Commit Graph

2106 Commits

Author SHA1 Message Date
Matthew Honnibal
f277bfdf0f
Add SpanGroup and Graph container types to represent arbitrary annotations (#6696)
* Draft out initial Spans data structure

* Initial span group commit

* Basic span group support on Doc

* Basic test for span group

* Compile span_group.pyx

* Draft addition of SpanGroup to DocBin

* Add deserialization for SpanGroup

* Add tests for serializing SpanGroup

* Fix serialization of SpanGroup

* Add EdgeC and GraphC structs

* Add draft Graph data structure

* Compile graph

* More work on Graph

* Update GraphC

* Upd graph

* Fix walk functions

* Let Graph take nodes and edges on construction

* Fix walking and getting

* Add graph tests

* Fix import

* Add module with the SpanGroups dict thingy

* Update test

* Rename 'span_groups' attribute

* Try to fix c++11 compilation

* Fix test

* Update DocBin

* Try to fix compilation

* Try to fix graph

* Improve SpanGroup docstrings

* Add doc.spans to documentation

* Fix serialization

* Tidy up and add docs

* Update docs [ci skip]

* Add SpanGroup.has_overlap

* WIP updated Graph API

* Start testing new Graph API

* Update Graph tests

* Update Graph

* Add docstring

Co-authored-by: Ines Montani <ines@ines.io>
2021-01-14 17:30:41 +11:00
Ines Montani
31a92b28ae
Merge pull request #6715 from adrianeboyd/feature/before-after-init-callbacks
Add initialize.before_init and after_init callbacks
2021-01-13 12:17:00 +11:00
Ines Montani
97d5a7ba99 Merge branch 'develop' of https://github.com/explosion/spaCy into develop 2021-01-13 12:03:02 +11:00
Ines Montani
8d6448ccf7 Add config resolver test 2021-01-13 12:02:59 +11:00
Adriane Boyd
a45d89f09a Add initialize.before_init and after_init callbacks
Add `initialize.before_init` and `initialize.after_init` callbacks to
the config. The `initialize.before_init` callback is a place to
implement one-time tokenizer customizations that are then saved with the
model.
2021-01-12 13:07:44 +01:00
Adriane Boyd
ad43cbb042
Sync missing and misaligned values in Tagger loss (#6689)
Use `None` for both missing and misaligned annotation in
`Tagger.get_loss`, reverting to the default missing value in the loss
function.
2021-01-10 11:30:37 +11:00
Sofie Van Landeghem
75d9019343
Fix types of Tok2Vec encoding architectures (#6442)
* fix TorchBiLSTMEncoder documentation

* ensure the types of the encoding Tok2vec layers are correct

* update references from v1 to v2 for the new architectures
2021-01-07 16:39:27 +11:00
Sofie Van Landeghem
8c1a23209f
Getting scores out of beam_parser (#6684)
* clean up of ner tests

* beam_parser tests

* implement get_beam_parses and scored_parses for the dep parser

* we don't have to add the parse if there are no arcs
2021-01-07 16:28:27 +11:00
Sofie Van Landeghem
402dbc5bae
Getting scores out of beam_ner (#6575)
* small fixes and formatting

* bring test_issue4313 up-to-date, currently fails

* formatting

* add get_beam_parses method back

* add scored_ents function

* delete tag map
2021-01-06 12:02:32 +01:00
Adriane Boyd
bf9096437e
Set default lemmas in retokenizer (#6667)
Instead of unsetting lemmas on retokenized tokens, set the default
lemmas to:

* merge: concatenate any existing lemmas with `SPACY` preserved
* split: use the new `ORTH` values if lemmas were previously set,
  otherwise leave unset
2021-01-06 12:29:44 +08:00
Adriane Boyd
0041dfbc7f
Use special matcher for exceptions with spaces (#6668)
Use the special cases phrase matcher for exceptions that include space
characters so that exceptions including spaces are supported.
2021-01-06 12:05:10 +08:00
Sofie Van Landeghem
afc5714d32
multi-label textcat component (#6474)
* multi-label textcat component

* formatting

* fix comment

* cleanup

* fix from #6481

* random edit to push the tests

* add explicit error when textcat is called with multi-label gold data

* fix error nr

* small fix
2021-01-06 13:07:14 +11:00
Ines Montani
81f018fb67
Merge pull request #6671 from explosion/chore/tidy-autoformat
Tidy up and auto-format
2021-01-05 14:45:31 +11:00
Ines Montani
224a3590e9
Merge pull request #6654 from svlandeg/chore/tests-cleanup
Unskipping tests
2021-01-05 13:53:40 +11:00
Ines Montani
c4993f16d0
Merge pull request #6651 from svlandeg/bugfix/cli_info 2021-01-05 13:44:26 +11:00
Ines Montani
991669c934 Tidy up and auto-format 2021-01-05 13:41:53 +11:00
Adriane Boyd
b57be94c78
Fix memory issues in Language.evaluate (#6386)
* Fix memory issues in Language.evaluate

Reset annotation in predicted docs before evaluating and store all data
in `examples`.

* Minor refactor to docs generator init

* Fix generator expression

* Fix final generator check

* Refactor pipeline loop

* Handle examples generator in Language.evaluate

* Add test with generator

* Use make_doc
2020-12-31 10:45:50 +11:00
svlandeg
a6a68da673 unskipping tests with python >= 3.6 2020-12-30 18:46:43 +01:00
svlandeg
d5ff0fecf8 add docs 2020-12-30 14:01:13 +01:00
svlandeg
c74ab6a313 fix imports 2020-12-30 12:40:12 +01:00
svlandeg
712a78b74a add simple unit test 2020-12-30 12:35:26 +01:00
Adriane Boyd
5ca57d8221
Add logger warning when serializing user hooks (#6595)
Add a warning that user hooks are lost on serialization.

Add a `user_hooks` exclude to skip the warning with pickle.
2020-12-29 11:54:32 +01:00
Adriane Boyd
cabd4ae5b1
Use logger.warning instead of logger.warn (#6596)
Use `logger.warning` instead of deprecated `logger.warn`.
2020-12-21 08:25:10 +08:00
Sofie Van Landeghem
282a3b49ea
Fix parser resizing when there is no upper layer (#6460)
* allow resizing of the parser model even when upper=False

* update from spacy.TransitionBasedParser.v1 to v2

* bugfix
2020-12-18 18:56:57 +08:00
Sofie Van Landeghem
0a923a7915
Tagger robustness (#6580)
* require labels in taggers

* ensure tagger works with incomplete data
2020-12-18 18:51:47 +08:00
Adriane Boyd
1ddf2f39c7
Switch converters to generator functions (#6547)
* Switch converters to generator functions

To reduce the memory usage when converting large corpora, refactor the
convert methods to be generator functions.

* Update tests
2020-12-15 16:47:16 +08:00
Matthew Honnibal
8656a08777
Add beam_parser and beam_ner components for v3 (#6369)
* Get basic beam tests working

* Get basic beam tests working

* Compile _beam_utils

* Remove prints

* Test beam density

* Beam parser seems to train

* Draft beam NER

* Upd beam

* Add hypothesis as dev dependency

* Implement missing is-gold-parse method

* Implement early update

* Fix state hashing

* Fix test

* Fix test

* Default to non-beam in parser constructor

* Improve oracle for beam

* Start refactoring beam

* Update test

* Refactor beam

* Update nn

* Refactor beam and weight by cost

* Update ner beam settings

* Update test

* Add __init__.pxd

* Upd test

* Fix test

* Upd test

* Fix test

* Remove ring buffer history from StateC

* WIP change arc-eager transitions

* Add state tests

* Support ternary sent start values

* Fix arc eager

* Fix NER

* Pass oracle cut size for beam

* Fix ner test

* Fix beam

* Improve StateC.clone

* Improve StateClass.borrow

* Work directly with StateC, not StateClass

* Remove print statements

* Fix state copy

* Improve state class

* Refactor parser oracles

* Fix arc eager oracle

* Fix arc eager oracle

* Use a vector to implement the stack

* Refactor state data structure

* Fix alignment of sent start

* Add get_aligned_sent_starts method

* Add test for ae oracle when bad sentence starts

* Fix sentence segment handling

* Avoid Reduce that inserts illegal sentence

* Update preset SBD test

* Fix test

* Remove prints

* Fix sent starts in Example

* Improve python API of StateClass

* Tweak comments and debug output of arc eager

* Upd test

* Fix state test

* Fix state test
2020-12-13 09:08:32 +08:00
Ines Montani
9d32e839d3 Merge branch 'develop' into feature/init-config-cpu-gpu 2020-12-10 08:50:53 +11:00
Ines Montani
f2571b5ec4
Merge pull request #6444 from adrianeboyd/chore/update-develop-from-master 2020-12-09 13:09:58 +11:00
Ines Montani
90171f2031
Merge pull request #6528 from svlandeg/feature/pipe_fill_config 2020-12-09 12:01:22 +11:00
Ines Montani
dfaef27f90
Merge pull request #6503 from adrianeboyd/feature/lemmatizer-rule-warning-pos
Warn on empty POS for the rule-based lemmatizer
2020-12-09 11:34:16 +11:00
Ines Montani
b85bd63eca Fix test 2020-12-09 11:24:01 +11:00
Ines Montani
febf71af28 Fix test 2020-12-09 11:23:07 +11:00
Ines Montani
1980203229 Merge branch 'master' into pr/6444 2020-12-09 11:09:40 +11:00
Ines Montani
05a2812ae0 Merge branch 'develop' into pr/6444 2020-12-09 11:04:03 +11:00
Sofie Van Landeghem
cfc72c2995
Bugfix multi-label textcat reproducibility (#6481)
* add test for multi-label textcat reproducibility

* remove positive_label

* fix lengths dtype

* fix comments

* remove comment that we should not have forgotten :-)
2020-12-09 06:29:15 +08:00
svlandeg
8f8a7f1733 returning config in init_config 2020-12-08 17:37:20 +01:00
Sofie Van Landeghem
2c27093c5f
require_cpu functionality (#6336)
* add require_cpu from Thinc 8.0.0rc2

* add docs

* fix test if cupy is not installed
2020-12-08 14:42:40 +08:00
Sofie Van Landeghem
f98a04434a
pretrain architectures (#6451)
* define new architectures for the pretraining objective

* add loss function as attr of the omdel

* cleanup

* cleanup

* shorten name

* fix typo

* remove unused error
2020-12-08 14:41:03 +08:00
Adriane Boyd
29b058ebdc
Fix spacy when retokenizing cases with affixes (#6475)
Preserve `token.spacy` corresponding to the span end token in the
original doc rather than adjusting for the current offset.

* If not modifying in place, this checks in the original document
(`doc.c` rather than `tokens`).
* If modifying in place, the document has not been modified past the
current span start position so the value at the current span end
position is valid.
2020-12-08 14:25:56 +08:00
Adriane Boyd
4448680750
Fix alignment for 1-to-1 tokens and lowercasing (#6476)
* When checking for token alignments, check not only that the tokens are
identical but that the character positions are both at the start of a
token.

  It's possible for the tokens to be identical even though the two
tokens aren't aligned one-to-one in a case like `["a'", "''"]` vs.
`["a", "''", "'"]`, where the middle tokens are identical but should not
be aligned on the token level at character position 2 since it's the
start of one token but the middle of another.

* Use the lowercased version of the token texts to create the
character-to-token alignment because lowercasing can change the string
length (e.g., for `İ`, see the not-a-bug bug report:
https://bugs.python.org/issue34723)
2020-12-08 14:25:16 +08:00
Adriane Boyd
d70950605c Warn on empty POS for the rule-based lemmatizer
Add a warning to the rule-based lemmatizer for any tokens without POS
annotation.
2020-12-04 11:46:15 +01:00
Sofie Van Landeghem
d6c616a125
Fixes in test suite (#6457)
* fix slow test for textcat readers

* cleanup test_issue5551

* add explicit score weight

* cleanup
2020-12-02 12:57:08 +01:00
Adriane Boyd
53c0fb7431
Only set NORM on Token in retokenizer (#6464)
* Only set NORM on Token in retokenizer

Instead of setting `NORM` on both the token and lexeme, set `NORM` only
on the token.

The retokenizer tries to set all possible attributes with
`Token/Lexeme.set_struct_attr` so that it doesn't have to enumerate
which attributes are available for each. `NORM` is the only attribute
that's stored on both and for most cases it doesn't make sense to set
the global norms based on a individual retokenization. For lexeme-only
attributes like `IS_STOP` there's no way to avoid the global side
effects, but I think that `NORM` would be better only on the token.

* Fix test
2020-11-30 09:35:42 +08:00
Adriane Boyd
03ae77e603
Add SPACY as a Matcher attribute (#6463) 2020-11-30 09:34:50 +08:00
Adriane Boyd
724831b066 Merge remote-tracking branch 'upstream/master' into chore/update-develop-from-master
* Update Macedonian for v3
* Update Turkish for v3
2020-11-25 11:49:34 +01:00
Sofie Van Landeghem
2af31a8c8d
Bugfix textcat reproducibility on GPU (#6411)
* add seed argument to ParametricAttention layer

* bump thinc to 7.4.3

* set thinc version range

Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
2020-11-23 12:29:35 +01:00
Adriane Boyd
320a8b1481
Add ent_id_ to strings serialized with Doc (#6353) 2020-11-10 20:16:07 +08:00
Adriane Boyd
a7e7d6c6c9
Ignore misaligned in Morphologizer.get_loss (#6363)
Fix bug where `Morphologizer.get_loss` treated misaligned annotation as
`EMPTY_MORPH` rather than ignoring it. Remove unneeded default `EMPTY_MORPH`
mappings.
2020-11-10 20:15:09 +08:00
Ines Montani
363ac73c72 Update docs [ci skip] 2020-11-09 12:43:26 +08:00