def output_sentence(sent):
label_key = 'cpos' if args['--coarse'] else 'pos'
for i in range(len(sent['word'])):
print >>data_out, "{label} '{name}-{sent_i}-{token_i}|".format(
label=normalize_label(sent[label_key][i]),
name=args['--name'],
sent_i=sent_i,
token_i=i+1),
print >>data_out, u" ".join(features_for_token(sent['word'], sent[label_key], i))
def output_sentence(sent):
label_key = 'cpos' if args['--coarse'] else 'pos'
if args['--feature-set'] == "dependency":
for i in xrange(len(sent["word"])):
try:
print >>data_out, u"{label} '{sent_i}-{token_i}{features}".format(
label=normalize_label(sent['dependency'][i]),
sent_i=sent_i,
token_i=i,
features=u" ".join(features_for_token(sent['word'], sent[label_key], i)))
#We need this check for sentences we skip
except IndexError:
continue
else:
for i in xrange(len(sent["word"])):
print >>data_out, "{label} '{name}-{sent_i}-{token_i}|".format(
label=normalize_label(sent[label_key][i]),
name=args['--name'],
sent_i=sent_i,
token_i=i + 1),
print >>data_out, u" ".join(
features_for_token(sent['word'], sent[label_key], i))
def write_sentences_to_file(sents, data_out):
for sent_i, sent in enumerate(sents):
tokens = [token for token, pos in sent]
tags = [pos for token, pos in sent]
if sent_i > 0:
print >> data_out, ""
for token_i in range(len(sent)):
print >> data_out, "{label} '{name}-{sent_i}-{token_i}|".format(
label=normalize_label(tags[token_i]),
name='brown',
sent_i=sent_i,
token_i=token_i),
print >> data_out, u" ".join(honnibal13(tokens, tags, token_i))