Commit Graph

120 Commits

Author SHA1 Message Date
Matthw Honnibal
5a272d9029 Add method to decode predicted characters 2019-10-21 03:56:15 +02:00
Matthw Honnibal
eba89f08bd Use chars loss in ClozeMultitask 2019-10-20 17:47:15 +02:00
Matthw Honnibal
e63f28079a Try 3 NER features 2019-10-07 16:51:03 +02:00
Matthw Honnibal
c8857181f8 Fix get labels for textcat 2019-10-07 16:50:15 +02:00
Matthw Honnibal
9dbaea1ab4 Use cosine loss in Cloze multitask 2019-10-06 19:23:46 +02:00
Sofie Van Landeghem
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
Ines Montani
b6670bf0c2 Use consistent spelling 2019-10-02 10:37:39 +02:00
Ines Montani
3297a19545 Warn in Tagger.begin_training if no lemma tables are available (#4351) 2019-10-01 15:13:55 +02:00
Sofie Van Landeghem
22b9e12159 Ensure the NER remains consistent after resizing (#4330)
* test and fix for second bug of issue 4042

* fix for first bug in 4042

* crashing test for Issue 4313

* forgot one instance of resize

* remove prints

* undo uncomment

* delete test for 4313 (uses third party lib)

* add fix for Issue 4313

* unit test for 4313
2019-09-27 20:57:13 +02:00
Matthew Honnibal
46c02d25b1 Merge changes to test_ner 2019-09-18 21:41:24 +02:00
tamuhey
875f3e5d8c remove redundant __call__ method in pipes.TextCategorizer (#4305)
* remove redundant __call__ method in pipes.TextCategorizer

Because the parent __call__ method behaves in the same way.

* fix: Pipe.__call__ arg

* fix: invalid arg in Pipe.__call__

* modified:   spacy/tests/regression/test_issue4278.py (#4278)

* deleted:    Pipfile
2019-09-18 21:31:27 +02:00
adrianeboyd
b5d999e510 Add textcat to train CLI (#4226)
* Add doc.cats to spacy.gold at the paragraph level

Support `doc.cats` as `"cats": [{"label": string, "value": number}]` in
the spacy JSON training format at the paragraph level.

* `spacy.gold.docs_to_json()` writes `docs.cats`

* `GoldCorpus` reads in cats in each `GoldParse`

* Update instances of gold_tuples to handle cats

Update iteration over gold_tuples / gold_parses to handle addition of
cats at the paragraph level.

* Add textcat to train CLI

* Add textcat options to train CLI
* Add textcat labels in `TextCategorizer.begin_training()`
* Add textcat evaluation to `Scorer`:
  * For binary exclusive classes with provided label: F1 for label
  * For 2+ exclusive classes: F1 macro average
  * For multilabel (not exclusive): ROC AUC macro average (currently
relying on sklearn)
* Provide user info on textcat evaluation settings, potential
incompatibilities
* Provide pipeline to Scorer in `Language.evaluate` for textcat config
* Customize train CLI output to include only metrics relevant to current
pipeline
* Add textcat evaluation to evaluate CLI

* Fix handling of unset arguments and config params

Fix handling of unset arguments and model confiug parameters in Scorer
initialization.

* Temporarily add sklearn requirement

* Remove sklearn version number

* Improve Scorer handling of models without textcats

* Fixing Scorer handling of models without textcats

* Update Scorer output for python 2.7

* Modify inf in Scorer for python 2.7

* Auto-format

Also make small adjustments to make auto-formatting with black easier and produce nicer results

* Move error message to Errors

* Update documentation

* Add cats to annotation JSON format [ci skip]

* Fix tpl flag and docs [ci skip]

* Switch to internal roc_auc_score

Switch to internal `roc_auc_score()` adapted from scikit-learn.

* Add AUCROCScore tests and improve errors/warnings

* Add tests for AUCROCScore and roc_auc_score
* Add missing error for only positive/negative values
* Remove unnecessary warnings and errors

* Make reduced roc_auc_score functions private

Because most of the checks and warnings have been stripped for the
internal functions and access is only intended through `ROCAUCScore`,
make the functions for roc_auc_score adapted from scikit-learn private.

* Check that data corresponds with multilabel flag

Check that the training instances correspond with the multilabel flag,
adding the multilabel flag if required.

* Add textcat score to early stopping check

* Add more checks to debug-data for textcat

* Add example training data for textcat

* Add more checks to textcat train CLI

* Check configuration when extending base model
* Fix typos

* Update textcat example data

* Provide licensing details and licenses for data
* Remove two labels with no positive instances from jigsaw-toxic-comment
data.


Co-authored-by: Ines Montani <ines@ines.io>
2019-09-15 22:31:31 +02:00
Ines Montani
16c2522791 Merge branch 'master' into develop 2019-09-14 16:42:01 +02:00
adrianeboyd
6942a6a69b Extend default punct for sentencizer (#4290)
Most of these characters are for languages / writing systems that aren't
supported by spacy, but I don't think it causes problems to include
them. In the UD evals, Hindi and Urdu improve a lot as expected (from
0-10% to 70-80%) and Persian improves a little (90% to 96%). Tamil
improves in combination with #4288.

The punctuation list is converted to a set internally because of its
increased length.

Sentence final punctuation generated with:

```
unichars -gas '[\p{Sentence_Break=STerm}\p{Sentence_Break=ATerm}]' '\p{Terminal_Punctuation}'
```

See: https://stackoverflow.com/a/9508766/461847

Fixes #4269.
2019-09-14 15:25:48 +02:00
Ines Montani
27106d6528 Merge branch 'master' into develop 2019-09-13 17:07:17 +02:00
Sofie Van Landeghem
2ae5db580e dim bugfix when incl_prior is False (#4285) 2019-09-13 16:30:05 +02:00
Ines Montani
3c3658ef9f Merge branch 'master' into develop 2019-09-12 18:03:01 +02:00
Ines Montani
228bbf506d Improve label properties on pipes 2019-09-12 18:02:44 +02:00
Ines Montani
655b434553 Merge branch 'master' into develop 2019-09-12 11:39:18 +02:00
tamuhey
71909cdf22 Fix iss4278 (#4279)
* fix: len(tuple) == 2

* (#4278) add fail test

* add contributor's aggreement
2019-09-12 10:44:49 +02:00
Ines Montani
8ebc3711dc Fix bug in Parser.labels and add test (#4275) 2019-09-11 18:29:35 +02:00
Matthew Honnibal
c308cf3e3e
Merge branch 'master' into feature/lemmatizer 2019-08-25 13:52:27 +02:00
Matthew Honnibal
bb911e5f4e Fix #3830: 'subtok' label being added even if learn_tokens=False (#4188)
* Prevent subtok label if not learning tokens

The parser introduces the subtok label to mark tokens that should be
merged during post-processing. Previously this happened even if we did
not have the --learn-tokens flag set. This patch passes the config
through to the parser, to prevent the problem.

* Make merge_subtokens a parser post-process if learn_subtokens

* Fix train script

* Add test for 3830: subtok problem

* Fix handlign of non-subtok in parser training
2019-08-23 17:54:00 +02:00
Ines Montani
f5d3afb1a3 Fix typo in docstrings [ci skip] 2019-08-22 16:24:15 +02:00
Matthew Honnibal
bcd08f20af Merge changes from master 2019-08-21 14:18:52 +02:00
adrianeboyd
8fe7bdd0fa Improve token pattern checking without validation (#4105)
* Fix typo in rule-based matching docs

* Improve token pattern checking without validation

Add more detailed token pattern checks without full JSON pattern validation and
provide more detailed error messages.

Addresses #4070 (also related: #4063, #4100).

* Check whether top-level attributes in patterns and attr for PhraseMatcher are
  in token pattern schema

* Check whether attribute value types are supported in general (as opposed to
  per attribute with full validation)

* Report various internal error types (OverflowError, AttributeError, KeyError)
  as ValueError with standard error messages

* Check for tagger/parser in PhraseMatcher pipeline for attributes TAG, POS,
  LEMMA, and DEP

* Add error messages with relevant details on how to use validate=True or nlp()
  instead of nlp.make_doc()

* Support attr=TEXT for PhraseMatcher

* Add NORM to schema

* Expand tests for pattern validation, Matcher, PhraseMatcher, and EntityRuler

* Remove unnecessary .keys()

* Rephrase error messages

* Add another type check to Matcher

Add another type check to Matcher for more understandable error messages
in some rare cases.

* Support phrase_matcher_attr=TEXT for EntityRuler

* Don't use spacy.errors in examples and bin scripts

* Fix error code

* Auto-format

Also try get Azure pipelines to finally start a build :(

* Update errors.py


Co-authored-by: Ines Montani <ines@ines.io>
Co-authored-by: Matthew Honnibal <honnibal+gh@gmail.com>
2019-08-21 14:00:37 +02:00
Ines Montani
f65e36925d Fix absolute imports and avoid importing from cli 2019-08-20 15:08:59 +02:00
Sofie Van Landeghem
0ba1b5eebc CLI scripts for entity linking (wikipedia & generic) (#4091)
* document token ent_kb_id

* document span kb_id

* update pipeline documentation

* prior and context weights as bool's instead

* entitylinker api documentation

* drop for both models

* finish entitylinker documentation

* small fixes

* documentation for KB

* candidate documentation

* links to api pages in code

* small fix

* frequency examples as counts for consistency

* consistent documentation about tensors returned by predict

* add entity linking to usage 101

* add entity linking infobox and KB section to 101

* entity-linking in linguistic features

* small typo corrections

* training example and docs for entity_linker

* predefined nlp and kb

* revert back to similarity encodings for simplicity (for now)

* set prior probabilities to 0 when excluded

* code clean up

* bugfix: deleting kb ID from tokens when entities were removed

* refactor train el example to use either model or vocab

* pretrain_kb example for example kb generation

* add to training docs for KB + EL example scripts

* small fixes

* error numbering

* ensure the language of vocab and nlp stay consistent across serialization

* equality with =

* avoid conflict in errors file

* add error 151

* final adjustements to the train scripts - consistency

* update of goldparse documentation

* small corrections

* push commit

* turn kb_creator into CLI script (wip)

* proper parameters for training entity vectors

* wikidata pipeline split up into two executable scripts

* remove context_width

* move wikidata scripts in bin directory, remove old dummy script

* refine KB script with logs and preprocessing options

* small edits

* small improvements to logging of EL CLI script
2019-08-13 15:38:59 +02:00
adrianeboyd
69aca7d839 Add validate option to EntityRuler (#4089)
* Add validate option to EntityRuler

* Add validate to EntityRuler, passed to Matcher and PhraseMatcher

* Add validate to usage and API docs

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

Co-Authored-By: Ines Montani <ines@ines.io>

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

Co-Authored-By: Ines Montani <ines@ines.io>
2019-08-07 00:40:53 +02:00
Matthew Honnibal
4632c597e7 Fix Pipe base class 2019-08-01 17:29:01 +02:00
Sofie Van Landeghem
7de3b129ab Resolve edge case when calling textcat.predict with empty doc (#4035)
* resolve edge case where no doc has tokens when calling textcat.predict

* more explicit value test
2019-07-30 14:58:01 +02:00
Matthew Honnibal
06eb428ed1 Make pipe base class a bit less presumptuous 2019-07-28 17:56:11 +02:00
Matthew Honnibal
16b5144095 Don't raise NotImplemented in Pipe.update 2019-07-28 17:54:11 +02:00
Matthew Honnibal
73e095923f 💫 Improve error message when model.from_bytes() dies (#4014)
* Improve error message when model.from_bytes() dies

When Thinc's model.from_bytes() is called with a mismatched model, often
we get a particularly ungraceful error,

e.g. "AttributeError: FunctionLayer has no attribute G"

This is because we're trying to load the parameters for something like
a LayerNorm layer, and the model architecture has some other layer there
instead. This is obviously terrible, especially since the error *type*
is wrong.

I've changed it to raise a ValueError. The error message is still
probably a bit terse, but it's hard to be sure exactly what's gone
wrong.

* Update spacy/pipeline/pipes.pyx

* Update spacy/pipeline/pipes.pyx

* Update spacy/pipeline/pipes.pyx

* Update spacy/syntax/nn_parser.pyx

* Update spacy/syntax/nn_parser.pyx

* Update spacy/pipeline/pipes.pyx

Co-Authored-By: Matthew Honnibal <honnibal+gh@gmail.com>

* Update spacy/pipeline/pipes.pyx

Co-Authored-By: Matthew Honnibal <honnibal+gh@gmail.com>


Co-authored-by: Ines Montani <ines@ines.io>
2019-07-24 11:27:34 +02:00
svlandeg
4e7ec1ed31 return fix 2019-07-23 14:23:58 +02:00
svlandeg
400ff342cf replace assert's with custom error messages 2019-07-23 11:52:48 +02:00
svlandeg
20389e4553 format and bugfix 2019-07-22 15:08:17 +02:00
svlandeg
41fb5204ba output tensors as part of predict 2019-07-19 14:47:36 +02:00
svlandeg
21176517a7 have gold.links correspond exactly to doc.ents 2019-07-19 12:36:15 +02:00
svlandeg
e1213eaf6a use original gold object in get_loss function 2019-07-18 13:35:10 +02:00
svlandeg
ec55d2fccd filter training data beforehand (+black formatting) 2019-07-18 10:22:24 +02:00
svlandeg
a63d15a142 code cleanup 2019-07-15 17:36:43 +02:00
svlandeg
60f299374f set default context width 2019-07-15 12:03:09 +02:00
Sofie Van Landeghem
c4c21cb428 more friendly textcat errors (#3946)
* more friendly textcat errors with require_model and require_labels

* update thinc version with recent bugfix
2019-07-10 19:39:38 +02:00
Ines Montani
40cd03fc35 Improve EntityRuler serialization 2019-07-10 12:25:45 +02:00
Ines Montani
570ab1f481 Fix handling of old entity ruler files
Expected an `entity_ruler.jsonl` file in the top-level model directory, so the path passed to from_disk by default (model path plus componentn name), but with the suffix ".jsonl".
2019-07-10 12:14:12 +02:00
Ines Montani
ea2050079b Auto-format 2019-07-10 12:03:05 +02:00
Ines Montani
f2ea3e3ea2
Merge branch 'master' into feature/nel-wiki 2019-07-09 21:57:47 +02:00
Ines Montani
547464609d Remove merge_subtokens from parser postprocessing for now 2019-07-09 21:50:30 +02:00
Joshua Smith
2eb925bd05 Added an argument to EntityRuler constructor to pass attrs to… (#3919)
* Perserve flags in EntityRuler

The EntityRuler (explosion/spaCy#3526) does not preserve
overwrite flags (or `ent_id_sep`) when serialized.  This
commit adds support for serialization/deserialization preserving
overwrite and ent_id_sep flags.

* add signed contributor agreement

* flake8 cleanup

mostly blank line issues.

* mark test from the issue as needing a model

The test from the issue needs some language model for serialization
but the test wasn't originally marked correctly.

* Adds `phrase_matcher_attr` to allow args to PhraseMatcher

This is an added arg to pass to the `PhraseMatcher`. For example,
this allows creation of a case insensitive phrase matcher when the
`EntityRuler` is created.  References explosion/spaCy#3822

* remove unneeded model loading

The model didn't need to be loaded, and I replaced it with
a change that doesn't require it (using existings fixtures)

* updated docstring for new argument

* updated docs to reflect new argument to the EntityRuler constructor

* change tempdir handling to be compatible with python 2.7

* return conflicted code to entityruler

Some stuff got cut out because of merge conflicts, this
returns that code for the phrase_matcher_attr.

* fixed typo in the code added back after conflicts

* flake8 compliance

When I deconflicted the branch there were some flake8 issues
introduced. This resolves the spacing problems.

* test changes:  attempts to fix flaky test in python3.5

These tests seem to be alittle flaky in 3.5 so I changed the check to avoid
the comparisons that seem to be fail sometimes.
2019-07-09 20:09:17 +02:00