explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2025-01-26 09:14:32 +03:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'develop' into pr/5767

Browse Source
This commit is contained in:
Ines Montani 2020-07-19 13:37:46 +02:00
commit 796f6c52d1
12 changed files with 194 additions and 93 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

72
bin/ud/run_eval.py
View File

@ -12,11 +12,11 @@ from ud_train import write_conllu
from spacy.lang.lex_attrs import word_shape
from spacy.util import get_lang_class
# All languages in spaCy - in UD format (note that Norwegian is 'no' instead of 'nb')
ALL_LANGUAGES = ("af, ar, bg, bn, ca, cs, da, de, el, en, es, et, fa, fi, fr,"
"ga, he, hi, hr, hu, id, is, it, ja, kn, ko, lt, lv, mr, no,"
# All languages in spaCy format (note that Norwegian is 'no' in UD - gets remapped later)
ALL_LANGUAGES = ("af, ar, bg, bn, ca, cs, da, de, el, en, es, et, eu, fa, fi, fr,"
"ga, gu, he, hi, hr, hu, hy, id, is, it, ja, kn, ko, lb, lij, lt, lv, ml, mr, nb,"
"nl, pl, pt, ro, ru, si, sk, sl, sq, sr, sv, ta, te, th, tl,"
"tr, tt, uk, ur, vi, zh")
"tr, tt, uk, ur, vi, yo, zh")
# Non-parsing tasks that will be evaluated (works for default models)
EVAL_NO_PARSE = ['Tokens', 'Words', 'Lemmas', 'Sentences', 'Feats']
@ -251,39 +251,43 @@ def main(out_path, ud_dir, check_parse=False, langs=ALL_LANGUAGES, exclude_train
# initialize all models with the multi-lang model
for lang in languages:
models[lang] = [multi] if multi else []
# add default models if we don't want to evaluate parsing info
if not check_parse:
# Norwegian is 'nb' in spaCy but 'no' in the UD corpora
if lang == 'no':
models['no'].append(load_default_model_sentencizer('nb'))
else:
models[lang].append(load_default_model_sentencizer(lang))
UD_lang = lang
# Norwegian is 'nb' in spaCy but 'no' in the UD corpora
if lang == "nb":
UD_lang = "no"
try:
models[UD_lang] = [multi] if multi else []
# add default models if we don't want to evaluate parsing info
if not check_parse:
models[UD_lang].append(load_default_model_sentencizer(lang))
except:
print(f"Exception initializing lang {lang} - skipping")
# language-specific trained models
if not exclude_trained_models:
if 'de' in models:
models['de'].append(load_model('de_core_news_sm'))
models['de'].append(load_model('de_core_news_md'))
if 'el' in models:
models['el'].append(load_model('el_core_news_sm'))
models['el'].append(load_model('el_core_news_md'))
if 'en' in models:
models['en'].append(load_model('en_core_web_sm'))
models['en'].append(load_model('en_core_web_md'))
models['en'].append(load_model('en_core_web_lg'))
if 'es' in models:
models['es'].append(load_model('es_core_news_sm'))
models['es'].append(load_model('es_core_news_md'))
if 'fr' in models:
models['fr'].append(load_model('fr_core_news_sm'))
models['fr'].append(load_model('fr_core_news_md'))
if 'it' in models:
models['it'].append(load_model('it_core_news_sm'))
if 'nl' in models:
models['nl'].append(load_model('nl_core_news_sm'))
if 'pt' in models:
models['pt'].append(load_model('pt_core_news_sm'))
news_languages = ["da", "de", "el", "es", "fr", "it", "ja", "lt", "nb", "nl", "pl", "pt", "ro"]
news_languages = ["nb"]
web_languages = ["en", "zh"]
sizes = ["sm", "md", "lg"]
for lang in web_languages:
UD_lang = lang
for size in sizes:
model_name = f'{lang}_core_web_{size}'
try:
models[UD_lang].append(load_model(model_name))
except Exception as e:
print(f"Error loading {model_name}: {e}")
for lang in news_languages:
UD_lang = lang
if lang == "nb":
UD_lang = "no"
for size in sizes:
model_name = f'{lang}_core_news_{size}'
try:
models[UD_lang].append(load_model(model_name))
except Exception as e:
print(f"Error loading {model_name}: {e}")
with out_path.open(mode='w', encoding='utf-8') as out_file:
run_all_evals(models, treebanks, out_file, check_parse, print_freq_tasks)

4
bin/ud/ud_train.py
View File

@ -303,7 +303,9 @@ def get_token_conllu(token, i):
feat_str = []
replacements = {"one": "1", "two": "2", "three": "3"}
for feat in features:
if not feat.startswith("begin") and not feat.startswith("end"):
if "=" in feat:
feat_str.append(feat)
elif not feat.startswith("begin") and not feat.startswith("end"):
key, value = feat.split("_", 1)
value = replacements.get(value, value)
feat_str.append("%s=%s" % (key, value.title()))

5
spacy/cli/debug_data.py
View File

@ -135,9 +135,8 @@ def debug_data(
morph_rules = {}
if morph_rules_path is not None:
morph_rules = srsly.read_json(morph_rules_path)
# Update tag map with provided mapping
nlp.vocab.morphology.tag_map.update(tag_map)
# Replace tag map with provided mapping
nlp.vocab.morphology.load_tag_map(tag_map)
# Load morph rules
nlp.vocab.morphology.load_morph_exceptions(morph_rules)

4
spacy/cli/train.py
View File

@ -124,8 +124,8 @@ def train(
)
nlp.begin_training(lambda: train_examples)
# Update tag map with provided mapping
nlp.vocab.morphology.tag_map.update(tag_map)
# Replace tag map with provided mapping
nlp.vocab.morphology.load_tag_map(tag_map)
# Load morph rules
nlp.vocab.morphology.load_morph_exceptions(morph_rules)

56
spacy/morphology.pyx
View File

@ -58,12 +58,23 @@ cdef class Morphology:
FEATURE_SEP = "|"
FIELD_SEP = "="
VALUE_SEP = ","
EMPTY_MORPH = "_"
EMPTY_MORPH = "_" # not an empty string so that the PreshMap key is not 0
def __init__(self, StringStore strings, tag_map, lemmatizer, exc=None):
self.mem = Pool()
self.strings = strings
self.tags = PreshMap()
self.load_tag_map(tag_map)
self.lemmatizer = lemmatizer
self._cache = PreshMapArray(self.n_tags)
self._exc = {}
if exc is not None:
self.load_morph_exceptions(exc)
def load_tag_map(self, tag_map):
self.tag_map = {}
self.reverse_index = {}
# Add special space symbol. We prefix with underscore, to make sure it
# always sorts to the end.
if '_SP' in tag_map:
@ -74,24 +85,14 @@ cdef class Morphology:
self.strings.add('_SP')
tag_map = dict(tag_map)
tag_map['_SP'] = space_attrs
self.tag_names = tuple(sorted(tag_map.keys()))
self.tag_map = {}
self.lemmatizer = lemmatizer
self.n_tags = len(tag_map)
self.reverse_index = {}
self._load_from_tag_map(tag_map)
self._cache = PreshMapArray(self.n_tags)
self._exc = {}
if exc is not None:
self.load_morph_exceptions(exc)
def _load_from_tag_map(self, tag_map):
for i, (tag_str, attrs) in enumerate(sorted(tag_map.items())):
attrs = _normalize_props(attrs)
self.add(attrs)
self.tag_map[tag_str] = dict(attrs)
self.reverse_index[self.strings.add(tag_str)] = i
self.tag_names = tuple(sorted(self.tag_map.keys()))
self.n_tags = len(self.tag_map)
self._cache = PreshMapArray(self.n_tags)
def __reduce__(self):
return (Morphology, (self.strings, self.tag_map, self.lemmatizer,
@ -114,13 +115,7 @@ cdef class Morphology:
if not isinstance(features, dict):
warnings.warn(Warnings.W100.format(feature=features))
features = {}
features = _normalize_props(features)
string_features = {self.strings.as_string(field): self.strings.as_string(values) for field, values in features.items()}
# normalized UFEATS string with sorted fields and values
norm_feats_string = self.FEATURE_SEP.join(sorted([
self.FIELD_SEP.join([field, values])
for field, values in string_features.items()
]))
# intified ("Field", "Field=Value") pairs
field_feature_pairs = []
for field in sorted(string_features):
@ -134,6 +129,7 @@ cdef class Morphology:
# the hash key for the tag is either the hash of the normalized UFEATS
# string or the hash of an empty placeholder (using the empty string
# would give a hash key of 0, which is not good for PreshMap)
norm_feats_string = self.normalize_features(features)
if norm_feats_string:
tag.key = self.strings.add(norm_feats_string)
else:
@ -141,6 +137,26 @@ cdef class Morphology:
self.insert(tag)
return tag.key
def normalize_features(self, features):
"""Create a normalized UFEATS string from a features string or dict.
features (Union[dict, str]): Features as dict or UFEATS string.
RETURNS (str): Features as normalized UFEATS string.
"""
if isinstance(features, str):
features = self.feats_to_dict(features)
if not isinstance(features, dict):
warnings.warn(Warnings.W100.format(feature=features))
features = {}
features = _normalize_props(features)
string_features = {self.strings.as_string(field): self.strings.as_string(values) for field, values in features.items()}
# normalized UFEATS string with sorted fields and values
norm_feats_string = self.FEATURE_SEP.join(sorted([
self.FIELD_SEP.join([field, values])
for field, values in string_features.items()
]))
return norm_feats_string or self.EMPTY_MORPH
cdef MorphAnalysisC create_morph_tag(self, field_feature_pairs) except *:
"""Creates a MorphAnalysisC from a list of intified
("Field", "Field=Value") tuples where fields with multiple values have

71
spacy/pipeline/morphologizer.pyx
View File

@ -23,29 +23,45 @@ from .defaults import default_morphologizer
@component("morphologizer", assigns=["token.morph", "token.pos"], default_model=default_morphologizer)
class Morphologizer(Tagger):
POS_FEAT = "POS"
def __init__(self, vocab, model, **cfg):
self.vocab = vocab
self.model = model
self._rehearsal_model = None
self.cfg = dict(sorted(cfg.items()))
self.cfg.setdefault("labels", {})
self.cfg.setdefault("morph_pos", {})
# to be able to set annotations without string operations on labels,
# store mappings from morph+POS labels to token-level annotations:
# 1) labels_morph stores a mapping from morph+POS->morph
self.cfg.setdefault("labels_morph", {})
# 2) labels_pos stores a mapping from morph+POS->POS
self.cfg.setdefault("labels_pos", {})
# add mappings for empty morph
self.cfg["labels_morph"][Morphology.EMPTY_MORPH] = Morphology.EMPTY_MORPH
self.cfg["labels_pos"][Morphology.EMPTY_MORPH] = POS_IDS[""]
@property
def labels(self):
return tuple(self.cfg["labels"].keys())
return tuple(self.cfg["labels_morph"].keys())
def add_label(self, label):
if not isinstance(label, str):
raise ValueError(Errors.E187)
if label in self.labels:
return 0
morph = Morphology.feats_to_dict(label)
norm_morph_pos = self.vocab.strings[self.vocab.morphology.add(morph)]
pos = morph.get("POS", "")
if norm_morph_pos not in self.cfg["labels"]:
self.cfg["labels"][norm_morph_pos] = norm_morph_pos
self.cfg["morph_pos"][norm_morph_pos] = POS_IDS[pos]
# normalize label
norm_label = self.vocab.morphology.normalize_features(label)
# extract separate POS and morph tags
label_dict = Morphology.feats_to_dict(label)
pos = label_dict.get(self.POS_FEAT, "")
if self.POS_FEAT in label_dict:
label_dict.pop(self.POS_FEAT)
# normalize morph string and add to morphology table
norm_morph = self.vocab.strings[self.vocab.morphology.add(label_dict)]
# add label mappings
if norm_label not in self.cfg["labels_morph"]:
self.cfg["labels_morph"][norm_label] = norm_morph
self.cfg["labels_pos"][norm_label] = POS_IDS[pos]
return 1
def begin_training(self, get_examples=lambda: [], pipeline=None, sgd=None,
@ -53,14 +69,16 @@ class Morphologizer(Tagger):
for example in get_examples():
for i, token in enumerate(example.reference):
pos = token.pos_
morph = token.morph
norm_morph = self.vocab.strings[self.vocab.morphology.add(morph)]
morph = token.morph_
# create and add the combined morph+POS label
morph_dict = Morphology.feats_to_dict(morph)
if pos:
morph["POS"] = pos
norm_morph_pos = self.vocab.strings[self.vocab.morphology.add(morph)]
if norm_morph_pos not in self.cfg["labels"]:
self.cfg["labels"][norm_morph_pos] = norm_morph
self.cfg["morph_pos"][norm_morph_pos] = POS_IDS[pos]
morph_dict[self.POS_FEAT] = pos
norm_label = self.vocab.strings[self.vocab.morphology.add(morph_dict)]
# add label->morph and label->POS mappings
if norm_label not in self.cfg["labels_morph"]:
self.cfg["labels_morph"][norm_label] = morph
self.cfg["labels_pos"][norm_label] = POS_IDS[pos]
self.set_output(len(self.labels))
self.model.initialize()
link_vectors_to_models(self.vocab)
@ -79,8 +97,8 @@ class Morphologizer(Tagger):
doc_tag_ids = doc_tag_ids.get()
for j, tag_id in enumerate(doc_tag_ids):
morph = self.labels[tag_id]
doc.c[j].morph = self.vocab.morphology.add(self.cfg["labels"][morph])
doc.c[j].pos = self.cfg["morph_pos"][morph]
doc.c[j].morph = self.vocab.morphology.add(self.cfg["labels_morph"][morph])
doc.c[j].pos = self.cfg["labels_pos"][morph]
doc.is_morphed = True
@ -94,14 +112,17 @@ class Morphologizer(Tagger):
for i in range(len(morphs)):
pos = pos_tags[i]
morph = morphs[i]
feats = Morphology.feats_to_dict(morph)
# POS may align (same value for multiple tokens) when morph
# doesn't, so if either is None, treat both as None here so that
# truths doesn't end up with an unknown morph+POS combination
if pos is None or morph is None:
pos = None
morph = None
label_dict = Morphology.feats_to_dict(morph)
if pos:
feats["POS"] = pos
if len(feats) > 0:
morph = self.vocab.strings[self.vocab.morphology.add(feats)]
if morph == "":
morph = Morphology.EMPTY_MORPH
eg_truths.append(morph)
label_dict[self.POS_FEAT] = pos
label = self.vocab.strings[self.vocab.morphology.add(label_dict)]
eg_truths.append(label)
truths.append(eg_truths)
d_scores, loss = loss_func(scores, truths)
if self.model.ops.xp.isnan(loss):

31
spacy/tests/pipeline/test_morphologizer.py
View File

@ -5,6 +5,7 @@ from spacy.gold import Example
from spacy.lang.en import English
from spacy.language import Language
from spacy.tests.util import make_tempdir
from spacy.morphology import Morphology
def test_label_types():
@ -23,9 +24,10 @@ TRAIN_DATA = [
"pos": ["NOUN", "VERB", "ADJ", "NOUN"],
},
),
# test combinations of morph+POS
(
"Eat blue ham",
{"morphs": ["Feat=V", "Feat=J", "Feat=N"], "pos": ["VERB", "ADJ", "NOUN"]},
{"morphs": ["Feat=V", "", ""], "pos": ["", "ADJ", ""]},
),
]
@ -38,7 +40,12 @@ def test_overfitting_IO():
for inst in TRAIN_DATA:
train_examples.append(Example.from_dict(nlp.make_doc(inst[0]), inst[1]))
for morph, pos in zip(inst[1]["morphs"], inst[1]["pos"]):
morphologizer.add_label(morph + "|POS=" + pos)
if morph and pos:
morphologizer.add_label(morph + Morphology.FEATURE_SEP + "POS" + Morphology.FIELD_SEP + pos)
elif pos:
morphologizer.add_label("POS" + Morphology.FIELD_SEP + pos)
elif morph:
morphologizer.add_label(morph)
nlp.add_pipe(morphologizer)
optimizer = nlp.begin_training()
@ -48,19 +55,27 @@ def test_overfitting_IO():
assert losses["morphologizer"] < 0.00001
# test the trained model
test_text = "I like blue eggs"
test_text = "I like blue ham"
doc = nlp(test_text)
gold_morphs = [
"Feat=N|POS=NOUN",
"Feat=V|POS=VERB",
"Feat=J|POS=ADJ",
"Feat=N|POS=NOUN",
"Feat=N",
"Feat=V",
"",
"",
]
gold_pos_tags = [
"NOUN",
"VERB",
"ADJ",
"",
]
assert [t.morph_ for t in doc] == gold_morphs
assert [t.pos_ for t in doc] == gold_pos_tags
# Also test the results are still the same after IO
with make_tempdir() as tmp_dir:
nlp.to_disk(tmp_dir)
nlp2 = util.load_model_from_path(tmp_dir)
doc2 = nlp2(test_text)
assert gold_morphs == [t.morph_ for t in doc2]
assert [t.morph_ for t in doc2] == gold_morphs
assert [t.pos_ for t in doc2] == gold_pos_tags

1
spacy/tests/pipeline/test_tagger.py
View File

@ -31,6 +31,7 @@ def test_overfitting_IO():
nlp.vocab.morphology.load_tag_map(TAG_MAP)
nlp.vocab.morphology.load_morph_exceptions(MORPH_RULES)
tagger = nlp.create_pipe("tagger")
nlp.vocab.morphology.load_tag_map(TAG_MAP)
train_examples = []
for t in TRAIN_DATA:
train_examples.append(Example.from_dict(nlp.make_doc(t[0]), t[1]))

5
website/gatsby-config.js
View File

@ -19,6 +19,7 @@ const isNightly = !!+process.env.SPACY_NIGHTLY || site.nightlyBranches.includes(
const favicon = isNightly ? `src/images/icon_nightly.png` : `src/images/icon.png`
const binderBranch = isNightly ? 'nightly' : site.binderBranch
const siteUrl = isNightly ? site.siteUrlNightly : site.siteUrl
const domain = isNightly ? site.domainNightly : site.domain
module.exports = {
siteMetadata: {
@ -148,6 +149,10 @@ module.exports = {
respectDNT: true,
},
},
{
resolve: `gatsby-plugin-plausible`,
options: { domain },
},
{
resolve: 'gatsby-plugin-robots-txt',
options: {

2
website/meta/site.json
View File

@ -3,7 +3,9 @@
"description": "spaCy is a free open-source library for Natural Language Processing in Python. It features NER, POS tagging, dependency parsing, word vectors and more.",
"slogan": "Industrial-strength Natural Language Processing in Python",
"siteUrl": "https://spacy.io",
"domain": "spacy.io",
"siteUrlNightly": "https://nightly.spacy.io",
"domainNightly": "nightly.spacy.io",
"nightlyBranches": ["nightly.spacy.io"],
"email": "contact@explosion.ai",
"company": "Explosion",

35
website/package-lock.json generated
View File

@ -12915,6 +12915,41 @@
}
}
},
"gatsby-plugin-plausible": {
"version": "0.0.6",
"resolved": "https://registry.npmjs.org/gatsby-plugin-plausible/-/gatsby-plugin-plausible-0.0.6.tgz",
"integrity": "sha512-qUdPQ3haeX2DIywGZ2boMpmFAnSbWzqS9cG9/OO0mWLigA0sDLWwGkpHIAvrfepgbB9U/roLtXflctBwOIxtcQ==",
"requires": {
"@babel/runtime": "^7.9.2",
"minimatch": "3.0.4",
"react": "^16.13.1"
},
"dependencies": {
"@babel/runtime": {
"version": "7.10.5",
"resolved": "https://registry.npmjs.org/@babel/runtime/-/runtime-7.10.5.tgz",
"integrity": "sha512-otddXKhdNn7d0ptoFRHtMLa8LqDxLYwTjB4nYgM1yy5N6gU/MUf8zqyyLltCH3yAVitBzmwK4us+DD0l/MauAg==",
"requires": {
"regenerator-runtime": "^0.13.4"
}
},
"react": {
"version": "16.13.1",
"resolved": "https://registry.npmjs.org/react/-/react-16.13.1.tgz",
"integrity": "sha512-YMZQQq32xHLX0bz5Mnibv1/LHb3Sqzngu7xstSM+vrkE5Kzr9xE0yMByK5kMoTK30YVJE61WfbxIFFvfeDKT1w==",
"requires": {
"loose-envify": "^1.1.0",
"object-assign": "^4.1.1",
"prop-types": "^15.6.2"
}
},
"regenerator-runtime": {
"version": "0.13.5",
"resolved": "https://registry.npmjs.org/regenerator-runtime/-/regenerator-runtime-0.13.5.tgz",
"integrity": "sha512-ZS5w8CpKFinUzOwW3c83oPeVXoNsrLsaCoLtJvAClH135j/R77RuymhiSErhm2lKcwSCIpmvIWSbDkIfAqKQlA=="
}
}
},
"gatsby-plugin-react-helmet": {
"version": "3.0.6",
"resolved": "https://registry.npmjs.org/gatsby-plugin-react-helmet/-/gatsby-plugin-react-helmet-3.0.6.tgz",

1
website/package.json
View File

@ -23,6 +23,7 @@
"gatsby-plugin-google-analytics": "^2.0.14",
"gatsby-plugin-manifest": "^2.0.17",
"gatsby-plugin-offline": "^2.0.24",
"gatsby-plugin-plausible": "0.0.6",
"gatsby-plugin-react-helmet": "^3.0.6",
"gatsby-plugin-react-svg": "^2.0.0",
"gatsby-plugin-robots-txt": "^1.5.1",