jjzffzjx_recommend.py 6.3 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153
  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. from common.redirect_url import get_redirect_url
  12. class JJZFFZJXRecommend(object):
  13. def __init__(self, platform, mode, rule_dict, user_list, env):
  14. self.platform = platform
  15. self.mode = mode
  16. self.rule_dict = rule_dict
  17. self.user_list = user_list
  18. self.env = env
  19. self.download_cnt = 0
  20. self.limit_flag = False
  21. self.mq = MQ(topic_name="topic_crawler_etl_" + self.env)
  22. def get_video_list(self, token):
  23. """
  24. 推荐流大约110条数据,目前暂时不会更新
  25. """
  26. headers = {
  27. 'Host': 'api.yaxun.top',
  28. 'ik': 'b326b5062b2f0e69046810717534cb09',
  29. 'xweb_xhr': '1',
  30. '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',
  31. 'token': token,
  32. 'Content-Type': 'application/json',
  33. 'Accept': '*/*',
  34. 'Sec-Fetch-Site': 'cross-site',
  35. 'Sec-Fetch-Mode': 'cors',
  36. 'Sec-Fetch-Dest': 'empty',
  37. 'Referer': 'https://servicewechat.com/wx479b517622b7875e/4/page-frame.html',
  38. 'Accept-Language': 'en-US,en;q=0.9'
  39. }
  40. page = 1
  41. while True:
  42. time.sleep(random.randint(1, 10))
  43. try:
  44. if self.limit_flag:
  45. message = "本轮已经抓取到足够的数据,自动退出\t{}".format(self.download_cnt)
  46. AliyunLogger.logging(
  47. code="2000",
  48. platform=self.platform,
  49. env=self.env,
  50. mode=self.mode,
  51. message=message
  52. )
  53. return
  54. else:
  55. params = {
  56. "s": "mobile/Video/getList",
  57. "cid": 1,
  58. "page": str(page),
  59. "api_version": 4,
  60. "appid": "wx479b517622b7875e",
  61. "version": "1.9.1",
  62. "env_version": "release",
  63. "scene": 1089
  64. }
  65. response = requests.get('https://api.yaxun.top/index.php', headers=headers, params=params, proxies=tunnel_proxies())
  66. video_list = response.json()['data']['list']
  67. page += 1
  68. if video_list:
  69. for index, video_obj in enumerate(video_list, 1):
  70. try:
  71. AliyunLogger.logging(
  72. code="1001",
  73. platform=self.platform,
  74. mode=self.mode,
  75. env=self.env,
  76. message="扫描到一条视频",
  77. data=video_obj
  78. )
  79. self.process_video_obj(video_obj)
  80. except Exception as e:
  81. AliyunLogger.logging(
  82. code="3000",
  83. platform=self.platform,
  84. mode=self.mode,
  85. env=self.env,
  86. message="抓取第{}条的时候出现问题, 报错信息是{}".format(index, e)
  87. )
  88. else:
  89. AliyunLogger.logging(
  90. code="2000",
  91. platform=self.platform,
  92. mode=self.mode,
  93. env=self.env,
  94. message="视频已经抓完",
  95. )
  96. return
  97. except Exception as e:
  98. AliyunLogger.logging(
  99. code="3000",
  100. platform=self.platform,
  101. mode=self.mode,
  102. env=self.env,
  103. message="抓取推荐页的时候出现错误, 报错信息是{}".format(e)
  104. )
  105. def process_video_obj(self, video_obj):
  106. trace_id = self.platform + str(uuid.uuid1())
  107. our_user = random.choice(self.user_list)
  108. ori_video_url = video_obj['video_url']
  109. new_video_url = get_redirect_url(ori_video_url)
  110. item = VideoItem()
  111. item.add_video_info("user_id", our_user["uid"])
  112. item.add_video_info("user_name", our_user["nick_name"])
  113. item.add_video_info("video_id", video_obj["id"])
  114. item.add_video_info("video_title", video_obj["title"])
  115. item.add_video_info("video_url", new_video_url)
  116. item.add_video_info("cover_url", video_obj["video_cover"])
  117. item.add_video_info("play_cnt", video_obj['visited'])
  118. item.add_video_info("share_cnt", video_obj['shared'])
  119. item.add_video_info("out_video_id", video_obj["id"])
  120. item.add_video_info("platform", self.platform)
  121. item.add_video_info("strategy", self.mode)
  122. item.add_video_info("session", "{}-{}".format(self.platform, int(time.time())))
  123. mq_obj = item.produce_item()
  124. pipeline = PiaoQuanPipeline(
  125. platform=self.platform,
  126. mode=self.mode,
  127. rule_dict=self.rule_dict,
  128. env=self.env,
  129. item=mq_obj,
  130. trace_id=trace_id,
  131. )
  132. if pipeline.process_item():
  133. self.download_cnt += 1
  134. self.mq.send_msg(mq_obj)
  135. AliyunLogger.logging(
  136. code="1002",
  137. platform=self.platform,
  138. mode=self.mode,
  139. env=self.env,
  140. message="成功发送至 ETL",
  141. data=mq_obj,
  142. trace_id=trace_id
  143. )
  144. if self.download_cnt >= int(
  145. self.rule_dict.get("videos_cnt", {}).get("min", 200)
  146. ):
  147. self.limit_flag = True