123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164 |
- import json
- import os
- import random
- import sys
- import time
- import uuid
- import requests
- sys.path.append(os.getcwd())
- from common.video_item import VideoItem
- from common import PiaoQuanPipeline, AliyunLogger, tunnel_proxies, get_redirect_url
- from common.mq import MQ
- class Zfdgnmxqyyspcheduling(object):
- def __init__(self, platform, mode, rule_dict, user_list, env):
- self.platform = platform
- self.mode = mode
- self.rule_dict = rule_dict
- self.user_list = user_list
- self.env = env
- self.download_cnt = 0
- self.mq = MQ(topic_name="topic_crawler_etl_" + self.env)
- self.limit_flag = False
- def get_videoList(self):
- headers = {
- 'Host': 'api.xiahong.top',
- 'Content-Type': 'application/json',
- 'Accept-Language': 'zh-cn',
- 'Accept': '*/*',
- 'User-Agent': 'Mozilla/5.0 (iPhone; CPU iPhone OS 11_3 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Mobile/15E217 MicroMessenger/6.8.0(0x16080000) NetType/WIFI Language/en Branch/Br_trunk MiniProgramEnv/Mac',
- 'Referer': 'https://servicewechat.com/wxd0e272b032f12c4b/2/page-frame.html',
- 'token': 'eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpYXQiOjE3MDI1MjQyNzcsIm5iZiI6MTcwMjUyNDI3NywiZXhwIjoxNzAyNTMxNDc3LCJkYXRhIjp7InVzZXJfaWQiOjI0NDc3MjQ5N319.jFp2L-QKLudo0sOtwfyk8noIwn8e54a7OQUX3GSn_nQ',
- 'ik': 'b326b5062b2f0e69046810717534cb09'
- }
- page = 1
- while True:
- time.sleep(random.randint(1, 10))
- try:
- if self.limit_flag:
- message = "本轮已经抓取到足够的数据,自动退出\t{}".format(self.download_cnt)
- AliyunLogger.logging(
- code="2000",
- platform=self.platform,
- env=self.env,
- mode=self.mode,
- message=message
- )
- return
- else:
- url = "https://api.xiahong.top/index.php?s=mobile/Home/getHomeList&page={}&appid=wxd0e272b032f12c4b&version=1.9.1&env_version=release&scene=1053".format(
- page)
- response = requests.post(url, headers=headers, proxies=tunnel_proxies())
- video_list = response.json()['data']['list']
- page += 1
- if video_list:
- for index, video_obj in enumerate(video_list, 1):
- try:
- AliyunLogger.logging(
- code="1001",
- platform=self.platform,
- mode=self.mode,
- env=self.env,
- message="扫描到一条视频",
- data=video_obj
- )
- self.process_video_obj(video_obj)
- except Exception as e:
- AliyunLogger.logging(
- code="3000",
- platform=self.platform,
- mode=self.mode,
- env=self.env,
- message="抓取第{}条的时候出现问题, 报错信息是{}".format(index, e)
- )
- else:
- AliyunLogger.logging(
- code="2000",
- platform=self.platform,
- mode=self.mode,
- env=self.env,
- message="视频已经抓完",
- )
- return
- except Exception as e:
- AliyunLogger.logging(
- code="3000",
- platform=self.platform,
- mode=self.mode,
- env=self.env,
- message="抓取推荐页的时候出现错误, 报错信息是{}".format(e)
- )
- def process_video_obj(self, video_obj):
- trace_id = self.platform + str(uuid.uuid1())
- video_id = video_obj['id']
- url = "https://api.xiahong.top/index.php?s=mobile/Video/getVideoInfo&vid={}&appid=wxd0e272b032f12c4b&version=1.9.1&scene=1053".format(
- video_id)
- headers = {
- 'Host': 'api.xiahong.top',
- 'Content-Type': 'application/json',
- 'Accept-Language': 'zh-cn',
- 'Accept': '*/*',
- 'User-Agent': 'Mozilla/5.0 (iPhone; CPU iPhone OS 11_3 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Mobile/15E217 MicroMessenger/6.8.0(0x16080000) NetType/WIFI Language/en Branch/Br_trunk MiniProgramEnv/Mac',
- 'Referer': 'https://servicewechat.com/wxd0e272b032f12c4b/2/page-frame.html',
- 'token': 'eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpYXQiOjE3MDI1MjQyNzcsIm5iZiI6MTcwMjUyNDI3NywiZXhwIjoxNzAyNTMxNDc3LCJkYXRhIjp7InVzZXJfaWQiOjI0NDc3MjQ5N319.jFp2L-QKLudo0sOtwfyk8noIwn8e54a7OQUX3GSn_nQ',
- 'ik': 'b326b5062b2f0e69046810717534cb09'
- }
- response = requests.post(url, headers=headers, proxies=tunnel_proxies())
- video_info = response.json()["data"]["video_info"]
- video_url = video_info['video_url']
- video_url = get_redirect_url(video_url)
- our_user = random.choice(self.user_list)
- item = VideoItem()
- item.add_video_info("user_id", our_user["uid"])
- item.add_video_info("user_name", our_user["nick_name"])
- item.add_video_info("video_id", video_id)
- item.add_video_info("video_title", video_info["title"])
- item.add_video_info("video_url", video_url)
- item.add_video_info("cover_url", video_obj["images"])
- item.add_video_info("avatar_url", video_obj["images"])
- item.add_video_info("out_video_id", video_id)
- item.add_video_info("platform", self.platform)
- item.add_video_info("strategy", self.mode)
- item.add_video_info("session", "{}-{}".format(self.platform, int(time.time())))
- mq_obj = item.produce_item()
- pipeline = PiaoQuanPipeline(
- platform=self.platform,
- mode=self.mode,
- rule_dict=self.rule_dict,
- env=self.env,
- item=mq_obj,
- trace_id=trace_id,
- )
- if pipeline.process_item():
- self.download_cnt += 1
- self.mq.send_msg(mq_obj)
- AliyunLogger.logging(
- code="1002",
- platform=self.platform,
- mode=self.mode,
- env=self.env,
- message="成功发送至 ETL",
- data=mq_obj,
- trace_id=trace_id
- )
- if self.download_cnt >= int(
- self.rule_dict.get("videos_cnt", {}).get("min", 200)
- ):
- self.limit_flag = True
- if __name__ == "__main__":
- ZL = Zfdgnmxqyyspcheduling(
- platform="Jxzfwncdhy",
- mode="recommend",
- env="dev",
- rule_dict={},
- user_list=[{'nick_name': "Ivring", 'uid': "1997"}, {'nick_name': "paul", 'uid': "1998"}]
- )
- ZL.get_videoList()
|