2016-12-12 20:13:33 +00:00
|
|
|
# Load NER
|
|
|
|
|
from __future__ import unicode_literals
|
|
|
|
|
import spacy
|
|
|
|
|
import pathlib
|
|
|
|
|
from spacy.pipeline import EntityRecognizer
|
|
|
|
|
from spacy.vocab import Vocab
|
|
|
|
|
|
|
|
|
|
def load_model(model_dir):
|
|
|
|
|
model_dir = pathlib.Path(model_dir)
|
|
|
|
|
nlp = spacy.load('en', parser=False, entity=False, add_vectors=False)
|
|
|
|
|
with (model_dir / 'vocab' / 'strings.json').open('r', encoding='utf8') as file_:
|
|
|
|
|
nlp.vocab.strings.load(file_)
|
|
|
|
|
nlp.vocab.load_lexemes(model_dir / 'vocab' / 'lexemes.bin')
|
2016-12-12 20:23:24 +00:00
|
|
|
ner = EntityRecognizer.load(model_dir, nlp.vocab, require=True)
|
2016-12-12 20:13:33 +00:00
|
|
|
return (nlp, ner)
|
|
|
|
|
|
|
|
|
|
(nlp, ner) = load_model('ner')
|
|
|
|
|
doc = nlp.make_doc('Who is Shaka Khan?')
|
|
|
|
|
nlp.tagger(doc)
|
|
|
|
|
ner(doc)
|
|
|
|
|
for word in doc:
|
|
|
|
|
print(word.text, word.orth, word.lower, word.tag_, word.ent_type_, word.ent_iob)
|
