def __init__(self):
"""初始化自动机，load词典，构建自动机"""
self.__acto = AhoCorasick()
with open("dict/SogouLabDic.dic", "r") as fin:
for line in fin:
line = line.strip()
tokens = line.split("\t")
if len(tokens) < 2:
continue
if len(tokens) == 2:
word = tokens[0].decode("utf8", "ignore")
pos_tag = u"ProperNoun"
self.__acto.insert(word, pos_tag)
elif len(tokens) == 3:
word = tokens[0].decode("utf8", "ignore")
pos_list = tokens[2].split(",")
sign = 0
for pos_tag in pos_list:
if len(pos_tag) == 0:
continue
sign = 1
pos_tag = pos_tag.decode("utf8", "ignore")
self.__acto.insert(word, pos_tag)
if sign == 0:
self.__acto.insert(word, u"ProperNoun")
self.__acto.build_fail()
def __init__(self):
"""初始化自动机，load词典，构建自动机"""
self.__acto = AhoCorasick()
with open("dict/SogouLabDic.dic", "r") as fin:
for line in fin:
line = line.strip()
tokens = line.split("\t")
if len(tokens) < 2:
continue
if len(tokens) == 2:
word = tokens[0].decode("utf8", "ignore")
pos_tag = u"ProperNoun"
self.__acto.insert(word, pos_tag)
elif len(tokens) == 3:
word = tokens[0].decode("utf8", "ignore")
pos_list = tokens[2].split(",")
sign = 0
for pos_tag in pos_list:
if len(pos_tag) == 0:
continue
sign = 1
pos_tag = pos_tag.decode("utf8", "ignore")
self.__acto.insert(word, pos_tag)
if sign == 0:
self.__acto.insert(word, u"ProperNoun")
self.__acto.build_fail()
class RawSeg(object):
"""基础切词，返回多条切词路径"""
def __init__(self):
"""初始化自动机，load词典，构建自动机"""
self.__acto = AhoCorasick()
with open("dict/SogouLabDic.dic", "r") as fin:
for line in fin:
line = line.strip()
tokens = line.split("\t")
if len(tokens) < 2:
continue
if len(tokens) == 2:
word = tokens[0].decode("utf8", "ignore")
pos_tag = u"ProperNoun"
self.__acto.insert(word, pos_tag)
elif len(tokens) == 3:
word = tokens[0].decode("utf8", "ignore")
pos_list = tokens[2].split(",")
sign = 0
for pos_tag in pos_list:
if len(pos_tag) == 0:
continue
sign = 1
pos_tag = pos_tag.decode("utf8", "ignore")
self.__acto.insert(word, pos_tag)
if sign == 0:
self.__acto.insert(word, u"ProperNoun")
self.__acto.build_fail()
def backtrace(self, ulist, N, cidx, wlist, M, widx, path_stk):
"""回溯法，寻找所有路径"""
if cidx >= N:
path = Path()
path.word_list = copy.deepcopy(path_stk)
return [path]
if widx >= M:
"""单词已经枚举光了"""
path.word_list = copy.deepcopy(path_stk)
while cidx < N:
word = Word(ulist[cidx], [u"Single"])
cidx += 1
path.word_list.append(word)
return [path]
path_list = []
if cidx < N:
word = Word(ulist[cidx], [u"Single"])
path_stk.append(word)
path_list += self.backtrace(ulist, N, cidx + 1, wlist, M, widx, path_stk)
path_stk.pop()
while widx < M:
if wlist[widx][0] < cidx:
widx += 1
continue
if wlist[widx][0] > cidx:
break
word = Word(ulist[wlist[widx][0]:wlist[widx][1]], wlist[widx][2])
path_stk.append(word)
path_list += self.backtrace(ulist, N, wlist[widx][1], wlist, M, widx + 1, path_stk)
path_stk.pop()
widx += 1
return path_list
def seg(self, query):
"""对query进行切词，切分路径list"""
assert isinstance(query, unicode)
word_idx_list = self.__acto.get_match_word_idx(query)
word_idx_list = sorted(word_idx_list, key=lambda x: x[0])
path_list = []
path_stk = []
return self.backtrace(query, len(query), 0, word_idx_list, len(word_idx_list), 0, path_stk)
class RawSeg(object):
"""基础切词，返回多条切词路径"""
def __init__(self):
"""初始化自动机，load词典，构建自动机"""
self.__acto = AhoCorasick()
with open("dict/SogouLabDic.dic", "r") as fin:
for line in fin:
line = line.strip()
tokens = line.split("\t")
if len(tokens) < 2:
continue
if len(tokens) == 2:
word = tokens[0].decode("utf8", "ignore")
pos_tag = u"ProperNoun"
self.__acto.insert(word, pos_tag)
elif len(tokens) == 3:
word = tokens[0].decode("utf8", "ignore")
pos_list = tokens[2].split(",")
sign = 0
for pos_tag in pos_list:
if len(pos_tag) == 0:
continue
sign = 1
pos_tag = pos_tag.decode("utf8", "ignore")
self.__acto.insert(word, pos_tag)
if sign == 0:
self.__acto.insert(word, u"ProperNoun")
self.__acto.build_fail()
def backtrace(self, ulist, N, cidx, wlist, M, widx, path_stk):
"""回溯法，寻找所有路径"""
if cidx >= N:
path = Path()
path.word_list = copy.deepcopy(path_stk)
return [path]
if widx >= M:
"""单词已经枚举光了"""
path.word_list = copy.deepcopy(path_stk)
while cidx < N:
word = Word(ulist[cidx], [u"Single"])
cidx += 1
path.word_list.append(word)
return [path]
path_list = []
if cidx < N:
word = Word(ulist[cidx], [u"Single"])
path_stk.append(word)
path_list += self.backtrace(ulist, N, cidx + 1, wlist, M, widx,
path_stk)
path_stk.pop()
while widx < M:
if wlist[widx][0] < cidx:
widx += 1
continue
if wlist[widx][0] > cidx:
break
word = Word(ulist[wlist[widx][0]:wlist[widx][1]], wlist[widx][2])
path_stk.append(word)
path_list += self.backtrace(ulist, N, wlist[widx][1], wlist, M,
widx + 1, path_stk)
path_stk.pop()
widx += 1
return path_list
def seg(self, query):
"""对query进行切词，切分路径list"""
assert isinstance(query, unicode)
word_idx_list = self.__acto.get_match_word_idx(query)
word_idx_list = sorted(word_idx_list, key=lambda x: x[0])
path_list = []
path_stk = []
return self.backtrace(query, len(query), 0, word_idx_list,
len(word_idx_list), 0, path_stk)
