCann
d4863ec156 TODO REVERT Mark dependency check as xfail
spacy-legacy is specified as a git checkout in requirements.txt while
this PR is in progress, which makes the consistency check here fail.
2022-12-29 16:31:07 +09:00
Paul O'Leary McCann
0892b8f5d1 Merge branch 'v4' into cleanup/move-legacy-entity-linker 2022-12-29 16:14:53 +09:00
Daniël de Kok
20b63943f5
Adjust to new Schedule class and pass scores to Optimizer (#12008)
* Adjust to new `Schedule` class and pass scores to `Optimizer`

Requires https://github.com/explosion/thinc/pull/804

* Bump minimum Thinc requirement to 9.0.0.dev1
2022-12-29 08:03:24 +01:00
Paul O'Leary McCann
e4bb40c564 Use same thinc version as v4 branch 2022-12-29 14:33:48 +09:00
Paul O'Leary McCann
12adc8eb66 Update thinc requirements
This makes it the same as it was before the merge, >=8.1.0,<8.2.0.
2022-12-29 14:00:01 +09:00
Paul O'Leary McCann
0638236725 Merge branch 'v4' into cleanup/move-legacy-entity-linker 2022-12-29 13:45:22 +09:00
Paul O'Leary McCann
7eef9a5f7a TODO REVERT Change thinc spec
Want to see what happens to the installed thinc spec with this change.
2022-12-29 13:39:52 +09:00
Paul O'Leary McCann
6d46067c7f Log requirements at start of build 2022-12-28 18:55:18 +09:00
Paul O'Leary McCann
e11b71739a Cat requirements.txt to confirm contents
In the branch, the thinc version spec is `thinc>=8.1.0,<8.2.0`. But in
the logs, it's clear that a development release of 9.0 is being
installed. It's not clear why that would happen.
2022-12-28 18:34:57 +09:00
Paul O'Leary McCann
41e2ca893a Try explicitly uninstalling spacy-legacy first 2022-12-28 17:40:09 +09:00
Paul O'Leary McCann
37aacec300 Try installing directly in workflow 2022-12-28 14:30:55 +09:00
Paul O'Leary McCann
d8d7f10b1d Add more logging 2022-12-28 13:19:58 +09:00
Paul O'Leary McCann
4f1286b18e Revert "Add comment to Python to trigger tests"
This reverts commit 11840fc598.
2022-12-28 12:53:09 +09:00
Paul O'Leary McCann
b7d890b557 Revert "TODO REVERT This is a commit with logic changes to trigger tests"
This reverts commit 689fae71f3.
2022-12-28 12:52:55 +09:00
Paul O'Leary McCann
ce5831b5d2 Remove pipe from YAML
Works locally, but possibly this is causing a quoting error or
something.
2022-12-27 16:47:53 +09:00
Paul O'Leary McCann
689fae71f3 TODO REVERT This is a commit with logic changes to trigger tests 2022-12-27 16:46:32 +09:00
Paul O'Leary McCann
11840fc598 Add comment to Python to trigger tests 2022-12-27 15:40:55 +09:00
Paul O'Leary McCann
7a9db60913 Modify requirements.txt to trigger tests 2022-12-27 15:32:14 +09:00
Paul O'Leary McCann
1d08dce32c Add temporary step to log installed spacy-legacy version 2022-12-27 15:23:05 +09:00
Daniël de Kok
d30ba9b7b8
Merge pull request #12015 from danieldk/chore/v4-merge-master-20221222
Merge master into v4
2022-12-22 11:22:33 +01:00
Daniël de Kok
2f08deea2a Fix fallout from a previous merge 2022-12-22 10:23:31 +01:00
Daniël de Kok
207565a788 Merge remote-tracking branch 'upstream/master' into chore/v4-merge-master-20221222 2022-12-22 10:08:54 +01:00
Paul O'Leary McCann
4201367cfa Use spacy-legacy pr for CI
This will need to be reverted before merging.
2022-12-22 14:40:22 +09:00
Raphael Mitsch
eef3d950b4
Fix SpanGroup and Span typing (#12009)
* Correct Span.label, Span.kb_id types. Fix SpanGroup.__iter__().

* Extend test.

* Rename test. Fix typo.

* Add comment.

* Fix types for Span.label, Span.kb_id, Span.char_span().

* Update spacy/tests/doc/test_span_group.py

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

* Update docs.

* Fix typo.

* Update spacy/tokens/span_group.pyx

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

Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
2022-12-21 18:54:27 +01:00
Paul O'Leary McCann
d43d92c365 Move Entity Linker v1 component to spacy-legacy
This is a follow up to #11889 that moves the component instead of
removing it.

In general, we never import from spacy-legacy in spaCy proper. However,
to use this component, that kind of import will be necessary. I was able
to test this without issues, but is this current import strategy
acceptable? Or should we put the component in a registry?
2022-12-21 15:15:52 +09:00
kadarakos
c223cd7a86
Add apply CLI (#11376)
* annotate cli first try

* add batch-size and n_process

* rename to apply

* typing fix

* handle file suffixes

* walk directories

* support jsonl

* typing fix

* remove debug

* make suffix optional for walk

* revert unrelated

* don't warn but raise

* better error message

* minor touch up

* Update spacy/tests/test_cli.py

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

* Update spacy/cli/apply.py

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

* Update spacy/cli/apply.py

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

* update tests and bugfix

* add force_overwrite

* typo

* fix adding .spacy suffix

* Update spacy/cli/apply.py

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

* Update spacy/cli/apply.py

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

* Update spacy/cli/apply.py

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

* store user data and rename cmd arg

* include test for user attr

* rename cmd arg

* better help message

* documentation

* prettier

* black

* link fix

* Update spacy/cli/apply.py

Co-authored-by: Paul O'Leary McCann <polm@dampfkraft.com>

* Update website/docs/api/cli.md

Co-authored-by: Paul O'Leary McCann <polm@dampfkraft.com>

* Update website/docs/api/cli.md

Co-authored-by: Paul O'Leary McCann <polm@dampfkraft.com>

* Update website/docs/api/cli.md

Co-authored-by: Paul O'Leary McCann <polm@dampfkraft.com>

* addressing reviews

* dont quit but warn

* prettier

Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
Co-authored-by: Paul O'Leary McCann <polm@dampfkraft.com>
2022-12-20 17:11:33 +01:00
Jos Polfliet
18ffe5bbd6
Update stop_words.py (#11997)
fix typo in "aangaande"
2022-12-19 16:17:49 +01:00
cfuerbachersparks
3a2b655a29
Update lexeme.md (#11994)
Change suffix_ string to end
2022-12-19 10:33:38 +01:00
Daniël de Kok
f9308aae13
Fix v4 branch to build against Thinc v9 (#11921)
* Move `thinc.extra.search` to `spacy.pipeline._parser_internals`

Backport of:
https://github.com/explosion/spaCy/pull/11317

Co-authored-by: Madeesh Kannan <shadeMe@users.noreply.github.com>

* Replace references to `thinc.backends.linalg` with `CBlas`

Backport of:
https://github.com/explosion/spaCy/pull/11292

Co-authored-by: Madeesh Kannan <shadeMe@users.noreply.github.com>

* Use cross entropy from `thinc.legacy`

* Require thinc>=9.0.0.dev0,<9.1.0

Co-authored-by: Madeesh Kannan <shadeMe@users.noreply.github.com>
2022-12-17 14:32:19 +01:00
Adriane Boyd
c9d9d6847f
Update build constraints for python 3.11 (#11981) 2022-12-15 10:55:01 +01:00
Adriane Boyd
e5c7f3b077
CI: Install thinc-apple-ops through extra (#11963) 2022-12-12 10:13:10 +01:00
Edward
ca75190a3d
Custom extensions for spans with equal boundaries (#11429)
* Init

* Fix return type for mypy

* adjust types and improve setting new attributes

* Add underscore changes to json conversion

* Add test and underscore changes to from_docs

* add underscore changes and test to span.to_doc

* update return values

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

* Add types to function

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

* adjust formatting

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

* shorten return type

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

* add helper function to improve readability

* Improve code and add comments

* rerun azure tests

* Fix tests for json conversion

Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
2022-12-12 08:55:53 +01:00
Adriane Boyd
0591e67265
Cast to uint64 for all array-based doc representations (#11933)
* Convert all individual values explicitly to uint64 for array-based doc representations

* Temporarily test with latest numpy v1.24.0rc

* Remove unnecessary conversion from attr_t

* Reduce number of individual casts

* Convert specifically from int32 to uint64

* Revert "Temporarily test with latest numpy v1.24.0rc"

This reverts commit eb0e3c5006.

* Also use int32 in tests
2022-12-12 08:45:35 +01:00
Adriane Boyd
8c291ace0c
Extend to wasabi v1.1 (#11945)
* Extend to wasabi v1.1

* Temporarily run mypy and tests with newest wasabi

* Temporarily skip check requirements test

* Revert "Temporarily skip check requirements test"

This reverts commit 44f4ce20a8.

* Revert "Temporarily run mypy and tests with newest wasabi"

This reverts commit e677a2257c.
2022-12-12 08:38:36 +01:00
github-actions[bot]
f22fc7a113
Auto-format code with black (#11955)
Co-authored-by: explosion-bot <explosion-bot@users.noreply.github.com>
2022-12-09 10:15:52 +01:00
Madeesh Kannan
f5aabaf7d6
Remove unused, experimental multi-task components (#11919)
* Remove experimental multi-task components

These are incomplete implementations and are not usable in their current state.

* Remove orphaned error message

* Switch ubuntu-latest to ubuntu-20.04 in main tests (#11928)

* Switch ubuntu-latest to ubuntu-20.04 in main tests

* Only use 20.04 for 3.6

* Revert "Switch ubuntu-latest to ubuntu-20.04 in main tests (#11928)"

This reverts commit 77c0fd7b17.

Co-authored-by: Paul O'Leary McCann <polm@dampfkraft.com>
2022-12-08 13:24:45 +01:00
Paul O'Leary McCann
d60997febb
Remove old model shortcuts (#11916)
* Remove old model shortcuts

* Remove error, docs warnings about shortcuts

* Fix import in util

Accidentally deleted the whole import and not just the old part...

* Change universe example to v3 style

* Switch ubuntu-latest to ubuntu-20.04 in main tests (#11928)

* Switch ubuntu-latest to ubuntu-20.04 in main tests

* Only use 20.04 for 3.6

* Update some model loading in Universe

* Add v2 tag to neuralcoref

* Use the spacy-version feature instead of a v2 tag

Co-authored-by: svlandeg <svlandeg@github.com>
2022-12-08 11:45:52 +01:00
Paul O'Leary McCann
6b9af38eeb
Remove all references to "begin_training" (#11943)
When v3 was released, `begin_training` was renamed to `initialize`.
There were warnings in the code and docs about that. This PR removes
them.
2022-12-08 11:43:52 +01:00
vincent d warmerdam
6d2ca1ab3a
Update custom solutions links (#11903)
* Update custom solutions

Will now point to https://explosion.ai/custom-solutions

* added-sidebar

* added-analysis-to-readme

* update-landing-page
2022-12-07 16:02:09 +01:00
Paul O'Leary McCann
73919336fb
Remove spacy-sentence-segmenter from Universe (#11932) 2022-12-07 15:56:03 +01:00
Paul O'Leary McCann
5c3a60e8f4
Add in errors used in the beam code that were removed at some point (#11935)
I don't think there's any way to use the beam code at the moment, but as
long as it's around the errors it refers to should also be present.
2022-12-07 15:52:35 +01:00
Sofie Van Landeghem
60379cec65
Merge pull request #11929 from svlandeg/copy_v4
sync v4 with latest master
2022-12-07 15:24:07 +01:00