spaCy/spacy/vocab.pyx

611 lines
24 KiB
Cython
Raw Normal View History

# cython: profile=True
from libc.string cimport memcpy
2018-12-06 17:53:16 +00:00
import srsly
2020-02-18 14:38:18 +00:00
from thinc.api import get_array_module
from .lexeme cimport EMPTY_LEXEME
from .lexeme cimport Lexeme
2015-01-31 05:38:58 +00:00
from .typedefs cimport attr_t
from .tokens.token cimport Token
from .attrs cimport PROB, LANG, ORTH, TAG, POS
from .structs cimport SerializedLexemeC
from .compat import copy_reg
from .errors import Errors
from .lemmatizer import Lemmatizer
from .attrs import intify_attrs, NORM
2017-08-19 16:50:16 +00:00
from .vectors import Vectors
Update spaCy for thinc 8.0.0 (#4920) * Add load_from_config function * Add train_from_config script * Merge configs and expose via spacy.config * Fix script * Suggest create_evaluation_callback * Hard-code for NER * Fix errors * Register command * Add TODO * Update train-from-config todos * Fix imports * Allow delayed setting of parser model nr_class * Get train-from-config working * Tidy up and fix scores and printing * Hide traceback if cancelled * Fix weighted score formatting * Fix score formatting * Make output_path optional * Add Tok2Vec component * Tidy up and add tok2vec_tensors * Add option to copy docs in nlp.update * Copy docs in nlp.update * Adjust nlp.update() for set_annotations * Don't shuffle pipes in nlp.update, decruft * Support set_annotations arg in component update * Support set_annotations in parser update * Add get_gradients method * Add get_gradients to parser * Update errors.py * Fix problems caused by merge * Add _link_components method in nlp * Add concept of 'listeners' and ControlledModel * Support optional attributes arg in ControlledModel * Try having tok2vec component in pipeline * Fix tok2vec component * Fix config * Fix tok2vec * Update for Example * Update for Example * Update config * Add eg2doc util * Update and add schemas/types * Update schemas * Fix nlp.update * Fix tagger * Remove hacks from train-from-config * Remove hard-coded config str * Calculate loss in tok2vec component * Tidy up and use function signatures instead of models * Support union types for registry models * Minor cleaning in Language.update * Make ControlledModel specifically Tok2VecListener * Fix train_from_config * Fix tok2vec * Tidy up * Add function for bilstm tok2vec * Fix type * Fix syntax * Fix pytorch optimizer * Add example configs * Update for thinc describe changes * Update for Thinc changes * Update for dropout/sgd changes * Update for dropout/sgd changes * Unhack gradient update * Work on refactoring _ml * Remove _ml.py module * WIP upgrade cli scripts for thinc * Move some _ml stuff to util * Import link_vectors from util * Update train_from_config * Import from util * Import from util * Temporarily add ml.component_models module * Move ml methods * Move typedefs * Update load vectors * Update gitignore * Move imports * Add PrecomputableAffine * Fix imports * Fix imports * Fix imports * Fix missing imports * Update CLI scripts * Update spacy.language * Add stubs for building the models * Update model definition * Update create_default_optimizer * Fix import * Fix comment * Update imports in tests * Update imports in spacy.cli * Fix import * fix obsolete thinc imports * update srsly pin * from thinc to ml_datasets for example data such as imdb * update ml_datasets pin * using STATE.vectors * small fix * fix Sentencizer.pipe * black formatting * rename Affine to Linear as in thinc * set validate explicitely to True * rename with_square_sequences to with_list2padded * rename with_flatten to with_list2array * chaining layernorm * small fixes * revert Optimizer import * build_nel_encoder with new thinc style * fixes using model's get and set methods * Tok2Vec in component models, various fixes * fix up legacy tok2vec code * add model initialize calls * add in build_tagger_model * small fixes * setting model dims * fixes for ParserModel * various small fixes * initialize thinc Models * fixes * consistent naming of window_size * fixes, removing set_dropout * work around Iterable issue * remove legacy tok2vec * util fix * fix forward function of tok2vec listener * more fixes * trying to fix PrecomputableAffine (not succesful yet) * alloc instead of allocate * add morphologizer * rename residual * rename fixes * Fix predict function * Update parser and parser model * fixing few more tests * Fix precomputable affine * Update component model * Update parser model * Move backprop padding to own function, for test * Update test * Fix p. affine * Update NEL * build_bow_text_classifier and extract_ngrams * Fix parser init * Fix test add label * add build_simple_cnn_text_classifier * Fix parser init * Set gpu off by default in example * Fix tok2vec listener * Fix parser model * Small fixes * small fix for PyTorchLSTM parameters * revert my_compounding hack (iterable fixed now) * fix biLSTM * Fix uniqued * PyTorchRNNWrapper fix * small fixes * use helper function to calculate cosine loss * small fixes for build_simple_cnn_text_classifier * putting dropout default at 0.0 to ensure the layer gets built * using thinc util's set_dropout_rate * moving layer normalization inside of maxout definition to optimize dropout * temp debugging in NEL * fixed NEL model by using init defaults ! * fixing after set_dropout_rate refactor * proper fix * fix test_update_doc after refactoring optimizers in thinc * Add CharacterEmbed layer * Construct tagger Model * Add missing import * Remove unused stuff * Work on textcat * fix test (again :)) after optimizer refactor * fixes to allow reading Tagger from_disk without overwriting dimensions * don't build the tok2vec prematuraly * fix CharachterEmbed init * CharacterEmbed fixes * Fix CharacterEmbed architecture * fix imports * renames from latest thinc update * one more rename * add initialize calls where appropriate * fix parser initialization * Update Thinc version * Fix errors, auto-format and tidy up imports * Fix validation * fix if bias is cupy array * revert for now * ensure it's a numpy array before running bp in ParserStepModel * no reason to call require_gpu twice * use CupyOps.to_numpy instead of cupy directly * fix initialize of ParserModel * remove unnecessary import * fixes for CosineDistance * fix device renaming * use refactored loss functions (Thinc PR 251) * overfitting test for tagger * experimental settings for the tagger: avoid zero-init and subword normalization * clean up tagger overfitting test * use previous default value for nP * remove toy config * bringing layernorm back (had a bug - fixed in thinc) * revert setting nP explicitly * remove setting default in constructor * restore values as they used to be * add overfitting test for NER * add overfitting test for dep parser * add overfitting test for textcat * fixing init for linear (previously affine) * larger eps window for textcat * ensure doc is not None * Require newer thinc * Make float check vaguer * Slop the textcat overfit test more * Fix textcat test * Fix exclusive classes for textcat * fix after renaming of alloc methods * fixing renames and mandatory arguments (staticvectors WIP) * upgrade to thinc==8.0.0.dev3 * refer to vocab.vectors directly instead of its name * rename alpha to learn_rate * adding hashembed and staticvectors dropout * upgrade to thinc 8.0.0.dev4 * add name back to avoid warning W020 * thinc dev4 * update srsly * using thinc 8.0.0a0 ! Co-authored-by: Matthew Honnibal <honnibal+gh@gmail.com> Co-authored-by: Ines Montani <ines@ines.io>
2020-01-29 16:06:46 +00:00
from .util import link_vectors_to_models
Bloom-filter backed Lookup Tables (#4268) * Improve load_language_data helper * WIP: Add Lookups implementation * Start moving lemma data over to JSON * WIP: move data over for more languages * Convert more languages * Fix lemmatizer fixtures in tests * Finish conversion * Auto-format JSON files * Fix test for now * Make sure tables are stored on instance * Update docstrings * Update docstrings and errors * Update test * Add Lookups.__len__ * Add serialization methods * Add Lookups.remove_table * Use msgpack for serialization to disk * Fix file exists check * Try using OrderedDict for everything * Update .flake8 [ci skip] * Try fixing serialization * Update test_lookups.py * Update test_serialize_vocab_strings.py * Lookups / Tables now work This implements the stubs in the Lookups/Table classes. Currently this is in Cython but with no type declarations, so that could be improved. * Add lookups to setup.py * Actually add lookups pyx The previous commit added the old py file... * Lookups work-in-progress * Move from pyx back to py * Add string based lookups, fix serialization * Update tests, language/lemmatizer to work with string lookups There are some outstanding issues here: - a pickling-related test fails due to the bloom filter - some custom lemmatizers (fr/nl at least) have issues More generally, there's a question of how to deal with the case where you have a string but want to use the lookup table. Currently the table allows access by string or id, but that's getting pretty awkward. * Change lemmatizer lookup method to pass (orth, string) * Fix token lookup * Fix French lookup * Fix lt lemmatizer test * Fix Dutch lemmatizer * Fix lemmatizer lookup test This was using a normal dict instead of a Table, so checks for the string instead of an integer key failed. * Make uk/nl/ru lemmatizer lookup methods consistent The mentioned tokenizers all have their own implementation of the `lookup` method, which accesses a `Lookups` table. The way that was called in `token.pyx` was changed so this should be updated to have the same arguments as `lookup` in `lemmatizer.py` (specificially (orth/id, string)). Prior to this change tests weren't failing, but there would probably be issues with normal use of a model. More tests should proably be added. Additionally, the language-specific `lookup` implementations seem like they might not be needed, since they handle things like lower-casing that aren't actually language specific. * Make recently added Greek method compatible * Remove redundant class/method Leftovers from a merge not cleaned up adequately.
2019-09-12 15:26:11 +00:00
from .lookups import Lookups
2017-10-27 19:07:59 +00:00
from . import util
2014-12-24 06:42:00 +00:00
cdef class Vocab:
"""A look-up table that allows you to access `Lexeme` objects. The `Vocab`
instance also provides access to the `StringStore`, and owns underlying
C-data that is shared between `Doc` objects.
DOCS: https://spacy.io/api/vocab
"""
def __init__(self, lex_attr_getters=None, tag_map=None, lemmatizer=None,
2019-09-16 13:16:41 +00:00
strings=tuple(), lookups=None, oov_prob=-20., vectors_name=None,
**deprecated_kwargs):
"""Create the vocabulary.
2017-10-27 17:45:19 +00:00
lex_attr_getters (dict): A dictionary mapping attribute IDs to
functions to compute them. Defaults to `None`.
tag_map (dict): Dictionary mapping fine-grained tags to coarse-grained
parts-of-speech, and optionally morphological attributes.
lemmatizer (object): A lemmatizer. Defaults to `None`.
strings (StringStore): StringStore that maps strings to integers, and
vice versa.
lookups (Lookups): Container for large lookup tables and dictionaries.
2020-05-24 15:20:58 +00:00
name (str): Optional name to identify the vectors table.
2017-10-27 17:45:19 +00:00
RETURNS (Vocab): The newly constructed object.
"""
lex_attr_getters = lex_attr_getters if lex_attr_getters is not None else {}
tag_map = tag_map if tag_map is not None else {}
if lookups in (None, True, False):
lookups = Lookups()
if lemmatizer in (None, True, False):
lemmatizer = Lemmatizer(lookups)
self.cfg = {'oov_prob': oov_prob}
self.mem = Pool()
self._by_orth = PreshMap()
self.strings = StringStore()
2017-06-02 08:57:25 +00:00
self.length = 0
if strings:
for string in strings:
_ = self[string]
self.lex_attr_getters = lex_attr_getters
self.morphology = Morphology(self.strings, tag_map, lemmatizer)
2019-09-16 13:16:41 +00:00
self.vectors = Vectors(name=vectors_name)
self.lookups = lookups
2016-12-21 17:04:41 +00:00
@property
def lang(self):
langfunc = None
if self.lex_attr_getters:
langfunc = self.lex_attr_getters.get(LANG, None)
return langfunc("_") if langfunc else ""
property writing_system:
"""A dict with information about the language's writing system. To get
the data, we use the vocab.lang property to fetch the Language class.
If the Language class is not loaded, an empty dict is returned.
"""
def __get__(self):
if not util.lang_class_is_loaded(self.lang):
return {}
lang_class = util.get_lang_class(self.lang)
return dict(lang_class.Defaults.writing_system)
def __len__(self):
"""The current number of lexemes stored.
RETURNS (int): The current number of lexemes stored.
"""
return self.length
def add_flag(self, flag_getter, int flag_id=-1):
"""Set a new boolean flag to words in the vocabulary.
2016-12-21 17:04:41 +00:00
The flag_getter function will be called over the words currently in the
2016-11-01 11:25:36 +00:00
vocab, and then applied to new words as they occur. You'll then be able
2017-10-27 17:45:19 +00:00
to access the flag value on each token using token.check_flag(flag_id).
See also: `Lexeme.set_flag`, `Lexeme.check_flag`, `Token.set_flag`,
`Token.check_flag`.
2020-05-24 15:20:58 +00:00
flag_getter (callable): A function `f(str) -> bool`, to get the
2017-10-27 17:45:19 +00:00
flag value.
flag_id (int): An integer between 1 and 63 (inclusive), specifying
the bit at which the flag will be stored. If -1, the lowest
available bit will be chosen.
RETURNS (int): The integer ID by which the flag value can be checked.
DOCS: https://spacy.io/api/vocab#add_flag
"""
if flag_id == -1:
for bit in range(1, 64):
if bit not in self.lex_attr_getters:
flag_id = bit
break
else:
raise ValueError(Errors.E062)
elif flag_id >= 64 or flag_id < 1:
raise ValueError(Errors.E063.format(value=flag_id))
for lex in self:
lex.set_flag(flag_id, flag_getter(lex.orth_))
self.lex_attr_getters[flag_id] = flag_getter
return flag_id
cdef const LexemeC* get(self, Pool mem, unicode string) except NULL:
2017-10-27 17:45:19 +00:00
"""Get a pointer to a `LexemeC` from the lexicon, creating a new
`Lexeme` if necessary using memory acquired from the given pool. If the
pool is the lexicon's own memory, the lexeme is saved in the lexicon.
"""
if string == "":
2015-07-25 22:18:30 +00:00
return &EMPTY_LEXEME
cdef LexemeC* lex
2018-08-15 21:43:34 +00:00
cdef hash_t key = self.strings[string]
lex = <LexemeC*>self._by_orth.get(key)
cdef size_t addr
if lex != NULL:
2018-08-15 21:43:34 +00:00
assert lex.orth in self.strings
if lex.orth != key:
raise KeyError(Errors.E064.format(string=lex.orth,
orth=key, orth_id=string))
return lex
else:
2015-08-22 20:04:34 +00:00
return self._new_lexeme(mem, string)
cdef const LexemeC* get_by_orth(self, Pool mem, attr_t orth) except NULL:
2017-10-27 17:45:19 +00:00
"""Get a pointer to a `LexemeC` from the lexicon, creating a new
`Lexeme` if necessary using memory acquired from the given pool. If the
pool is the lexicon's own memory, the lexeme is saved in the lexicon.
"""
2015-07-26 17:26:41 +00:00
if orth == 0:
2015-07-26 16:39:27 +00:00
return &EMPTY_LEXEME
cdef LexemeC* lex
lex = <LexemeC*>self._by_orth.get(orth)
if lex != NULL:
return lex
2015-08-22 20:04:34 +00:00
else:
return self._new_lexeme(mem, self.strings[orth])
2015-08-22 20:04:34 +00:00
cdef const LexemeC* _new_lexeme(self, Pool mem, unicode string) except NULL:
if len(string) < 3 or self.length < 10000:
mem = self.mem
cdef bint is_oov = mem is not self.mem
lex = <LexemeC*>mem.alloc(sizeof(LexemeC), 1)
lex.orth = self.strings.add(string)
2015-08-26 17:21:46 +00:00
lex.length = len(string)
2017-11-15 13:23:58 +00:00
if self.vectors is not None:
lex.id = self.vectors.key2row.get(lex.orth, 0)
else:
lex.id = 0
if self.lex_attr_getters is not None:
for attr, func in self.lex_attr_getters.items():
value = func(string)
if isinstance(value, unicode):
value = self.strings.add(value)
2015-08-26 17:21:46 +00:00
if attr == PROB:
lex.prob = value
elif value is not None:
2015-08-26 17:21:46 +00:00
Lexeme.set_struct_attr(lex, attr, value)
if not is_oov:
self._add_lex_to_vocab(lex.orth, lex)
if lex == NULL:
raise ValueError(Errors.E085.format(string=string))
return lex
cdef int _add_lex_to_vocab(self, hash_t key, const LexemeC* lex) except -1:
self._by_orth.set(lex.orth, <void*>lex)
self.length += 1
def __contains__(self, key):
"""Check whether the string or int key has an entry in the vocabulary.
2016-11-01 11:25:36 +00:00
2020-05-24 15:20:58 +00:00
string (str): The ID string.
RETURNS (bool) Whether the string has an entry in the vocabulary.
DOCS: https://spacy.io/api/vocab#contains
"""
cdef hash_t int_key
if isinstance(key, bytes):
int_key = self.strings[key.decode("utf8")]
elif isinstance(key, unicode):
2018-08-15 21:43:34 +00:00
int_key = self.strings[key]
else:
int_key = key
lex = self._by_orth.get(int_key)
2017-01-11 09:18:22 +00:00
return lex is not NULL
def __iter__(self):
"""Iterate over the lexemes in the vocabulary.
2016-11-01 11:25:36 +00:00
YIELDS (Lexeme): An entry in the vocabulary.
DOCS: https://spacy.io/api/vocab#iter
"""
2017-10-31 01:00:26 +00:00
cdef attr_t key
cdef size_t addr
2017-10-31 01:00:26 +00:00
for key, addr in self._by_orth.items():
lex = Lexeme(self, key)
yield lex
2017-10-27 17:45:19 +00:00
def __getitem__(self, id_or_string):
"""Retrieve a lexeme, given an int ID or a unicode string. If a
previously unseen unicode string is given, a new lexeme is created and
stored.
id_or_string (int or unicode): The integer ID of a word, or its unicode
string. If `int >= Lexicon.size`, `IndexError` is raised. If
`id_or_string` is neither an int nor a unicode string, `ValueError`
is raised.
RETURNS (Lexeme): The lexeme indicated by the given ID.
EXAMPLE:
>>> apple = nlp.vocab.strings["apple"]
>>> assert nlp.vocab[apple] == nlp.vocab[u"apple"]
DOCS: https://spacy.io/api/vocab#getitem
"""
cdef attr_t orth
2017-10-31 01:00:26 +00:00
if isinstance(id_or_string, unicode):
orth = self.strings.add(id_or_string)
else:
orth = id_or_string
return Lexeme(self, orth)
cdef const TokenC* make_fused_token(self, substrings) except NULL:
cdef int i
tokens = <TokenC*>self.mem.alloc(len(substrings) + 1, sizeof(TokenC))
for i, props in enumerate(substrings):
2017-10-27 17:45:19 +00:00
props = intify_attrs(props, strings_map=self.strings,
_do_deprecated=True)
token = &tokens[i]
# Set the special tokens up to have arbitrary attributes
2017-10-27 17:45:19 +00:00
lex = <LexemeC*>self.get_by_orth(self.mem, props[ORTH])
token.lex = lex
2017-10-27 17:45:19 +00:00
if TAG in props:
self.morphology.assign_tag(token, props[TAG])
elif POS in props:
# Don't allow POS to be set without TAG -- this causes problems,
# see #1773
props.pop(POS)
for attr_id, value in props.items():
Token.set_struct_attr(token, attr_id, value)
# NORM is the only one that overlaps between the two
# (which is maybe not great?)
if attr_id != NORM:
Lexeme.set_struct_attr(lex, attr_id, value)
return tokens
2016-12-21 17:04:41 +00:00
2017-05-30 21:34:50 +00:00
@property
def vectors_length(self):
2017-08-22 17:46:35 +00:00
return self.vectors.data.shape[1]
2017-05-30 21:34:50 +00:00
2017-10-31 17:25:08 +00:00
def reset_vectors(self, *, width=None, shape=None):
2017-05-30 21:34:50 +00:00
"""Drop the current vector table. Because all vectors must be the same
width, you have to call this to change the size of the vectors.
"""
2017-10-31 17:25:08 +00:00
if width is not None and shape is not None:
raise ValueError(Errors.E065.format(width=width, shape=shape))
2017-10-31 17:25:08 +00:00
elif shape is not None:
self.vectors = Vectors(shape=shape)
else:
width = width if width is not None else self.vectors.data.shape[1]
self.vectors = Vectors(shape=(self.vectors.shape[0], width))
2017-10-31 01:00:26 +00:00
2017-10-31 17:25:08 +00:00
def prune_vectors(self, nr_row, batch_size=1024):
2017-10-31 01:00:26 +00:00
"""Reduce the current vector table to `nr_row` unique entries. Words
mapped to the discarded vectors will be remapped to the closest vector
among those remaining.
For example, suppose the original table had vectors for the words:
['sat', 'cat', 'feline', 'reclined']. If we prune the vector table to,
two rows, we would discard the vectors for 'feline' and 'reclined'.
These words would then be remapped to the closest remaining vector
-- so "feline" would have the same vector as "cat", and "reclined"
would have the same vector as "sat".
The similarities are judged by cosine. The original vectors may
be large, so the cosines are calculated in minibatches, to reduce
memory usage.
nr_row (int): The number of rows to keep in the vector table.
batch_size (int): Batch of vectors for calculating the similarities.
Larger batch sizes might be faster, while temporarily requiring
more memory.
RETURNS (dict): A dictionary keyed by removed words mapped to
`(string, score)` tuples, where `string` is the entry the removed
word was mapped to, and `score` the similarity score between the
two words.
DOCS: https://spacy.io/api/vocab#prune_vectors
2017-10-31 01:00:26 +00:00
"""
xp = get_array_module(self.vectors.data)
2017-10-31 17:25:08 +00:00
# Make prob negative so it sorts by rank ascending
# (key2row contains the rank)
priority = [(-lex.prob, self.vectors.key2row[lex.orth], lex.orth)
for lex in self if lex.orth in self.vectors.key2row]
priority.sort()
indices = xp.asarray([i for (prob, i, key) in priority], dtype="i")
keys = xp.asarray([key for (prob, i, key) in priority], dtype="uint64")
2017-10-31 17:25:08 +00:00
keep = xp.ascontiguousarray(self.vectors.data[indices[:nr_row]])
toss = xp.ascontiguousarray(self.vectors.data[indices[nr_row:]])
self.vectors = Vectors(data=keep, keys=keys, name=self.vectors.name)
💫 Port master changes over to develop (#2979) * Create aryaprabhudesai.md (#2681) * Update _install.jade (#2688) Typo fix: "models" -> "model" * Add FAC to spacy.explain (resolves #2706) * Remove docstrings for deprecated arguments (see #2703) * When calling getoption() in conftest.py, pass a default option (#2709) * When calling getoption() in conftest.py, pass a default option This is necessary to allow testing an installed spacy by running: pytest --pyargs spacy * Add contributor agreement * update bengali token rules for hyphen and digits (#2731) * Less norm computations in token similarity (#2730) * Less norm computations in token similarity * Contributor agreement * Remove ')' for clarity (#2737) Sorry, don't mean to be nitpicky, I just noticed this when going through the CLI and thought it was a quick fix. That said, if this was intention than please let me know. * added contributor agreement for mbkupfer (#2738) * Basic support for Telugu language (#2751) * Lex _attrs for polish language (#2750) * Signed spaCy contributor agreement * Added polish version of english lex_attrs * Introduces a bulk merge function, in order to solve issue #653 (#2696) * Fix comment * Introduce bulk merge to increase performance on many span merges * Sign contributor agreement * Implement pull request suggestions * Describe converters more explicitly (see #2643) * Add multi-threading note to Language.pipe (resolves #2582) [ci skip] * Fix formatting * Fix dependency scheme docs (closes #2705) [ci skip] * Don't set stop word in example (closes #2657) [ci skip] * Add words to portuguese language _num_words (#2759) * Add words to portuguese language _num_words * Add words to portuguese language _num_words * Update Indonesian model (#2752) * adding e-KTP in tokenizer exceptions list * add exception token * removing lines with containing space as it won't matter since we use .split() method in the end, added new tokens in exception * add tokenizer exceptions list * combining base_norms with norm_exceptions * adding norm_exception * fix double key in lemmatizer * remove unused import on punctuation.py * reformat stop_words to reduce number of lines, improve readibility * updating tokenizer exception * implement is_currency for lang/id * adding orth_first_upper in tokenizer_exceptions * update the norm_exception list * remove bunch of abbreviations * adding contributors file * Fixed spaCy+Keras example (#2763) * bug fixes in keras example * created contributor agreement * Adding French hyphenated first name (#2786) * Fix typo (closes #2784) * Fix typo (#2795) [ci skip] Fixed typo on line 6 "regcognizer --> recognizer" * Adding basic support for Sinhala language. (#2788) * adding Sinhala language package, stop words, examples and lex_attrs. * Adding contributor agreement * Updating contributor agreement * Also include lowercase norm exceptions * Fix error (#2802) * Fix error ValueError: cannot resize an array that references or is referenced by another array in this way. Use the resize function * added spaCy Contributor Agreement * Add charlax's contributor agreement (#2805) * agreement of contributor, may I introduce a tiny pl languge contribution (#2799) * Contributors agreement * Contributors agreement * Contributors agreement * Add jupyter=True to displacy.render in documentation (#2806) * Revert "Also include lowercase norm exceptions" This reverts commit 70f4e8adf37cfcfab60be2b97d6deae949b30e9e. * Remove deprecated encoding argument to msgpack * Set up dependency tree pattern matching skeleton (#2732) * Fix bug when too many entity types. Fixes #2800 * Fix Python 2 test failure * Require older msgpack-numpy * Restore encoding arg on msgpack-numpy * Try to fix version pin for msgpack-numpy * Update Portuguese Language (#2790) * Add words to portuguese language _num_words * Add words to portuguese language _num_words * Portuguese - Add/remove stopwords, fix tokenizer, add currency symbols * Extended punctuation and norm_exceptions in the Portuguese language * Correct error in spacy universe docs concerning spacy-lookup (#2814) * Update Keras Example for (Parikh et al, 2016) implementation (#2803) * bug fixes in keras example * created contributor agreement * baseline for Parikh model * initial version of parikh 2016 implemented * tested asymmetric models * fixed grevious error in normalization * use standard SNLI test file * begin to rework parikh example * initial version of running example * start to document the new version * start to document the new version * Update Decompositional Attention.ipynb * fixed calls to similarity * updated the README * import sys package duh * simplified indexing on mapping word to IDs * stupid python indent error * added code from https://github.com/tensorflow/tensorflow/issues/3388 for tf bug workaround * Fix typo (closes #2815) [ci skip] * Update regex version dependency * Set version to 2.0.13.dev3 * Skip seemingly problematic test * Remove problematic test * Try previous version of regex * Revert "Remove problematic test" This reverts commit bdebbef45552d698d390aa430b527ee27830f11b. * Unskip test * Try older version of regex * 💫 Update training examples and use minibatching (#2830) <!--- Provide a general summary of your changes in the title. --> ## Description Update the training examples in `/examples/training` to show usage of spaCy's `minibatch` and `compounding` helpers ([see here](https://spacy.io/usage/training#tips-batch-size) for details). The lack of batching in the examples has caused some confusion in the past, especially for beginners who would copy-paste the examples, update them with large training sets and experienced slow and unsatisfying results. ### Types of change enhancements ## 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. * Visual C++ link updated (#2842) (closes #2841) [ci skip] * New landing page * Add contribution agreement * Correcting lang/ru/examples.py (#2845) * Correct some grammatical inaccuracies in lang\ru\examples.py; filled Contributor Agreement * Correct some grammatical inaccuracies in lang\ru\examples.py * Move contributor agreement to separate file * Set version to 2.0.13.dev4 * Add Persian(Farsi) language support (#2797) * Also include lowercase norm exceptions * Remove in favour of https://github.com/explosion/spaCy/graphs/contributors * Rule-based French Lemmatizer (#2818) <!--- Provide a general summary of your changes in the title. --> ## Description <!--- Use this section to describe your changes. If your changes required testing, include information about the testing environment and the tests you ran. If your test fixes a bug reported in an issue, don't forget to include the issue number. If your PR is still a work in progress, that's totally fine – just include a note to let us know. --> Add a rule-based French Lemmatizer following the english one and the excellent PR for [greek language optimizations](https://github.com/explosion/spaCy/pull/2558) to adapt the Lemmatizer class. ### Types of change <!-- What type of change does your PR cover? Is it a bug fix, an enhancement or new feature, or a change to the documentation? --> - Lemma dictionary used can be found [here](http://infolingu.univ-mlv.fr/DonneesLinguistiques/Dictionnaires/telechargement.html), I used the XML version. - Add several files containing exhaustive list of words for each part of speech - Add some lemma rules - Add POS that are not checked in the standard Lemmatizer, i.e PRON, DET, ADV and AUX - Modify the Lemmatizer class to check in lookup table as a last resort if POS not mentionned - Modify the lemmatize function to check in lookup table as a last resort - Init files are updated so the model can support all the functionalities mentioned above - Add words to tokenizer_exceptions_list.py in respect to regex used in tokenizer_exceptions.py ## 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. * Set version to 2.0.13 * Fix formatting and consistency * Update docs for new version [ci skip] * Increment version [ci skip] * Add info on wheels [ci skip] * Adding "This is a sentence" example to Sinhala (#2846) * Add wheels badge * Update badge [ci skip] * Update README.rst [ci skip] * Update murmurhash pin * Increment version to 2.0.14.dev0 * Update GPU docs for v2.0.14 * Add wheel to setup_requires * Import prefer_gpu and require_gpu functions from Thinc * Add tests for prefer_gpu() and require_gpu() * Update requirements and setup.py * Workaround bug in thinc require_gpu * Set version to v2.0.14 * Update push-tag script * Unhack prefer_gpu * Require thinc 6.10.6 * Update prefer_gpu and require_gpu docs [ci skip] * Fix specifiers for GPU * Set version to 2.0.14.dev1 * Set version to 2.0.14 * Update Thinc version pin * Increment version * Fix msgpack-numpy version pin * Increment version * Update version to 2.0.16 * Update version [ci skip] * Redundant ')' in the Stop words' example (#2856) <!--- Provide a general summary of your changes in the title. --> ## Description <!--- Use this section to describe your changes. If your changes required testing, include information about the testing environment and the tests you ran. If your test fixes a bug reported in an issue, don't forget to include the issue number. If your PR is still a work in progress, that's totally fine – just include a note to let us know. --> ### Types of change <!-- What type of change does your PR cover? Is it a bug fix, an enhancement or new feature, or a change to the documentation? --> ## Checklist <!--- Before you submit the PR, go over this checklist and make sure you can tick off all the boxes. [] -> [x] --> - [ ] I have submitted the spaCy Contributor Agreement. - [ ] I ran the tests, and all new and existing tests passed. - [ ] My changes don't require a change to the documentation, or if they do, I've added all required information. * Documentation improvement regarding joblib and SO (#2867) Some documentation improvements ## Description 1. Fixed the dead URL to joblib 2. Fixed Stack Overflow brand name (with space) ### Types of change Documentation ## 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. * raise error when setting overlapping entities as doc.ents (#2880) * Fix out-of-bounds access in NER training The helper method state.B(1) gets the index of the first token of the buffer, or -1 if no such token exists. Normally this is safe because we pass this to functions like state.safe_get(), which returns an empty token. Here we used it directly as an array index, which is not okay! This error may have been the cause of out-of-bounds access errors during training. Similar errors may still be around, so much be hunted down. Hunting this one down took a long time...I printed out values across training runs and diffed, looking for points of divergence between runs, when no randomness should be allowed. * Change PyThaiNLP Url (#2876) * Fix missing comma * Add example showing a fix-up rule for space entities * Set version to 2.0.17.dev0 * Update regex version * Revert "Update regex version" This reverts commit 62358dd867d15bc6a475942dff34effba69dd70a. * Try setting older regex version, to align with conda * Set version to 2.0.17 * Add spacy-js to universe [ci-skip] * Add spacy-raspberry to universe (closes #2889) * Add script to validate universe json [ci skip] * Removed space in docs + added contributor indo (#2909) * - removed unneeded space in documentation * - added contributor info * Allow input text of length up to max_length, inclusive (#2922) * Include universe spec for spacy-wordnet component (#2919) * feat: include universe spec for spacy-wordnet component * chore: include spaCy contributor agreement * Minor formatting changes [ci skip] * Fix image [ci skip] Twitter URL doesn't work on live site * Check if the word is in one of the regular lists specific to each POS (#2886) * 💫 Create random IDs for SVGs to prevent ID clashes (#2927) Resolves #2924. ## Description Fixes problem where multiple visualizations in Jupyter notebooks would have clashing arc IDs, resulting in weirdly positioned arc labels. Generating a random ID prefix so even identical parses won't receive the same IDs for consistency (even if effect of ID clash isn't noticable here.) ### Types of change bug fix ## 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. * Fix typo [ci skip] * fixes symbolic link on py3 and windows (#2949) * fixes symbolic link on py3 and windows during setup of spacy using command python -m spacy link en_core_web_sm en closes #2948 * Update spacy/compat.py Co-Authored-By: cicorias <cicorias@users.noreply.github.com> * Fix formatting * Update universe [ci skip] * Catalan Language Support (#2940) * Catalan language Support * Ddding Catalan to documentation * Sort languages alphabetically [ci skip] * Update tests for pytest 4.x (#2965) <!--- Provide a general summary of your changes in the title. --> ## Description - [x] Replace marks in params for pytest 4.0 compat ([see here](https://docs.pytest.org/en/latest/deprecations.html#marks-in-pytest-mark-parametrize)) - [x] Un-xfail passing tests (some fixes in a recent update resolved a bunch of issues, but tests were apparently never updated here) ### Types of change <!-- What type of change does your PR cover? Is it a bug fix, an enhancement or new feature, or a change to the documentation? --> ## 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. * Fix regex pin to harmonize with conda (#2964) * Update README.rst * Fix bug where Vocab.prune_vector did not use 'batch_size' (#2977) Fixes #2976 * Fix typo * Fix typo * Remove duplicate file * Require thinc 7.0.0.dev2 Fixes bug in gpu_ops that would use cupy instead of numpy on CPU * Add missing import * Fix error IDs * Fix tests
2018-11-29 15:30:29 +00:00
syn_keys, syn_rows, scores = self.vectors.most_similar(toss, batch_size=batch_size)
2017-10-31 10:40:46 +00:00
remap = {}
2017-10-31 17:25:08 +00:00
for i, key in enumerate(keys[nr_row:]):
self.vectors.add(key, row=syn_rows[i][0])
2017-10-31 17:25:08 +00:00
word = self.strings[key]
synonym = self.strings[syn_keys[i][0]]
score = scores[i][0]
2017-10-31 17:25:08 +00:00
remap[word] = (synonym, score)
2017-10-31 01:00:26 +00:00
link_vectors_to_models(self)
return remap
2017-05-30 21:34:50 +00:00
def get_vector(self, orth, minn=None, maxn=None):
2017-10-27 17:45:19 +00:00
"""Retrieve a vector for a word in the vocabulary. Words can be looked
up by string or int ID. If no vectors data is loaded, ValueError is
raised.
If `minn` is defined, then the resulting vector uses Fasttext's
subword features by average over ngrams of `orth`.
2020-05-24 16:51:10 +00:00
orth (int / str): The hash value of a word, or its unicode string.
minn (int): Minimum n-gram length used for Fasttext's ngram computation.
Defaults to the length of `orth`.
maxn (int): Maximum n-gram length used for Fasttext's ngram computation.
Defaults to the length of `orth`.
2017-10-27 17:45:19 +00:00
RETURNS (numpy.ndarray): A word vector. Size
and shape determined by the `vocab.vectors` instance. Usually, a
numpy ndarray of shape (300,) and dtype float32.
DOCS: https://spacy.io/api/vocab#get_vector
2017-05-28 09:45:32 +00:00
"""
if isinstance(orth, str):
2017-08-19 17:52:25 +00:00
orth = self.strings.add(orth)
word = self[orth].orth_
2017-08-22 17:46:35 +00:00
if orth in self.vectors.key2row:
return self.vectors[orth]
# Assign default ngram limits to minn and maxn which is the length of the word.
if minn is None:
minn = len(word)
if maxn is None:
maxn = len(word)
xp = get_array_module(self.vectors.data)
vectors = xp.zeros((self.vectors_length,), dtype="f")
# Fasttext's ngram computation taken from
# https://github.com/facebookresearch/fastText
ngrams_size = 0;
for i in range(len(word)):
ngram = ""
if (word[i] and 0xC0) == 0x80:
continue
n = 1
j = i
while (j < len(word) and n <= maxn):
if n > maxn:
break
ngram += word[j]
j = j + 1
while (j < len(word) and (word[j] and 0xC0) == 0x80):
ngram += word[j]
j = j + 1
if (n >= minn and not (n == 1 and (i == 0 or j == len(word)))):
if self.strings[ngram] in self.vectors.key2row:
vectors = xp.add(self.vectors[self.strings[ngram]], vectors)
ngrams_size += 1
n = n + 1
if ngrams_size > 0:
vectors = vectors * (1.0/ngrams_size)
return vectors
2017-05-28 09:45:32 +00:00
2017-05-30 21:34:50 +00:00
def set_vector(self, orth, vector):
2017-10-27 17:45:19 +00:00
"""Set a vector for a word in the vocabulary. Words can be referenced
by string or int ID.
2020-05-24 16:51:10 +00:00
orth (int / str): The word.
vector (numpy.ndarray[ndim=1, dtype='float32']): The vector to set.
DOCS: https://spacy.io/api/vocab#set_vector
2017-05-30 21:34:50 +00:00
"""
if isinstance(orth, str):
2017-10-31 17:25:08 +00:00
orth = self.strings.add(orth)
if self.vectors.is_full and orth not in self.vectors:
new_rows = max(100, int(self.vectors.shape[0]*1.3))
if self.vectors.shape[1] == 0:
width = vector.size
else:
width = self.vectors.shape[1]
self.vectors.resize((new_rows, width))
lex = self[orth] # Adds words to vocab
2017-10-31 17:25:08 +00:00
self.vectors.add(orth, vector=vector)
2017-08-19 18:35:33 +00:00
self.vectors.add(orth, vector=vector)
2017-05-30 21:34:50 +00:00
2017-05-28 09:45:32 +00:00
def has_vector(self, orth):
2017-10-27 17:45:19 +00:00
"""Check whether a word has a vector. Returns False if no vectors have
been loaded. Words can be looked up by string or int ID.
2020-05-24 16:51:10 +00:00
orth (int / str): The word.
RETURNS (bool): Whether the word has a vector.
DOCS: https://spacy.io/api/vocab#has_vector
"""
if isinstance(orth, str):
2017-08-19 17:52:25 +00:00
orth = self.strings.add(orth)
return orth in self.vectors
2017-05-28 09:45:32 +00:00
def to_disk(self, path, exclude=tuple(), **kwargs):
"""Save the current state to a directory.
2020-05-24 16:51:10 +00:00
path (str / Path): A path to a directory, which will be created if
it doesn't exist.
exclude (list): String names of serialization fields to exclude.
DOCS: https://spacy.io/api/vocab#to_disk
"""
path = util.ensure_path(path)
if not path.exists():
path.mkdir()
setters = ["strings", "lexemes", "vectors"]
exclude = util.get_serialization_exclude(setters, exclude, kwargs)
if "strings" not in exclude:
self.strings.to_disk(path / "strings.json")
if "lexemes" not in exclude:
with (path / "lexemes.bin").open("wb") as file_:
file_.write(self.lexemes_to_bytes())
if "vectors" not in "exclude" and self.vectors is not None:
2017-08-19 19:27:35 +00:00
self.vectors.to_disk(path)
if "lookups" not in "exclude" and self.lookups is not None:
self.lookups.to_disk(path)
def from_disk(self, path, exclude=tuple(), **kwargs):
"""Loads state from a directory. Modifies the object in place and
returns it.
2020-05-24 16:51:10 +00:00
path (str / Path): A path to a directory.
exclude (list): String names of serialization fields to exclude.
RETURNS (Vocab): The modified `Vocab` object.
DOCS: https://spacy.io/api/vocab#to_disk
"""
path = util.ensure_path(path)
getters = ["strings", "lexemes", "vectors"]
exclude = util.get_serialization_exclude(getters, exclude, kwargs)
if "strings" not in exclude:
self.strings.from_disk(path / "strings.json") # TODO: add exclude?
if "lexemes" not in exclude:
with (path / "lexemes.bin").open("rb") as file_:
self.lexemes_from_bytes(file_.read())
if "vectors" not in exclude:
if self.vectors is not None:
self.vectors.from_disk(path, exclude=["strings"])
if self.vectors.name is not None:
link_vectors_to_models(self)
if "lookups" not in exclude:
self.lookups.from_disk(path)
2017-05-28 21:34:12 +00:00
return self
2016-11-01 11:25:36 +00:00
def to_bytes(self, exclude=tuple(), **kwargs):
"""Serialize the current state to a binary string.
exclude (list): String names of serialization fields to exclude.
RETURNS (bytes): The serialized form of the `Vocab` object.
DOCS: https://spacy.io/api/vocab#to_bytes
"""
2017-08-18 18:46:56 +00:00
def deserialize_vectors():
if self.vectors is None:
return None
else:
return self.vectors.to_bytes()
getters = {
"strings": lambda: self.strings.to_bytes(),
"lexemes": lambda: self.lexemes_to_bytes(),
"vectors": deserialize_vectors,
"lookups": lambda: self.lookups.to_bytes()
}
exclude = util.get_serialization_exclude(getters, exclude, kwargs)
2017-05-29 08:14:20 +00:00
return util.to_bytes(getters, exclude)
def from_bytes(self, bytes_data, exclude=tuple(), **kwargs):
"""Load state from a binary string.
bytes_data (bytes): The data to load from.
exclude (list): String names of serialization fields to exclude.
RETURNS (Vocab): The `Vocab` object.
DOCS: https://spacy.io/api/vocab#from_bytes
"""
2017-08-18 18:46:56 +00:00
def serialize_vectors(b):
if self.vectors is None:
return None
else:
return self.vectors.from_bytes(b)
setters = {
"strings": lambda b: self.strings.from_bytes(b),
"lexemes": lambda b: self.lexemes_from_bytes(b),
"vectors": lambda b: serialize_vectors(b),
"lookups": lambda b: self.lookups.from_bytes(b)
}
exclude = util.get_serialization_exclude(setters, exclude, kwargs)
2017-06-02 08:56:40 +00:00
util.from_bytes(bytes_data, setters, exclude)
if self.vectors.name is not None:
link_vectors_to_models(self)
2017-06-02 08:56:40 +00:00
return self
2017-05-28 21:34:12 +00:00
def lexemes_to_bytes(self):
cdef hash_t key
cdef size_t addr
2017-03-11 18:43:09 +00:00
cdef LexemeC* lexeme = NULL
cdef SerializedLexemeC lex_data
cdef int size = 0
for key, addr in self._by_orth.items():
if addr == 0:
continue
size += sizeof(lex_data.data)
byte_string = b"\0" * size
byte_ptr = <unsigned char*>byte_string
cdef int j
cdef int i = 0
for key, addr in self._by_orth.items():
if addr == 0:
continue
lexeme = <LexemeC*>addr
lex_data = Lexeme.c_to_bytes(lexeme)
for j in range(sizeof(lex_data.data)):
byte_ptr[i] = lex_data.data[j]
i += 1
return byte_string
2016-11-01 11:25:36 +00:00
def lexemes_from_bytes(self, bytes bytes_data):
"""Load the binary vocabulary data from the given string."""
cdef LexemeC* lexeme
2017-03-07 19:25:12 +00:00
cdef hash_t key
cdef unicode py_str
cdef int i = 0
cdef int j = 0
cdef SerializedLexemeC lex_data
chunk_size = sizeof(lex_data.data)
cdef void* ptr
cdef unsigned char* bytes_ptr = bytes_data
for i in range(0, len(bytes_data), chunk_size):
lexeme = <LexemeC*>self.mem.alloc(1, sizeof(LexemeC))
for j in range(sizeof(lex_data.data)):
lex_data.data[j] = bytes_ptr[i+j]
Lexeme.c_from_bytes(lexeme, lex_data)
prev_entry = self._by_orth.get(lexeme.orth)
if prev_entry != NULL:
memcpy(prev_entry, lexeme, sizeof(LexemeC))
continue
ptr = self.strings._map.get(lexeme.orth)
if ptr == NULL:
continue
2017-03-07 19:25:12 +00:00
py_str = self.strings[lexeme.orth]
if self.strings[py_str] != lexeme.orth:
raise ValueError(Errors.E086.format(string=py_str,
orth_id=lexeme.orth,
hash_id=self.strings[py_str]))
2017-03-07 19:25:12 +00:00
self._by_orth.set(lexeme.orth, lexeme)
self.length += 1
2017-11-14 18:15:04 +00:00
def _reset_cache(self, keys, strings):
# I'm not sure this made sense. Disable it for now.
raise NotImplementedError
2017-11-14 18:15:04 +00:00
2017-03-07 19:25:12 +00:00
def pickle_vocab(vocab):
sstore = vocab.strings
2017-11-23 16:19:50 +00:00
vectors = vocab.vectors
2017-03-07 19:25:12 +00:00
morph = vocab.morphology
length = vocab.length
data_dir = vocab.data_dir
2018-12-06 17:53:16 +00:00
lex_attr_getters = srsly.pickle_dumps(vocab.lex_attr_getters)
lexemes_data = vocab.lexemes_to_bytes()
2017-03-07 19:25:12 +00:00
return (unpickle_vocab,
2017-11-23 16:19:50 +00:00
(sstore, vectors, morph, data_dir, lex_attr_getters, lexemes_data, length))
2017-03-07 19:25:12 +00:00
2017-11-23 16:19:50 +00:00
def unpickle_vocab(sstore, vectors, morphology, data_dir,
2017-10-27 17:45:19 +00:00
lex_attr_getters, bytes lexemes_data, int length):
2017-03-07 19:25:12 +00:00
cdef Vocab vocab = Vocab()
vocab.length = length
2017-11-23 16:19:50 +00:00
vocab.vectors = vectors
2017-03-07 19:25:12 +00:00
vocab.strings = sstore
vocab.morphology = morphology
vocab.data_dir = data_dir
2018-12-06 17:53:16 +00:00
vocab.lex_attr_getters = srsly.pickle_loads(lex_attr_getters)
vocab.lexemes_from_bytes(lexemes_data)
2017-03-07 19:25:12 +00:00
vocab.length = length
return vocab
copy_reg.pickle(Vocab, pickle_vocab, unpickle_vocab)