diff --git a/smoothnlp/unittest/test6_Query_analysis.py b/smoothnlp/unittest/test6_Query_analysis.py new file mode 100644 index 0000000..d908b32 --- /dev/null +++ b/smoothnlp/unittest/test6_Query_analysis.py @@ -0,0 +1,165 @@ +import sys +import unittest +import requests +from smoothnlp import config,nlp + +class MainTest_Query(unittest.TestCase): + def setUp(self): + global failed_counter + global number_test_case + failed_counter = 0 + number_test_case = 0 + + def GET_Query_msg(self,input_): + API_Key = "#############请在这里输入您的API,具体获取方式请参考（https://github.com/smoothnlp/SmoothNLP/tree/master/tutorials/Query%E8%A7%A3%E6%9E%90）说明文档#############" + #格式参考：API_Key = "F3AFbm8YOzl2tTnPzPk884GCL5rNDiME"(仅为事例，该API不可用) + config.setNLP_Path("/pro/nlp/query") ## 使用pro接口 + config.setApiKey(API_Key) ## 设置自己的apikey + response_data = nlp.parseq(input_) + text = input_ + return response_data + + def Error_Message(self,input_text,result_word_1,result_word_2,want_word_1,want_word_2): + global failed_counter + global number_test_case + if result_word_1 != want_word_1: + failed_counter += 1 + print(input_text+'：error:!!!!!【result_word_1: '+str(result_word_1)+' != want_word_1: '+str(want_word_1)+'】!!!!!') + number_test_case += 1 + if result_word_2 != want_word_2: + print(input_text+'：error:!!!!!【result_word_2: '+str(result_word_2)+' != want_word_2: '+str(want_word_2)+'】!!!!!') + else: + number_test_case += 1 + + def entityQuery(self,input_text,want_word_1,want_word_2): + """测试entityQuery：""" + global failed_counter + global number_test_case + entity = [] + response_data = self.GET_Query_msg(input_text) + try: + for i in range(len(response_data['entities'])): + entity.append(response_data['entities'][i]['entity']) + if want_word_1 not in entity: + result_word_1 = entity + if want_word_2 in entity: + failed_counter += 1 + print(input_text+'：error:!!!!!【result_word_1: '+str(result_word_1)+' != want_word_1: '+str(want_word_1)+'】!!!!!') + if want_word_2 not in entity: + result_word_2 = entity + if want_word_1 in entity: + failed_counter += 1 + print(input_text+'：error:!!!!!【result_word_2: '+str(result_word_2)+' != want_word_2: '+str(want_word_2)+'】!!!!!') + if want_word_1 not in entity: + if want_word_2 not in entity: + failed_counter += 1 + number_test_case += 1 + except: + failed_counter += 1 + number_test_case += 1 + print("This text does not contain some attributes =",input_text) + + def describeQuery(self,input_text,want_word_1,want_word_2): + """测试describeQuery：""" + global failed_counter + global number_test_case + response_data = self.GET_Query_msg(input_text) + try: + word_1 = response_data['relationships'][0]['entitiy_pair'][0]['describe'] + word_2 = response_data['relationships'][0]['entitiy_pair'][1]['describe'] + self.Error_Message(input_text,word_1,word_2,want_word_1,want_word_2) + except: + failed_counter += 1 + number_test_case += 1 + print("This text does not contain some attributes =",input_text) + + def typeQuery(self,input_text,want_word_1): + """测试typeQuery：""" + global failed_counter + global number_test_case + response_data = self.GET_Query_msg(input_text) + try: + result_word_1 = response_data['relationships'][0]['type'] + if result_word_1 != want_word_1: + failed_counter += 1 + print(input_text+'：error:!!!!!【result_word_1: '+str(result_word_1)+' != want_word_1: '+str(want_word_1)+'】!!!!!') + number_test_case += 1 + else: + number_test_case += 1 + except: + failed_counter += 1 + number_test_case += 1 + print("This text does not contain some attributes =",input_text) + +###########################实体抽取 + def test1_Entity_extract(self): + """抽取文本中的对象实体（Entity + Entity）""" + print('\n######',sys._getframe().f_code.co_name,'测试结果：######') + self.entityQuery("陈冠希和新女友",'陈冠希','女友') + self.entityQuery("如何评价王瑞恩与毕志飞的直播辩论？",'王瑞恩','毕志飞') + self.entityQuery("奚梦瑶晒与老公何猷君合照双双素颜出镜超接地",'奚梦瑶','何猷君') + self.entityQuery("冯小刚与郑爽聚餐暖心安慰希望你的心是暖的",'冯小刚','郑爽') + self.entityQuery("欣赏乐曲梁山伯与祝英台",'梁山伯','祝英台') + self.entityQuery("宋仲基推倒与宋慧乔新房",'宋仲基','宋慧乔') + + self.entityQuery("尤文愿与C罗续约至39岁",'尤文','C罗') + self.entityQuery("泰勒和侃爷事件",'泰勒','侃爷') + self.entityQuery("特朗普宣布华盛顿州为重大灾区",'特朗普','华盛顿州') + + self.entityQuery("腾讯和任天堂合作的背后是什么",'腾讯','任天堂') + self.entityQuery("腾讯云IoT与意法半导体宣布在物联网方面展开合作",'腾讯云','意法半导体') + self.entityQuery("青海泰丰先行与北汽集团等5家单位签署战略投资协议",'青海泰丰','北汽集团') + accuracy_rate = 1-failed_counter/number_test_case + self.assertGreaterEqual(accuracy_rate,0.2) + print(accuracy_rate) + print("本功能共测试了%s条"%number_test_case) +###########################对象描述 + def test2_describe_extract(self): + """抽取文本中的Entity 与 修饰;""" + print('\n######',sys._getframe().f_code.co_name,'测试结果：######') + self.describeQuery("疫情下影视互联网新机会'",'','新') + self.describeQuery("久历职场的老油条",'','老') + self.describeQuery("陈冠希和前女友",'','前') + + self.describeQuery("网曝baby和儿子的可爱对话",'','可爱') + self.describeQuery("无论是即将迈出校门象牙塔的稚嫩社会新人",'','稚嫩') + + self.describeQuery("钟南山的这张车票",'','这张') + self.describeQuery("客厅和阳台应该分开",'','应该') + self.describeQuery("教育形成了我们特有的思维",'','特有的') + self.describeQuery("南非因疫情取消与日本国奥比赛",'','取消') + accuracy_rate = 1-failed_counter/number_test_case + self.assertGreaterEqual(accuracy_rate,0.2) + print(accuracy_rate) + print("本功能共测试了%s条"%number_test_case) +###########################对象之间的关系 + def test3_Entity_extract(self): + """抽取文本中entity对象之间的关系""" + print('\n######',sys._getframe().f_code.co_name,'测试结果：######') + self.typeQuery("陈冠希和新女友",'conjunction') + self.typeQuery("恋爱和婚姻的心理学知识",'conjunction') + self.typeQuery("新媒体和数字媒体的区别",'conjunction') + self.typeQuery("家长和老师应该是最好的搭档",'conjunction') + self.typeQuery("时彩公司和农业基地签约仪式",'conjunction') + self.typeQuery("音乐剧和歌剧之间的差别是什么",'conjunction') + self.typeQuery("这与金钱和权力无关只是我和你在一起",'conjunction') + + self.typeQuery("陈冠希的新女友",'associate') + self.typeQuery("了解新冠肺炎的传播途径",'associate') + self.typeQuery("那人生的意义到底是什么呢",'associate') + self.typeQuery("关于此次新型冠状病毒的认识",'associate') + self.typeQuery("新一线城市的经济将面临怎样的挑战",'associate') + self.typeQuery("2个新冠肺炎逝者的最后几天，看完真的很想哭",'associate') + + accuracy_rate = 1-failed_counter/number_test_case + self.assertGreaterEqual(accuracy_rate,0.2) + print(accuracy_rate) + print("本功能共测试了%s条"%number_test_case) + print("已结束") + + +#若在jupyter-notebook中run上述代码请将下行取消注释 +#unittest.main(argv=['ignored', '-v'], exit=False) + +if __name__ == '__main__': + unittest.main()