Commit Graph

163 Commits

Author SHA1 Message Date
Matthew Honnibal
a338c6f8f6 Fix JSON segmentation bug that affected French
Fix a bug in the JSON streaming code that GoldCorpus uses. Escaped
slashes were being handled incorrectly. This bug caused low scores for
French in the early v2.1.0 alphas, because most of the data was not
being read in.

Fittingly, the document that triggered the bug was a Wikipedia article about
Perl. Parsing perl remains difficult!
2018-12-08 10:41:24 +01:00
Ines Montani
5b2741f751 Remove unused cytoolz / itertools imports 2018-12-03 02:12:07 +01:00
Ines Montani
f37863093a 💫 Replace ujson, msgpack and dill/pickle/cloudpickle with srsly (#3003)
Remove hacks and wrappers, keep code in sync across our libraries and move spaCy a few steps closer to only depending on packages with binary wheels 🎉

See here: https://github.com/explosion/srsly

    Serialization is hard, especially across Python versions and multiple platforms. After dealing with many subtle bugs over the years (encodings, locales, large files) our libraries like spaCy and Prodigy have steadily grown a number of utility functions to wrap the multiple serialization formats we need to support (especially json, msgpack and pickle). These wrapping functions ended up duplicated across our codebases, so we wanted to put them in one place.

    At the same time, we noticed that having a lot of small dependencies was making maintainence harder, and making installation slower. To solve this, we've made srsly standalone, by including the component packages directly within it. This way we can provide all the serialization utilities we need in a single binary wheel.

    srsly currently includes forks of the following packages:

        ujson
        msgpack
        msgpack-numpy
        cloudpickle



* WIP: replace json/ujson with srsly

* Replace ujson in examples

Use regular json instead of srsly to make code easier to read and follow

* Update requirements

* Fix imports

* Fix typos

* Replace msgpack with srsly

* Fix warning
2018-12-03 01:28:22 +01:00
Matthew Honnibal
9536ee787c Add comma deletion to data noising 2018-12-01 13:42:18 +00:00
Ines Montani
37c7c85a86 💫 New JSON helpers, training data internals & CLI rewrite (#2932)
* Support nowrap setting in util.prints

* Tidy up and fix whitespace

* Simplify script and use read_jsonl helper

* Add JSON schemas (see #2928)

* Deprecate Doc.print_tree

Will be replaced with Doc.to_json, which will produce a unified format

* Add Doc.to_json() method (see #2928)

Converts Doc objects to JSON using the same unified format as the training data. Method also supports serializing selected custom attributes in the doc._. space.

* Remove outdated test

* Add write_json and write_jsonl helpers

* WIP: Update spacy train

* Tidy up spacy train

* WIP: Use wasabi for formatting

* Add GoldParse helpers for JSON format

* WIP: add debug-data command

* Fix typo

* Add missing import

* Update wasabi pin

* Add missing import

* 💫 Refactor CLI (#2943)

To be merged into #2932.

## Description
- [x] refactor CLI To use [`wasabi`](https://github.com/ines/wasabi)
- [x] use [`black`](https://github.com/ambv/black) for auto-formatting
- [x] add `flake8` config
- [x] move all messy UD-related scripts to `cli.ud`
- [x] make converters function that take the opened file and return the converted data (instead of having them handle the IO)

### Types of change
enhancement

## Checklist
<!--- Before you submit the PR, go over this checklist and make sure you can
tick off all the boxes. [] -> [x] -->
- [x] I have submitted the spaCy Contributor Agreement.
- [x] I ran the tests, and all new and existing tests passed.
- [x] My changes don't require a change to the documentation, or if they do, I've added all required information.

* Update wasabi pin

* Delete old test

* Update errors

* Fix typo

* Tidy up and format remaining code

* Fix formatting

* Improve formatting of messages

* Auto-format remaining code

* Add tok2vec stuff to spacy.train

* Fix typo

* Update wasabi pin

* Fix path checks for when train() is called as function

* Reformat and tidy up pretrain script

* Update argument annotations

* Raise error if model language doesn't match lang

* Document new train command
2018-11-30 20:16:14 +01:00
Matthew Honnibal
61e435610e
💫 Feature/improve pretraining (#2971)
* Improve spacy pretrain script

* Implement BERT-style 'masked language model' objective. Much better
results.

* Improve logging.

* Add length cap for documents, to avoid memory errors.

* Require thinc 7.0.0.dev1

* Require thinc 7.0.0.dev1

* Add argument for using pretrained vectors

* Fix defaults

* Fix syntax error

* Improve spacy pretrain script

* Implement BERT-style 'masked language model' objective. Much better
results.

* Improve logging.

* Add length cap for documents, to avoid memory errors.

* Require thinc 7.0.0.dev1

* Require thinc 7.0.0.dev1

* Add argument for using pretrained vectors

* Fix defaults

* Fix syntax error

* Tweak pretraining script

* Fix data limits in spacy.gold

* Fix pretrain script
2018-11-28 18:04:58 +01:00
Matthew Honnibal
0fdb25b958
Fix msgpack error 2018-11-27 19:35:55 +01:00
Matthew Honnibal
b9ef8ac616 Fix GoldParse class when no entities 2018-09-27 15:14:27 +02:00
Matthew Honnibal
3b6b018904 Fix loading of gold morphology 2018-09-26 21:01:48 +02:00
Matthew Honnibal
fb0abddd9e Call morph morphology in GoldParse 2018-09-25 21:34:53 +02:00
Matthew Honnibal
834dfb0e9d Add morph attribute to GoldParse 2018-09-25 21:32:05 +02:00
Matthew Honnibal
66a3f2ba21 Lower-case text before alignment 2018-08-16 00:42:36 +02:00
Matthew Honnibal
a9fb6d5511 Fix docs2jsonl function 2018-08-14 14:03:48 +02:00
Matthew Honnibal
2a5a61683e Add function to get train format from Doc objects
Our JSON training format is annoying to work with, and we've wanted to
retire it for some time. In the meantime, we can at least add some
missing functions to make it easier to live with.

This patch adds a function that generates the JSON format from a list
of Doc objects, one per paragraph. This should be a convenient way to handle
a lot of data conversions: whatever format you have the source
information in, you can use it to setup a Doc object. This approach
should offer better future-proofing as well. Hopefully, we can steadily
rewrite code that is sensitive to the current data-format, so that it
instead goes through this function. Then when we change the data format,
we won't have such a problem.
2018-08-14 13:13:10 +02:00
Matthew Honnibal
4336397ecb Update develop from master 2018-08-14 03:04:28 +02:00
Xiaoquan Kong
f0c9652ed1 New Feature: display more detail when Error E067 (#2639)
* Fix off-by-one error

* Add verbose option

* Update verbose option

* Update documents for verbose option
2018-08-07 10:45:29 +02:00
ines
4a62486340 Merge branch 'master' into develop 2018-05-30 13:01:01 +02:00
Maciej
c7d53348d7 Fix bug in CLI iob and ner converter (#2392) (fixes #2385)
* issue_2385 add tests for iob_to_biluo converter function

* issue_2385 fix and modify iob_to_biluo function to accept either iob or biluo tags in cli.converter

* issue_2385 add test to fix b char bug

* add contributor agreement

* fill contributor agreement
2018-05-30 12:28:44 +02:00
Matthew Honnibal
8661218fe8
Refactor parser (#2308)
* Work on refactoring greedy parser

* Compile updated parser

* Fix refactored parser

* Update test

* Fix refactored parser

* Fix refactored parser

* Readd beam search after refactor

* Fix beam search after refactor

* Fix parser

* Fix beam parsing

* Support oracle segmentation in ud-train CLI command

* Avoid relying on final gold check in beam search

* Add a keyword argument sink to GoldParse

* Bug fixes to beam search after refactor

* Avoid importing fused token symbol in ud-run-test, untl that's added

* Avoid importing fused token symbol in ud-run-test, untl that's added

* Don't modify Token in global scope

* Fix error in beam gradient calculation

* Default to beam_update_prob 1

* Set a more aggressive threshold on the max violn update

* Disable some tests to figure out why CI fails

* Disable some tests to figure out why CI fails

* Add some diagnostics to travis.yml to try to figure out why build fails

* Tell Thinc to link against system blas on Travis

* Point thinc to libblas on Travis

* Try running sudo=true for travis

* Unhack travis.sh

* Restore beam_density argument for parser beam

* Require thinc 6.11.1.dev16

* Revert hacks to tests

* Revert hacks to travis.yml

* Update thinc requirement

* Fix parser model loading

* Fix size limits in training data

* Add missing name attribute for parser

* Fix appveyor for Windows
2018-05-15 22:17:29 +02:00
Matthew Honnibal
bf19f22340 Allow gold.sent_starts to be set from Python 2018-05-07 15:51:34 +02:00
Jens Dahl Møllerhøj
b9290397fb rename SP to _SP (#2289) 2018-05-03 18:33:49 +02:00
Matthew Honnibal
2c4a6d66fa Merge master into develop. Big merge, many conflicts -- need to review 2018-04-29 14:49:26 +02:00
Ines Montani
3141e04822
💫 New system for error messages and warnings (#2163)
* Add spacy.errors module

* Update deprecation and user warnings

* Replace errors and asserts with new error message system

* Remove redundant asserts

* Fix whitespace

* Add messages for print/util.prints statements

* Fix typo

* Fix typos

* Move CLI messages to spacy.cli._messages

* Add decorator to display error code with message

An implementation like this is nice because it only modifies the string when it's retrieved from the containing class – so we don't have to worry about manipulating tracebacks etc.

* Remove unused link in spacy.about

* Update errors for invalid pipeline components

* Improve error for unknown factories

* Add displaCy warnings

* Update formatting consistency

* Move error message to spacy.errors

* Update errors and check if doc returned by component is None
2018-04-03 15:50:31 +02:00
Matthew Honnibal
1f7229f40f Revert "Merge branch 'develop' of https://github.com/explosion/spaCy into develop"
This reverts commit c9ba3d3c2d, reversing
changes made to 92c26a35d4.
2018-03-27 19:23:02 +02:00
ines
c699aec089 Add offsets_from_biluo_tags helper and tests (see #1626) 2017-11-26 16:38:01 +01:00
Matthew Honnibal
86ddf692a1 Fix bug in limit calculation on dev data 2017-11-14 01:37:10 +01:00
Matthew Honnibal
1cab703bba Move minibatch function to util 2017-11-06 23:45:36 +01:00
ines
d96e72f656 Tidy up rest 2017-10-27 21:07:59 +02:00
ines
a6135336f5 Tidy up gold 2017-10-27 17:02:55 +02:00
Matthew Honnibal
6e552c9d83 Prune number of non-projective labels more aggressiely 2017-10-11 02:46:44 -05:00
Matthew Honnibal
563f46f026 Fix multi-label support for text classification
The TextCategorizer class is supposed to support multi-label
text classification, and allow training data to contain missing
values.

For this to work, the gradient of the loss should be 0 when labels
are missing. Instead, there was no way to actually denote "missing"
in the GoldParse class, and so the TextCategorizer class treated
the label set within gold.cats as complete.

To fix this, we change GoldParse.cats to be a dict instead of a list.
The GoldParse.cats dict should map to floats, with 1. denoting
'present' and 0. denoting 'absent'. Gradients are zeroed for categories
absent from the gold.cats dict. A nice bonus is that you can also set
values between 0 and 1 for partial membership. You can also set numeric
values, if you're using a text classification model that uses an
appropriate loss function.

Unfortunately this is a breaking change; although the functionality
was only recently introduced and hasn't been properly documented
yet. I've updated the example script accordingly.
2017-10-05 18:43:02 -05:00
Matthew Honnibal
ba23d63c35 Fix minibatch function, for fixed batch size 2017-09-14 13:37:41 +02:00
Matthew Honnibal
4bb6bc3f9e Add support for sent_start to GoldParse 2017-08-25 20:03:14 -05:00
Matthew Honnibal
84b7ed49e4 Ensure updates aren't made if no gold available 2017-08-20 14:41:38 +02:00
Matthew Honnibal
ec63f4fe7b Add option to control how missing entities are handled when getting NER tags 2017-07-29 21:58:37 +02:00
Matthew Honnibal
9bae0ddc50 Fix minibatching 2017-07-22 20:14:49 +02:00
Matthew Honnibal
ed6c85fa3c Fix loading of text categories in GoldParse 2017-07-22 20:04:03 +02:00
Matthew Honnibal
7ea50182a5 Add support for text-classification labels to GoldParse 2017-07-20 00:17:47 +02:00
Matthew Honnibal
ebb6c49cd5 Make alignment case-insensitive for gold 2017-06-04 20:26:42 -05:00
Matthew Honnibal
fc4dd62e84 Merge branch 'develop' of https://github.com/explosion/spaCy into develop 2017-06-04 20:19:05 -05:00
Matthew Honnibal
a053b1218e Fix item counting during training 2017-06-04 20:18:20 -05:00
Matthew Honnibal
9bc4a26213 Add option of data augmentation noise 2017-06-04 20:16:57 -05:00
Matthew Honnibal
f6955a459c Fix prev commit 2017-06-03 14:38:37 -05:00
Matthew Honnibal
468ca6c760 Merge branch 'develop' of https://github.com/explosion/spaCy into develop 2017-06-03 14:33:51 -05:00
Matthew Honnibal
c647a0d33e Fix training counter for gold preprocessing 2017-06-03 14:33:39 -05:00
Matthew Honnibal
e62f46d39f Clarify gold.pyx slightly 2017-06-03 13:28:52 -05:00
Matthew Honnibal
be4a640f0c Fix arc eager label costs for uint64 2017-05-30 20:37:58 +02:00
Matthew Honnibal
84e66ca6d4 WIP on stringstore change. 27 failures 2017-05-28 14:06:40 +02:00
Matthew Honnibal
d06f235fc9 Fix conflict on convert.py 2017-05-26 11:33:29 -05:00
Matthew Honnibal
2e587c6417 Export iob_to_biluo utility 2017-05-26 11:32:55 -05:00
Matthew Honnibal
daac3e3573 Always shuffle gold data, and support length cap 2017-05-26 11:30:52 -05:00
Matthew Honnibal
3a6e59cc53 Add minibatch function in spacy.gold 2017-05-25 17:15:09 -05:00
Matthew Honnibal
3959d778ac Revert "Revert "WIP on improving parser efficiency""
This reverts commit 532afef4a8.
2017-05-23 03:06:53 -05:00
Matthew Honnibal
532afef4a8 Revert "WIP on improving parser efficiency"
This reverts commit bdaac7ab44.
2017-05-23 03:05:25 -05:00
Matthew Honnibal
bdaac7ab44 WIP on improving parser efficiency 2017-05-23 02:59:31 -05:00
Matthew Honnibal
c9760b2104 Support sentence limits in GoldCorpus 2017-05-22 10:40:46 -05:00
ines
54f04a9fe0 Update API docs with changes in spacy.gold and spacy.language 2017-05-22 12:29:30 +02:00
Matthew Honnibal
2a5eb9f61e Make nonproj methods top-level functions, instead of class methods 2017-05-22 04:51:08 -05:00
Matthew Honnibal
025d9bbc37 Fix handling of non-projective deps 2017-05-22 04:51:08 -05:00
Matthew Honnibal
f13d6c7359 Support gold preprocessing and single gold files 2017-05-22 04:51:08 -05:00
Matthew Honnibal
5db89053aa Merge docstrings 2017-05-21 13:46:23 -05:00
Matthew Honnibal
432b3499b3 Fix memory leak 2017-05-21 13:38:46 -05:00
Matthew Honnibal
4803b3b69e Add GoldCorpus class, to manage data streaming 2017-05-21 09:06:17 -05:00
ines
075f5ff87a Update docstrings and API docs for GoldParse 2017-05-21 13:53:46 +02:00
Matthew Honnibal
fc8d3a112c Add util.env_opt support: Can set hyper params through environment variables. 2017-05-18 04:36:53 -05:00
Matthew Honnibal
793430aa7a Get spaCy train command working with neural network
* Integrate models into pipeline
* Add basic serialization (maybe incorrect)
* Fix pickle on vocab
2017-05-17 12:04:50 +02:00
Matthew Honnibal
89a4f262fc Fix training methods 2017-04-16 13:00:37 -05:00
ines
e1efd589c3 Fix json imports and use ujson 2017-04-15 12:13:34 +02:00
ines
958b12dec8 Use pathlib instead of os.path 2017-04-15 12:13:00 +02:00
ines
d24589aa72 Clean up imports, unused code, whitespace, docstrings 2017-04-15 12:05:47 +02:00
ines
561f2a3eb4 Use consistent formatting for docstrings 2017-04-15 11:59:21 +02:00
Raphaël Bournhonesque
f332bf05be Remove unused import statements 2017-03-21 21:08:54 +01:00
Matthew Honnibal
2611ac2a89 Fix scorer bug for NER, related to ambiguity between missing annotations and misaligned tokens 2017-03-16 09:38:28 -05:00
Matthew Honnibal
3d4e389d23 Whitespace 2017-03-15 09:29:42 -05:00
Matthew Honnibal
159e8c46e1 Merge old training fixes with newer state 2016-11-25 09:16:36 -06:00
Matthew Honnibal
cc7e607a8a Fix gold.pyx for 1.0 2016-11-25 08:57:59 -06:00
Matthew Honnibal
b86f8af0c1 Fix doc strings 2016-11-01 12:25:36 +01:00
Matthew Honnibal
f5fe4f595b Fix json loading, for Python 3. 2016-10-20 21:23:26 +02:00
Matthew Honnibal
52b48b415e Fix GoldParse class 2016-10-16 11:41:36 +02:00
Matthew Honnibal
0317cea0ad Fix GoldParse 2016-10-15 23:55:07 +02:00
Matthew Honnibal
a48aa15384 Improve the API for the GoldParse class. 2016-10-15 23:53:29 +02:00
Matthew Honnibal
e07fe92b27 Draft a refactored init for the GoldParse class 2016-10-15 22:09:52 +02:00
Matthew Honnibal
86ae665c78 Add function for entity->biluo transformation 2016-10-15 21:51:04 +02:00
Matthew Honnibal
645d99523a Move merge_sents method into spacy.gold 2016-10-13 03:24:29 +02:00
Matthew Honnibal
ea23b64cc8 Refactor training, with new spacy.train module. Defaults still a little awkward. 2016-10-09 12:24:24 +02:00
Wolfgang Seeker
b6b96b233c don't require read_json_file to expect particular annotations 2016-05-02 15:29:30 +02:00
Wolfgang Seeker
4d7f393fae don't require json-files to have syntactic annotation 2016-04-22 16:32:27 +02:00
Henning Peters
6215272786 remove ujson as default non-dev dependency (still works as fallback if installed), because ujson doesn't ship wheels 2016-04-12 11:28:07 +02:00
Wolfgang Seeker
690c5acabf adjust train.py to train both english and german models 2016-03-03 15:21:00 +01:00
Wolfgang Seeker
3448cb40a4 integrated pseudo-projective parsing into parser
- nonproj.pyx holds a class PseudoProjectivity which currently holds
  all functionality to implement Nivre & Nilsson 2005's pseudo-projective
  parsing using the HEAD decoration scheme
- changed lefts/rights in Token to account for possible non-projective
  structures
2016-03-01 10:09:08 +01:00
Wolfgang Seeker
4b2297d5d4 add class PseudoProjective for pseudo-projective parsing
PseudoProjective() implements the algorithm from Nivre & Nilsson 2005
using their HEAD decoration scheme.
2016-02-24 11:26:25 +01:00
Wolfgang Seeker
8d531c958b replace tests for non-projectivity
- add functions to find non-projective edges
- add test file for non-projectivity functions
2016-02-22 14:40:40 +01:00
Matthew Honnibal
83dccf0fd7 * Use io module insteads of deprecated codecs module 2015-10-10 14:13:01 +11:00
alvations
8caedba42a caught more codecs.open -> io.open 2015-09-30 20:20:09 +02:00
Matthew Honnibal
7606d9936f * Python3 correction for GoldParse 2015-07-28 14:44:53 +02:00
Matthew Honnibal
f4809e562f * Allow json to be used as a fallback if ujson is not available 2015-07-25 18:11:36 +02:00
Matthew Honnibal
2ae0b439b2 * Fix space check in gold.pyx 2015-07-14 00:10:27 +02:00
Matthew Honnibal
89a91ad726 * Add SPACE part-of-speech tag, and train tagger to assign it. Also train tagger not to make whitespace an entity 2015-07-09 13:30:41 +02:00
Matthew Honnibal
43ef5ddea5 * Ensure root albel is spelled ROOT, for backwards compatibility 2015-06-23 04:14:03 +02:00
Matthew Honnibal
46fb24e9fd * Add cycle-checking code in gold.pyx 2015-06-23 00:02:22 +02:00