|
|
@@ -0,0 +1,71 @@
|
|
|
+"""
|
|
|
+@author: luojunhui
|
|
|
+"""
|
|
|
+import json
|
|
|
+import pandas as pd
|
|
|
+from applications.search import hksp_search
|
|
|
+
|
|
|
+
|
|
|
+def ab_test(title):
|
|
|
+ """
|
|
|
+ :param title:
|
|
|
+ :return:
|
|
|
+ """
|
|
|
+ recall_list = hksp_search(title)
|
|
|
+ if recall_list:
|
|
|
+ best_0 = recall_list[0]
|
|
|
+ print(best_0['title'], best_0['playcnt'], best_0['like'] if best_0.get('like') else 0, best_0['comment'] if best_0.get("comment") else 0)
|
|
|
+ ab_1 = sorted(recall_list, reverse=True, key=lambda x: int(x['playcnt']))[0]
|
|
|
+ ab_2 = sorted(recall_list, reverse=True, key=lambda x: int(x['like'] if x.get('like') else 0))[0]
|
|
|
+ ab_3 = sorted(recall_list, reverse=True, key=lambda x: int(x['comment'] if x.get("comment") else 0))[0]
|
|
|
+ ab_4 = sorted(recall_list, reverse=True, key=lambda x: (int(x['like'] if x.get('like') else 0) / int(x['playcnt'])))[0]
|
|
|
+ # print(ab_1['title'], ab_1['playcnt'], ab_1['like'] if ab_1.get('like') else 0, ab_1['comment'] if ab_1.get("comment") else 0)
|
|
|
+ # print(ab_2['title'], ab_2['playcnt'], ab_2['like'] if ab_2.get('like') else 0, ab_2['comment'] if ab_2.get("comment") else 0)
|
|
|
+ # print(ab_3['title'], ab_3['playcnt'], ab_3['like'] if ab_3.get('like') else 0, ab_3['comment'] if ab_3.get("comment") else 0)
|
|
|
+ # print(ab_4['title'], ab_4['playcnt'], ab_4['like'] if ab_4.get('like') else 0, ab_4['comment'] if ab_4.get("comment") else 0)
|
|
|
+ return [best_0, ab_1, ab_2, ab_3, ab_4]
|
|
|
+ else:
|
|
|
+ return []
|
|
|
+
|
|
|
+
|
|
|
+title_list = [
|
|
|
+ "菲再闯仁爱礁，中国海警船掀了炮衣，好话说尽，1号令立即实施",
|
|
|
+ "乌克兰遭受严重损失！俄罗斯在西方国家大使馆展示摧毁的设备",
|
|
|
+ "中方出重拳了，一句话把以色列挂在火上烤，就看它能否接住",
|
|
|
+ "终于合作！中东国家大团结，以色列要完蛋",
|
|
|
+ "【2】毛远新从监狱刑满释放后，对女儿十分愧疚，回忆往事时他偷偷落泪",
|
|
|
+ "【2】韩国政局要变天，中国一年前的警示应验了，尹锡悦终究自食其果",
|
|
|
+ "金正恩乘专列过江，为何故意绕开中国？原因有两个，值得我们警惕",
|
|
|
+ "中方出手！给巴勒斯坦打去电话，“战狼”已至中东！",
|
|
|
+ "中国不欠犹太人的，华春莹用双语发文，西方该感恩没资历道德绑架"
|
|
|
+]
|
|
|
+
|
|
|
+with open("result.json", encoding="utf-8") as f:
|
|
|
+ title_dict = json.loads(f.read())
|
|
|
+
|
|
|
+ooo = []
|
|
|
+for line in title_list:
|
|
|
+ print(line)
|
|
|
+ c_title = title_dict[line]['c_title']
|
|
|
+ c_keys = "# ".join(title_dict[line]["keys"])
|
|
|
+ result = ab_test(line)
|
|
|
+ if result:
|
|
|
+ for index, item in enumerate(result):
|
|
|
+ temp = [line, c_title, c_keys, "ab_{}".format(index), item['title'], item.get('playcnt', None), item.get('like', None), item.get('comment', None), item['playurl']]
|
|
|
+ ooo.append(temp)
|
|
|
+ else:
|
|
|
+ sub_result = ab_test(title_dict[line]['c_title'])
|
|
|
+ if sub_result:
|
|
|
+ for index, item in enumerate(sub_result):
|
|
|
+ temp = [line, c_title, c_keys, "ab_{}".format(index), item['title'], item.get('playcnt', None), item.get('like', None), item.get('comment', None), item['playurl']]
|
|
|
+ print(temp)
|
|
|
+ ooo.append(temp)
|
|
|
+ else:
|
|
|
+ ss_result = ab_test(title_dict[line]['keys'][0])
|
|
|
+ for index, item in enumerate(ss_result):
|
|
|
+ temp = [line, c_title, c_keys, "ab_{}".format(index), item['title'], item.get('playcnt', None), item.get('like', None), item.get('comment', None), item['playurl']]
|
|
|
+ print(temp)
|
|
|
+ ooo.append(temp)
|
|
|
+
|
|
|
+df = pd.DataFrame(ooo, columns=['article_title', 'kimi_content_summary', 'kimi_content_keys', 'ab_test', 'out_title', 'views', 'like', 'comment', 'videoUrl'])
|
|
|
+df.to_excel("baidu_test.xlsx", index=False)
|
罗俊辉