def qa(question):
#初始化jieba分词器
T.jieba_initialize()
#切换到语料库所在工作目录
mybot_path = './'
# os.chdir(mybot_path)
mybot = aiml.Kernel()
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] +
"/resources/std-startup.xml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] + "/resources/bye.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] + "/resources/tools.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] + "/resources/bad.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] + "/resources/funny.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] +
"/resources/OrdinaryQuestion.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] +
"/resources/Common conversation.aiml")
# mybot.respond('Load Doc Snake')
#载入百科属性列表
input_message = question
if len(input_message) > 60:
print(mybot.respond("句子长度过长"))
elif input_message.strip() == '':
print(mybot.respond("无"))
print(input_message)
message = T.wordSegment(input_message)
# 去标点
print('word Seg:' + message)
print('词性：')
words = T.postag(input_message)
if message == 'q':
exit()
else:
response = mybot.respond(message)
print("=======")
print(response)
print("=======")
if response == "":
ans = mybot.respond('找不到答案')
# print(robot_id + ":" + ans)
print("{0}:{1}".format(robot_id, ans))
# 百科搜索
elif response[0] == '#':
# 匹配百科
if response.__contains__("searchbaike"):
print("search from baike")
print(response)
res = response.split(':')
#实体
entity = str(res[1]).replace(" ", "")
#属性
attr = str(res[2]).replace(" ", "")
print(entity + '<---->' + attr)
ans = baike.query(entity, attr)
# 如果命中答案
if type(ans) == list:
print("{0}:{1}".format(robot_id, QAT.ptranswer(ans,
False)))
elif ans.decode('utf-8').__contains__(u'::找不到'):
#百度摘要+Bing摘要
print("通用搜索")
log.info("通用搜索")
ans = search_summary.kwquery(input_message)
# 匹配不到模版，通用查询
elif response.__contains__("NoMatchingTemplate"):
print("NoMatchingTemplate")
ans = search_summary.kwquery(input_message)
if len(ans) == 0:
ans = mybot.respond('找不到答案')
logs.info("{0}:{1}".format(robot_id, ans))
elif len(ans) > 1:
logs.info(sys.exc_info())
logs.info("不确定候选答案")
logs.info("[{0}][func:{1}][line:{2}]:不确定候选答案".format(
sys._getframe().f_code.co_filename,
sys._getframe().f_code.co_name,
sys._getframe().f_lineno))
print(robot_id + ': ')
for a in ans:
print(a)
# print(a.encode("utf8"))
else:
print('{0}:{1}'.format(robot_id, ans[0]))
# 匹配模版
else:
print("{}: {}".format(robot_id, response))
def __inner_job():
start = time.time()
text_binary = analyze_current_screen_text(
directory=data_directory, compress_level=image_compress_level[0])
keywords = get_text_from_image(image_data=text_binary, )
if not keywords:
print("text not recognize")
return
true_flag, question, answers = parse_question_and_answer(keywords)
#questions=question.decode('unicode-escape')
#new_ans=[]
#for ans in answers:
# new_ans.append(ans.decode('unicode-escape'))
print('-' * 72)
print(question)
print('-' * 72)
print("\n".join(answers))
# notice browser
if enable_chrome:
with question_obj.get_lock():
question_obj.value = question
keyboard.press("space")
search_question = pre_process_question(question)
summary = baidu_count(search_question, answers, timeout=timeout)
summary_li = sorted(summary.items(),
key=operator.itemgetter(1),
reverse=True)
data = [("选项", "同比")]
for a, w in summary_li:
data.append((a, w))
table = AsciiTable(data)
print(table.table)
print("*" * 72)
if true_flag:
print("肯定回答(**)： ", summary_li[0][0])
print("否定回答( )： ", summary_li[-1][0])
else:
print("肯定回答( )： ", summary_li[0][0])
print("否定回答(**)： ", summary_li[-1][0])
print("*" * 72)
##############################################################
input_message = question
if len(input_message) > 60:
print(mybot.respond("句子长度过长"))
elif input_message.strip() == '':
print(mybot.respond("无"))
#print(input_message)
message = T.wordSegment(input_message)
# 去标点
#print('word Seg:' + message)
#print('词性：')
words = T.postag(input_message)
if message == 'q':
exit()
else:
response = mybot.respond(message)
#print("=======")
#print(response)
#print("=======")
if response == "":
ans = mybot.respond('找不到答案')
print('Eric：' + ans)
# 百科搜索
elif response[0] == '#':
# 匹配百科
if response.__contains__("searchbaike"):
#print("searchbaike")
#print(response)
res = response.split(':')
# 实体
entity = str(res[1]).replace(" ", "")
# 属性
attr = str(res[2]).replace(" ", "")
#print(entity + '<---->' + attr)
ans = baike.query(entity, attr)
# 如果命中答案
if type(ans) == list:
print('Eric：' + QAT.ptranswer(ans, False))
elif ans.decode('utf-8').__contains__(u'::找不到'):
# 百度摘要+Bing摘要
print("通用搜索")
ans = search_summary.kwquery(input_message)
# 匹配不到模版，通用查询
elif response.__contains__("NoMatchingTemplate"):
#print("NoMatchingTemplate")
ans = search_summary.kwquery(input_message, answers)
if len(ans) == 0:
print('Eric：' + '找不到答案')
elif len(ans) > 1:
print("不确定候选答案")
print('Eric: ')
for a in ans:
print(a)
else:
print('Eric：' + ans[0])
# 匹配模版
else:
print('Eric：' + response)
end = time.time()
print("use {0} 秒".format(end - start))
save_screen(directory=data_directory)
def main():
# 初始化jb分词器
T.jieba_initialize()
# 切换到语料库所在工作目录
mybot_path = './'
os.chdir(mybot_path)
# 加载AIML的规则
mybot = aiml.Kernel()
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] +
"/resources/std-startup.xml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] + "/resources/abc.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] +
"/resources/bot_profile.aiml")
# mybot.learn(os.path.split(os.path.realpath(__file__))[0] + "/resources/general.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] + "/resources/infor.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] + "/resources/main.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] +
"/resources/new07281.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] +
"/resources/salutations.aiml")
mybot.learn(
os.path.split(os.path.realpath(__file__))[0] +
"/resources/virus0727.aiml")
# mybot.learn(os.path.split(os.path.realpath(__file__))[0] + "/resources/zextra_weibao.aiml")
# mybot.learn(os.path.split(os.path.realpath(__file__))[0] + "/resources/bye.aiml")
# mybot.learn(os.path.split(os.path.realpath(__file__))[0] + "/resources/tools.aiml")
# mybot.learn(os.path.split(os.path.realpath(__file__))[0] + "/resources/bad.aiml")
# mybot.learn(os.path.split(os.path.realpath(__file__))[0] + "/resources/funny.aiml")
# mybot.learn(os.path.split(os.path.realpath(__file__))[0] + "/resources/OrdinaryQuestion.aiml")
# mybot.learn(os.path.split(os.path.realpath(__file__))[0] + "/resources/Common conversation.aiml")
# mybot.respond('Load Doc Snake')
#载入百科属性列表
# print '''
# .----------------. .-----------------. .----------------. .----------------. .----------------.
# | .--------------. || .--------------. || .--------------. || .--------------. || .--------------. |
# | | _______ | || | ____ _____ | || | __ | || | ___ ____ | || | _________ | |
# | | / ___ | | || ||_ \|_ _| | || | / \ | || | |_ ||_ _| | || | |_ ___ | | |
# | | | (__ \_| | || | | \ | | | || | / /\ \ | || | | |_/ / | || | | |_ \_| | |
# | | '.___`-. | || | | |\ \| | | || | / /__\ \ | || | | __'. | || | | _| _ | |
# | | |`\____) | | || | _| |_\ |_ | || | _/ / \ \_ | || | _| | \ \_ | || | _| |___/ | | |
# | | |_______.' | || ||_____|\____| | || ||____| |____|| || | |____||____| | || | |_________| | |
# | | | || | | || | | || | | || | | |
# | '--------------' || '--------------' || '--------------' || '--------------' || '--------------' |
# '----------------' '----------------' '----------------' '----------------' '----------------'
# Eric：你好，我是Eric。╭(╯^╰)╮
# '''
print('泰康小康：你好，我是小康。╭(╯^╰)╮')
# 对问题的处理流程
while True:
# 输入这个问题
input_message = raw_input("您想问什么 >> ")
# 对输入问题进行简单的处理：
# 忽略过长（超过60）的问题
# 忽略空问题
if len(input_message) > 60:
print(mybot.respond("句子长度过长"))
continue
elif input_message.strip() == '':
print(mybot.respond("无"))
continue
# 利用Tools工具对问题进行处理
print(input_message)
message = T.wordSegment(input_message)
# 去标点
print('word Seg:' + message)
# print('词性：')
words = T.postag(input_message)
# 退出
if message == 'q':
exit()
# 返回信息的优先级
else:
# 首先是AIML的模板匹配
response = mybot.respond(message)
print("=======")
print(response)
print(len(response.decode('utf8')))
print("=======")
def find_ans(question=''):
global raw
log = '答案来源：'
cnt = 0
input_message = question
if len(input_message) > 60:
return (mybot.respond("句子长度过长"), log)
elif input_message.strip() == '':
return (mybot.respond("无"), log)
# 检索本地知识库得到答案，代码在QA1文件夹
cnt += 1
log += str(cnt) + ':检索本地知识库\n'
old_client = Client()
ans, log_tmp = old_client.qa_find_ans(input_message)
log += log_tmp
if ans != "不知道~":
return (clean_str(ans), log)
log += '本地知识库找不到答案或答案不确定\n'
message = T.wordSegment(input_message)
response = mybot.respond(message)
# log += 'AIML模板返回内容:' + response + '\n'
if response == "":
ans = mybot.respond('不知道~')
return (ans, log)
elif response[0] == '#':
cnt += 1
log += str(cnt) + ':匹配不到问句模板\n'
if response.__contains__("searchbaike"):
res = response.split(':')
entity = str(res[1]).replace(" ","")
attr = str(res[2]).replace(" ","")
cnt += 1
log += str(cnt) + ':匹配到实体属性模板,' + '实体:' + entity + ' 属性:' + attr + '\n'
ans = baike.query(entity, attr)
if type(ans) == list:
cnt += 1
log += str(cnt) + ':来自百科Infobox\n'
return (QAT.ptranswer(ans,False), log)
elif '-找不到' in ans:
cnt += 1
log += str(cnt) + ':百科Infobox查询不到:' + ans + '\n'
cnt += 1
log += str(cnt) + '来自搜索\n'
(ans, tmplog) = search_summary.kwquery(input_message)
log += tmplog
elif response.__contains__("NoMatchingTemplate"):
cnt += 1
log += str(cnt) + ':匹配不到实体关系模板\n'
cnt += 1
log += str(cnt) + ':来自搜索\n'
(ans,tmplog) = search_summary.kwquery(input_message)
log += tmplog
if len(ans) == 0:
cnt += 1
log += str(cnt) + ':未查询到答案\n'
return (mybot.respond('不知道~'), log)
elif len(ans) >1:
cnt += 1
log += str(cnt) + ':返回百度摘要\n'
if raw == False and ('什么是' in question or '是什么' in question):
result = "给你找到几篇新闻："
for a in ans:
result += a + '\n'
return (result, log)
else:
raw = False
question = question.replace("是什么", "").replace("什么是", "")
ans2, log2 = find_ans(question + "是什么")
ans1, log1 = find_ans("什么是" + question)
if "给你找到几篇新闻" not in ans1:
return (ans1, log1)
else:
return (ans2, log2)
else:
return (clean_str(ans[0]), log)
# 直接匹配问句模版
else:
cnt += 1
log += str(cnt) + ':匹配问句模板\n'
return (clean_str(response), log)