def predict(generator, text):
"""
Args:
generator: 生成器
text:
Return:
response:
predict
"""
model_text, topic_dict = \
preprocessing_for_one_conversation(text.strip(), \
topic_generalization=True)
src, tgt, cue = model_text.split('\t')
cue = cue.split('\1')
response = generator.interact(src, cue)
topic_list = sorted(topic_dict.items(),
key=lambda item: len(item[1]),
reverse=True)
for key, value in topic_list:
response = response.replace(key, value)
return response
def predict(model, text):
"""
predict
"""
model_text, candidates = \
preprocessing_for_one_conversation(text.strip(),
candidate_num=50,
use_knowledge=True,
topic_generalization=True,
for_predict=True)
for i, text_ in enumerate(model_text):
score = interact.predict(model, text_, task_name="match_kn_gene")
candidates[i] = [candidates[i], score]
candidate_legal = sorted(candidates,
key=lambda item: item[1],
reverse=True)
return candidate_legal[0][0]
def predict(model, text):
"""
predict
"""
model_text, topic_dict = \
preprocessing_for_one_conversation(text.strip(), topic_generalization=True)
if isinstance(model_text, unicode):
model_text = model_text.encode('utf-8')
response = network.predict(model, model_text)
topic_list = sorted(topic_dict.items(),
key=lambda item: len(item[1]),
reverse=True)
for key, value in topic_list:
response = response.replace(key, value)
return response