def ask(self, question):
# Repetition
if self.previn and clean_string(question) == clean_string(self.previn[0]):
res = random.choice(self.repetitiondb)
res = res.replace('----', "'%s'"%' '.join(self.prevout[0]))
yield self.macro.subst(res)
return
answers = []
lang = detect_lang(question, self.lastlang)
if lang is None:
lang = self.lastlang
#print 'Language:', lang
#print 'Last Language:', self.lastlang
if lang != self.lastlang:
tran = translate(question, lang).encode('utf-8')
lasttran = translate(question, self.lastlang).encode('utf-8')
lang, prob, tran = max([(lang, ratio_correct(tran), tran),
(self.lastlang, ratio_correct(lasttran), lasttran)],
key=lambda x: x[1])
question = tran
#print 'Chosen:', lang, 'prob:', prob
#print 'Question:', question
is_foreign = lang != 'en'
#if is_foreign:
# question = language.translate(question, lang).encode('utf-8')
if question is None:
return
#check = check_blank(question)
#if check is not None:
# return [check]
if check_for_end(question):
self.running = False
return
#question = self._clean_text(question)
for sentence in sentence_split(question):
answer = None
for handler in self.handlers:
if handler.check(sentence):
try:
ans = handler.answer(sentence)
except HALcannotHandle:
continue
else:
answer = ans
break
if answer is None:
answer = random.choice(self.generic)
answer = self.macro.subst(answer)
if self.semantics:
lang, inter, answer = HALtran.transform(answer)
if self.rndname:
answer = rndname(answer, self.user)
if is_foreign:
#print 'Answer:', answer
answer = translate(answer, 'en', lang)
yield answer
answers.append(answer)
#answers = [self.macro.subst(answer) for answer in answers]
self.do_speech(' '.join(answers))
self.previn.appendleft(question)
self.prevout.appendleft(answers)
def wikipedia(key):
bing = urllib2.urlopen('http://www.bing.com/search?q=%s'%quote_plus(key+' wikipedia')).read()
#result = re.findall('http://en.wikipedia.org/wiki/([^"#]+?)"', bing)[:2]
result = re.findall('http://en.wikipedia.org/wiki/([^"#]+?)"', bing)
if not result:
return None
result = result[0].replace('_', ' ')
#print result
pages = {}
#for page in result:
for page in [result]:
head = None
buf = ''
a = cached_fetch_wiki(page)
if a is None:
continue
page, p = a
#print page
if p is None:
continue
for line in p.split('\n'):
if line and line[0] == '=':
if head is not None:
id = '%s %s'%(page, head)
else:
id = page
pages[id] = buf
head = line.replace('=', '').strip()
buf = ''
elif line and line[0] in '|<{}[]~!@#$%^&*()_+=-\\`/:;.,<>?':
continue
else:
buf += line + '\n'
if head is not None:
id = '%s %s'%(page, head)
else:
id = page
pages[id] = buf
if not pages:
return None
selected = most_relavent(pages.keys(), key)
#print 'Selected:', selected
# page = clean_wikipedia(pages[selected])
# sentences = [i for i in sentence_split(page) if i]
# return ''.join(sentences)
content = pages[selected]
firstline = ''
for line in content.split('\n'):
if not line:
continue
if line[0] in ' |<{}[]~!@#$%^&*()_+=-\\`/:;.,<>?':
continue
if line[-1] in ':':
continue
firstline = line
break
sentences = sentence_split(clean_wikipedia(firstline))
#sentences = resentence.split(clean_wikipedia(firstline))
sentences = [i for i in sentences if i and i[0] != '.']
str = ' '.join(sentences[:2]).strip()
if str and str[-1] not in '!.,?:':
str += '.'
return str
