def msgpreprocess(self, text):
at = False
for t in jieba.cut(text, HMM=False):
if t == '@':
at = True
elif at:
yield '@' + t
at = False
elif t.lower() not in self.stopwords:
# t.isidentifier() and
yield t
def msgpreprocess(self, text):
at = False
for t in jieba.cut(text, HMM=False):
if t == '@':
at = True
elif at:
yield '@' + t
at = False
elif t.lower() not in self.stopwords:
# t.isidentifier() and
yield t
def getsayingbytext(text='', mode='r'):
global SAY_P
with SAY_LCK:
text = (mode + ' '.join(mosesproxy.cut(zhconv.convert(text, 'zh-hans'), HMM=False)[:60]).strip()).encode('utf-8') + b'\n'
try:
SAY_P.stdin.write(text)
SAY_P.stdin.flush()
say = SAY_P.stdout.readline().strip().decode('utf-8')
except BrokenPipeError:
SAY_P = subprocess.Popen(SAY_CMD, stdin=subprocess.PIPE, stdout=subprocess.PIPE, cwd='vendor')
SAY_P.stdin.write(text)
SAY_P.stdin.flush()
say = SAY_P.stdout.readline().strip().decode('utf-8')
return say
def cmd_cut(tinput, lang):
if lang == 'c':
return ' '.join(mosesproxy.jiebazhc.cut(tinput, HMM=False))
else:
return ' '.join(mosesproxy.cut(tinput, HMM=False))