Commit Graph

9151 Commits

Author SHA1 Message Date
Daniël de Kok
8f84e6ea8a Add type annotations for activations in predict/set_annotations 2022-08-30 10:07:33 +02:00
Daniël de Kok
8c2652d788 Remove TrainablePipe.activations
We do not need to enumerate the activations anymore since `store_activations` is
`bool`.
2022-08-29 16:47:13 +02:00
Daniël de Kok
aea53378dc Make the TrainablePipe.store_activations property a bool
This means that we can also bring back `store_activations` setter.
2022-08-29 16:41:11 +02:00
Daniël de Kok
1cfbb934ed TextCat.predict: return dict 2022-08-05 14:40:42 +02:00
Daniël de Kok
c792019350
Apply suggestions from code review
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
2022-08-05 14:17:07 +02:00
Daniël de Kok
75d76cb2a3 set_annotations: add type annotations 2022-08-05 14:09:30 +02:00
Daniël de Kok
ce36f345db _store_activations: make kwarg-only, remove doc_scores_lens arg 2022-08-05 12:00:01 +02:00
Daniël de Kok
57caae8393 EntityLinker: add type annotations to _add_activations 2022-08-05 11:48:04 +02:00
Daniël de Kok
230264daa0 Revert "Use dict comprehension suggested by @svlandeg"
This reverts commit 6e7b958f70.
2022-08-05 10:33:20 +02:00
Daniël de Kok
6e7b958f70 Use dict comprehension suggested by @svlandeg 2022-08-04 17:13:52 +02:00
Daniël de Kok
51f72e41ec Replace store_activations setter by set_store_activations method
Setters that take a different type than what the getter returns are still
problematic for MyPy. Replace the setter by a method, so that type inference
works everywhere.
2022-08-04 15:14:22 +02:00
Daniël de Kok
288d27e17e Merge remote-tracking branch 'upstream/v4' into store-activations 2022-08-01 09:30:04 +02:00
Daniël de Kok
e581eeac34
precompute_hiddens/Parser: look up CPU ops once (v4) (#11068)
* precompute_hiddens/Parser: look up CPU ops once

* precompute_hiddens: make cpu_ops private
2022-07-29 15:12:19 +02:00
Daniël de Kok
1ff683a50b Merge remote-tracking branch 'upstream/master' into merge-master-v4-20220728 2022-07-28 13:53:59 +02:00
ninjalu
95a1b8aca6
add additional REL_OP (#10371)
* add additional  REL_OP

* change to condition and new rel_op symbols

* add operators to docs

* add the anchor while we're in here

* add tests

Co-authored-by: Peter Baumgartner <5107405+pmbaumgartner@users.noreply.github.com>
2022-07-27 13:16:44 +02:00
Edward
360a702ecd
Add parent argument (#11210) 2022-07-26 14:35:18 +02:00
Adriane Boyd
5c2a00cef0
Set version to v3.4.1 (#11209) 2022-07-26 12:52:38 +02:00
Daniël de Kok
4ee8a06149
Fix compatibility with CuPy 9.x (#11194)
After the precomputable affine table of shape [nB, nF, nO, nP] is
computed, padding with shape [1, nF, nO, nP] is assigned to the first
row of the precomputed affine table. However, when we are indexing the
precomputed table, we get a row of shape [nF, nO, nP]. CuPy versions
before 10.0 cannot paper over this shape difference.

This change fixes compatibility with CuPy < 10.0 by squeezing the first
dimension of the padding before assignment.
2022-07-26 10:52:01 +02:00
Adriane Boyd
e5990db713 Revert "Temporarily skip tests that require models/compat"
This reverts commit d9320db7db.
2022-07-25 18:12:18 +02:00
Madeesh Kannan
ba18d2913d
Morphology/Morphologizer optimizations and refactoring (#11024)
* `Morphology`: Refactor to use C types, reduce allocations, remove unused code

* `Morphologzier`: Avoid unnecessary sorting of morpho features

* `Morphologizer`: Remove execessive reallocations of labels, improve hash lookups of labels, coerce `numpy` numeric types to native ints
Update docs

* Remove unused method

* Replace `unique_ptr` usage with `shared_ptr`

* Add type annotations to internal Python methods, rename `hash` variable, fix typos

* Add comment to clarify implementation detail

* Fix return type

* `Morphology`: Stop early when splitting fields and values
2022-07-15 11:14:08 +02:00
Nicolai Bjerre Pedersen
2fa983aa2e
Fix span typings (#11119)
Add id, id_ to span.pyi.
2022-07-12 13:47:35 +02:00
Peter Baumgartner
36cb2029a9
displaCy Spans Vertical Alignment Fix 2 (#11092)
* add in span render slot fix

* fix spacing off by one

* rm demo

* adjust comments

* fix whitespace and overlap issue
2022-07-08 19:20:13 +02:00
github-actions[bot]
e7fd06bdbe
Auto-format code with black (#11099)
Co-authored-by: explosion-bot <explosion-bot@users.noreply.github.com>
2022-07-08 18:43:25 +09:00
Daniël de Kok
403b1f1312 Add store_activations to docstrings. 2022-07-06 16:15:13 +02:00
Daniël de Kok
5eeb2e8f90 Start errors/warnings at higher numbers to avoid merge conflicts
Between the master and v4 branches.
2022-07-06 15:56:21 +02:00
Daniël de Kok
a06cbae70d
precompute_hiddens/Parser: do not look up CPU ops (3.4) (#11069)
* precompute_hiddens/Parser: do not look up CPU ops

`get_ops("cpu")` is quite expensive. To avoid this, we want to cache the
result as in #11068. However, for 3.x we do not want to change the ABI.
So we avoid the expensive lookup by using NumpyOps. This should have a
minimal impact, since `get_ops("cpu")` was only used when the model ops
were `CupyOps`. If the ops are `AppleOps`, we are still passing through
the correct BLAS implementation.

* _NUMPY_OPS -> NUMPY_OPS
2022-07-05 10:53:42 +02:00
Madeesh Kannan
d36d66b7ca
Increase test deadline to 30 minutes to prevent spurious test failures (#11070)
* Increase test deadline to 30 minutes to prevent spurious test failures

* Reduce deadline to 2 minutes
2022-07-04 18:37:09 +02:00
kadarakos
5240baccfe
dont use get_array_module (#11056) 2022-07-04 17:15:33 +02:00
Raphael Mitsch
e9eb59699f
NEL confidence threshold (#11016)
* Add base for NEL abstention threshold mechanism.

* Add abstention threshold to entity linker. Add test.

* Fix entity linking tests.

* Changed abstention default threshold from 0 to None.

* Fix default values for abstention thresholds.

* Fix mypy errors.

* Replace assertion with raise of proper error code.

* Simplify threshold check. Remove thresholding from EntityLinker_v1.

* Rename test.

* Update spacy/pipeline/entity_linker.py

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

* Update spacy/pipeline/entity_linker.py

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

* Make E1043 configurable.

* Update docs.

* Rephrase description in docs. Adjusting error code message.

Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
2022-07-04 17:05:21 +02:00
Madeesh Kannan
59c763eec1
StringStore-related optimizations (#10938)
* `strings`: More roubust type checking of keys/IDs, coerce `int`-like types to `hash_t`

* Preserve existing public API behaviour

* Fix return type

* Replace `bool` with `bint`, rename to `_try_coerce_to_hash`, replace `id` with `hash`

* Avoid unnecessary re-encoding and re-calculation of strings and hashs respectively

* Rename variables named `hash`
Add comment on early return
2022-07-04 15:04:03 +02:00
explosion-bot
7e55a51314 Auto-format code with black 2022-07-01 08:04:32 +00:00
Madeesh Kannan
eaf66e7431
Add NVTX ranges to TrainablePipe components (#10965)
* `TrainablePipe`: Add NVTX range decorator

* Annotate `TrainablePipe` subclasses with NVTX ranges

* Export function signature to allow introspection of args in tests

* Revert "Annotate `TrainablePipe` subclasses with NVTX ranges"

This reverts commit d8684f7372.

* Revert "Export function signature to allow introspection of args in tests"

This reverts commit f4405ca3ad.

* Revert "`TrainablePipe`: Add NVTX range decorator"

This reverts commit 26536eb6b8.

* Add `spacy.pipes_with_nvtx_range` pipeline callback

* Show warnings for all missing user-defined pipe functions that need to be annotated
Fix imports, typos

* Rename `DEFAULT_ANNOTATABLE_PIPE_METHODS` to `DEFAULT_NVTX_ANNOTATABLE_PIPE_METHODS`
Reorder import

* Walk model nodes directly whilst applying NVTX ranges
Ignore pipe method wrapper when applying range
2022-06-30 11:28:12 +02:00
Adriane Boyd
3fe9f47de4
Revert "disable failing test because Stanford servers are down (#11015)" (#11054)
This reverts commit f8116078ce.
2022-06-30 11:24:54 +02:00
Shen Qin
be00db6645
Addition of min_max quantifier in matcher {n,m} (#10981)
* Min_max_operators
1. Modified API and Usage for spaCy website to include min_max operator
2. Modified matcher.pyx to include min_max function {n,m} and its variants
3. Modified schemas.py to include min_max validation error
4. Added test cases to test_matcher_api.py, test_matcher_logic.py and test_pattern_validation.py

* attempt to fix mypy/pydantic compat issue

* formatting

* Update spacy/tests/matcher/test_pattern_validation.py

Co-authored-by: Source-Shen <82353723+Source-Shen@users.noreply.github.com>
Co-authored-by: svlandeg <svlandeg@github.com>
Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
2022-06-30 11:01:58 +02:00
Daniël de Kok
0ff14aabce
vectors: avoid expensive comparisons between numpy ints and Python ints (#10992)
* vectors: avoid expensive comparisons between numpy ints and Python ints

* vectors: avoid failure on lists of ints

* Convert another numpy int to Python
2022-06-29 12:58:31 +02:00
Peter Baumgartner
dd038b536c
fix to horizontal space (#10994) 2022-06-28 20:42:40 +02:00
Adriane Boyd
24f4908fce
Update vector handling in similarity methods (#11013)
Distinguish between vectors that are 0 vs. missing vectors when warning
about missing vectors.

Update `Doc.has_vector` to match `Span.has_vector` and
`Token.has_vector` for cases where the vocab has vectors but none of the
tokens in the container have vectors.
2022-06-28 19:50:47 +02:00
Madeesh Kannan
1d5cad0b42
Example.get_aligned_parse: Handle unit and zero length vectors correctly (#11026)
* `Example.get_aligned_parse`: Do not squeeze gold token idx vector
Correctly handle zero-size vectors passed to `np.vectorize`

* Add tests

* Use `Doc` ctor to initialize attributes

* Remove unintended change

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

* Remove unused import

Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
2022-06-28 19:42:58 +02:00
Richard Hudson
a9559e7435
Handle Cyrillic combining diacritics (#10837)
* Handle Russian, Ukrainian and Bulgarian

* Corrections

* Correction

* Correction to comment

* Changes based on review

* Correction

* Reverted irrelevant change in punctuation.py

* Remove unnecessary group

* Reverted accidental change
2022-06-28 15:35:32 +02:00
Zackere
8ffff18ac4
Try cloning repo from main & master (#10843)
* Try cloning repo from main & master

* fixup! Try cloning repo from main & master

* fixup! fixup! Try cloning repo from main & master

* refactor clone and check for repo:branch existence

* spacing fix

* make mypy happy

* type util function

* Update spacy/cli/project/clone.py

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

Co-authored-by: Peter Baumgartner <5107405+pmbaumgartner@users.noreply.github.com>
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
2022-06-28 09:11:15 -04:00
Daniël de Kok
df513beddb Merge remote-tracking branch 'upstream/v4' into store-activations 2022-06-27 19:21:33 +02:00
Daniël de Kok
1605ef7319 Merge remote-tracking branch 'upstream/master' into merge-master-v4-20220627-2 2022-06-27 17:45:45 +02:00
Daniël de Kok
508b96fdc7 Merge remote-tracking branch 'upstream/master' into store-activations 2022-06-27 14:14:44 +02:00
github-actions[bot]
4155a59d47
Auto-format code with black (#11022)
Co-authored-by: explosion-bot <explosion-bot@users.noreply.github.com>
2022-06-27 09:35:35 +02:00
Adriane Boyd
738b38064f
Merge pull request #11021 from adrianeboyd/chore/v3.4.0
Set version to v3.4.0
2022-06-24 14:54:16 +02:00
Madeesh Kannan
8f1ba4de58
Backport parser/alignment optimizations from feature/refactor-parser (#10952) 2022-06-24 13:39:52 +02:00
Adriane Boyd
d9320db7db Temporarily skip tests that require models/compat 2022-06-24 11:20:53 +02:00
Daniël de Kok
c8a12c5d85 Extend tagger and senter tests
So that they, like the other tests, also check that we get no
activations if no activations were requested.
2022-06-24 10:18:00 +02:00
Adriane Boyd
bffe54d02b Set version to v3.4.0 2022-06-24 08:48:58 +02:00
Daniël de Kok
009a96085b parser/ner: do not currently support returning activations 2022-06-23 15:53:20 +02:00