增加吉祥幸福渠道

application/config/topic_group_queue.py

@@ -11,7 +11,8 @@ class TopicGroup(object):
             ('hhxxzfdn', 'recommend', 'huanhuanxixizhufudaonew'),
             ('bqzf', 'recommend', 'boqingzhufu'),
             ('syzf', 'recommend', 'sharkzhufu'),
-            ("xsdd", 'recommend', 'xishiduoduo')
+            ("xsdd", 'recommend', 'xishiduoduo'),
+            ("jxxf", 'recommend', 'jixiangxingfu')
         ]
 
     def produce(self):

application/pipeline/pipeline.py

@@ -128,6 +128,8 @@ class PiaoQuanPipeline(object):
         """
         if self.platform == "zhuwanwufusunew" and self.mode == "recommend":
             return True
+        if self.platform == "jixiangxingfu" and self.mode == "recommend":
+            return True
         out_id = self.item["out_video_id"]
         sql = f""" select 1 from crawler_video where platform = "{self.platform}" and out_video_id="{out_id}"; """
         repeat_video = self.mysql.select(sql=sql)

spider/crawler_online/jixiangxingfu.py

@@ -0,0 +1,151 @@
+import os
+import random
+import sys
+import time
+import uuid
+from datetime import datetime
+
+import requests
+
+from application.common import Feishu
+
+sys.path.append(os.getcwd())
+
+from application.items import VideoItem
+from application.pipeline import PiaoQuanPipeline
+from application.common.messageQueue import MQ
+from application.common.proxies import tunnel_proxies
+from application.common.log import AliyunLogger
+from application.common.mysql import MysqlHelper
+
+
+
+class JXXFRecommend(object):
+    """
+    吉祥幸福-欢快吉祥早安祝福
+    """
+
+    def __init__(self, platform, mode, rule_dict, user_list, env="prod"):
+        self.limit_flag = False
+        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.expire_flag = False
+        self.aliyun_log = AliyunLogger(mode=self.mode, platform=self.platform)
+        self.mysql = MysqlHelper(mode=self.mode, platform=self)
+
+    def get_recommend_list(self):
+        """
+        获取推荐页视频
+        """
+        headers = {
+            'Host': 'api.huanqiwl.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/wx6692a24ad2a88bfb/3/page-frame.html'
+        }
+        for i in range(10):
+            time.sleep(random.randint(1, 10))
+            url = f"https://api.huanqiwl.top/index.php?s=mobile/Video/getList&cid=1&page={i}&api_version=4&appid=wx6692a24ad2a88bfb&version=1.9.5&env_version=release&scene=1053"
+            payload = {}
+            response = requests.request("GET", url, headers=headers, data=payload, proxies=tunnel_proxies())
+            for index, video_obj in enumerate(response.json()['data']['list'], 1):
+                try:
+                    self.aliyun_log.logging(
+                        code="1001", message="扫描到一条视频", data=video_obj
+                    )
+                    self.process_video_obj(video_obj)
+                except Exception as e:
+                    self.aliyun_log.logging(
+                        code="3000",
+                        message="抓取单条视频失败, 该视频位于第{}页第{}条报错原因是{}".format(
+                            i, index, e
+                        ),
+                    )
+                except Exception as e:
+                    self.aliyun_log.logging(
+                        code="3000", message="抓取第{}页的时候失败, 报错原因是{}".format(i, e)
+                    )
+            if self.limit_flag:
+                return
+            time.sleep(random.randint(5, 10))
+
+    def process_video_obj(self, video_obj):
+        """
+        处理视频
+        :param video_obj:
+        """
+        time.sleep(random.randint(3, 8))
+        trace_id = self.platform + str(uuid.uuid1())
+        our_user = random.choice(self.user_list)
+        item = VideoItem()
+        item.add_video_info("video_id", video_obj["id"])
+        item.add_video_info("video_title", video_obj["title"])
+        item.add_video_info("play_cnt", 0)
+        item.add_video_info("publish_time_stamp", int(time.time()))
+        item.add_video_info("out_user_id", video_obj["id"])
+        item.add_video_info("cover_url", video_obj["images"])
+        item.add_video_info("like_cnt", 0)
+        item.add_video_info("video_url", video_obj["video_url"])
+        item.add_video_info("out_video_id", video_obj["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())))
+        item.add_video_info("user_id", our_user["uid"])
+        item.add_video_info("user_name", our_user["nick_name"])
+        # 获取当前时间
+        current_time = datetime.now()
+        formatted_time = current_time.strftime("%Y-%m-%d %H:%M:%S")
+        values = [[
+            video_obj["id"],
+            formatted_time,
+            video_obj["title"],
+            video_obj["images"],
+            video_obj["video_url"]
+        ]]
+        Feishu.insert_columns(self.platform, 'jixiangxingfu', "L0KXHh", "ROWS", 1, 2)
+        time.sleep(0.5)
+        Feishu.update_values(self.platform, 'jixiangxingfu', "L0KXHh", "A2:Z2", values)
+        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)
+            self.aliyun_log.logging(code="1002", 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
+
+
+
+
+    def run(self):
+        """
+        先抓取推荐列表的视频， 等待 2 分钟后抓取 detail 页面，等待 5 分钟后，抓取账号视频
+        """
+        self.get_recommend_list()
+
+
+if __name__ == '__main__':
+    J = JXXFRecommend(
+        platform="jixiangxingfu",
+        mode="recommend",
+        rule_dict={},
+        user_list=[{'uid': "123456", 'nick_name': "xiaoxiao"}],
+
+    )
+    J.get_recommend_list()