spaCy/bin/ner_tag.py

35 lines
831 B
Python

import codecs
import plac
from spacy.en import English
def main(text_loc):
with codecs.open(text_loc, 'r', 'utf8') as file_:
text = file_.read()
NLU = English()
for paragraph in text.split('\n\n'):
tokens = NLU(paragraph)
ent_starts = {}
ent_ends = {}
for span in tokens.ents:
ent_starts[span.start] = span.label_
ent_ends[span.end] = span.label_
output = []
for token in tokens:
if token.i in ent_starts:
output.append('<%s>' % ent_starts[token.i])
output.append(token.orth_)
if (token.i+1) in ent_ends:
output.append('</%s>' % ent_ends[token.i+1])
output.append('\n\n')
print ' '.join(output)
if __name__ == '__main__':
plac.call(main)