def preprocess(text):
#TODO とりあえず最初の文の最初の名詞を利用, なければ最初の形態素
sentences = api.sentences(text)['sentences']
for sentence in api.sentences(text)['sentences']:
morphs = api.morph(sentence)['morphs']
for morph in morphs:
if u'名詞' in morph['pos']:
seed = morph
break
else:
seed = morphs[1]
return seed
def twitter_based(text):
seed = generate_seed(text)
# TODO とりあえずツイート検索結果の最初のツイートを利用
tweet_example = api.search_tweets(seed['norm_surface'], limit=1)
# 空ならNoneを返す
if tweet_example['count'] == 0:
return None
# TODO とりあえず先頭の文を利用
# 形態素列書き換え
sent = api.sentences(tweet_example['texts'][0])['sentences'][0]
return postprocess(sent)
def generate_seed(text):
max_length = 0
for sentence in api.sentences(text)['sentences']:
morphs = api.morph(sentence)['morphs']
for morph in morphs:
if u'固有' in morph['pos'] or u'名詞' in morph['pos']:
seed = morph
break
else:
seed = morphs[1]
return seed
def scenario_based(text):
sent = api.sentences(text)
text = []
for s in sent['sentences']:
morphs = api.morph(s)
query = list()
for morph in morphs['morphs']:
query.append(u'{}:{}'.format(morph['surface'], morph['pos']))
texts = api.trigger(scenario_file,query)
for t in texts['texts']:
text.append(t)
if len(text)>0:
r = random.randint(0,len(text)-1)
return text[r]
return None
# -*- coding: utf-8 -*-
import api
api = api.API('https://52.68.75.108', 'secret', 'js2015cps')
print
print '文分割'
s = '日本語文字列を文単位で分割する。複数文を渡すと、文ごとに区切ってくれる。'
for sentence in api.sentences(s)['sentences']:
print sentence
print
print 'ツイート検索'
print '=' * 20
for text in api.search_tweets('検索')['texts']:
print text
print
print 'リプライ検索'
print '=' * 20
for text in api.search_reply('検索')['texts']:
print text
print
print 'マルコフ連鎖'
print '=' * 20
seed = {'norm_surface': "今日", 'pos': "名詞"}
for morph in api.markov_chain(seed)['morphs']:
print morph,
if __name__ == "__main__":
api = get_api()
reps = api.get_reply()
print json.dumps(reps, ensure_ascii=False, indent=4)
for rep in reps["replies"]:
print "======================================================"
if reps["grade"] == 0:
scenario_file = "scenario_c09.txt"
elif reps["grade"] == 1:
scenario_file = "scenario_c09.txt"
else:
scenario_file = "scenario_c09.txt"
sent = api.sentences(rep["text"])
text = []
for s in sent["sentences"]:
print "-------------------------------------------------------"
print s
morphs = api.morph(s)
print json.dumps(morphs, ensure_ascii=False, indent=4)
query = list()
for morph in morphs["morphs"]:
query.append(u"{}:{}".format(morph["surface"], morph["pos"]))
print json.dumps(query, ensure_ascii=False, indent=4)
texts = api.trigger(scenario_file, query)
print json.dumps(texts, ensure_ascii=False, indent=4)
for t in texts["texts"]:
text.append(t)
print t
