* Refactor the Scorer to improve flexibility
Refactor the `Scorer` to improve flexibility for arbitrary pipeline
components.
* Individual pipeline components provide their own `evaluate` methods
that score a list of `Example`s and return a dictionary of scores
* `Scorer` is initialized either:
* with a provided pipeline containing components to be scored
* with a default pipeline containing the built-in statistical
components (senter, tagger, morphologizer, parser, ner)
* `Scorer.score` evaluates a list of `Example`s and returns a dictionary
of scores referring to the scores provided by the components in the
pipeline
Significant differences:
* `tags_acc` is renamed to `tag_acc` to be consistent with `token_acc`
and the new `morph_acc`, `pos_acc`, and `lemma_acc`
* Scoring is no longer cumulative: `Scorer.score` scores a list of
examples rather than a single example and does not retain any state
about previously scored examples
* PRF values in the returned scores are no longer multiplied by 100
* Add kwargs to Morphologizer.evaluate
* Create generalized scoring methods in Scorer
* Generalized static scoring methods are added to `Scorer`
* Methods require an attribute (either on Token or Doc) that is
used to key the returned scores
Naming differences:
* `uas`, `las`, and `las_per_type` in the scores dict are renamed to
`dep_uas`, `dep_las`, and `dep_las_per_type`
Scoring differences:
* `Doc.sents` is now scored as spans rather than on sentence-initial
token positions so that `Doc.sents` and `Doc.ents` can be scored with
the same method (this lowers scores since a single incorrect sentence
start results in two incorrect spans)
* Simplify / extend hasattr check for eval method
* Add hasattr check to tokenizer scoring
* Simplify to hasattr check for component scoring
* Reset Example alignment if docs are set
Reset the Example alignment if either doc is set in case the
tokenization has changed.
* Add PRF tokenization scoring for tokens as spans
Add PRF scores for tokens as character spans. The scores are:
* token_acc: # correct tokens / # gold tokens
* token_p/r/f: PRF for (token.idx, token.idx + len(token))
* Add docstring to Scorer.score_tokenization
* Rename component.evaluate() to component.score()
* Update Scorer API docs
* Update scoring for positive_label in textcat
* Fix TextCategorizer.score kwargs
* Update Language.evaluate docs
* Update score names in default config
* Update POS tests to reflect current behavior (it is not entirely clear
whether the AUX/VERB mapping is indeed the desired behavior?)
* Switch to `from_config` initialization in subtoken test
* `MorphAnalysis.get` returns only the field values
* Move `_normalize_props` inside `Morphology` as
`Morphology.normalize_attrs` and simplify
* Simplify POS field detection/conversion
* Convert all non-POS features to strings
* `Morphology` returns an empty string for a missing morph to align
with the FEATS string returned for an existing morph
* Remove unused `list_to_feats`
* Update with WIP
* Update with WIP
* Update with pipeline serialization
* Update types and pipe factories
* Add deep merge, tidy up and add tests
* Fix pipe creation from config
* Don't validate default configs on load
* Update spacy/language.py
Co-authored-by: Ines Montani <ines@ines.io>
* Adjust factory/component meta error
* Clean up factory args and remove defaults
* Add test for failing empty dict defaults
* Update pipeline handling and methods
* provide KB as registry function instead of as object
* small change in test to make functionality more clear
* update example script for EL configuration
* Fix typo
* Simplify test
* Simplify test
* splitting pipes.pyx into separate files
* moving default configs to each component file
* fix batch_size type
* removing default values from component constructors where possible (TODO: test 4725)
* skip instead of xfail
* Add test for config -> nlp with multiple instances
* pipeline.pipes -> pipeline.pipe
* Tidy up, document, remove kwargs
* small cleanup/generalization for Tok2VecListener
* use DEFAULT_UPSTREAM field
* revert to avoid circular imports
* Fix tests
* Replace deprecated arg
* Make model dirs require config
* fix pickling of keyword-only arguments in constructor
* WIP: clean up and integrate full config
* Add helper to handle function args more reliably
Now also includes keyword-only args
* Fix config composition and serialization
* Improve config debugging and add visual diff
* Remove unused defaults and fix type
* Remove pipeline and factories from meta
* Update spacy/default_config.cfg
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
* Update spacy/default_config.cfg
* small UX edits
* avoid printing stack trace for debug CLI commands
* Add support for language-specific factories
* specify the section of the config which holds the model to debug
* WIP: add Language.from_config
* Update with language data refactor WIP
* Auto-format
* Add backwards-compat handling for Language.factories
* Update morphologizer.pyx
* Fix morphologizer
* Update and simplify lemmatizers
* Fix Japanese tests
* Port over tagger changes
* Fix Chinese and tests
* Update to latest Thinc
* WIP: xfail first Russian lemmatizer test
* Fix component-specific overrides
* fix nO for output layers in debug_model
* Fix default value
* Fix tests and don't pass objects in config
* Fix deep merging
* Fix lemma lookup data registry
Only load the lookups if an entry is available in the registry (and if spacy-lookups-data is installed)
* Add types
* Add Vocab.from_config
* Fix typo
* Fix tests
* Make config copying more elegant
* Fix pipe analysis
* Fix lemmatizers and is_base_form
* WIP: move language defaults to config
* Fix morphology type
* Fix vocab
* Remove comment
* Update to latest Thinc
* Add morph rules to config
* Tidy up
* Remove set_morphology option from tagger factory
* Hack use_gpu
* Move [pipeline] to top-level block and make [nlp.pipeline] list
Allows separating component blocks from component order – otherwise, ordering the config would mean a changed component order, which is bad. Also allows initial config to define more components and not use all of them
* Fix use_gpu and resume in CLI
* Auto-format
* Remove resume from config
* Fix formatting and error
* [pipeline] -> [components]
* Fix types
* Fix tagger test: requires set_morphology?
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com>
Co-authored-by: Matthew Honnibal <honnibal+gh@gmail.com>
* step_through tests: skip instead of xfail
* test_empty_doc should be fixed with new Thinc version
* remove outdated test (there are other misaligned tests now)
* xfail reason
* fix test according to french exceptions
* clarified some skipped tests
* skip ukranian test instead of xfail
* skip instead of xfail
* skip + reason instead of xfail
* removed obsolete tests referring to removed "set_frozen" functionality
* fix test 999
* remove unused AlignmentError
* remove xfail where possible, skip otherwise
* increment thinc release for empty_doc test
* Refactor Chinese tokenizer configuration
Refactor `ChineseTokenizer` configuration so that it uses a single
`segmenter` setting to choose between character segmentation, jieba, and
pkuseg.
* replace `use_jieba`, `use_pkuseg`, `require_pkuseg` with the setting
`segmenter` with the supported values: `char`, `jieba`, `pkuseg`
* make the default segmenter plain character segmentation `char` (no
additional libraries required)
* Fix Chinese serialization test to use char default
* Warn if attempting to customize other segmenter
Add a warning if `Chinese.pkuseg_update_user_dict` is called when
another segmenter is selected.
* Improve tag map initialization and updating
Generalize tag map initialization and updating so that the tag map can
be loaded correctly prior to loading a `Corpus` with `spacy debug-data`
and `spacy train`.
* normalize provided tag map as necessary
* use the same method for initializing and updating the tag map
* Replace rather than update tag map
Replace rather than update tag map when loading a custom tag map.
Updating the tag map is problematic due to the sorted list of tag names
and the fact that the tag map will contain lingering/unwanted tags from
the default tag map.
* Update CLI scripts
* Reinitialize cache after loading new tag map
Reinitialize the cache with the right size after loading a new tag map.
* update `Morphologizer.begin_training` for use with `Example`
* make init and begin_training more consistent
* add `Morphology.normalize_features` to normalize outside of
`Morphology.add`
* make sure `get_loss` doesn't create unknown labels when the POS and
morph alignments differ
Serialize `morph_rules` with the tagger alongside the `tag_map`.
Use `Morphology.load_tag_map` and `Morphology.load_morph_exceptions` to
load these settings rather than reinitializing the morphology each time
they are changed.
Remove corpus-specific tag maps from the language data for languages
without custom tokenizers. For languages with custom word segmenters
that also provide tags (Japanese and Korean), the tag maps for the
custom tokenizers are kept as the default.
The default tag maps for languages without custom tokenizers are now the
default tag map from `lang/tag_map/py`, UPOS -> UPOS.
* Add morph to morphology in Doc.from_array
Add morphological analyses to morphology table in `Doc.from_array`.
* Use separate vocab in DocBin roundtrip test
* adding debug-model to print the internals for debugging purposes
* expend debug-model script with 4 stages: before, init, train, predict
* avoid enforcing to have a seed in the train script
* small fixes
* add keyword separator for update functions and drop unused "state"
* few more Example tests and various small fixes
* consistently return losses after update call
* eliminate unused tensors field across pipe components
* fix name
* fix arg name
* Add initial reproducibility tests
* failing test for default_text_classifier (WIP)
* track trouble to underlying tok2vec layer
* add regression test for Issue 5551
* tests go green with https://github.com/explosion/thinc/pull/359
* update test
* adding fixed seeds to HashEmbed layers, seems to fix the reproducility issue
Co-authored-by: Matthew Honnibal <honnibal+gh@gmail.com>
* Make project command a submodule
* Update with WIP
* Add helper for joining commands
* Update docstrins, formatting and types
* Update assets and add support for copying local files
* Fix type
* Update success messages
* Fix get_loss for None alignments in senter
When converting the `sent_start` values back to `SentenceRecognizer`
labels, handle `None` alignments.
* Handle SENT_START as -1
Handle SENT_START as -1 (or -1 converted to uint64) by treating any
values other than 1 the same as 0 in `SentenceRecognizer.get_loss`.
* remove _convert_examples
* fix test_gold, raise TypeError if tuples are used instead of Example's
* throwing proper errors when the wrong type of objects are passed
* fix deprectated format in tests
* fix deprectated format in parser tests
* fix tests for NEL, morph, senter, tagger, textcat
* update regression tests with new Example format
* use make_doc
* more fixes to nlp.update calls
* few more small fixes for rehearse and evaluate
* only import ml_datasets if really necessary
