123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169 |
- import os
- import json
- 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
- from common.mq import MQ
- from common.redis_db import get_data, store_data
- class TFuQiWangRecommend(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_video_list(self):
- base_url = "https://jdfqw.wentingyou.cn/index.php/v111/index/index"
- headers = {
- 'chatkey': 'wx79ef316d416e5da1',
- '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',
- 'content-type': 'application/x-www-form-urlencoded',
- 'xweb_xhr': '1',
- 'vision': '1.1.0',
- 'token': '',
- 'accept': '*/*',
- 'referer': 'https://servicewechat.com/wx79ef316d416e5da1/1/page-frame.html',
- 'accept-language': 'en-US,en;q=0.9'
- }
- page_index = 1
- while True:
- time.sleep(random.randint(1, 10))
- try:
- if self.limit_flag:
- AliyunLogger.logging(
- code="2000",
- platform=self.platform,
- mode=self.mode,
- env=self.env,
- message="本轮已经抓取到足够的数据,自动退出\t{}".format(self.download_cnt),
- )
- return
- else:
- obj = {
- "cid": "",
- "page": 1,
- "is_ads": 1
- }
- params = {
- "parameter": json.dumps(obj)
- }
- response = requests.get(
- url=base_url,
- headers=headers,
- params=params
- )
- video_list = response.json()['data']['list']
- if video_list:
- for index, video_obj in enumerate(video_list, 1):
- try:
- if video_obj.get("title"):
- 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,
- data=video_obj,
- message="抓取第{}条的时候出现问题, 报错信息是{}".format(index, e),
- )
- page_index += 1
- else:
- AliyunLogger.logging(
- code="2000",
- platform=self.platform,
- mode=self.mode,
- env=self.env,
- message="已经抓完了,自动退出"
- )
- return
- except Exception as e:
- AliyunLogger.logging(
- code="3000",
- platform=self.platform,
- mode=self.mode,
- env=self.env,
- message="抓取第{}页时候出现错误, 报错信息是{}".format(page_index + 1, e),
- )
- def process_video_obj(self, video_obj):
- trace_id = self.platform + str(uuid.uuid1())
- 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_obj["nid"])
- item.add_video_info("video_title", video_obj["title"])
- item.add_video_info("publish_time_stamp", int(video_obj['update_time']))
- item.add_video_info("video_url", video_obj["video_url"])
- item.add_video_info("cover_url", video_obj["video_cover"])
- item.add_video_info("out_video_id", video_obj["nid"])
- 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()
- value = get_data(self.platform, video_obj["nid"])
- if int(value) == 1:
- AliyunLogger.logging(
- code="2004",
- platform=self.platform,
- mode=self.mode,
- env=self.env,
- message="redis重复视频")
- 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)
- store_data(self.platform, video_obj["nid"], False, 12)
- time.sleep(60 * random.randint(1, 5))
- AliyunLogger.logging(
- code="1002",
- platform=self.platform,
- mode=self.mode,
- env=self.env,
- message="成功发送至 ETL",
- data=mq_obj,
- )
- if self.download_cnt >= int(
- self.rule_dict.get("videos_cnt", {}).get("min", 200)
- ):
- self.limit_flag = True
- if __name__ == '__main__':
- S = TFuQiWangRecommend(
- platform="jingdianfuqiwang",
- mode="recommend",
- env="dev",
- rule_dict={},
- user_list=[{'nick_name': "Ivring", 'uid': "1997"}, {'nick_name': "paul", 'uid': "1998"}]
- )
- S.get_video_list()
|