explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2024-12-25 17:36:30 +03:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'master' of https://github.com/explosion/spaCy

Browse Source
This commit is contained in:
ines 2017-04-16 12:07:04 +02:00
commit d7229967b0
48 changed files with 888 additions and 765 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
examples/train_ner_standalone.py → examples/training/train_ner_standalone.py
View File

@ -1,3 +1,4 @@
#!/usr/bin/env python
'''Example of training a named entity recognition system from scratch using spaCy
This example is written to be self-contained and reasonably transparent.
@ -81,7 +82,7 @@ def load_vocab(path):
def init_ner_model(vocab, features=None):
if features is None:
features = tuple(EntityRecognizer.feature_templates)
return BeamEntityRecognizer(vocab, features=features)
return EntityRecognizer(vocab, features=features)
def save_ner_model(model, path):
@ -99,7 +100,7 @@ def save_ner_model(model, path):
def load_ner_model(vocab, path):
return BeamEntityRecognizer.load(path, vocab)
return EntityRecognizer.load(path, vocab)
class Pipeline(object):
@ -110,18 +111,21 @@ class Pipeline(object):
raise IOError("Cannot load pipeline from %s\nDoes not exist" % path)
if not path.is_dir():
raise IOError("Cannot load pipeline from %s\nNot a directory" % path)
vocab = load_vocab(path / 'vocab')
vocab = load_vocab(path)
tokenizer = Tokenizer(vocab, {}, None, None, None)
ner_model = load_ner_model(vocab, path / 'ner')
return cls(vocab, tokenizer, ner_model)
def __init__(self, vocab=None, tokenizer=None, ner_model=None):
def __init__(self, vocab=None, tokenizer=None, entity=None):
if vocab is None:
self.vocab = init_vocab()
vocab = init_vocab()
if tokenizer is None:
tokenizer = Tokenizer(vocab, {}, None, None, None)
if ner_model is None:
self.entity = init_ner_model(self.vocab)
if entity is None:
entity = init_ner_model(self.vocab)
self.vocab = vocab
self.tokenizer = tokenizer
self.entity = entity
self.pipeline = [self.entity]
def __call__(self, input_):
@ -173,7 +177,7 @@ class Pipeline(object):
save_ner_model(self.entity, path / 'ner')
def train(nlp, train_examples, dev_examples, nr_epoch=5):
def train(nlp, train_examples, dev_examples, ctx, nr_epoch=5):
next_epoch = train_examples
print("Iter", "Loss", "P", "R", "F")
for i in range(nr_epoch):
@ -186,14 +190,17 @@ def train(nlp, train_examples, dev_examples, nr_epoch=5):
next_epoch.append((input_, annot))
random.shuffle(next_epoch)
scores = nlp.evaluate(dev_examples)
precision = '%.2f' % scores['ents_p']
recall = '%.2f' % scores['ents_r']
f_measure = '%.2f' % scores['ents_f']
print(i, int(loss), precision, recall, f_measure)
report_scores(i, loss, scores)
nlp.average_weights()
scores = nlp.evaluate(dev_examples)
print("After averaging")
print(scores['ents_p'], scores['ents_r'], scores['ents_f'])
report_scores(channels, i+1, loss, scores)
def report_scores(i, loss, scores):
precision = '%.2f' % scores['ents_p']
recall = '%.2f' % scores['ents_r']
f_measure = '%.2f' % scores['ents_f']
print('%d %s %s %s' % (int(loss), precision, recall, f_measure))
def read_examples(path):
@ -221,15 +228,17 @@ def read_examples(path):
train_loc=("Path to your training data", "positional", None, Path),
dev_loc=("Path to your development data", "positional", None, Path),
)
def main(model_dir, train_loc, dev_loc, nr_epoch=10):
def main(model_dir=Path('/home/matt/repos/spaCy/spacy/data/de-1.0.0'),
train_loc=None, dev_loc=None, nr_epoch=30):
train_examples = read_examples(train_loc)
dev_examples = read_examples(dev_loc)
nlp = Pipeline()
nlp = Pipeline.load(model_dir)
train(nlp, train_examples, list(dev_examples), nr_epoch)
train(nlp, train_examples, list(dev_examples), ctx, nr_epoch)
nlp.save(model_dir)
if __name__ == '__main__':
plac.call(main)
main()

74
examples/training/train_new_entity_type.py Normal file
View File

@ -0,0 +1,74 @@
from __future__ import unicode_literals, print_function
import json
import pathlib
import random
import spacy
from spacy.pipeline import EntityRecognizer
from spacy.gold import GoldParse
from spacy.tagger import Tagger
try:
unicode
except:
unicode = str
def train_ner(nlp, train_data, output_dir):
# Add new words to vocab.
for raw_text, _ in train_data:
doc = nlp.make_doc(raw_text)
for word in doc:
_ = nlp.vocab[word.orth]
for itn in range(20):
random.shuffle(train_data)
for raw_text, entity_offsets in train_data:
gold = GoldParse(doc, entities=entity_offsets)
doc = nlp.make_doc(raw_text)
nlp.tagger(doc)
loss = nlp.entity.update(doc, gold)
nlp.end_training()
nlp.save_to_directory(output_dir)
def main(model_name, output_directory=None):
nlp = spacy.load(model_name)
train_data = [
(
"Horses are too tall and they pretend to care about your feelings",
[(0, 6, 'ANIMAL')],
),
(
"horses are too tall and they pretend to care about your feelings",
[(0, 6, 'ANIMAL')]
),
(
"horses pretend to care about your feelings",
[(0, 6, 'ANIMAL')]
),
(
"they pretend to care about your feelings, those horses",
[(48, 54, 'ANIMAL')]
)
]
nlp.entity.add_label('ANIMAL')
if output_directory is not None:
output_directory = pathlib.Path(output_directory)
ner = train_ner(nlp, train_data, output_directory)
doc = nlp('Do you like horses?')
for ent in doc.ents:
print(ent.label_, ent.text)
nlp2 = spacy.load('en', path=output_directory)
nlp2.entity.add_label('ANIMAL')
doc2 = nlp2('Do you like horses?')
for ent in doc2.ents:
print(ent.label_, ent.text)
if __name__ == '__main__':
import plac
plac.call(main)

2
fabfile.py vendored
View File

@ -14,7 +14,7 @@ VENV_DIR = path.join(PWD, ENV)
def env(lang='python2.7'):
if path.exists(VENV_DIR):
local('rm -rf {env}'.format(env=VENV_DIR))
local('virtualenv -p {lang} {env}'.format(lang=lang, env=VENV_DIR))
local('python -m virtualenv -p {lang} {env}'.format(lang=lang, env=VENV_DIR))
def install():

20
spacy/__init__.py
View File

@ -1,27 +1,13 @@
# coding: utf8
from __future__ import unicode_literals, print_function
from __future__ import unicode_literals
import json
from pathlib import Path
from .util import set_lang_class, get_lang_class, parse_package_meta
from .deprecated import resolve_model_name
from .cli import info
from . import en
from . import de
from . import zh
from . import es
from . import it
from . import hu
from . import fr
from . import pt
from . import nl
from . import sv
from . import fi
from . import bn
from . import he
from .about import *
from . import en, de, zh, es, it, hu, fr, pt, nl, sv, fi, bn, he
set_lang_class(en.English.lang, en.English)

12
spacy/__main__.py
View File

@ -14,8 +14,9 @@ from spacy.cli import convert as cli_convert
class CLI(object):
"""Command-line interface for spaCy"""
"""
Command-line interface for spaCy
"""
commands = ('download', 'link', 'info', 'package', 'train', 'model', 'convert')
@plac.annotations(
@ -29,7 +30,6 @@ class CLI(object):
can be shortcut, model name or, if --direct flag is set, full model name
with version.
"""
cli_download(model, direct)
@ -44,7 +44,6 @@ class CLI(object):
either the name of a pip package, or the local path to the model data
directory. Linking models allows loading them via spacy.load(link_name).
"""
cli_link(origin, link_name, force)
@ -58,7 +57,6 @@ class CLI(object):
speficied as an argument, print model information. Flag --markdown
prints details in Markdown for easy copy-pasting to GitHub issues.
"""
cli_info(model, markdown)
@ -73,7 +71,6 @@ class CLI(object):
installation files. A new directory will be created in the specified
output directory, and model data will be copied over.
"""
cli_package(input_dir, output_dir, force)
@ -93,7 +90,6 @@ class CLI(object):
"""
Train a model. Expects data in spaCy's JSON format.
"""
cli_train(lang, output_dir, train_data, dev_data, n_iter, not no_tagger,
not no_parser, not no_ner, parser_L1)
@ -108,7 +104,6 @@ class CLI(object):
"""
Initialize a new model and its data directory.
"""
cli_model(lang, model_dir, freqs_data, clusters_data, vectors_data)
@plac.annotations(
@ -122,7 +117,6 @@ class CLI(object):
Convert files into JSON format for use with train command and other
experiment management functions.
"""
cli_convert(input_file, output_dir, n_sents, morphology)

9
spacy/attrs.pyx
View File

@ -1,3 +1,7 @@
# coding: utf8
from __future__ import unicode_literals
IDS = {
"": NULL_ATTR,
"IS_ALPHA": IS_ALPHA,
@ -92,7 +96,8 @@ NAMES = [key for key, value in sorted(IDS.items(), key=lambda item: item[1])]
def intify_attrs(stringy_attrs, strings_map=None, _do_deprecated=False):
'''Normalize a dictionary of attributes, converting them to ints.
"""
Normalize a dictionary of attributes, converting them to ints.
Arguments:
stringy_attrs (dict):
@ -105,7 +110,7 @@ def intify_attrs(stringy_attrs, strings_map=None, _do_deprecated=False):
inty_attrs (dict):
Attributes dictionary with keys and optionally values converted to
ints.
'''
"""
inty_attrs = {}
if _do_deprecated:
if 'F' in stringy_attrs:

3
spacy/cfile.pyx
View File

@ -1,3 +1,6 @@
# coding: utf8
from __future__ import unicode_literals
from libc.stdio cimport fopen, fclose, fread, fwrite
from libc.string cimport memcpy

1
spacy/cli/convert.py
View File

@ -1,7 +1,6 @@
# coding: utf8
from __future__ import unicode_literals
import io
from pathlib import Path
from .converters import conllu2json

4
spacy/cli/converters/conllu2json.py
View File

@ -2,12 +2,12 @@
from __future__ import unicode_literals
import json
from ...gold import read_json_file, merge_sents
from ... import util
def conllu2json(input_path, output_path, n_sents=10, use_morphology=False):
"""Convert conllu files into JSON format for use with train cli.
"""
Convert conllu files into JSON format for use with train cli.
use_morphology parameter enables appending morphology to tags, which is
useful for languages such as Spanish, where UD tags are not so rich.
"""

1
spacy/cli/download.py
View File

@ -1,7 +1,6 @@
# coding: utf8
from __future__ import unicode_literals
import pip
import requests
import os
import subprocess

3
spacy/cli/info.py
View File

@ -18,7 +18,6 @@ def info(model=None, markdown=False):
else:
data['source'] = str(model_path)
print_info(data, "model " + model, markdown)
else:
data = get_spacy_data()
print_info(data, "spaCy", markdown)
@ -26,10 +25,8 @@ def info(model=None, markdown=False):
def print_info(data, title, markdown):
title = "Info about {title}".format(title=title)
if markdown:
util.print_markdown(data, title=title)
else:
util.print_table(data, title=title)

30
spacy/cli/link.py
View File

@ -4,6 +4,7 @@ from __future__ import unicode_literals
import pip
from pathlib import Path
import importlib
from ..compat import unicode_, symlink_to
from .. import util
@ -20,7 +21,6 @@ def link_package(package_name, link_name, force=False):
# Python's installation and import rules are very complicated.
pkg = importlib.import_module(package_name)
package_path = Path(pkg.__file__).parent.parent
meta = get_meta(package_path, package_name)
model_name = package_name + '-' + meta['version']
model_path = package_path / package_name / model_name
@ -43,23 +43,17 @@ def symlink(model_path, link_name, force):
elif link_path.exists():
link_path.unlink()
# Add workaround for Python 2 on Windows (see issue #909)
if util.is_python2() and util.is_windows():
import subprocess
command = ['mklink', '/d', unicode(link_path), unicode(model_path)]
try:
subprocess.call(command, shell=True)
except:
# This is quite dirty, but just making sure other Windows-specific
# errors are caught so users at least see a proper error message.
util.sys_exit(
"Creating a symlink in spacy/data failed. You can still import "
"the model as a Python package and call its load() method, or "
"create the symlink manually:",
"{a} --> {b}".format(a=unicode(model_path), b=unicode(link_path)),
title="Error: Couldn't link model to '{l}'".format(l=link_name))
else:
link_path.symlink_to(model_path)
try:
symlink_to(link_path, model_path)
except:
# This is quite dirty, but just making sure other errors are caught so
# users at least see a proper message.
util.sys_exit(
"Creating a symlink in spacy/data failed. You can still import "
"the model as a Python package and call its load() method, or "
"create the symlink manually:",
"{a} --> {b}".format(a=unicode_(model_path), b=unicode_(link_path)),
title="Error: Couldn't link model to '{l}'".format(l=link_name))
util.print_msg(
"{a} --> {b}".format(a=model_path.as_posix(), b=link_path.as_posix()),

21
spacy/cli/package.py
View File

@ -1,20 +1,13 @@
# coding: utf8
from __future__ import unicode_literals
import json
import shutil
import requests
from pathlib import Path
import six
from .. import about
from ..compat import unicode_, json_dumps
from .. import util
if six.PY2:
json_dumps = lambda data: json.dumps(data, indent=2).decode("utf8")
elif six.PY3:
json_dumps = lambda data: json.dumps(data, indent=2)
def package(input_dir, output_dir, force):
input_path = Path(input_dir)
@ -32,31 +25,31 @@ def package(input_dir, output_dir, force):
package_path = main_path / model_name
create_dirs(package_path, force)
shutil.copytree(input_path.as_posix(), (package_path / model_name_v).as_posix())
shutil.copytree(unicode_(input_path), unicode_(package_path / model_name_v))
create_file(main_path / 'meta.json', json_dumps(meta))
create_file(main_path / 'setup.py', template_setup)
create_file(main_path / 'MANIFEST.in', template_manifest)
create_file(package_path / '__init__.py', template_init)
util.print_msg(
main_path.as_posix(),
unicode_(main_path),
"To build the package, run `python setup.py sdist` in that directory.",
title="Successfully created package {p}".format(p=model_name_v))
def check_dirs(input_path, output_path):
if not input_path.exists():
util.sys_exit(input_path.as_poisx(), title="Model directory not found")
util.sys_exit(unicode_(input_path.as_poisx), title="Model directory not found")
if not output_path.exists():
util.sys_exit(output_path.as_posix(), title="Output directory not found")
util.sys_exit(unicode_(output_path), title="Output directory not found")
def create_dirs(package_path, force):
if package_path.exists():
if force:
shutil.rmtree(package_path.as_posix())
shutil.rmtree(unicode_(package_path.as_posix))
else:
util.sys_exit(package_path.as_posix(),
util.sys_exit(unicode_(package_path.as_posix),
"Please delete the directory and try again.",
title="Package directory already exists")
Path.mkdir(package_path, parents=True)

3
spacy/cli/train.py
View File

@ -5,8 +5,6 @@ import json
from pathlib import Path
from ..scorer import Scorer
from ..tagger import Tagger
from ..syntax.parser import Parser
from ..gold import GoldParse, merge_sents
from ..gold import read_json_file as read_gold_json
from .. import util
@ -60,7 +58,6 @@ def train_model(Language, train_data, dev_data, output_path, tagger_cfg, parser_
print("Itn.\tN weight\tN feats\tUAS\tNER F.\tTag %\tToken %")
with Language.train(output_path, train_data, tagger_cfg, parser_cfg, entity_cfg) as trainer:
loss = 0
for itn, epoch in enumerate(trainer.epochs(n_iter, augment_data=None)):
for doc, gold in epoch:
trainer.update(doc, gold)

54
spacy/compat.py Normal file
View File

@ -0,0 +1,54 @@
# coding: utf8
from __future__ import unicode_literals
import six
import sys
import ujson
try:
import cPickle as pickle
except ImportError:
import pickle
try:
import copy_reg
except ImportError:
import copyreg as copy_reg
is_python2 = six.PY2
is_python3 = six.PY3
is_windows = sys.platform.startswith('win')
is_linux = sys.platform.startswith('linux')
is_osx = sys.platform == 'darwin'
if is_python2:
bytes_ = str
unicode_ = unicode
basestring_ = basestring
input_ = raw_input
json_dumps = lambda data: ujson.dumps(data, indent=2).decode('utf8')
elif is_python3:
bytes_ = bytes
unicode_ = str
basestring_ = str
input_ = input
json_dumps = lambda data: ujson.dumps(data, indent=2)
def symlink_to(orig, dest):
if is_python2 and is_windows:
import subprocess
subprocess.call(['mklink', '/d', unicode(orig), unicode(dest)], shell=True)
else:
orig.symlink_to(dest)
def is_config(python2=None, python3=None, windows=None, linux=None, osx=None):
return ((python2 == None or python2 == is_python2) and
(python3 == None or python3 == is_python3) and
(windows == None or windows == is_windows) and
(linux == None or linux == is_linux) and
(osx == None or osx == is_osx))

57
spacy/deprecated.py
View File

@ -1,16 +1,14 @@
# coding: utf8
from __future__ import unicode_literals
from pathlib import Path
from . import about
from . import util
from .cli import download
from .cli import link
try:
basestring
except NameError:
basestring = str
def read_lang_data(package):
tokenization = package.load_json(('tokenizer', 'specials.json'))
with package.open(('tokenizer', 'prefix.txt'), default=None) as file_:
@ -36,7 +34,8 @@ def align_tokens(ref, indices): # Deprecated, surely?
def detokenize(token_rules, words): # Deprecated?
"""To align with treebanks, return a list of "chunks", where a chunk is a
"""
To align with treebanks, return a list of "chunks", where a chunk is a
sequence of tokens that are separated by whitespace in actual strings. Each
chunk should be a tuple of token indices, e.g.
@ -57,10 +56,30 @@ def detokenize(token_rules, words): # Deprecated?
return positions
def fix_glove_vectors_loading(overrides):
"""Special-case hack for loading the GloVe vectors, to support deprecated
<1.0 stuff. Phase this out once the data is fixed."""
def match_best_version(target_name, target_version, path):
path = util.ensure_path(path)
if path is None or not path.exists():
return None
matches = []
for data_name in path.iterdir():
name, version = split_data_name(data_name.parts[-1])
if name == target_name:
matches.append((tuple(float(v) for v in version.split('.')), data_name))
if matches:
return Path(max(matches)[1])
else:
return None
def split_data_name(name):
return name.split('-', 1) if '-' in name else (name, '')
def fix_glove_vectors_loading(overrides):
"""
Special-case hack for loading the GloVe vectors, to support deprecated
<1.0 stuff. Phase this out once the data is fixed.
"""
if 'data_dir' in overrides and 'path' not in overrides:
raise ValueError("The argument 'data_dir' has been renamed to 'path'")
if overrides.get('path') is False:
@ -68,18 +87,16 @@ def fix_glove_vectors_loading(overrides):
if overrides.get('path') in (None, True):
data_path = util.get_data_path()
else:
path = overrides['path']
if isinstance(path, basestring):
path = Path(path)
path = util.ensure_path(overrides['path'])
data_path = path.parent
vec_path = None
if 'add_vectors' not in overrides:
if 'vectors' in overrides:
vec_path = util.match_best_version(overrides['vectors'], None, data_path)
vec_path = match_best_version(overrides['vectors'], None, data_path)
if vec_path is None:
return overrides
else:
vec_path = util.match_best_version('en_glove_cc_300_1m_vectors', None, data_path)
vec_path = match_best_version('en_glove_cc_300_1m_vectors', None, data_path)
if vec_path is not None:
vec_path = vec_path / 'vocab' / 'vec.bin'
if vec_path is not None:
@ -88,13 +105,13 @@ def fix_glove_vectors_loading(overrides):
def resolve_model_name(name):
"""If spaCy is loaded with 'de', check if symlink already exists. If
"""
If spaCy is loaded with 'de', check if symlink already exists. If
not, user have upgraded from older version and have old models installed.
Check if old model directory exists and if so, return that instead and create
shortcut link. If English model is found and no shortcut exists, raise error
and tell user to install new model.
"""
if name == 'en' or name == 'de':
versions = ['1.0.0', '1.1.0']
data_path = Path(util.get_data_path())
@ -117,9 +134,11 @@ def resolve_model_name(name):
class ModelDownload():
"""Replace download modules within en and de with deprecation warning and
"""
Replace download modules within en and de with deprecation warning and
download default language model (using shortcut). Use classmethods to allow
importing ModelDownload as download and calling download.en() etc."""
importing ModelDownload as download and calling download.en() etc.
"""
@classmethod
def load(self, lang):

6
spacy/en/__init__.py
View File

@ -11,12 +11,6 @@ from ..deprecated import fix_glove_vectors_loading
from .language_data import *
try:
basestring
except NameError:
basestring = str
class English(Language):
lang = 'en'

34
spacy/gold.pyx
View File

@ -1,13 +1,11 @@
# cython: profile=True
# coding: utf8
from __future__ import unicode_literals, print_function
import io
import json
import re
import os
from os import path
import ujson as json
import ujson
from pathlib import Path
from .syntax import nonproj
@ -141,12 +139,13 @@ def _min_edit_path(cand_words, gold_words):
def read_json_file(loc, docs_filter=None):
if path.isdir(loc):
for filename in os.listdir(loc):
yield from read_json_file(path.join(loc, filename))
loc = Path(loc)
if loc.is_dir():
for filename in loc.iterdir():
yield from read_json_file(loc / filename)
else:
with io.open(loc, 'r', encoding='utf8') as file_:
docs = json.load(file_)
docs = ujson.load(file_)
for doc in docs:
if docs_filter is not None and not docs_filter(doc):
continue
@ -220,7 +219,8 @@ cdef class GoldParse:
def __init__(self, doc, annot_tuples=None, words=None, tags=None, heads=None,
deps=None, entities=None, make_projective=False):
"""Create a GoldParse.
"""
Create a GoldParse.
Arguments:
doc (Doc):
@ -302,7 +302,8 @@ cdef class GoldParse:
self.heads = proj_heads
def __len__(self):
"""Get the number of gold-standard tokens.
"""
Get the number of gold-standard tokens.
Returns (int): The number of gold-standard tokens.
"""
@ -310,13 +311,16 @@ cdef class GoldParse:
@property
def is_projective(self):
"""Whether the provided syntactic annotations form a projective dependency
tree."""
"""
Whether the provided syntactic annotations form a projective dependency
tree.
"""
return not nonproj.is_nonproj_tree(self.heads)
def biluo_tags_from_offsets(doc, entities):
'''Encode labelled spans into per-token tags, using the Begin/In/Last/Unit/Out
"""
Encode labelled spans into per-token tags, using the Begin/In/Last/Unit/Out
scheme (biluo).
Arguments:
@ -347,7 +351,7 @@ def biluo_tags_from_offsets(doc, entities):
tags = biluo_tags_from_offsets(doc, entities)
assert tags == ['O', 'O', 'U-LOC', 'O']
'''
"""
starts = {token.idx: token.i for token in doc}
ends = {token.idx+len(token): token.i for token in doc}
biluo = ['-' for _ in doc]

178
spacy/language.py
View File

@ -1,39 +1,26 @@
from __future__ import absolute_import
from __future__ import unicode_literals
import pathlib
# coding: utf8
from __future__ import absolute_import, unicode_literals
from contextlib import contextmanager
import shutil
import ujson
try:
basestring
except NameError:
basestring = str
try:
unicode
except NameError:
unicode = str
from .tokenizer import Tokenizer
from .vocab import Vocab
from .tagger import Tagger
from .matcher import Matcher
from . import attrs
from . import orth
from . import util
from . import language_data
from .lemmatizer import Lemmatizer
from .train import Trainer
from .attrs import TAG, DEP, ENT_IOB, ENT_TYPE, HEAD, PROB, LANG, IS_STOP
from .syntax.parser import get_templates
from .syntax.nonproj import PseudoProjectivity
from .pipeline import DependencyParser, EntityRecognizer
from .syntax.arc_eager import ArcEager
from .syntax.ner import BiluoPushDown
from .compat import unicode_
from .attrs import IS_STOP
from . import attrs
from . import orth
from . import util
from . import language_data
class BaseDefaults(object):
@ -150,25 +137,15 @@ class BaseDefaults(object):
return pipeline
token_match = language_data.TOKEN_MATCH
prefixes = tuple(language_data.TOKENIZER_PREFIXES)
suffixes = tuple(language_data.TOKENIZER_SUFFIXES)
infixes = tuple(language_data.TOKENIZER_INFIXES)
tag_map = dict(language_data.TAG_MAP)
tokenizer_exceptions = {}
parser_features = get_templates('parser')
entity_features = get_templates('ner')
tagger_features = Tagger.feature_templates # TODO -- fix this
stop_words = set()
lemma_rules = {}
lemma_exc = {}
lemma_index = {}
@ -202,53 +179,42 @@ class BaseDefaults(object):
class Language(object):
'''A text-processing pipeline. Usually you'll load this once per process, and
"""
A text-processing pipeline. Usually you'll load this once per process, and
pass the instance around your program.
'''
"""
Defaults = BaseDefaults
lang = None
@classmethod
@contextmanager
def train(cls, path, gold_tuples, *configs):
if isinstance(path, basestring):
path = pathlib.Path(path)
tagger_cfg, parser_cfg, entity_cfg = configs
dep_model_dir = path / 'deps'
ner_model_dir = path / 'ner'
pos_model_dir = path / 'pos'
if dep_model_dir.exists():
shutil.rmtree(str(dep_model_dir))
if ner_model_dir.exists():
shutil.rmtree(str(ner_model_dir))
if pos_model_dir.exists():
shutil.rmtree(str(pos_model_dir))
dep_model_dir.mkdir()
ner_model_dir.mkdir()
pos_model_dir.mkdir()
def setup_directory(cls, path, **configs):
for name, config in configs.items():
directory = path / name
if directory.exists():
shutil.rmtree(str(directory))
directory.mkdir()
with (directory / 'config.json').open('wb') as file_:
data = ujson.dumps(config, indent=2)
if isinstance(data, unicode_):
data = data.encode('utf8')
file_.write(data)
if not (path / 'vocab').exists():
(path / 'vocab').mkdir()
@classmethod
@contextmanager
def train(cls, path, gold_tuples, **configs):
if parser_cfg['pseudoprojective']:
# preprocess training data here before ArcEager.get_labels() is called
gold_tuples = PseudoProjectivity.preprocess_training_data(gold_tuples)
parser_cfg['actions'] = ArcEager.get_actions(gold_parses=gold_tuples)
entity_cfg['actions'] = BiluoPushDown.get_actions(gold_parses=gold_tuples)
for subdir in ('deps', 'ner', 'pos'):
if subdir not in configs:
configs[subdir] = {}
configs['deps']['actions'] = ArcEager.get_actions(gold_parses=gold_tuples)
configs['ner']['actions'] = BiluoPushDown.get_actions(gold_parses=gold_tuples)
with (dep_model_dir / 'config.json').open('wb') as file_:
data = ujson.dumps(parser_cfg)
if isinstance(data, unicode):
data = data.encode('utf8')
file_.write(data)
with (ner_model_dir / 'config.json').open('wb') as file_:
data = ujson.dumps(entity_cfg)
if isinstance(data, unicode):
data = data.encode('utf8')
file_.write(data)
with (pos_model_dir / 'config.json').open('wb') as file_:
data = ujson.dumps(tagger_cfg)
if isinstance(data, unicode):
data = data.encode('utf8')
file_.write(data)
cls.setup_directory(path, **configs)
self = cls(
path=path,
@ -269,14 +235,14 @@ class Language(object):
self.entity = self.Defaults.create_entity(self)
self.pipeline = self.Defaults.create_pipeline(self)
yield Trainer(self, gold_tuples)
self.end_training(path=path)
self.end_training()
self.save_to_directory(path, deps=self.parser.cfg, ner=self.entity.cfg,
pos=self.tagger.cfg)
def __init__(self, **overrides):
if 'data_dir' in overrides and 'path' not in overrides:
raise ValueError("The argument 'data_dir' has been renamed to 'path'")
path = overrides.get('path', True)
if isinstance(path, basestring):
path = pathlib.Path(path)
path = util.ensure_path(overrides.get('path', True))
if path is True:
path = util.get_data_path() / self.lang
if not path.exists() and 'path' not in overrides:
@ -322,7 +288,8 @@ class Language(object):
self.pipeline = [self.tagger, self.parser, self.matcher, self.entity]
def __call__(self, text, tag=True, parse=True, entity=True):
"""Apply the pipeline to some text. The text can span multiple sentences,
"""
Apply the pipeline to some text. The text can span multiple sentences,
and can contain arbtrary whitespace. Alignment into the original string
is preserved.
@ -352,7 +319,8 @@ class Language(object):
return doc
def pipe(self, texts, tag=True, parse=True, entity=True, n_threads=2, batch_size=1000):
'''Process texts as a stream, and yield Doc objects in order.
"""
Process texts as a stream, and yield Doc objects in order.
Supports GIL-free multi-threading.
@ -361,7 +329,7 @@ class Language(object):
tag (bool)
parse (bool)
entity (bool)
'''
"""
skip = {self.tagger: not tag, self.parser: not parse, self.entity: not entity}
stream = (self.make_doc(text) for text in texts)
for proc in self.pipeline:
@ -373,51 +341,35 @@ class Language(object):
for doc in stream:
yield doc
def end_training(self, path=None):
if path is None:
path = self.path
elif isinstance(path, basestring):
path = pathlib.Path(path)
def save_to_directory(self, path):
configs = {
'pos': self.tagger.cfg if self.tagger else {},
'deps': self.parser.cfg if self.parser else {},
'ner': self.entity.cfg if self.entity else {},
}
if self.tagger:
self.tagger.model.end_training()
self.tagger.model.dump(str(path / 'pos' / 'model'))
if self.parser:
self.parser.model.end_training()
self.parser.model.dump(str(path / 'deps' / 'model'))
if self.entity:
self.entity.model.end_training()
self.entity.model.dump(str(path / 'ner' / 'model'))
self.setup_directory(path, **configs)
strings_loc = path / 'vocab' / 'strings.json'
with strings_loc.open('w', encoding='utf8') as file_:
self.vocab.strings.dump(file_)
self.vocab.dump(path / 'vocab' / 'lexemes.bin')
# TODO: Word vectors?
if self.tagger:
tagger_freqs = list(self.tagger.freqs[TAG].items())
else:
tagger_freqs = []
self.tagger.model.dump(str(path / 'pos' / 'model'))
if self.parser:
dep_freqs = list(self.parser.moves.freqs[DEP].items())
head_freqs = list(self.parser.moves.freqs[HEAD].items())
else:
dep_freqs = []
head_freqs = []
self.parser.model.dump(str(path / 'deps' / 'model'))
if self.entity:
entity_iob_freqs = list(self.entity.moves.freqs[ENT_IOB].items())
entity_type_freqs = list(self.entity.moves.freqs[ENT_TYPE].items())
else:
entity_iob_freqs = []
entity_type_freqs = []
with (path / 'vocab' / 'serializer.json').open('wb') as file_:
data = ujson.dumps([
(TAG, tagger_freqs),
(DEP, dep_freqs),
(ENT_IOB, entity_iob_freqs),
(ENT_TYPE, entity_type_freqs),
(HEAD, head_freqs)
])
if isinstance(data, unicode):
data = data.encode('utf8')
file_.write(data)
self.entity.model.dump(str(path / 'ner' / 'model'))
def end_training(self, path=None):
if self.tagger:
self.tagger.model.end_training()
if self.parser:
self.parser.model.end_training()
if self.entity:
self.entity.model.end_training()
# NB: This is slightly different from before --- we no longer default
# to taking nlp.path
if path is not None:
self.save_to_directory(path)

17
spacy/lemmatizer.py
View File

@ -1,13 +1,8 @@
from __future__ import unicode_literals, print_function
import codecs
import pathlib
import ujson as json
# coding: utf8
from __future__ import unicode_literals
from .symbols import POS, NOUN, VERB, ADJ, PUNCT
from .symbols import VerbForm_inf, VerbForm_none
from .symbols import Number_sing
from .symbols import Degree_pos
from .symbols import VerbForm_inf, VerbForm_none, Number_sing, Degree_pos
class Lemmatizer(object):
@ -38,8 +33,10 @@ class Lemmatizer(object):
return lemmas
def is_base_form(self, univ_pos, morphology=None):
'''Check whether we're dealing with an uninflected paradigm, so we can
avoid lemmatization entirely.'''
"""
Check whether we're dealing with an uninflected paradigm, so we can
avoid lemmatization entirely.
"""
morphology = {} if morphology is None else morphology
others = [key for key in morphology if key not in (POS, 'number', 'pos', 'verbform')]
true_morph_key = morphology.get('morph', 0)

25
spacy/lexeme.pyx
View File

@ -1,4 +1,7 @@
# cython: embedsignature=True
# coding: utf8
from __future__ import unicode_literals, print_function
from libc.math cimport sqrt
from cpython.ref cimport Py_INCREF
from cymem.cymem cimport Pool
@ -9,14 +12,11 @@ from cython.view cimport array as cvarray
cimport numpy as np
np.import_array()
from libc.string cimport memset
import numpy
from .orth cimport word_shape
from .typedefs cimport attr_t, flags_t
import numpy
from .attrs cimport IS_ALPHA, IS_ASCII, IS_DIGIT, IS_LOWER, IS_PUNCT, IS_SPACE
from .attrs cimport IS_TITLE, IS_UPPER, LIKE_URL, LIKE_NUM, LIKE_EMAIL, IS_STOP
from .attrs cimport IS_BRACKET
@ -30,13 +30,15 @@ memset(&EMPTY_LEXEME, 0, sizeof(LexemeC))
cdef class Lexeme:
"""An entry in the vocabulary. A Lexeme has no string context --- it's a
"""
An entry in the vocabulary. A Lexeme has no string context --- it's a
word-type, as opposed to a word token. It therefore has no part-of-speech
tag, dependency parse, or lemma (lemmatization depends on the part-of-speech
tag).
"""
def __init__(self, Vocab vocab, int orth):
"""Create a Lexeme object.
"""
Create a Lexeme object.
Arguments:
vocab (Vocab): The parent vocabulary
@ -80,7 +82,8 @@ cdef class Lexeme:
return self.c.orth
def set_flag(self, attr_id_t flag_id, bint value):
"""Change the value of a boolean flag.
"""
Change the value of a boolean flag.
Arguments:
flag_id (int): The attribute ID of the flag to set.
@ -89,7 +92,8 @@ cdef class Lexeme:
Lexeme.c_set_flag(self.c, flag_id, value)
def check_flag(self, attr_id_t flag_id):
"""Check the value of a boolean flag.
"""
Check the value of a boolean flag.
Arguments:
flag_id (int): The attribute ID of the flag to query.
@ -98,7 +102,8 @@ cdef class Lexeme:
return True if Lexeme.c_check_flag(self.c, flag_id) else False
def similarity(self, other):
'''Compute a semantic similarity estimate. Defaults to cosine over vectors.
"""
Compute a semantic similarity estimate. Defaults to cosine over vectors.
Arguments:
other:
@ -106,7 +111,7 @@ cdef class Lexeme:
Token and Lexeme objects.
Returns:
score (float): A scalar similarity score. Higher is more similar.
'''
"""
if self.vector_norm == 0 or other.vector_norm == 0:
return 0.0
return numpy.dot(self.vector, other.vector) / (self.vector_norm * other.vector_norm)

39
spacy/matcher.pyx
View File

@ -1,7 +1,10 @@
# cython: profile=True
# cython: infer_types=True
# coding: utf8
from __future__ import unicode_literals
import ujson
from .typedefs cimport attr_t
from .typedefs cimport hash_t
from .attrs cimport attr_id_t
@ -52,12 +55,6 @@ from .attrs import FLAG36 as L9_ENT
from .attrs import FLAG35 as L10_ENT
try:
import ujson as json
except ImportError:
import json
cpdef enum quantifier_t:
_META
ONE
@ -164,7 +161,7 @@ def _convert_strings(token_specs, string_store):
def merge_phrase(matcher, doc, i, matches):
'''Callback to merge a phrase on match'''
ent_id, label, start, end = matches[i]
span = doc[start : end]
span = doc[start : end]
span.merge(ent_type=label, ent_id=ent_id)
@ -180,7 +177,8 @@ cdef class Matcher:
@classmethod
def load(cls, path, vocab):
'''Load the matcher and patterns from a file path.
"""
Load the matcher and patterns from a file path.
Arguments:
path (Path):
@ -189,16 +187,17 @@ cdef class Matcher:
The vocabulary that the documents to match over will refer to.
Returns:
Matcher: The newly constructed object.
'''
"""
if (path / 'gazetteer.json').exists():
with (path / 'gazetteer.json').open('r', encoding='utf8') as file_:
patterns = json.load(file_)
patterns = ujson.load(file_)
else:
patterns = {}
return cls(vocab, patterns)
def __init__(self, vocab, patterns={}):
"""Create the Matcher.
"""
Create the Matcher.
Arguments:
vocab (Vocab):
@ -227,7 +226,8 @@ cdef class Matcher:
def add_entity(self, entity_key, attrs=None, if_exists='raise',
acceptor=None, on_match=None):
"""Add an entity to the matcher.
"""
Add an entity to the matcher.
Arguments:
entity_key (unicode or int):
@ -264,7 +264,8 @@ cdef class Matcher:
self._callbacks[entity_key] = on_match
def add_pattern(self, entity_key, token_specs, label=""):
"""Add a pattern to the matcher.
"""
Add a pattern to the matcher.
Arguments:
entity_key (unicode or int):
@ -307,7 +308,8 @@ cdef class Matcher:
return entity_key
def has_entity(self, entity_key):
"""Check whether the matcher has an entity.
"""
Check whether the matcher has an entity.
Arguments:
entity_key (string or int): The entity key to check.
@ -318,7 +320,8 @@ cdef class Matcher:
return entity_key in self._entities
def get_entity(self, entity_key):
"""Retrieve the attributes stored for an entity.
"""
Retrieve the attributes stored for an entity.
Arguments:
entity_key (unicode or int): The entity to retrieve.
@ -332,7 +335,8 @@ cdef class Matcher:
return None
def __call__(self, Doc doc, acceptor=None):
"""Find all token sequences matching the supplied patterns on the Doc.
"""
Find all token sequences matching the supplied patterns on the Doc.
Arguments:
doc (Doc):
@ -445,7 +449,8 @@ cdef class Matcher:
return matches
def pipe(self, docs, batch_size=1000, n_threads=2):
"""Match a stream of documents, yielding them in turn.
"""
Match a stream of documents, yielding them in turn.
Arguments:
docs: A stream of documents.

15
spacy/morphology.pyx
View File

@ -1,13 +1,9 @@
# cython: infer_types
# coding: utf8
from __future__ import unicode_literals
from libc.string cimport memset
try:
import ujson as json
except ImportError:
import json
from .parts_of_speech cimport ADJ, VERB, NOUN, PUNCT
from .attrs cimport POS, IS_SPACE
from .parts_of_speech import IDS as POS_IDS
@ -16,7 +12,9 @@ from .attrs import LEMMA, intify_attrs
def _normalize_props(props):
'''Transform deprecated string keys to correct names.'''
"""
Transform deprecated string keys to correct names.
"""
out = {}
for key, value in props.items():
if key == POS:
@ -98,13 +96,14 @@ cdef class Morphology:
flags[0] &= ~(one << flag_id)
def add_special_case(self, unicode tag_str, unicode orth_str, attrs, force=False):
'''Add a special-case rule to the morphological analyser. Tokens whose
"""
Add a special-case rule to the morphological analyser. Tokens whose
tag and orth match the rule will receive the specified properties.
Arguments:
tag (unicode): The part-of-speech tag to key the exception.
orth (unicode): The word-form to key the exception.
'''
"""
tag = self.strings[tag_str]
tag_id = self.reverse_index[tag]
orth = self.strings[orth_str]

8
spacy/multi_words.py
View File

@ -1,8 +0,0 @@
class RegexMerger(object):
def __init__(self, regexes):
self.regexes = regexes
def __call__(self, tokens):
for tag, entity_type, regex in self.regexes:
for m in regex.finditer(tokens.string):
tokens.merge(m.start(), m.end(), tag, m.group(), entity_type)

3
spacy/orth.pyx
View File

@ -1,6 +1,7 @@
# coding: utf8
# cython: infer_types=True
# coding: utf8
from __future__ import unicode_literals
import unicodedata
import re

1
spacy/parts_of_speech.pyx
View File

@ -1,3 +1,4 @@
# coding: utf8
from __future__ import unicode_literals

39
spacy/pipeline.pyx
View File

@ -1,3 +1,6 @@
# coding: utf8
from __future__ import unicode_literals
from .syntax.parser cimport Parser
from .syntax.beam_parser cimport BeamParser
from .syntax.ner cimport BiluoPushDown
@ -11,44 +14,40 @@ from .attrs import DEP, ENT_TYPE
cdef class EntityRecognizer(Parser):
"""Annotate named entities on Doc objects."""
"""
Annotate named entities on Doc objects.
"""
TransitionSystem = BiluoPushDown
feature_templates = get_feature_templates('ner')
def add_label(self, label):
for action in self.moves.action_types:
self.moves.add_action(action, label)
if 'actions' in self.cfg:
self.cfg['actions'].setdefault(action,
{}).setdefault(label, True)
Parser.add_label(self, label)
if isinstance(label, basestring):
label = self.vocab.strings[label]
# Set label into serializer. Super hacky :(
for attr, freqs in self.vocab.serializer_freqs:
if attr == ENT_TYPE and label not in freqs:
freqs.append([label, 1])
# Super hacky :(
self.vocab._serializer = None
cdef class BeamEntityRecognizer(BeamParser):
"""Annotate named entities on Doc objects."""
"""
Annotate named entities on Doc objects.
"""
TransitionSystem = BiluoPushDown
feature_templates = get_feature_templates('ner')
def add_label(self, label):
for action in self.moves.action_types:
self.moves.add_action(action, label)
if 'actions' in self.cfg:
self.cfg['actions'].setdefault(action,
{}).setdefault(label, True)
Parser.add_label(self, label)
if isinstance(label, basestring):
label = self.vocab.strings[label]
# Set label into serializer. Super hacky :(
for attr, freqs in self.vocab.serializer_freqs:
if attr == ENT_TYPE and label not in freqs:
freqs.append([label, 1])
# Super hacky :(
self.vocab._serializer = None
@ -58,11 +57,7 @@ cdef class DependencyParser(Parser):
feature_templates = get_feature_templates('basic')
def add_label(self, label):
for action in self.moves.action_types:
self.moves.add_action(action, label)
if 'actions' in self.cfg:
self.cfg['actions'].setdefault(action,
{}).setdefault(label, True)
Parser.add_label(self, label)
if isinstance(label, basestring):
label = self.vocab.strings[label]
for attr, freqs in self.vocab.serializer_freqs:
@ -78,11 +73,7 @@ cdef class BeamDependencyParser(BeamParser):
feature_templates = get_feature_templates('basic')
def add_label(self, label):
for action in self.moves.action_types:
self.moves.add_action(action, label)
if 'actions' in self.cfg:
self.cfg['actions'].setdefault(action,
{}).setdefault(label, True)
Parser.add_label(self, label)
if isinstance(label, basestring):
label = self.vocab.strings[label]
for attr, freqs in self.vocab.serializer_freqs:

9
spacy/scorer.py
View File

@ -1,12 +1,13 @@
from __future__ import division
from __future__ import print_function
from __future__ import unicode_literals
# coding: utf8
from __future__ import division, print_function, unicode_literals
from .gold import tags_to_entities
class PRFScore(object):
"""A precision / recall / F score"""
"""
A precision / recall / F score
"""
def __init__(self):
self.tp = 0
self.fp = 0

36
spacy/strings.pyx
View File

@ -1,12 +1,11 @@
# cython: infer_types=True
# coding: utf8
from __future__ import unicode_literals, absolute_import
cimport cython
from libc.string cimport memcpy
from libc.stdint cimport uint64_t, uint32_t
from murmurhash.mrmr cimport hash64, hash32
from preshed.maps cimport map_iter, key_t
from .typedefs cimport hash_t
@ -73,13 +72,16 @@ cdef Utf8Str _allocate(Pool mem, const unsigned char* chars, uint32_t length) ex
cdef class StringStore:
'''Map strings to and from integer IDs.'''
"""
Map strings to and from integer IDs.
"""
def __init__(self, strings=None, freeze=False):
'''Create the StringStore.
"""
Create the StringStore.
Arguments:
strings: A sequence of unicode strings to add to the store.
'''
"""
self.mem = Pool()
self._map = PreshMap()
self._oov = PreshMap()
@ -104,7 +106,8 @@ cdef class StringStore:
return (StringStore, (list(self),))
def __len__(self):
"""The number of strings in the store.
"""
The number of strings in the store.
Returns:
int The number of strings in the store.
@ -112,8 +115,9 @@ cdef class StringStore:
return self.size-1
def __getitem__(self, object string_or_id):
"""Retrieve a string from a given integer ID, or vice versa.
"""
Retrieve a string from a given integer ID, or vice versa.
Arguments:
string_or_id (bytes or unicode or int):
The value to encode.
@ -149,17 +153,18 @@ cdef class StringStore:
raise TypeError(type(string_or_id))
utf8str = self._intern_utf8(byte_string, len(byte_string))
if utf8str is NULL:
# TODO: We need to use 32 bit here, for compatibility with the
# TODO: We need to use 32 bit here, for compatibility with the
# vocabulary values. This makes birthday paradox probabilities
# pretty bad.
# We could also get unlucky here, and hash into a value that
# collides with the 'real' strings.
# collides with the 'real' strings.
return hash32_utf8(byte_string, len(byte_string))
else:
return utf8str - self.c
def __contains__(self, unicode string not None):
"""Check whether a string is in the store.
"""
Check whether a string is in the store.
Arguments:
string (unicode): The string to check.
@ -172,7 +177,8 @@ cdef class StringStore:
return self._map.get(key) is not NULL
def __iter__(self):
"""Iterate over the strings in the store, in order.
"""
Iterate over the strings in the store, in order.
Yields: unicode A string in the store.
"""
@ -230,7 +236,8 @@ cdef class StringStore:
return &self.c[self.size-1]
def dump(self, file_):
"""Save the strings to a JSON file.
"""
Save the strings to a JSON file.
Arguments:
file_ (buffer): The file to save the strings.
@ -244,7 +251,8 @@ cdef class StringStore:
file_.write(string_data)
def load(self, file_):
"""Load the strings from a JSON file.
"""
Load the strings from a JSON file.
Arguments:
file_ (buffer): The file from which to load the strings.

1
spacy/symbols.pyx
View File

@ -1,3 +1,4 @@
# coding: utf8
from __future__ import unicode_literals
IDS = {

8
spacy/syntax/_parse_features.pyx
View File

@ -7,17 +7,17 @@ out of "context") is in features/extractor.pyx
The atomic feature names are listed in a big enum, so that the feature tuples
can refer to them.
"""
from libc.string cimport memset
# coding: utf-8
from __future__ import unicode_literals
from libc.string cimport memset
from itertools import combinations
from cymem.cymem cimport Pool
from ..structs cimport TokenC
from .stateclass cimport StateClass
from ._state cimport StateC
from cymem.cymem cimport Pool
cdef inline void fill_token(atom_t* context, const TokenC* token) nogil:
if token is NULL:

46
spacy/syntax/arc_eager.pyx
View File

@ -1,29 +1,26 @@
# cython: profile=True
# cython: cdivision=True
# cython: infer_types=True
# coding: utf-8
from __future__ import unicode_literals
from cpython.ref cimport PyObject, Py_INCREF, Py_XDECREF
import ctypes
import os
from ..structs cimport TokenC
from libc.stdint cimport uint32_t
from libc.string cimport memcpy
from cymem.cymem cimport Pool
from .stateclass cimport StateClass
from ._state cimport StateC, is_space_token
from .nonproj import PseudoProjectivity
from .nonproj import is_nonproj_tree
from .transition_system cimport do_func_t, get_cost_func_t
from .transition_system cimport move_cost_func_t, label_cost_func_t
from ..gold cimport GoldParse
from ..gold cimport GoldParseC
from ..attrs cimport TAG, HEAD, DEP, ENT_IOB, ENT_TYPE, IS_SPACE
from ..lexeme cimport Lexeme
from libc.stdint cimport uint32_t
from libc.string cimport memcpy
from cymem.cymem cimport Pool
from .stateclass cimport StateClass
from ._state cimport StateC, is_space_token
from .nonproj import PseudoProjectivity
from .nonproj import is_nonproj_tree
from ..structs cimport TokenC
DEF NON_MONOTONIC = True
@ -317,17 +314,20 @@ cdef class ArcEager(TransitionSystem):
def get_actions(cls, **kwargs):
actions = kwargs.get('actions',
{
SHIFT: {'': True},
REDUCE: {'': True},
RIGHT: {},
LEFT: {},
BREAK: {'ROOT': True}})
SHIFT: [''],
REDUCE: [''],
RIGHT: [],
LEFT: [],
BREAK: ['ROOT']})
seen_actions = set()
for label in kwargs.get('left_labels', []):
if label.upper() != 'ROOT':
actions[LEFT][label] = True
if (LEFT, label) not in seen_actions:
actions[LEFT].append(label)
for label in kwargs.get('right_labels', []):
if label.upper() != 'ROOT':
actions[RIGHT][label] = True
if (RIGHT, label) not in seen_actions:
actions[RIGHT].append(label)
for raw_text, sents in kwargs.get('gold_parses', []):
for (ids, words, tags, heads, labels, iob), ctnts in sents:
@ -336,9 +336,11 @@ cdef class ArcEager(TransitionSystem):
label = 'ROOT'
if label != 'ROOT':
if head < child:
actions[RIGHT][label] = True
if (RIGHT, label) not in seen_actions:
actions[RIGHT].append(label)
elif head > child:
actions[LEFT][label] = True
if (LEFT, label) not in seen_actions:
actions[LEFT].append(label)
return actions
property action_types:

31
spacy/syntax/beam_parser.pyx
View File

@ -1,50 +1,34 @@
"""
MALT-style dependency parser
"""
# cython: profile=True
# cython: experimental_cpp_class_def=True
# cython: cdivision=True
# cython: infer_types=True
"""
MALT-style dependency parser
"""
from __future__ import unicode_literals
# coding: utf-8
from __future__ import unicode_literals, print_function
cimport cython
from cpython.ref cimport PyObject, Py_INCREF, Py_XDECREF
from libc.stdint cimport uint32_t, uint64_t
from libc.string cimport memset, memcpy
from libc.stdlib cimport rand
from libc.math cimport log, exp, isnan, isinf
import random
import os.path
from os import path
import shutil
import json
import math
from cymem.cymem cimport Pool, Address
from murmurhash.mrmr cimport real_hash64 as hash64
from thinc.typedefs cimport weight_t, class_t, feat_t, atom_t, hash_t
from util import Config
from thinc.linear.features cimport ConjunctionExtracter
from thinc.structs cimport FeatureC, ExampleC
from thinc.extra.search cimport Beam
from thinc.extra.search cimport MaxViolation
from thinc.extra.search cimport Beam, MaxViolation
from thinc.extra.eg cimport Example
from thinc.extra.mb cimport Minibatch
from ..structs cimport TokenC
from ..tokens.doc cimport Doc
from ..strings cimport StringStore
from .transition_system cimport TransitionSystem, Transition
from ..gold cimport GoldParse
from . import _parse_features
from ._parse_features cimport CONTEXT_SIZE
from ._parse_features cimport fill_context
@ -266,4 +250,3 @@ def is_gold(StateClass state, GoldParse gold, StringStore strings):
id_, word, tag, head, dep, ner = gold.orig_annot[gold.cand_to_gold[i]]
truth.add((id_, head, dep))
return truth == predicted

11
spacy/syntax/iterators.pyx
View File

@ -1,9 +1,14 @@
from spacy.parts_of_speech cimport NOUN, PROPN, PRON
# coding: utf-8
from __future__ import unicode_literals
from ..parts_of_speech cimport NOUN, PROPN, PRON
def english_noun_chunks(obj):
'''Detect base noun phrases from a dependency parse.
Works on both Doc and Span.'''
"""
Detect base noun phrases from a dependency parse.
Works on both Doc and Span.
"""
labels = ['nsubj', 'dobj', 'nsubjpass', 'pcomp', 'pobj',
'attr', 'ROOT', 'root']
doc = obj.doc # Ensure works on both Doc and Span.

55
spacy/syntax/ner.pyx
View File

@ -1,17 +1,16 @@
# coding: utf-8
from __future__ import unicode_literals
from .transition_system cimport Transition
from .transition_system cimport do_func_t
from ..structs cimport TokenC, Entity
from thinc.typedefs cimport weight_t
from ..gold cimport GoldParseC
from ..gold cimport GoldParse
from ..attrs cimport ENT_TYPE, ENT_IOB
from .stateclass cimport StateClass
from ._state cimport StateC
from .transition_system cimport Transition
from .transition_system cimport do_func_t
from ..structs cimport TokenC, Entity
from ..gold cimport GoldParseC
from ..gold cimport GoldParse
from ..attrs cimport ENT_TYPE, ENT_IOB
cdef enum:
@ -21,6 +20,7 @@ cdef enum:
LAST
UNIT
OUT
ISNT
N_MOVES
@ -31,6 +31,7 @@ MOVE_NAMES[IN] = 'I'
MOVE_NAMES[LAST] = 'L'
MOVE_NAMES[UNIT] = 'U'
MOVE_NAMES[OUT] = 'O'
MOVE_NAMES[ISNT] = 'x'
cdef do_func_t[N_MOVES] do_funcs
@ -54,16 +55,20 @@ cdef class BiluoPushDown(TransitionSystem):
def get_actions(cls, **kwargs):
actions = kwargs.get('actions',
{
MISSING: {'': True},
BEGIN: {},
IN: {},
LAST: {},
UNIT: {},
OUT: {'': True}
MISSING: [''],
BEGIN: [],
IN: [],
LAST: [],
UNIT: [],
OUT: ['']
})
seen_entities = set()
for entity_type in kwargs.get('entity_types', []):
if entity_type in seen_entities:
continue
seen_entities.add(entity_type)
for action in (BEGIN, IN, LAST, UNIT):
actions[action][entity_type] = True
actions[action].append(entity_type)
moves = ('M', 'B', 'I', 'L', 'U')
for raw_text, sents in kwargs.get('gold_parses', []):
for (ids, words, tags, heads, labels, biluo), _ in sents:
@ -72,8 +77,10 @@ cdef class BiluoPushDown(TransitionSystem):
if ner_tag.count('-') != 1:
raise ValueError(ner_tag)
_, label = ner_tag.split('-')
for move_str in ('B', 'I', 'L', 'U'):
actions[moves.index(move_str)][label] = True
if label not in seen_entities:
seen_entities.add(label)
for move_str in ('B', 'I', 'L', 'U'):
actions[moves.index(move_str)].append(label)
return actions
property action_types:
@ -111,11 +118,17 @@ cdef class BiluoPushDown(TransitionSystem):
label = 0
elif '-' in name:
move_str, label_str = name.split('-', 1)
# Hacky way to denote 'not this entity'
if label_str.startswith('!'):
label_str = label_str[1:]
move_str = 'x'
label = self.strings[label_str]
else:
move_str = name
label = 0
move = MOVE_NAMES.index(move_str)
if move == ISNT:
return Transition(clas=0, move=ISNT, label=label, score=0)
for i in range(self.n_moves):
if self.c[i].move == move and self.c[i].label == label:
return self.c[i]
@ -225,6 +238,9 @@ cdef class Begin:
elif g_act == BEGIN:
# B, Gold B --> Label match
return label != g_tag
# Support partial supervision in the form of "not this label"
elif g_act == ISNT:
return label == g_tag
else:
# B, Gold I --> False (P)
# B, Gold L --> False (P)
@ -359,6 +375,9 @@ cdef class Unit:
elif g_act == UNIT:
# U, Gold U --> True iff tag match
return label != g_tag
# Support partial supervision in the form of "not this label"
elif g_act == ISNT:
return label == g_tag
else:
# U, Gold B --> False
# U, Gold I --> False
@ -388,7 +407,7 @@ cdef class Out:
cdef int g_act = gold.ner[s.B(0)].move
cdef int g_tag = gold.ner[s.B(0)].label
if g_act == MISSING:
if g_act == MISSING or g_act == ISNT:
return 0
elif g_act == BEGIN:
# O, Gold B --> False

5
spacy/syntax/nonproj.pyx
View File

@ -1,8 +1,9 @@
# coding: utf-8
from __future__ import unicode_literals
from copy import copy
from ..tokens.doc cimport Doc
from spacy.attrs import DEP, HEAD
from ..attrs import DEP, HEAD
def ancestors(tokenid, heads):
@ -201,5 +202,3 @@ class PseudoProjectivity:
filtered_sents.append(((ids,words,tags,heads,filtered_labels,iob), ctnts))
filtered.append((raw_text, filtered_sents))
return filtered

126
spacy/syntax/parser.pyx
View File

@ -1,58 +1,46 @@
# cython: infer_types=True
"""
MALT-style dependency parser
"""
# coding: utf-8
# cython: infer_types=True
from __future__ import unicode_literals
from collections import Counter
import ujson
cimport cython
cimport cython.parallel
from cpython.ref cimport PyObject, Py_INCREF, Py_XDECREF
from cpython.exc cimport PyErr_CheckSignals
from libc.stdint cimport uint32_t, uint64_t
from libc.string cimport memset, memcpy
from libc.stdlib cimport malloc, calloc, free
import os.path
from collections import Counter
from os import path
import shutil
import json
import sys
from .nonproj import PseudoProjectivity
from cymem.cymem cimport Pool, Address
from murmurhash.mrmr cimport hash64
from thinc.typedefs cimport weight_t, class_t, feat_t, atom_t, hash_t
from thinc.linear.avgtron cimport AveragedPerceptron
from thinc.linalg cimport VecVec
from thinc.structs cimport SparseArrayC
from thinc.structs cimport SparseArrayC, FeatureC, ExampleC
from thinc.extra.eg cimport Example
from cymem.cymem cimport Pool, Address
from murmurhash.mrmr cimport hash64
from preshed.maps cimport MapStruct
from preshed.maps cimport map_get
from thinc.structs cimport FeatureC
from thinc.structs cimport ExampleC
from thinc.extra.eg cimport Example
from util import Config
from ..structs cimport TokenC
from ..tokens.doc cimport Doc
from ..strings cimport StringStore
from .transition_system import OracleError
from .transition_system cimport TransitionSystem, Transition
from ..gold cimport GoldParse
from . import _parse_features
from ._parse_features cimport CONTEXT_SIZE
from ._parse_features cimport fill_context
from .stateclass cimport StateClass
from ._state cimport StateC
from .nonproj import PseudoProjectivity
from .transition_system import OracleError
from .transition_system cimport TransitionSystem, Transition
from ..structs cimport TokenC
from ..tokens.doc cimport Doc
from ..strings cimport StringStore
from ..gold cimport GoldParse
USE_FTRL = True
USE_FTRL = False
DEBUG = False
def set_debug(val):
global DEBUG
@ -80,7 +68,9 @@ cdef class ParserModel(AveragedPerceptron):
return nr_feat
def update(self, Example eg, itn=0):
'''Does regression on negative cost. Sort of cute?'''
"""
Does regression on negative cost. Sort of cute?
"""
self.time += 1
cdef int best = arg_max_if_gold(eg.c.scores, eg.c.costs, eg.c.nr_class)
cdef int guess = eg.guess
@ -132,10 +122,13 @@ cdef class ParserModel(AveragedPerceptron):
cdef class Parser:
"""Base class of the DependencyParser and EntityRecognizer."""
"""
Base class of the DependencyParser and EntityRecognizer.
"""
@classmethod
def load(cls, path, Vocab vocab, TransitionSystem=None, require=False, **cfg):
"""Load the statistical model from the supplied path.
"""
Load the statistical model from the supplied path.
Arguments:
path (Path):
@ -148,10 +141,16 @@ cdef class Parser:
The newly constructed object.
"""
with (path / 'config.json').open() as file_:
cfg = json.load(file_)
cfg = ujson.load(file_)
# TODO: remove this shim when we don't have to support older data
if 'labels' in cfg and 'actions' not in cfg:
cfg['actions'] = cfg.pop('labels')
# TODO: remove this shim when we don't have to support older data
for action_name, labels in dict(cfg['actions']).items():
# We need this to be sorted
if isinstance(labels, dict):
labels = list(sorted(labels.keys()))
cfg['actions'][action_name] = labels
self = cls(vocab, TransitionSystem=TransitionSystem, model=None, **cfg)
if (path / 'model').exists():
self.model.load(str(path / 'model'))
@ -161,7 +160,8 @@ cdef class Parser:
return self
def __init__(self, Vocab vocab, TransitionSystem=None, ParserModel model=None, **cfg):
"""Create a Parser.
"""
Create a Parser.
Arguments:
vocab (Vocab):
@ -186,12 +186,18 @@ cdef class Parser:
self.model.learn_rate = cfg.get('learn_rate', 0.001)
self.cfg = cfg
# TODO: This is a pretty hacky fix to the problem of adding more
# labels. The issue is they come in out of order, if labels are
# added during training
for label in cfg.get('extra_labels', []):
self.add_label(label)
def __reduce__(self):
return (Parser, (self.vocab, self.moves, self.model), None, None)
def __call__(self, Doc tokens):
"""Apply the entity recognizer, setting the annotations onto the Doc object.
"""
Apply the entity recognizer, setting the annotations onto the Doc object.
Arguments:
doc (Doc): The document to be processed.
@ -208,7 +214,8 @@ cdef class Parser:
self.moves.finalize_doc(tokens)
def pipe(self, stream, int batch_size=1000, int n_threads=2):
"""Process a stream of documents.
"""
Process a stream of documents.
Arguments:
stream: The sequence of documents to process.
@ -296,7 +303,8 @@ cdef class Parser:
return 0
def update(self, Doc tokens, GoldParse gold, itn=0):
"""Update the statistical model.
"""
Update the statistical model.
Arguments:
doc (Doc):
@ -334,15 +342,17 @@ cdef class Parser:
self.moves.finalize_state(stcls.c)
return loss
def step_through(self, Doc doc):
"""Set up a stepwise state, to introspect and control the transition sequence.
def step_through(self, Doc doc, GoldParse gold=None):
"""
Set up a stepwise state, to introspect and control the transition sequence.
Arguments:
doc (Doc): The document to step through.
gold (GoldParse): Optional gold parse
Returns (StepwiseState):
A state object, to step through the annotation process.
"""
return StepwiseState(self, doc)
return StepwiseState(self, doc, gold=gold)
def from_transition_sequence(self, Doc doc, sequence):
"""Control the annotations on a document by specifying a transition sequence
@ -360,18 +370,28 @@ cdef class Parser:
def add_label(self, label):
# Doesn't set label into serializer -- subclasses override it to do that.
for action in self.moves.action_types:
self.moves.add_action(action, label)
added = self.moves.add_action(action, label)
if added:
# Important that the labels be stored as a list! We need the
# order, or the model goes out of synch
self.cfg.setdefault('extra_labels', []).append(label)
cdef class StepwiseState:
cdef readonly StateClass stcls
cdef readonly Example eg
cdef readonly Doc doc
cdef readonly GoldParse gold
cdef readonly Parser parser
def __init__(self, Parser parser, Doc doc):
def __init__(self, Parser parser, Doc doc, GoldParse gold=None):
self.parser = parser
self.doc = doc
if gold is not None:
self.gold = gold
self.parser.moves.preprocess_gold(self.gold)
else:
self.gold = GoldParse(doc)
self.stcls = StateClass.init(doc.c, doc.length)
self.parser.moves.initialize_state(self.stcls.c)
self.eg = Example(
@ -406,6 +426,24 @@ cdef class StepwiseState:
return [self.doc.vocab.strings[self.stcls.c._sent[i].dep]
for i in range(self.stcls.c.length)]
@property
def costs(self):
"""
Find the action-costs for the current state.
"""
if not self.gold:
raise ValueError("Can't set costs: No GoldParse provided")
self.parser.moves.set_costs(self.eg.c.is_valid, self.eg.c.costs,
self.stcls, self.gold)
costs = {}
for i in range(self.parser.moves.n_moves):
if not self.eg.c.is_valid[i]:
continue
transition = self.parser.moves.c[i]
name = self.parser.moves.move_name(transition.move, transition.label)
costs[name] = self.eg.c.costs[i]
return costs
def predict(self):
self.eg.reset()
self.eg.c.nr_feat = self.parser.model.set_featuresC(self.eg.c.atoms, self.eg.c.features,

8
spacy/syntax/stateclass.pyx
View File

@ -1,5 +1,9 @@
# coding: utf-8
from __future__ import unicode_literals
from libc.string cimport memcpy, memset
from libc.stdint cimport uint32_t
from ..vocab cimport EMPTY_LEXEME
from ..structs cimport Entity
from ..lexeme cimport Lexeme
@ -28,6 +32,6 @@ cdef class StateClass:
top = words[self.S(0)] + '_%d' % self.S_(0).head
second = words[self.S(1)] + '_%d' % self.S_(1).head
third = words[self.S(2)] + '_%d' % self.S_(2).head
n0 = words[self.B(0)]
n1 = words[self.B(1)]
n0 = words[self.B(0)]
n1 = words[self.B(1)]
return ' '.join((third, second, top, '|', n0, n1))

7
spacy/syntax/transition_system.pyx
View File

@ -1,4 +1,8 @@
# cython: infer_types=True
# coding: utf-8
from __future__ import unicode_literals
from cpython.ref cimport PyObject, Py_INCREF, Py_XDECREF
from cymem.cymem cimport Pool
from thinc.typedefs cimport weight_t
from collections import defaultdict
@ -6,7 +10,6 @@ from collections import defaultdict
from ..structs cimport TokenC
from .stateclass cimport StateClass
from ..attrs cimport TAG, HEAD, DEP, ENT_TYPE, ENT_IOB
from cpython.ref cimport PyObject, Py_INCREF, Py_XDECREF
cdef weight_t MIN_SCORE = -90000
@ -32,7 +35,7 @@ cdef class TransitionSystem:
self.c = <Transition*>self.mem.alloc(self._size, sizeof(Transition))
for action, label_strs in sorted(labels_by_action.items()):
for label_str in sorted(label_strs):
for label_str in label_strs:
self.add_action(int(action), label_str)
self.root_label = self.strings['ROOT']
self.freqs = {} if _freqs is None else _freqs

18
spacy/syntax/util.py
View File

@ -1,18 +0,0 @@
from os import path
import json
class Config(object):
def __init__(self, **kwargs):
for key, value in kwargs.items():
setattr(self, key, value)
def get(self, attr, default=None):
return self.__dict__.get(attr, default)
@classmethod
def write(cls, model_dir, name, **kwargs):
open(path.join(model_dir, '%s.json' % name), 'w').write(json.dumps(kwargs))
@classmethod
def read(cls, model_dir, name):
return cls(**json.load(open(path.join(model_dir, '%s.json' % name))))

31
spacy/tagger.pyx
View File

@ -1,5 +1,7 @@
import json
import pathlib
# coding: utf8
from __future__ import unicode_literals
import ujson
from collections import defaultdict
from cymem.cymem cimport Pool
@ -12,8 +14,8 @@ from thinc.linalg cimport VecVec
from .tokens.doc cimport Doc
from .attrs cimport TAG
from .gold cimport GoldParse
from .attrs cimport *
from . import util
cpdef enum:
@ -106,10 +108,13 @@ cdef inline void _fill_from_token(atom_t* context, const TokenC* t) nogil:
cdef class Tagger:
"""Annotate part-of-speech tags on Doc objects."""
"""
Annotate part-of-speech tags on Doc objects.
"""
@classmethod
def load(cls, path, vocab, require=False):
"""Load the statistical model from the supplied path.
"""
Load the statistical model from the supplied path.
Arguments:
path (Path):
@ -123,10 +128,10 @@ cdef class Tagger:
"""
# TODO: Change this to expect config.json when we don't have to
# support old data.
path = path if not isinstance(path, basestring) else pathlib.Path(path)
path = util.ensure_path(path)
if (path / 'templates.json').exists():
with (path / 'templates.json').open('r', encoding='utf8') as file_:
templates = json.load(file_)
templates = ujson.load(file_)
elif require:
raise IOError(
"Required file %s/templates.json not found when loading Tagger" % str(path))
@ -142,7 +147,8 @@ cdef class Tagger:
return self
def __init__(self, Vocab vocab, TaggerModel model=None, **cfg):
"""Create a Tagger.
"""
Create a Tagger.
Arguments:
vocab (Vocab):
@ -180,7 +186,8 @@ cdef class Tagger:
tokens._py_tokens = [None] * tokens.length
def __call__(self, Doc tokens):
"""Apply the tagger, setting the POS tags onto the Doc object.
"""
Apply the tagger, setting the POS tags onto the Doc object.
Arguments:
doc (Doc): The tokens to be tagged.
@ -208,7 +215,8 @@ cdef class Tagger:
tokens._py_tokens = [None] * tokens.length
def pipe(self, stream, batch_size=1000, n_threads=2):
"""Tag a stream of documents.
"""
Tag a stream of documents.
Arguments:
stream: The sequence of documents to tag.
@ -225,7 +233,8 @@ cdef class Tagger:
yield doc
def update(self, Doc tokens, GoldParse gold, itn=0):
"""Update the statistical model, with tags supplied for the given document.
"""
Update the statistical model, with tags supplied for the given document.
Arguments:
doc (Doc):

65
spacy/tokenizer.pyx
View File

@ -1,17 +1,11 @@
# cython: embedsignature=True
# coding: utf8
from __future__ import unicode_literals
import pathlib
import ujson
from cython.operator cimport dereference as deref
from cython.operator cimport preincrement as preinc
try:
import ujson as json
except ImportError:
import json
from cymem.cymem cimport Pool
from preshed.maps cimport PreshMap
@ -23,12 +17,15 @@ from .tokens.doc cimport Doc
cdef class Tokenizer:
"""Segment text, and create Doc objects with the discovered segment boundaries."""
"""
Segment text, and create Doc objects with the discovered segment boundaries.
"""
@classmethod
def load(cls, path, Vocab vocab, rules=None, prefix_search=None, suffix_search=None,
infix_finditer=None, token_match=None):
'''Load a Tokenizer, reading unsupplied components from the path.
"""
Load a Tokenizer, reading unsupplied components from the path.
Arguments:
path (Path):
The path to load from.
@ -45,13 +42,11 @@ cdef class Tokenizer:
infix_finditer:
Signature of re.compile(string).finditer
Returns Tokenizer
'''
if isinstance(path, basestring):
path = pathlib.Path(path)
"""
path = util.ensure_path(path)
if rules is None:
with (path / 'tokenizer' / 'specials.json').open('r', encoding='utf8') as file_:
rules = json.load(file_)
rules = ujson.load(file_)
if prefix_search in (None, True):
with (path / 'tokenizer' / 'prefix.txt').open() as file_:
entries = file_.read().split('\n')
@ -67,8 +62,9 @@ cdef class Tokenizer:
return cls(vocab, rules, prefix_search, suffix_search, infix_finditer, token_match)
def __init__(self, Vocab vocab, rules, prefix_search, suffix_search, infix_finditer, token_match=None):
'''Create a Tokenizer, to create Doc objects given unicode text.
"""
Create a Tokenizer, to create Doc objects given unicode text.
Arguments:
vocab (Vocab):
A storage container for lexical types.
@ -85,7 +81,7 @@ cdef class Tokenizer:
to find infixes.
token_match:
A boolean function matching strings that becomes tokens.
'''
"""
self.mem = Pool()
self._cache = PreshMap()
self._specials = PreshMap()
@ -107,7 +103,7 @@ cdef class Tokenizer:
self.token_match)
return (self.__class__, args, None, None)
cpdef Doc tokens_from_list(self, list strings):
return Doc(self.vocab, words=strings)
#raise NotImplementedError(
@ -117,7 +113,8 @@ cdef class Tokenizer:
@cython.boundscheck(False)
def __call__(self, unicode string):
"""Tokenize a string.
"""
Tokenize a string.
Arguments:
string (unicode): The string to tokenize.
@ -170,7 +167,8 @@ cdef class Tokenizer:
return tokens
def pipe(self, texts, batch_size=1000, n_threads=2):
"""Tokenize a stream of texts.
"""
Tokenize a stream of texts.
Arguments:
texts: A sequence of unicode texts.
@ -270,7 +268,7 @@ cdef class Tokenizer:
cache_hit = self._try_cache(hash_string(string), tokens)
if cache_hit:
pass
elif self.token_match and self.token_match(string):
elif self.token_match and self.token_match(string):
# We're always saying 'no' to spaces here -- the caller will
# fix up the outermost one, with reference to the original.
# See Issue #859
@ -324,7 +322,8 @@ cdef class Tokenizer:
self._cache.set(key, cached)
def find_infix(self, unicode string):
"""Find internal split points of the string, such as hyphens.
"""
Find internal split points of the string, such as hyphens.
string (unicode): The string to segment.
@ -337,7 +336,8 @@ cdef class Tokenizer:
return list(self.infix_finditer(string))
def find_prefix(self, unicode string):
"""Find the length of a prefix that should be segmented from the string,
"""
Find the length of a prefix that should be segmented from the string,
or None if no prefix rules match.
Arguments:
@ -350,7 +350,8 @@ cdef class Tokenizer:
return (match.end() - match.start()) if match is not None else 0
def find_suffix(self, unicode string):
"""Find the length of a suffix that should be segmented from the string,
"""
Find the length of a suffix that should be segmented from the string,
or None if no suffix rules match.
Arguments:
@ -363,13 +364,15 @@ cdef class Tokenizer:
return (match.end() - match.start()) if match is not None else 0
def _load_special_tokenization(self, special_cases):
'''Add special-case tokenization rules.
'''
"""
Add special-case tokenization rules.
"""
for chunk, substrings in sorted(special_cases.items()):
self.add_special_case(chunk, substrings)
def add_special_case(self, unicode string, substrings):
'''Add a special-case tokenization rule.
"""
Add a special-case tokenization rule.
Arguments:
string (unicode): The string to specially tokenize.
@ -378,7 +381,7 @@ cdef class Tokenizer:
attributes. The ORTH fields of the attributes must exactly match
the string when they are concatenated.
Returns None
'''
"""
substrings = list(substrings)
cached = <_Cached*>self.mem.alloc(1, sizeof(_Cached))
cached.length = len(substrings)

93
spacy/tokens/doc.pyx
View File

@ -1,15 +1,18 @@
# coding: utf8
from __future__ import unicode_literals
cimport cython
cimport numpy as np
import numpy
import numpy.linalg
import struct
from libc.string cimport memcpy, memset
from libc.stdint cimport uint32_t
from libc.math cimport sqrt
import numpy
import numpy.linalg
import struct
cimport numpy as np
import six
import warnings
from .span cimport Span
from .token cimport Token
from ..lexeme cimport Lexeme
from ..lexeme cimport EMPTY_LEXEME
from ..typedefs cimport attr_t, flags_t
@ -19,11 +22,10 @@ from ..attrs cimport POS, LEMMA, TAG, DEP, HEAD, SPACY, ENT_IOB, ENT_TYPE
from ..parts_of_speech cimport CCONJ, PUNCT, NOUN
from ..parts_of_speech cimport univ_pos_t
from ..lexeme cimport Lexeme
from .span cimport Span
from .token cimport Token
from ..serialize.bits cimport BitArray
from ..util import normalize_slice
from ..syntax.iterators import CHUNKERS
from ..compat import is_config
DEF PADDING = 5
@ -76,7 +78,7 @@ cdef class Doc:
"""
def __init__(self, Vocab vocab, words=None, spaces=None, orths_and_spaces=None):
'''
"""
Create a Doc object.
Aside: Implementation
@ -97,7 +99,7 @@ cdef class Doc:
A list of boolean values, of the same length as words. True
means that the word is followed by a space, False means it is not.
If None, defaults to [True]*len(words)
'''
"""
self.vocab = vocab
size = 20
self.mem = Pool()
@ -158,7 +160,7 @@ cdef class Doc:
self.is_parsed = True
def __getitem__(self, object i):
'''
"""
doc[i]
Get the Token object at position i, where i is an integer.
Negative indexing is supported, and follows the usual Python
@ -172,7 +174,7 @@ cdef class Doc:
are not supported, as `Span` objects must be contiguous (cannot have gaps).
You can use negative indices and open-ended ranges, which have their
normal Python semantics.
'''
"""
if isinstance(i, slice):
start, stop = normalize_slice(len(self), i.start, i.stop, i.step)
return Span(self, start, stop, label=0)
@ -186,7 +188,7 @@ cdef class Doc:
return Token.cinit(self.vocab, &self.c[i], i, self)
def __iter__(self):
'''
"""
for token in doc
Iterate over `Token` objects, from which the annotations can
be easily accessed. This is the main way of accessing Token
@ -194,7 +196,7 @@ cdef class Doc:
Python. If faster-than-Python speeds are required, you can
instead access the annotations as a numpy array, or access the
underlying C data directly from Cython.
'''
"""
cdef int i
for i in range(self.length):
if self._py_tokens[i] is not None:
@ -203,10 +205,10 @@ cdef class Doc:
yield Token.cinit(self.vocab, &self.c[i], i, self)
def __len__(self):
'''
"""
len(doc)
The number of tokens in the document.
'''
"""
return self.length
def __unicode__(self):
@ -216,7 +218,7 @@ cdef class Doc:
return u''.join([t.text_with_ws for t in self]).encode('utf-8')
def __str__(self):
if six.PY3:
if is_config(python3=True):
return self.__unicode__()
return self.__bytes__()
@ -228,7 +230,8 @@ cdef class Doc:
return self
def similarity(self, other):
'''Make a semantic similarity estimate. The default estimate is cosine
"""
Make a semantic similarity estimate. The default estimate is cosine
similarity using an average of word vectors.
Arguments:
@ -237,7 +240,7 @@ cdef class Doc:
Return:
score (float): A scalar similarity score. Higher is more similar.
'''
"""
if 'similarity' in self.user_hooks:
return self.user_hooks['similarity'](self, other)
if self.vector_norm == 0 or other.vector_norm == 0:
@ -245,9 +248,9 @@ cdef class Doc:
return numpy.dot(self.vector, other.vector) / (self.vector_norm * other.vector_norm)
property has_vector:
'''
"""
A boolean value indicating whether a word vector is associated with the object.
'''
"""
def __get__(self):
if 'has_vector' in self.user_hooks:
return self.user_hooks['has_vector'](self)
@ -255,11 +258,11 @@ cdef class Doc:
return any(token.has_vector for token in self)
property vector:
'''
"""
A real-valued meaning representation. Defaults to an average of the token vectors.
Type: numpy.ndarray[ndim=1, dtype='float32']
'''
"""
def __get__(self):
if 'vector' in self.user_hooks:
return self.user_hooks['vector'](self)
@ -294,17 +297,21 @@ cdef class Doc:
return self.text
property text:
'''A unicode representation of the document text.'''
"""
A unicode representation of the document text.
"""
def __get__(self):
return u''.join(t.text_with_ws for t in self)
property text_with_ws:
'''An alias of Doc.text, provided for duck-type compatibility with Span and Token.'''
"""
An alias of Doc.text, provided for duck-type compatibility with Span and Token.
"""
def __get__(self):
return self.text
property ents:
'''
"""
Yields named-entity `Span` objects, if the entity recognizer
has been applied to the document. Iterate over the span to get
individual Token objects, or access the label:
@ -318,7 +325,7 @@ cdef class Doc:
assert ents[0].label_ == 'PERSON'
assert ents[0].orth_ == 'Best'
assert ents[0].text == 'Mr. Best'
'''
"""
def __get__(self):
cdef int i
cdef const TokenC* token
@ -382,13 +389,13 @@ cdef class Doc:
self.c[start].ent_iob = 3
property noun_chunks:
'''
"""
Yields base noun-phrase #[code Span] objects, if the document
has been syntactically parsed. A base noun phrase, or
'NP chunk', is a noun phrase that does not permit other NPs to
be nested within it – so no NP-level coordination, no prepositional
phrases, and no relative clauses. For example:
'''
phrases, and no relative clauses.
"""
def __get__(self):
if not self.is_parsed:
raise ValueError(
@ -496,7 +503,8 @@ cdef class Doc:
return output
def count_by(self, attr_id_t attr_id, exclude=None, PreshCounter counts=None):
"""Produce a dict of {attribute (int): count (ints)} frequencies, keyed
"""
Produce a dict of {attribute (int): count (ints)} frequencies, keyed
by the values of the given attribute ID.
Example:
@ -563,8 +571,9 @@ cdef class Doc:
self.c[i] = parsed[i]
def from_array(self, attrs, array):
'''Write to a `Doc` object, from an `(M, N)` array of attributes.
'''
"""
Write to a `Doc` object, from an `(M, N)` array of attributes.
"""
cdef int i, col
cdef attr_id_t attr_id
cdef TokenC* tokens = self.c
@ -603,19 +612,23 @@ cdef class Doc:
return self
def to_bytes(self):
'''Serialize, producing a byte string.'''
"""
Serialize, producing a byte string.
"""
byte_string = self.vocab.serializer.pack(self)
cdef uint32_t length = len(byte_string)
return struct.pack('I', length) + byte_string
def from_bytes(self, data):
'''Deserialize, loading from bytes.'''
"""
Deserialize, loading from bytes.
"""
self.vocab.serializer.unpack_into(data[4:], self)
return self
@staticmethod
def read_bytes(file_):
'''
"""
A static method, used to read serialized #[code Doc] objects from
a file. For example:
@ -630,7 +643,7 @@ cdef class Doc:
for byte_string in Doc.read_bytes(file_):
docs.append(Doc(nlp.vocab).from_bytes(byte_string))
assert len(docs) == 2
'''
"""
keep_reading = True
while keep_reading:
try:
@ -644,7 +657,8 @@ cdef class Doc:
yield n_bytes_str + data
def merge(self, int start_idx, int end_idx, *args, **attributes):
"""Retokenize the document, such that the span at doc.text[start_idx : end_idx]
"""
Retokenize the document, such that the span at doc.text[start_idx : end_idx]
is merged into a single token. If start_idx and end_idx do not mark start
and end token boundaries, the document remains unchanged.
@ -658,7 +672,6 @@ cdef class Doc:
token (Token):
The newly merged token, or None if the start and end indices did
not fall at token boundaries.
"""
cdef unicode tag, lemma, ent_type
if len(args) == 3:

72
spacy/tokens/span.pyx
View File

@ -1,26 +1,31 @@
# coding: utf8
from __future__ import unicode_literals
from collections import defaultdict
cimport numpy as np
import numpy
import numpy.linalg
cimport numpy as np
from libc.math cimport sqrt
import six
from .doc cimport token_by_start, token_by_end
from ..structs cimport TokenC, LexemeC
from ..typedefs cimport flags_t, attr_t, hash_t
from ..attrs cimport attr_id_t
from ..parts_of_speech cimport univ_pos_t
from ..util import normalize_slice
from .doc cimport token_by_start, token_by_end
from ..attrs cimport IS_PUNCT, IS_SPACE
from ..lexeme cimport Lexeme
from ..compat import is_config
cdef class Span:
"""A slice from a Doc object."""
"""
A slice from a Doc object.
"""
def __cinit__(self, Doc doc, int start, int end, int label=0, vector=None,
vector_norm=None):
'''Create a Span object from the slice doc[start : end]
"""
Create a Span object from the slice doc[start : end]
Arguments:
doc (Doc): The parent document.
@ -30,7 +35,7 @@ cdef class Span:
vector (ndarray[ndim=1, dtype='float32']): A meaning representation of the span.
Returns:
Span The newly constructed object.
'''
"""
if not (0 <= start <= end <= len(doc)):
raise IndexError
@ -68,7 +73,7 @@ cdef class Span:
return self.end - self.start
def __repr__(self):
if six.PY3:
if is_config(python3=True):
return self.text
return self.text.encode('utf-8')
@ -89,7 +94,8 @@ cdef class Span:
yield self.doc[i]
def merge(self, *args, **attributes):
"""Retokenize the document, such that the span is merged into a single token.
"""
Retokenize the document, such that the span is merged into a single token.
Arguments:
**attributes:
@ -102,7 +108,8 @@ cdef class Span:
return self.doc.merge(self.start_char, self.end_char, *args, **attributes)
def similarity(self, other):
'''Make a semantic similarity estimate. The default estimate is cosine
"""
Make a semantic similarity estimate. The default estimate is cosine
similarity using an average of word vectors.
Arguments:
@ -111,7 +118,7 @@ cdef class Span:
Return:
score (float): A scalar similarity score. Higher is more similar.
'''
"""
if 'similarity' in self.doc.user_span_hooks:
self.doc.user_span_hooks['similarity'](self, other)
if self.vector_norm == 0.0 or other.vector_norm == 0.0:
@ -133,11 +140,12 @@ cdef class Span:
self.end = end + 1
property sent:
'''The sentence span that this span is a part of.
"""
The sentence span that this span is a part of.
Returns:
Span The sentence this is part of.
'''
"""
def __get__(self):
if 'sent' in self.doc.user_span_hooks:
return self.doc.user_span_hooks['sent'](self)
@ -198,13 +206,13 @@ cdef class Span:
return u''.join([t.text_with_ws for t in self])
property noun_chunks:
'''
"""
Yields base noun-phrase #[code Span] objects, if the document
has been syntactically parsed. A base noun phrase, or
'NP chunk', is a noun phrase that does not permit other NPs to
be nested within it – so no NP-level coordination, no prepositional
phrases, and no relative clauses. For example:
'''
"""
def __get__(self):
if not self.doc.is_parsed:
raise ValueError(
@ -223,17 +231,16 @@ cdef class Span:
yield span
property root:
"""The token within the span that's highest in the parse tree. If there's a tie, the earlist is prefered.
"""
The token within the span that's highest in the parse tree. If there's a
tie, the earlist is prefered.
Returns:
Token: The root token.
i.e. has the
shortest path to the root of the sentence (or is the root itself).
If multiple words are equally high in the tree, the first word is taken.
For example:
i.e. has the shortest path to the root of the sentence (or is the root
itself). If multiple words are equally high in the tree, the first word
is taken. For example:
>>> toks = nlp(u'I like New York in Autumn.')
@ -303,7 +310,8 @@ cdef class Span:
return self.doc[root]
property lefts:
"""Tokens that are to the left of the span, whose head is within the Span.
"""
Tokens that are to the left of the span, whose head is within the Span.
Yields: Token A left-child of a token of the span.
"""
@ -314,7 +322,8 @@ cdef class Span:
yield left
property rights:
"""Tokens that are to the right of the Span, whose head is within the Span.
"""
Tokens that are to the right of the Span, whose head is within the Span.
Yields: Token A right-child of a token of the span.
"""
@ -325,7 +334,8 @@ cdef class Span:
yield right
property subtree:
"""Tokens that descend from tokens in the span, but fall outside it.
"""
Tokens that descend from tokens in the span, but fall outside it.
Yields: Token A descendant of a token within the span.
"""
@ -337,7 +347,9 @@ cdef class Span:
yield from word.subtree
property ent_id:
'''An (integer) entity ID. Usually assigned by patterns in the Matcher.'''
"""
An (integer) entity ID. Usually assigned by patterns in the Matcher.
"""
def __get__(self):
return self.root.ent_id
@ -345,9 +357,11 @@ cdef class Span:
# TODO
raise NotImplementedError(
"Can't yet set ent_id from Span. Vote for this feature on the issue "
"tracker: http://github.com/spacy-io/spaCy")
"tracker: http://github.com/explosion/spaCy/issues")
property ent_id_:
'''A (string) entity ID. Usually assigned by patterns in the Matcher.'''
"""
A (string) entity ID. Usually assigned by patterns in the Matcher.
"""
def __get__(self):
return self.root.ent_id_
@ -355,7 +369,7 @@ cdef class Span:
# TODO
raise NotImplementedError(
"Can't yet set ent_id_ from Span. Vote for this feature on the issue "
"tracker: http://github.com/spacy-io/spaCy")
"tracker: http://github.com/explosion/spaCy/issues")
property orth_:
def __get__(self):
@ -397,5 +411,5 @@ cdef int _count_words_to_root(const TokenC* token, int sent_length) except -1:
raise RuntimeError(
"Array bounds exceeded while searching for root word. This likely "
"means the parse tree is in an invalid state. Please report this "
"issue here: http://github.com/honnibal/spaCy/")
"issue here: http://github.com/explosion/spaCy/issues")
return n

103
spacy/tokens/token.pyx
View File

@ -1,5 +1,5 @@
# coding: utf8
# cython: infer_types=True
# coding: utf8
from __future__ import unicode_literals
from libc.string cimport memcpy
@ -8,20 +8,15 @@ from cpython.mem cimport PyMem_Malloc, PyMem_Free
from cython.view cimport array as cvarray
cimport numpy as np
np.import_array()
import numpy
import six
from ..typedefs cimport hash_t
from ..lexeme cimport Lexeme
from .. import parts_of_speech
from ..attrs cimport LEMMA
from ..attrs cimport ID, ORTH, NORM, LOWER, SHAPE, PREFIX, SUFFIX, LENGTH, CLUSTER
from ..attrs cimport POS, LEMMA, TAG, DEP
from ..parts_of_speech cimport CCONJ, PUNCT
from ..attrs cimport IS_ALPHA, IS_ASCII, IS_DIGIT, IS_LOWER, IS_PUNCT, IS_SPACE
from ..attrs cimport IS_BRACKET
from ..attrs cimport IS_QUOTE
@ -29,12 +24,13 @@ from ..attrs cimport IS_LEFT_PUNCT
from ..attrs cimport IS_RIGHT_PUNCT
from ..attrs cimport IS_TITLE, IS_UPPER, LIKE_URL, LIKE_NUM, LIKE_EMAIL, IS_STOP
from ..attrs cimport IS_OOV
from ..lexeme cimport Lexeme
from ..compat import is_config
cdef class Token:
"""An individual token --- i.e. a word, punctuation symbol, whitespace, etc.
"""
An individual token --- i.e. a word, punctuation symbol, whitespace, etc.
"""
def __cinit__(self, Vocab vocab, Doc doc, int offset):
self.vocab = vocab
@ -46,7 +42,9 @@ cdef class Token:
return hash((self.doc, self.i))
def __len__(self):
'''Number of unicode characters in token.text'''
"""
Number of unicode characters in token.text.
"""
return self.c.lex.length
def __unicode__(self):
@ -56,7 +54,7 @@ cdef class Token:
return self.text.encode('utf8')
def __str__(self):
if six.PY3:
if is_config(python3=True):
return self.__unicode__()
return self.__bytes__()
@ -83,27 +81,30 @@ cdef class Token:
raise ValueError(op)
cpdef bint check_flag(self, attr_id_t flag_id) except -1:
'''Check the value of a boolean flag.
"""
Check the value of a boolean flag.
Arguments:
flag_id (int): The ID of the flag attribute.
Returns:
is_set (bool): Whether the flag is set.
'''
"""
return Lexeme.c_check_flag(self.c.lex, flag_id)
def nbor(self, int i=1):
'''Get a neighboring token.
"""
Get a neighboring token.
Arguments:
i (int): The relative position of the token to get. Defaults to 1.
Returns:
neighbor (Token): The token at position self.doc[self.i+i]
'''
"""
return self.doc[self.i+i]
def similarity(self, other):
'''Compute a semantic similarity estimate. Defaults to cosine over vectors.
"""
Compute a semantic similarity estimate. Defaults to cosine over vectors.
Arguments:
other:
@ -111,7 +112,7 @@ cdef class Token:
Token and Lexeme objects.
Returns:
score (float): A scalar similarity score. Higher is more similar.
'''
"""
if 'similarity' in self.doc.user_token_hooks:
return self.doc.user_token_hooks['similarity'](self)
if self.vector_norm == 0 or other.vector_norm == 0:
@ -209,9 +210,9 @@ cdef class Token:
self.c.dep = label
property has_vector:
'''
"""
A boolean value indicating whether a word vector is associated with the object.
'''
"""
def __get__(self):
if 'has_vector' in self.doc.user_token_hooks:
return self.doc.user_token_hooks['has_vector'](self)
@ -223,11 +224,11 @@ cdef class Token:
return False
property vector:
'''
"""
A real-valued meaning representation.
Type: numpy.ndarray[ndim=1, dtype='float32']
'''
"""
def __get__(self):
if 'vector' in self.doc.user_token_hooks:
return self.doc.user_token_hooks['vector'](self)
@ -245,6 +246,7 @@ cdef class Token:
property repvec:
def __get__(self):
raise AttributeError("repvec was renamed to vector in v0.100")
property has_repvec:
def __get__(self):
raise AttributeError("has_repvec was renamed to has_vector in v0.100")
@ -265,7 +267,8 @@ cdef class Token:
property lefts:
def __get__(self):
"""The leftward immediate children of the word, in the syntactic
"""
The leftward immediate children of the word, in the syntactic
dependency parse.
"""
cdef int nr_iter = 0
@ -282,8 +285,10 @@ cdef class Token:
property rights:
def __get__(self):
"""The rightward immediate children of the word, in the syntactic
dependency parse."""
"""
The rightward immediate children of the word, in the syntactic
dependency parse.
"""
cdef const TokenC* ptr = self.c + (self.c.r_edge - self.i)
tokens = []
cdef int nr_iter = 0
@ -300,19 +305,21 @@ cdef class Token:
yield t
property children:
'''A sequence of the token's immediate syntactic children.
"""
A sequence of the token's immediate syntactic children.
Yields: Token A child token such that child.head==self
'''
"""
def __get__(self):
yield from self.lefts
yield from self.rights
property subtree:
'''A sequence of all the token's syntactic descendents.
"""
A sequence of all the token's syntactic descendents.
Yields: Token A descendent token such that self.is_ancestor(descendent)
'''
"""
def __get__(self):
for word in self.lefts:
yield from word.subtree
@ -321,26 +328,29 @@ cdef class Token:
yield from word.subtree
property left_edge:
'''The leftmost token of this token's syntactic descendents.
"""
The leftmost token of this token's syntactic descendents.
Returns: Token The first token such that self.is_ancestor(token)
'''
"""
def __get__(self):
return self.doc[self.c.l_edge]
property right_edge:
'''The rightmost token of this token's syntactic descendents.
"""
The rightmost token of this token's syntactic descendents.
Returns: Token The last token such that self.is_ancestor(token)
'''
"""
def __get__(self):
return self.doc[self.c.r_edge]
property ancestors:
'''A sequence of this token's syntactic ancestors.
"""
A sequence of this token's syntactic ancestors.
Yields: Token A sequence of ancestor tokens such that ancestor.is_ancestor(self)
'''
"""
def __get__(self):
cdef const TokenC* head_ptr = self.c
# guard against infinite loop, no token can have
@ -356,25 +366,29 @@ cdef class Token:
return self.is_ancestor(descendant)
def is_ancestor(self, descendant):
'''Check whether this token is a parent, grandparent, etc. of another
"""
Check whether this token is a parent, grandparent, etc. of another
in the dependency tree.
Arguments:
descendant (Token): Another token.
Returns:
is_ancestor (bool): Whether this token is the ancestor of the descendant.
'''
"""
if self.doc is not descendant.doc:
return False
return any( ancestor.i == self.i for ancestor in descendant.ancestors )
property head:
'''The syntactic parent, or "governor", of this token.
"""
The syntactic parent, or "governor", of this token.
Returns: Token
'''
"""
def __get__(self):
"""The token predicted by the parser to be the head of the current token."""
"""
The token predicted by the parser to be the head of the current token.
"""
return self.doc[self.i + self.c.head]
def __set__(self, Token new_head):
# this function sets the head of self to new_head
@ -467,10 +481,11 @@ cdef class Token:
self.c.head = rel_newhead_i
property conjuncts:
'''A sequence of coordinated tokens, including the token itself.
"""
A sequence of coordinated tokens, including the token itself.
Yields: Token A coordinated token
'''
"""
def __get__(self):
"""Get a list of conjoined words."""
cdef Token word
@ -501,7 +516,9 @@ cdef class Token:
return iob_strings[self.c.ent_iob]
property ent_id:
'''An (integer) entity ID. Usually assigned by patterns in the Matcher.'''
"""
An (integer) entity ID. Usually assigned by patterns in the Matcher.
"""
def __get__(self):
return self.c.ent_id
@ -509,7 +526,9 @@ cdef class Token:
self.c.ent_id = key
property ent_id_:
'''A (string) entity ID. Usually assigned by patterns in the Matcher.'''
"""
A (string) entity ID. Usually assigned by patterns in the Matcher.
"""
def __get__(self):
return self.vocab.strings[self.c.ent_id]

11
spacy/train.py
View File

@ -1,15 +1,16 @@
from __future__ import absolute_import
from __future__ import unicode_literals
# coding: utf8
from __future__ import absolute_import, unicode_literals
import random
import tqdm
from .gold import GoldParse
from .gold import GoldParse, merge_sents
from .scorer import Scorer
from .gold import merge_sents
class Trainer(object):
'''Manage training of an NLP pipeline.'''
"""
Manage training of an NLP pipeline.
"""
def __init__(self, nlp, gold_tuples):
self.nlp = nlp
self.gold_tuples = gold_tuples

94
spacy/util.py
View File

@ -1,29 +1,18 @@
# coding: utf8
from __future__ import unicode_literals, print_function
import os
import io
import json
import ujson
import re
import os.path
import pathlib
from pathlib import Path
import sys
import textwrap
try:
basestring
except NameError:
basestring = str
try:
raw_input
except NameError: # Python 3
raw_input = input
from .compat import basestring_, unicode_, input_
LANGUAGES = {}
_data_path = pathlib.Path(__file__).parent / 'data'
_data_path = Path(__file__).parent / 'data'
def set_lang_class(name, cls):
@ -47,9 +36,14 @@ def get_data_path(require_exists=True):
def set_data_path(path):
global _data_path
if isinstance(path, basestring):
path = pathlib.Path(path)
_data_path = path
_data_path = ensure_path(path)
def ensure_path(path):
if isinstance(path, basestring_):
return Path(path)
else:
return path
def or_(val1, val2):
@ -61,41 +55,8 @@ def or_(val1, val2):
return val2
def match_best_version(target_name, target_version, path):
path = path if not isinstance(path, basestring) else pathlib.Path(path)
if path is None or not path.exists():
return None
matches = []
for data_name in path.iterdir():
name, version = split_data_name(data_name.parts[-1])
if name == target_name and constraint_match(target_version, version):
matches.append((tuple(float(v) for v in version.split('.')), data_name))
if matches:
return pathlib.Path(max(matches)[1])
else:
return None
def split_data_name(name):
return name.split('-', 1) if '-' in name else (name, '')
def constraint_match(constraint_string, version):
# From http://github.com/spacy-io/sputnik
if not constraint_string:
return True
constraints = [c.strip() for c in constraint_string.split(',') if c.strip()]
for c in constraints:
if not re.match(r'[><=][=]?\d+(\.\d+)*', c):
raise ValueError('invalid constraint: %s' % c)
return all(semver.match(version, c) for c in constraints)
def read_regex(path):
path = path if not isinstance(path, basestring) else pathlib.Path(path)
path = ensure_path(path)
with path.open() as file_:
entries = file_.read().split('\n')
expression = '|'.join(['^' + re.escape(piece) for piece in entries if piece.strip()])
@ -152,21 +113,11 @@ def check_renamed_kwargs(renamed, kwargs):
raise TypeError("Keyword argument %s now renamed to %s" % (old, new))
def is_windows():
"""Check if user is on Windows."""
return sys.platform.startswith('win')
def is_python2():
"""Check if Python 2 is used."""
return sys.version.startswith('2.')
def parse_package_meta(package_path, package, require=True):
location = os.path.join(str(package_path), package, 'meta.json')
if os.path.isfile(location):
with io.open(location, encoding='utf8') as f:
meta = json.load(f)
location = package_path / package / 'meta.json'
if location.is_file():
with location.open('r', encoding='utf8') as f:
meta = ujson.load(f)
return meta
elif require:
raise IOError("Could not read meta.json from %s" % location)
@ -181,7 +132,7 @@ def get_raw_input(description, default=False):
additional = ' (default: {d})'.format(d=default) if default else ''
prompt = ' {d}{a}: '.format(d=description, a=additional)
user_input = raw_input(prompt)
user_input = input_(prompt)
return user_input
@ -209,10 +160,9 @@ def print_markdown(data, **kwargs):
which will be converted to a list of tuples."""
def excl_value(value):
# don't print value if it contains absolute path of directory
# (i.e. personal info that shouldn't need to be shared)
# other conditions can be included here if necessary
if str(pathlib.Path(__file__).parent) in value:
# don't print value if it contains absolute path of directory (i.e.
# personal info). Other conditions can be included here if necessary.
if unicode_(Path(__file__).parent) in value:
return True
if type(data) == dict:

119
spacy/vocab.pyx
View File

@ -1,41 +1,29 @@
# coding: utf8
from __future__ import unicode_literals
import bz2
import ujson
import re
from libc.string cimport memset
from libc.stdint cimport int32_t
from libc.math cimport sqrt
from pathlib import Path
import bz2
import ujson as json
import re
try:
import cPickle as pickle
except ImportError:
import pickle
from cymem.cymem cimport Address
from .lexeme cimport EMPTY_LEXEME
from .lexeme cimport Lexeme
from .strings cimport hash_string
from .typedefs cimport attr_t
from .cfile cimport CFile, StringCFile
from .lemmatizer import Lemmatizer
from .attrs import intify_attrs
from .tokens.token cimport Token
from . import attrs
from . import symbols
from cymem.cymem cimport Address
from .serialize.packer cimport Packer
from .attrs cimport PROB, LANG
from .compat import copy_reg, pickle
from .lemmatizer import Lemmatizer
from .attrs import intify_attrs
from . import util
try:
import copy_reg
except ImportError:
import copyreg as copy_reg
from . import attrs
from . import symbols
DEF MAX_VEC_SIZE = 100000
@ -48,8 +36,9 @@ EMPTY_LEXEME.vector = EMPTY_VEC
cdef class Vocab:
'''A map container for a language's LexemeC structs.
'''
"""
A map container for a language's LexemeC structs.
"""
@classmethod
def load(cls, path, lex_attr_getters=None, lemmatizer=True,
tag_map=True, serializer_freqs=True, oov_prob=True, **deprecated_kwargs):
@ -72,8 +61,7 @@ cdef class Vocab:
Returns:
Vocab: The newly constructed vocab object.
"""
if isinstance(path, basestring):
path = Path(path)
path = util.ensure_path(path)
util.check_renamed_kwargs({'get_lex_attr': 'lex_attr_getters'}, deprecated_kwargs)
if 'vectors' in deprecated_kwargs:
raise AttributeError(
@ -81,7 +69,7 @@ cdef class Vocab:
"Install vectors after loading.")
if tag_map is True and (path / 'vocab' / 'tag_map.json').exists():
with (path / 'vocab' / 'tag_map.json').open('r', encoding='utf8') as file_:
tag_map = json.load(file_)
tag_map = ujson.load(file_)
elif tag_map is True:
tag_map = None
if lex_attr_getters is not None \
@ -94,12 +82,12 @@ cdef class Vocab:
lemmatizer = Lemmatizer.load(path)
if serializer_freqs is True and (path / 'vocab' / 'serializer.json').exists():
with (path / 'vocab' / 'serializer.json').open('r', encoding='utf8') as file_:
serializer_freqs = json.load(file_)
serializer_freqs = ujson.load(file_)
else:
serializer_freqs = None
with (path / 'vocab' / 'strings.json').open('r', encoding='utf8') as file_:
strings_list = json.load(file_)
strings_list = ujson.load(file_)
cdef Vocab self = cls(lex_attr_getters=lex_attr_getters, tag_map=tag_map,
lemmatizer=lemmatizer, serializer_freqs=serializer_freqs,
strings=strings_list)
@ -108,7 +96,8 @@ cdef class Vocab:
def __init__(self, lex_attr_getters=None, tag_map=None, lemmatizer=None,
serializer_freqs=None, strings=tuple(), **deprecated_kwargs):
'''Create the vocabulary.
"""
Create the vocabulary.
lex_attr_getters (dict):
A dictionary mapping attribute IDs to functions to compute them.
@ -123,7 +112,7 @@ cdef class Vocab:
Returns:
Vocab: The newly constructed vocab object.
'''
"""
util.check_renamed_kwargs({'get_lex_attr': 'lex_attr_getters'}, deprecated_kwargs)
lex_attr_getters = lex_attr_getters if lex_attr_getters is not None else {}
@ -172,17 +161,19 @@ cdef class Vocab:
return langfunc('_') if langfunc else ''
def __len__(self):
"""The current number of lexemes stored."""
"""
The current number of lexemes stored.
"""
return self.length
def resize_vectors(self, int new_size):
'''
"""
Set vectors_length to a new size, and allocate more memory for the Lexeme
vectors if necessary. The memory will be zeroed.
Arguments:
new_size (int): The new size of the vectors.
'''
"""
cdef hash_t key
cdef size_t addr
if new_size > self.vectors_length:
@ -193,7 +184,8 @@ cdef class Vocab:
self.vectors_length = new_size
def add_flag(self, flag_getter, int flag_id=-1):
'''Set a new boolean flag to words in the vocabulary.
"""
Set a new boolean flag to words in the vocabulary.
The flag_setter function will be called over the words currently in the
vocab, and then applied to new words as they occur. You'll then be able
@ -213,7 +205,7 @@ cdef class Vocab:
Returns:
flag_id (int): The integer ID by which the flag value can be checked.
'''
"""
if flag_id == -1:
for bit in range(1, 64):
if bit not in self.lex_attr_getters:
@ -234,9 +226,11 @@ cdef class Vocab:
return flag_id
cdef const LexemeC* get(self, Pool mem, unicode string) except NULL:
'''Get a pointer to a LexemeC from the lexicon, creating a new Lexeme
"""
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.'''
is the lexicon's own memory, the lexeme is saved in the lexicon.
"""
if string == u'':
return &EMPTY_LEXEME
cdef LexemeC* lex
@ -252,9 +246,11 @@ cdef class Vocab:
return self._new_lexeme(mem, string)
cdef const LexemeC* get_by_orth(self, Pool mem, attr_t orth) except NULL:
'''Get a pointer to a LexemeC from the lexicon, creating a new Lexeme
"""
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.'''
is the lexicon's own memory, the lexeme is saved in the lexicon.
"""
if orth == 0:
return &EMPTY_LEXEME
cdef LexemeC* lex
@ -297,30 +293,33 @@ cdef class Vocab:
self.length += 1
def __contains__(self, unicode string):
'''Check whether the string has an entry in the vocabulary.
"""
Check whether the string has an entry in the vocabulary.
Arguments:
string (unicode): The ID string.
Returns:
bool Whether the string has an entry in the vocabulary.
'''
"""
key = hash_string(string)
lex = self._by_hash.get(key)
return lex is not NULL
def __iter__(self):
'''Iterate over the lexemes in the vocabulary.
"""
Iterate over the lexemes in the vocabulary.
Yields: Lexeme An entry in the vocabulary.
'''
"""
cdef attr_t orth
cdef size_t addr
for orth, addr in self._by_orth.items():
yield Lexeme(self, orth)
def __getitem__(self, id_or_string):
'''Retrieve a lexeme, given an int ID or a unicode string. If a previously
"""
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.
Arguments:
@ -332,7 +331,7 @@ cdef class Vocab:
Returns:
lexeme (Lexeme): The lexeme indicated by the given ID.
'''
"""
cdef attr_t orth
if type(id_or_string) == unicode:
orth = self.strings[id_or_string]
@ -355,7 +354,8 @@ cdef class Vocab:
return tokens
def dump(self, loc=None):
"""Save the lexemes binary data to the given location, or
"""
Save the lexemes binary data to the given location, or
return a byte-string with the data if loc is None.
Arguments:
@ -392,14 +392,15 @@ cdef class Vocab:
return fp.string_data()
def load_lexemes(self, loc):
'''Load the binary vocabulary data from the given location.
"""
Load the binary vocabulary data from the given location.
Arguments:
loc (Path): The path to load from.
Returns:
None
'''
"""
fp = CFile(loc, 'rb',
on_open_error=lambda: IOError('LexemeCs file not found at %s' % loc))
cdef LexemeC* lexeme = NULL
@ -440,8 +441,9 @@ cdef class Vocab:
fp.close()
def _deserialize_lexemes(self, CFile fp):
'''Load the binary vocabulary data from the given CFile.
'''
"""
Load the binary vocabulary data from the given CFile.
"""
cdef LexemeC* lexeme = NULL
cdef hash_t key
cdef unicode py_str
@ -494,13 +496,14 @@ cdef class Vocab:
fp.close()
def dump_vectors(self, out_loc):
'''Save the word vectors to a binary file.
"""
Save the word vectors to a binary file.
Arguments:
loc (Path): The path to save to.
Returns:
None
'''
"""
cdef int32_t vec_len = self.vectors_length
cdef int32_t word_len
cdef bytes word_str
@ -522,7 +525,8 @@ cdef class Vocab:
out_file.close()
def load_vectors(self, file_):
"""Load vectors from a text-based file.
"""
Load vectors from a text-based file.
Arguments:
file_ (buffer): The file to read from. Entries should be separated by newlines,
@ -561,7 +565,8 @@ cdef class Vocab:
return vec_len
def load_vectors_from_bin_loc(self, loc):
"""Load vectors from the location of a binary file.
"""
Load vectors from the location of a binary file.
Arguments:
loc (unicode): The path of the binary file to load from.