* Tagger: use unnormalized probabilities for inference
Using unnormalized softmax avoids use of the relatively expensive exp function,
which can significantly speed up non-transformer models (e.g. I got a speedup
of 27% on a German tagging + parsing pipeline).
* Add spacy.Tagger.v2 with configurable normalization
Normalization of probabilities is disabled by default to improve
performance.
* Update documentation, models, and tests to spacy.Tagger.v2
* Move Tagger.v1 to spacy-legacy
* docs/architectures: run prettier
* Unnormalized softmax is now a Softmax_v2 option
* Require thinc 8.0.14 and spacy-legacy 3.0.9
* Partial fix of entity linker batching
* Add import
* Better name
* Add `use_gold_ents` option, docs
* Change to v2, create stub v1, update docs etc.
* Fix error type
Honestly no idea what the right type to use here is.
ConfigValidationError seems wrong. Maybe a NotImplementedError?
* Make mypy happy
* Add hacky fix for init issue
* Add legacy pipeline entity linker
* Fix references to class name
* Add __init__.py for legacy
* Attempted fix for loss issue
* Remove placeholder V1
* formatting
* slightly more interesting train data
* Handle batches with no usable examples
This adds a test for batches that have docs but not entities, and a
check in the component that detects such cases and skips the update step
as thought the batch were empty.
* Remove todo about data verification
Check for empty data was moved further up so this should be OK now - the
case in question shouldn't be possible.
* Fix gradient calculation
The model doesn't know which entities are not in the kb, so it generates
embeddings for the context of all of them.
However, the loss does know which entities aren't in the kb, and it
ignores them, as there's no sensible gradient.
This has the issue that the gradient will not be calculated for some of
the input embeddings, which causes a dimension mismatch in backprop.
That should have caused a clear error, but with numpyops it was causing
nans to happen, which is another problem that should be addressed
separately.
This commit changes the loss to give a zero gradient for entities not in
the kb.
* add failing test for v1 EL legacy architecture
* Add nasty but simple working check for legacy arch
* Clarify why init hack works the way it does
* Clarify use_gold_ents use case
* Fix use gold ents related handling
* Add tests for no gold ents and fix other tests
* Use aligned ents function (not working)
This doesn't actually work because the "aligned" ents are gold-only. But
if I have a different function that returns the intersection, *then*
this will work as desired.
* Use proper matching ent check
This changes the process when gold ents are not used so that the
intersection of ents in the pred and gold is used.
* Move get_matching_ents to Example
* Use model attribute to check for legacy arch
* Rename flag
* bump spacy-legacy to lower 3.0.9
Co-authored-by: svlandeg <svlandeg@github.com>
* Setup debug data for spancat
* Add check for missing labels
* Add low-level data warning error
* Improve logic when compiling the gold train data
* Implement check for negative examples
* Remove breakpoint
* Remove ws_ents and missing entity checks
* Fix mypy errors
* Make variable name spans_key consistent
* Rename pipeline -> component for consistency
* Account for missing labels per spans_key
* Cleanup variable names for consistency
* Improve brevity of conditional statements
* Remove unused variables
* Include spans_key as an argument for _get_examples
* Add a conditional check for spans_key
* Update spancat debug data based on new API
- Instead of using _get_labels_from_model(), I'm now using
_get_labels_from_spancat() (cf. https://github.com/explosion/spaCy/pull10079)
- The way information is displayed was also changed (text -> table)
* Rename model_labels to ensure mypy works
* Update wording on warning messages
Use "span type" instead of "entity type" in wording the warning messages.
This is because Spans aren't necessarily entities.
* Update component type into a Literal
This is to make it clear that the component parameter should only accept
either 'spancat' or 'ner'.
* Update checks to include actual model span_keys
Instead of looking at everything in the data, we only check those
span_keys from the actual spancat component. Instead of doing the filter
inside the for-loop, I just made another dictionary,
data_labels_in_component to hold this value.
* Update spacy/cli/debug_data.py
* Show label counts only when verbose is True
Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
* Fix debug data check for ents that cross sents
* Use aligned sent starts to have the same indices for the NER and sent
start annotation
* Add a temporary, insufficient hack for the case where a
sentence-initial reference token is split into multiple tokens in the
predicted doc, since `Example.get_aligned("SENT_START")` currently
aligns `True` to all the split tokens.
* Improve test example
* Use Example.get_aligned_sent_starts
* Add test for crossing entity
So that overriding `paths.vectors` works consistently in generated
configs, set vectors model in `paths.vectors` and always refer to this
path in `initialize.vectors`.
* Determine labels by factory name in debug data
For all components, return labels for all components with the
corresponding factory name rather than for only the default name.
For `spancat`, return labels as a dict keyed by `spans_key`.
* Refactor for typing
* Add test
* Use assert instead of cast, removed unneeded arg
* Mark test as slow
* Check for assets with size of 0 bytes
* Update spacy/cli/project/assets.py
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
* Use internal names for factories
If a component factory is registered like `@French.factory(...)` instead
of `@Language.factory(...)`, the name in the factories registry will be
prefixed with the language code. However in the nlp.config object the
factory will be listed without the language code. The `add_pipe` code
has fallback logic to handle this, but packaging code and the registry
itself don't.
This change makes it so that the factory name in nlp.config is the
language-specific form. It's not clear if this will break anything else,
but it does seem to fix the inconsistency and resolve the specific user
issue that brought this to our attention.
* Change approach to use fallback in package lookup
This adds fallback logic to the package lookup, so it doesn't have to
touch the way the config is built. It seems to fix the tests too.
* Remove unecessary line
* Add test
Thsi also adds an assert that seems to have been forgotten.
* Add section for spacy.cli.train.train
* Add link from training page to train function
* Ensure path in train helper
* Update docs
Co-authored-by: Ines Montani <ines@ines.io>
* Add support for fasttext-bloom hash-only vectors
Overview:
* Extend `Vectors` to have two modes: `default` and `ngram`
* `default` is the default mode and equivalent to the current
`Vectors`
* `ngram` supports the hash-only ngram tables from `fasttext-bloom`
* Extend `spacy.StaticVectors.v2` to handle both modes with no changes
for `default` vectors
* Extend `spacy init vectors` to support ngram tables
The `ngram` mode **only** supports vector tables produced by this
fork of fastText, which adds an option to represent all vectors using
only the ngram buckets table and which uses the exact same ngram
generation algorithm and hash function (`MurmurHash3_x64_128`).
`fasttext-bloom` produces an additional `.hashvec` table, which can be
loaded by `spacy init vectors --fasttext-bloom-vectors`.
https://github.com/adrianeboyd/fastText/tree/feature/bloom
Implementation details:
* `Vectors` now includes the `StringStore` as `Vectors.strings` so that
the API can stay consistent for both `default` (which can look up from
`str` or `int`) and `ngram` (which requires `str` to calculate the
ngrams).
* In ngram mode `Vectors` uses a default `Vectors` object as a cache
since the ngram vectors lookups are relatively expensive.
* The default cache size is the same size as the provided ngram vector
table.
* Once the cache is full, no more entries are added. The user is
responsible for managing the cache in cases where the initial
documents are not representative of the texts.
* The cache can be resized by setting `Vectors.ngram_cache_size` or
cleared with `vectors._ngram_cache.clear()`.
* The API ends up a bit split between methods for `default` and for
`ngram`, so functions that only make sense for `default` or `ngram`
include warnings with custom messages suggesting alternatives where
possible.
* `Vocab.vectors` becomes a property so that the string stores can be
synced when assigning vectors to a vocab.
* `Vectors` serializes its own config settings as `vectors.cfg`.
* The `Vectors` serialization methods have added support for `exclude`
so that the `Vocab` can exclude the `Vectors` strings while serializing.
Removed:
* The `minn` and `maxn` options and related code from
`Vocab.get_vector`, which does not work in a meaningful way for default
vector tables.
* The unused `GlobalRegistry` in `Vectors`.
* Refactor to use reduce_mean
Refactor to use reduce_mean and remove the ngram vectors cache.
* Rename to floret
* Rename to floret in error messages
* Use --vectors-mode in CLI, vector init
* Fix vectors mode in init
* Remove unused var
* Minor API and docstrings adjustments
* Rename `--vectors-mode` to `--mode` in `init vectors` CLI
* Rename `Vectors.get_floret_vectors` to `Vectors.get_batch` and support
both modes.
* Minor updates to Vectors docstrings.
* Update API docs for Vectors and init vectors CLI
* Update types for StaticVectors
* Remove some old version refs in the docs
* Remove warning
* Update spacy/matcher/matcher.pyx
* Remove all references to the punctuation warning
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
* 🚨 Ignore all existing Mypy errors
* 🏗 Add Mypy check to CI
* Add types-mock and types-requests as dev requirements
* Add additional type ignore directives
* Add types packages to dev-only list in reqs test
* Add types-dataclasses for python 3.6
* Add ignore to pretrain
* 🏷 Improve type annotation on `run_command` helper
The `run_command` helper previously declared that it returned an
`Optional[subprocess.CompletedProcess]`, but it isn't actually possible
for the function to return `None`. These changes modify the type
annotation of the `run_command` helper and remove all now-unnecessary
`# type: ignore` directives.
* 🔧 Allow variable type redefinition in limited contexts
These changes modify how Mypy is configured to allow variables to have
their type automatically redefined under certain conditions. The Mypy
documentation contains the following example:
```python
def process(items: List[str]) -> None:
# 'items' has type List[str]
items = [item.split() for item in items]
# 'items' now has type List[List[str]]
...
```
This configuration change is especially helpful in reducing the number
of `# type: ignore` directives needed to handle the common pattern of:
* Accepting a filepath as a string
* Overwriting the variable using `filepath = ensure_path(filepath)`
These changes enable redefinition and remove all `# type: ignore`
directives rendered redundant by this change.
* 🏷 Add type annotation to converters mapping
* 🚨 Fix Mypy error in convert CLI argument verification
* 🏷 Improve type annotation on `resolve_dot_names` helper
* 🏷 Add type annotations for `Vocab` attributes `strings` and `vectors`
* 🏷 Add type annotations for more `Vocab` attributes
* 🏷 Add loose type annotation for gold data compilation
* 🏷 Improve `_format_labels` type annotation
* 🏷 Fix `get_lang_class` type annotation
* 🏷 Loosen return type of `Language.evaluate`
* 🏷 Don't accept `Scorer` in `handle_scores_per_type`
* 🏷 Add `string_to_list` overloads
* 🏷 Fix non-Optional command-line options
* 🙈 Ignore redefinition of `wandb_logger` in `loggers.py`
* ➕ Install `typing_extensions` in Python 3.8+
The `typing_extensions` package states that it should be used when
"writing code that must be compatible with multiple Python versions".
Since SpaCy needs to support multiple Python versions, it should be used
when newer `typing` module members are required. One example of this is
`Literal`, which is available starting with Python 3.8.
Previously SpaCy tried to import `Literal` from `typing`, falling back
to `typing_extensions` if the import failed. However, Mypy doesn't seem
to be able to understand what `Literal` means when the initial import
means. Therefore, these changes modify how `compat` imports `Literal` by
always importing it from `typing_extensions`.
These changes also modify how `typing_extensions` is installed, so that
it is a requirement for all Python versions, including those greater
than or equal to 3.8.
* 🏷 Improve type annotation for `Language.pipe`
These changes add a missing overload variant to the type signature of
`Language.pipe`. Additionally, the type signature is enhanced to allow
type checkers to differentiate between the two overload variants based
on the `as_tuple` parameter.
Fixes#8772
* ➖ Don't install `typing-extensions` in Python 3.8+
After more detailed analysis of how to implement Python version-specific
type annotations using SpaCy, it has been determined that by branching
on a comparison against `sys.version_info` can be statically analyzed by
Mypy well enough to enable us to conditionally use
`typing_extensions.Literal`. This means that we no longer need to
install `typing_extensions` for Python versions greater than or equal to
3.8! 🎉
These changes revert previous changes installing `typing-extensions`
regardless of Python version and modify how we import the `Literal` type
to ensure that Mypy treats it properly.
* resolve mypy errors for Strict pydantic types
* refactor code to avoid missing return statement
* fix types of convert CLI command
* avoid list-set confustion in debug_data
* fix typo and formatting
* small fixes to avoid type ignores
* fix types in profile CLI command and make it more efficient
* type fixes in projects CLI
* put one ignore back
* type fixes for render
* fix render types - the sequel
* fix BaseDefault in language definitions
* fix type of noun_chunks iterator - yields tuple instead of span
* fix types in language-specific modules
* 🏷 Expand accepted inputs of `get_string_id`
`get_string_id` accepts either a string (in which case it returns its
ID) or an ID (in which case it immediately returns the ID). These
changes extend the type annotation of `get_string_id` to indicate that
it can accept either strings or IDs.
* 🏷 Handle override types in `combine_score_weights`
The `combine_score_weights` function allows users to pass an `overrides`
mapping to override data extracted from the `weights` argument. Since it
allows `Optional` dictionary values, the return value may also include
`Optional` dictionary values.
These changes update the type annotations for `combine_score_weights` to
reflect this fact.
* 🏷 Fix tokenizer serialization method signatures in `DummyTokenizer`
* 🏷 Fix redefinition of `wandb_logger`
These changes fix the redefinition of `wandb_logger` by giving a
separate name to each `WandbLogger` version. For
backwards-compatibility, `spacy.train` still exports `wandb_logger_v3`
as `wandb_logger` for now.
* more fixes for typing in language
* type fixes in model definitions
* 🏷 Annotate `_RandomWords.probs` as `NDArray`
* 🏷 Annotate `tok2vec` layers to help Mypy
* 🐛 Fix `_RandomWords.probs` type annotations for Python 3.6
Also remove an import that I forgot to move to the top of the module 😅
* more fixes for matchers and other pipeline components
* quick fix for entity linker
* fixing types for spancat, textcat, etc
* bugfix for tok2vec
* type annotations for scorer
* add runtime_checkable for Protocol
* type and import fixes in tests
* mypy fixes for training utilities
* few fixes in util
* fix import
* 🐵 Remove unused `# type: ignore` directives
* 🏷 Annotate `Language._components`
* 🏷 Annotate `spacy.pipeline.Pipe`
* add doc as property to span.pyi
* small fixes and cleanup
* explicit type annotations instead of via comment
Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com>
Co-authored-by: svlandeg <svlandeg@github.com>
* Handle spacy-legacy in package CLI for dependencies
* Implement legacy backoff in spacy registry.find
* Remove unused import
* Update and format test
* Fix surprises when asking for the root of a git repo
In the case of the first asset I wanted to get from git, the data I
wanted was the entire repository. I tried leaving "path" blank, which
gave a less-than-helpful error, and then I tried `path: "/"`, which
started copying my entire filesystem into the project. The path I should
have used was "".
I've made two changes to make this smoother for others:
- The 'path' within a git clone defaults to ""
- If the path points outside of the tmpdir that the git clone goes
into, we fail with an error
Signed-off-by: Elia Robyn Speer <elia@explosion.ai>
* use a descriptive error instead of a default
plus some minor fixes from PR review
Signed-off-by: Elia Robyn Speer <elia@explosion.ai>
* check for None values in assets
Signed-off-by: Elia Robyn Speer <elia@explosion.ai>
Co-authored-by: Elia Robyn Speer <elia@explosion.ai>
* avoid msg var impliciteness
* rename local msg
* Add CI tests for debug data and train
* Adjust debug data CLI test
Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
* Draft spancat model
* Add spancat model
* Add test for extract_spans
* Add extract_spans layer
* Upd extract_spans
* Add spancat model
* Add test for spancat model
* Upd spancat model
* Update spancat component
* Upd spancat
* Update spancat model
* Add quick spancat test
* Import SpanCategorizer
* Fix SpanCategorizer component
* Import SpanGroup
* Fix span extraction
* Fix import
* Fix import
* Upd model
* Update spancat models
* Add scoring, update defaults
* Update and add docs
* Fix type
* Update spacy/ml/extract_spans.py
* Auto-format and fix import
* Fix comment
* Fix type
* Fix type
* Update website/docs/api/spancategorizer.md
* Fix comment
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
* Better defense
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
* Fix labels list
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
* Update spacy/ml/extract_spans.py
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
* Update spacy/pipeline/spancat.py
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
* Set annotations during update
* Set annotations in spancat
* fix imports in test
* Update spacy/pipeline/spancat.py
* replace MaxoutLogistic with LinearLogistic
* fix config
* various small fixes
* remove set_annotations parameter in update
* use our beloved tupley format with recent support for doc.spans
* bugfix to allow renaming the default span_key (scores weren't showing up)
* use different key in docs example
* change defaults to better-working parameters from project (WIP)
* register spacy.extract_spans.v1 for legacy purposes
* Upd dev version so can build wheel
* layers instead of architectures for smaller building blocks
* Update website/docs/api/spancategorizer.md
Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
* Update website/docs/api/spancategorizer.md
Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
* Include additional scores from overrides in combined score weights
* Parameterize spans key in scoring
Parameterize the `SpanCategorizer` `spans_key` for scoring purposes so
that it's possible to evaluate multiple `spancat` components in the same
pipeline.
* Use the (intentionally very short) default spans key `sc` in the
`SpanCategorizer`
* Adjust the default score weights to include the default key
* Adjust the scorer to use `spans_{spans_key}` as the prefix for the
returned score
* Revert addition of `attr_name` argument to `score_spans` and adjust
the key in the `getter` instead.
Note that for `spancat` components with a custom `span_key`, the score
weights currently need to be modified manually in
`[training.score_weights]` for them to be available during training. To
suppress the default score weights `spans_sc_p/r/f` during training, set
them to `null` in `[training.score_weights]`.
* Update website/docs/api/scorer.md
* Fix scorer for spans key containing underscore
* Increment version
* Add Spans to Evaluate CLI (#8439)
* Add Spans to Evaluate CLI
* Change to spans_key
* Add spans per_type output
Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
* Fix spancat GPU issues (#8455)
* Fix GPU issues
* Require thinc >=8.0.6
* Switch to glorot_uniform_init
* Fix and test ngram suggester
* Include final ngram in doc for all sizes
* Fix ngrams for docs of the same length as ngram size
* Handle batches of docs that result in no ngrams
* Add tests
Co-authored-by: Ines Montani <ines@ines.io>
Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com>
Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com>
Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
Co-authored-by: Nirant <NirantK@users.noreply.github.com>
* Use minor version for compatibility check
* Use minor version of compatibility table
* Soften warning message about incompatible models
* Add test for presence of current version in compatibility table
* Add test for download compatibility table
* Use minor version of lower pin in error message if possible
* Fall back to spacy_git_version if available
* Fix unknown version string
* Copy rather than move files to top-level of package
* Add all files to `MANIFEST.in` (primarily for older versions of pip)
* Include the `README.md` contents as `long_description` in the setup
* implement textcat resizing for TextCatCNN
* resizing textcat in-place
* simplify code
* ensure predictions for old textcat labels remain the same after resizing (WIP)
* fix for softmax
* store softmax as attr
* fix ensemble weight copy and cleanup
* restructure slightly
* adjust documentation, update tests and quickstart templates to use latest versions
* extend unit test slightly
* revert unnecessary edits
* fix typo
* ensemble architecture won't be resizable for now
* use resizable layer (WIP)
* revert using resizable layer
* resizable container while avoid shape inference trouble
* cleanup
* ensure model continues training after resizing
* use fill_b parameter
* use fill_defaults
* resize_layer callback
* format
* bump thinc to 8.0.4
* bump spacy-legacy to 3.0.6
The behavior of `spacy.Corpus.v1` is unexpected enough for `max_length
!= 0` that `0` is a better default for users creating a new config with
the quickstart.
If not, documents are skipped, sometimes the entire corpus is skipped,
and sometimes documents are (quite unexpectedly for your average user)
split into sentences.
* Check for unsupported cats values
* Only show labels if train/dev mismatched
* Don't show label counts (only counting positive labels seems odd)
* Use warnings for mismatched train/dev labels
This came up in #7878, but if --resume-path is a directory then loading
the weights will fail. On Linux this will give a straightforward error
message, but on Windows it gives "Permission Denied", which is
confusing.
* Fix percent unk display
This was showing (ratio %), so 10% would show as 0.10%. Fix by
multiplying ration by 100.
Might want to add a warning if this is over a threshold.
* Only show whole-integer percents
* Add empty lines at the end of Python files
* Only prepend the lang code if it's not there already
* Update spacy/cli/package.py
* fix whitespace stripping
* Replace negative rows with 0 in StaticVectors
Replace negative row indices with 0-vectors in `StaticVectors`.
* Increase versions related to StaticVectors
* Increase versions of all architctures and layers related to
`StaticVectors`
* Improve efficiency of 0-vector operations
Parallel `spacy-legacy` PR: https://github.com/explosion/spacy-legacy/pull/5
* Update config defaults to new versions
* Update docs
* Terminology: deprecated vs obsolete
Typically, deprecated is used for functionality that is bound to become unavailable but that can still be used. Obsolete is used for features that have been removed. In E941, I think what is meant is "obsolete" since loading a model by a shortcut simply does not work anymore (and throws an error). This is different from downloading a model with a shortcut, which is deprecated but still works.
In light of this, perhaps all other error codes should be checked as well.
* clarify that the link command is removed and not just deprecated
Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com>
* Update debug data further for v3
* Remove new/existing label distinction (new labels are not immediately
distinguishable because the pipeline is already initialized)
* Warn on missing labels in training data for all components except parser
* Separate textcat and textcat_multilabel sections
* Add section for morphologizer
* Reword missing label warnings
* add multi-label textcat to menu
* add infobox on textcat API
* add info to v3 migration guide
* small edits
* further fixes in doc strings
* add infobox to textcat architectures
* add textcat_multilabel to overview of built-in components
* spelling
* fix unrelated warn msg
* Add textcat_multilabel to quickstart [ci skip]
* remove separate documentation page for multilabel_textcategorizer
* small edits
* positive label clarification
* avoid duplicating information in self.cfg and fix textcat.score
* fix multilabel textcat too
* revert threshold to storage in cfg
* revert threshold stuff for multi-textcat
Co-authored-by: Ines Montani <ines@ines.io>
* Add hint for --gpu-id to CLI device info
If the user has `cupy` and an available GPU, add a hint about using
`--gpu-id 0` to the CLI output.
* Undo change to original CPU message
Now that `nlp.evaluate()` does not modify the examples, rerun the
pipeline on the (limited) texts in order to provide the predicted
annotation in the displacy output option.
* add capture argument to project_run and run_commands
* git bump to 3.0.1
* Set version to 3.0.1.dev0
Co-authored-by: Matthew Honnibal <honnibal+gh@gmail.com>
When `--no-cache-dir` is present, it prevents caching to properly function.
If the user still wants to do this, there is the possibility to pass options with `user_pip_args`.
But you should not enforce options like these. In my case this is preventing some docker build (using buildkit caching) to have proper caching of models.
