def __timeEx(self):
"""
:param target: 输入文本字符串
:param timeBase: 输入基准时间
:return: TimeUnit[]时间表达式类型数组
"""
startline = -1
endline = -1
rpointer = 0
temp = []
match = self.pattern.finditer(self.target)
for m in match:
startline = m.start()
if startline == endline:
rpointer -= 1
temp[rpointer] = temp[rpointer] + m.group()
else:
temp.append(m.group())
endline = m.end()
rpointer += 1
res = []
# 时间上下文： 前一个识别出来的时间会是下一个时间的上下文，用于处理：周六3点到5点这样的多个时间的识别，第二个5点应识别到是周六的。
contextTp = TimePoint()
for i in range(0, rpointer):
res.append(TimeUnit(temp[i], self, contextTp))
contextTp = res[i].tp
res = self.__filterTimeUnit(res)
return res
def __timeEx(self):
"""
:param target: 输入文本字符串
:param timeBase: 输入基准时间
:return: TimeUnit[]时间表达式类型数组
"""
res = []
sents = cut_sent(
re.sub(
'http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@.&+]|[!*\(\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+',
'[url]', self.target)).split('\n')
for sent in sents:
startline = -1
endline = -1
rpointer = 0
temp = []
match = self.pattern.finditer(sent)
for m in match:
# print(m.group())
startline = m.start()
if startline == endline:
rpointer -= 1
temp[rpointer] = temp[rpointer] + m.group()
else:
temp.append(m.group())
endline = m.end()
rpointer += 1
if rpointer > 0:
if sent.find('将') != -1:
isPreferFuture = True
# print('future!!')
else:
isPreferFuture = False
self.isPreferFuture = isPreferFuture
# 时间上下文： 前一个识别出来的时间会是下一个时间的上下文，用于处理：周六3点到5点这样的多个时间的识别，第二个5点应识别到是周六的。
contextTp = TimePoint()
for i in range(0, rpointer):
res.append(
TimeUnit(temp[i], self, contextTp, sent, isPreferFuture))
contextTp = res[i].tp
res = self.__filterTimeUnit(res)
return res
def __timeEx(self):
"""
:param target: 输入文本字符串
:param timeBase: 输入基准时间
:return: TimeUnit[]时间表达式类型数组
"""
startline = -1
endline = -1
rpointer = 0
temp = []
match = self.pattern.finditer(self.target)
for m in match:
startline = m.start()
if startline == endline:
rpointer -= 1
temp[rpointer] = temp[rpointer] + m.group()
else:
temp.append(m.group())
endline = m.end()
rpointer += 1
res = []
# 时间上下文： 前一个识别出来的时间会是下一个时间的上下文，用于处理：周六3点到5点这样的多个时间的识别，第二个5点应识别到是周六的。
contextTp = TimePoint()
# print(self.timeBase)
# print('temp',temp)
for i in range(0, rpointer):
# 这里是一个类嵌套了一个类
# y = temp[i]
# ss = TimeUnit(temp[i], self, contextTp)
res.append(TimeUnit(temp[i], self, contextTp))
# res[i].tp.tunit[3] = -1
contextTp = res[i].tp
self.timeBase = arrow.get(arrow.now()).format('YYYY-M-D-H-m-s')
# print(self.nowTime.year)
# print(contextTp.tunit)
res = self.__filterTimeUnit(res)
return [temp, res]
def __timeEx(self):
"""
:param target: 输入文本字符串
:param timeBase: 输入基准时间
:return: TimeUnit[]时间表达式类型数组
"""
startline = -1
endline = -1
rpointer = 0
time_extractor = []
match = self.pattern.finditer(self.target)
for m in match:
startline = m.start()
if startline == endline:
rpointer -= 1
time_extractor[rpointer] = time_extractor[rpointer] + m.group()
else:
time_extractor.append(m.group())
endline = m.end()
rpointer += 1
res = []
# 时间上下文： 前一个识别出来的时间会是下一个时间的上下文，用于处理：周六3点到5点这样的多个时间的识别，第二个5点应识别到是周六的。
contextTp = TimePoint()
# print(self.timeBase)
print('time_extractor:', time_extractor)
for i in range(0, rpointer):
# 这里是一个类嵌套了一个类
res.append(TimeUnit(time_extractor[i], self, contextTp))
# res[i].tp.tunit[3] = -1
contextTp = res[i].tp
# print(self.nowTime.year)
# print(contextTp.tunit)
res = self.__filterTimeUnit(res)
return res