zhufuyiqifengfa_recommend.py 6.1 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150
  1. import os
  2. import random
  3. import sys
  4. import time
  5. import uuid
  6. import requests
  7. sys.path.append(os.getcwd())
  8. from common.video_item import VideoItem
  9. from common import PiaoQuanPipeline, AliyunLogger, tunnel_proxies
  10. from common.mq import MQ
  11. class YiQiFengFaRecommend(object):
  12. def __init__(self, platform, mode, rule_dict, user_list, env):
  13. self.platform = platform
  14. self.mode = mode
  15. self.rule_dict = rule_dict
  16. self.user_list = user_list
  17. self.env = env
  18. self.download_cnt = 0
  19. self.limit_flag = False
  20. self.mq = MQ(topic_name="topic_crawler_etl_" + self.env)
  21. def get_video_list(self, token):
  22. """
  23. 推荐流大约110条数据,目前暂时不会更新
  24. """
  25. headers = {
  26. 'Host': 'api.xiahong.top',
  27. 'ik': 'b326b5062b2f0e69046810717534cb09',
  28. 'xweb_xhr': '1',
  29. 'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/107.0.0.0 Safari/537.36 MicroMessenger/6.8.0(0x16080000) NetType/WIFI MiniProgramEnv/Mac MacWechat/WMPF MacWechat/3.8.5(0x13080510)XWEB/1100',
  30. 'token': token,
  31. 'Content-Type': 'application/json',
  32. 'Accept': '*/*',
  33. 'Sec-Fetch-Site': 'cross-site',
  34. 'Sec-Fetch-Mode': 'cors',
  35. 'Sec-Fetch-Dest': 'empty',
  36. 'Referer': 'https://servicewechat.com/wxe5389a260a0a4ee2/2/page-frame.html',
  37. 'Accept-Language': 'en-US,en;q=0.9',
  38. }
  39. page = 1
  40. while True:
  41. time.sleep(random.randint(1, 10))
  42. try:
  43. if self.limit_flag:
  44. message = "本轮已经抓取到足够的数据,自动退出\t{}".format(self.download_cnt)
  45. AliyunLogger.logging(
  46. code="2000",
  47. platform=self.platform,
  48. env=self.env,
  49. mode=self.mode,
  50. message=message
  51. )
  52. return
  53. else:
  54. params = {
  55. 's': 'mobile/Video/getList',
  56. 'cid': '1',
  57. 'page': str(page),
  58. 'api_version': '4',
  59. 'appid': 'wxe5389a260a0a4ee2',
  60. 'version': '1.9.1',
  61. 'env_version': 'release',
  62. 'scene': '1089'
  63. }
  64. response = requests.get('https://api.xiahong.top/index.php', headers=headers, params=params, proxies=tunnel_proxies())
  65. video_list = response.json()['data']['list']
  66. page += 1
  67. if video_list:
  68. for index, video_obj in enumerate(video_list, 1):
  69. try:
  70. AliyunLogger.logging(
  71. code="1001",
  72. platform=self.platform,
  73. mode=self.mode,
  74. env=self.env,
  75. message="扫描到一条视频",
  76. data=video_obj
  77. )
  78. self.process_video_obj(video_obj)
  79. except Exception as e:
  80. AliyunLogger.logging(
  81. code="3000",
  82. platform=self.platform,
  83. mode=self.mode,
  84. env=self.env,
  85. message="抓取第{}条的时候出现问题, 报错信息是{}".format(index, e)
  86. )
  87. else:
  88. AliyunLogger.logging(
  89. code="2000",
  90. platform=self.platform,
  91. mode=self.mode,
  92. env=self.env,
  93. message="视频已经抓完",
  94. )
  95. return
  96. except Exception as e:
  97. AliyunLogger.logging(
  98. code="3000",
  99. platform=self.platform,
  100. mode=self.mode,
  101. env=self.env,
  102. message="抓取推荐页的时候出现错误, 报错信息是{}".format(e)
  103. )
  104. def process_video_obj(self, video_obj):
  105. trace_id = self.platform + str(uuid.uuid1())
  106. our_user = random.choice(self.user_list)
  107. item = VideoItem()
  108. item.add_video_info("user_id", our_user["uid"])
  109. item.add_video_info("user_name", our_user["nick_name"])
  110. item.add_video_info("video_id", video_obj["id"])
  111. item.add_video_info("video_title", video_obj["title"])
  112. item.add_video_info("video_url", video_obj['video_url'])
  113. item.add_video_info("cover_url", video_obj["video_cover"])
  114. item.add_video_info("play_cnt", video_obj['visited'])
  115. item.add_video_info("share_cnt", video_obj['shared'])
  116. item.add_video_info("out_video_id", video_obj["id"])
  117. item.add_video_info("platform", self.platform)
  118. item.add_video_info("strategy", self.mode)
  119. item.add_video_info("session", "{}-{}".format(self.platform, int(time.time())))
  120. mq_obj = item.produce_item()
  121. pipeline = PiaoQuanPipeline(
  122. platform=self.platform,
  123. mode=self.mode,
  124. rule_dict=self.rule_dict,
  125. env=self.env,
  126. item=mq_obj,
  127. trace_id=trace_id,
  128. )
  129. if pipeline.process_item():
  130. self.download_cnt += 1
  131. self.mq.send_msg(mq_obj)
  132. AliyunLogger.logging(
  133. code="1002",
  134. platform=self.platform,
  135. mode=self.mode,
  136. env=self.env,
  137. message="成功发送至 ETL",
  138. data=mq_obj,
  139. trace_id=trace_id
  140. )
  141. if self.download_cnt >= int(
  142. self.rule_dict.get("videos_cnt", {}).get("min", 200)
  143. ):
  144. self.limit_flag = True