crawler_zhihu/main/zhihu_follow.py

# -*- coding: utf-8 -*-
# @Author: wangkun
# @Time: 2022/10/31
import os
import sys
import time
import requests
import urllib3

sys.path.append(os.getcwd())
from main.common import Common
from main.feishu_lib import Feishu
from main.zhihu_follow_publish import Publish

proxies = {'http': None, 'https': None}


class ZhihuFollow:
    offset = 0

    @classmethod
    def get_users_from_feishu(cls, log_type):
        try:
            user_sheet = Feishu.get_values_batch(log_type, 'zhihu', '4NTla6')
            user_dict = {}
            for i in range(1, len(user_sheet)):
                user_name = user_sheet[i][0]
                url_token = user_sheet[i][1]
                referer = user_sheet[i][2]
                our_id = user_sheet[i][3]
                if user_name is None or url_token is None or our_id is None:
                    pass
                else:
                    user_dict[user_name] = str(url_token) + ',' + str(referer) + ',' + str(our_id)
            return user_dict
        except Exception as e:
            Common.logger(log_type).error('get_users_from_feishu异常:{}\n', e)

    @classmethod
    def get_follow_feeds(cls, log_type, url_token, referer, our_uid, env):
        while True:
            try:
                # url = "https://www.zhihu.com//api//v4//members//" + str(url_token) + "//zvideos?"
                url = f"https://www.zhihu.com//api//v4//members//{url_token}//zvideos?offset={cls.offset}&limit=20&similar_aggregation=true&include=similar_zvideo%2Ccreation_relationship%2Creaction_instruction"
                payload = {}
                headers = {
                    'authority': 'www.zhihu.com',
                    'accept': '*/*',
                    'accept-language': 'zh-CN,zh;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6',
                    'cache-control': 'no-cache',
                    'cookie': '_zap=246102fb-af66-40c3-a5a5-9901921d5a71; d_c0=AHCWVw5U5hWPTqifPR-jYwskMnmcUFEgHzQ=|1669014326; q_c1=40c865e7cbed4099b5d090229d3096f5|1669983925000|1669983925000; _xsrf=05151c3d-2d05-47fe-98bc-7b01dae731ba; Hm_lvt_98beee57fd2ef70ccdd5ca52b9740c49=1675166229; arialoadData=false; z_c0=2|1:0|10:1675650146|4:z_c0|80:MS4xVFdsTlB3QUFBQUFtQUFBQVlBSlZUVDFJeG1TWWVaZEJSVWdsSWdRalloaWlGaVlqYlFrVmpRPT0=|c0945918804e3c623699052665e50a2452bceb732d25e544df0ca9419e50fa6d; SESSIONID=AgocnyI3witm93R7LqTR2y59rcyJQ9p0QONL8jD8laf; JOID=V1sUA0OVH-cf-lXUM5BktSnhBDki0nqJdJkqnwHvcJFzy2uABvN4DnnxUdc6mEsEpn2HejuruTnxTM_CAaxxrJQ=; osd=UVodBEiTHu4Y8VPVOpdvsyjoAzIk03OOf58rlgbkdpB6zGCGB_p_BX_wWNAxnkoNoXaBezKssj_wRcjJB614q58=; tst=v; Hm_lpvt_98beee57fd2ef70ccdd5ca52b9740c49=1675650941; unlock_ticket=AGBWwkpQuhUmAAAAYAJVTYVu4GP28ZqvlQjQejVT8YHwQEuzKp3jPQ==; KLBRSID=b5ffb4aa1a842930a6f64d0a8f93e9bf|1675650942|1675650143; KLBRSID=b5ffb4aa1a842930a6f64d0a8f93e9bf|1675651010|1675650143',
                    'pragma': 'no-cache',
                    'referer': f'https://www.zhihu.com/people/{url_token}/zvideos',
                    'sec-ch-ua': '"Not_A Brand";v="99", "Microsoft Edge";v="109", "Chromium";v="109"',
                    'sec-ch-ua-mobile': '?0',
                    'sec-ch-ua-platform': '"macOS"',
                    'sec-fetch-dest': 'empty',
                    'sec-fetch-mode': 'cors',
                    'sec-fetch-site': 'same-origin',
                    'user-agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/109.0.0.0 Safari/537.36 Edg/109.0.1518.52',
                    'x-ab-param': '',
                    'x-ab-pb': 'CmQIABsAPwBHALQAaQFqAXQBOwLMAtcC2AK3A9YEEQVRBYsFjAWeBTAGMQbrBicHdAh2CHkIPwlgCfQJBApJCmUKawq+Cv4KQwtxC4cLjQvXC+AL5QvmCzgMcQyPDKwMwwzJDPgMEjIBAAAAAAAAAAADAAAABAAAAAABAAABAAAABgABAwAAAAAAAAEAAAUCAQAAAgYAAAIAAA==',
                    'x-requested-with': 'fetch',
                    'x-zse-93': '101_3_3.0',
                    'x-zse-96': '2.0_YgMIhebWasNsCXw=FiiAUcdqV88f=NTzIsgp2cwr0d2+eelDIHhqN3Tru+P9NAs6',
                    'x-zst-81': '3_2.0aR_sn77yn6O92wOB8hPZn79qE72xcXFZ16fyQArZ39Sm7820XM20cL_1kwxYUqwT16P0EiUZbR2x-LOmwhp1tD_I-JOfgGXTzJO1ADRZ0cHsTJXII820Eer0c4nVDJH8zGCBADwMuukRo4Cqm4w0riRO70CB70O83uPmgbgmhufXiqomKbO1FJLYiRnxEL2ZZrxmDucmqhPXnXFMTAoTF6RhRuLPFXwMxBwBogLLLhSVPgSfRcxqjhOY2BSVibNfX9eLXJp9cTgVxrOB4h2C2uoTv0SpiBFMvBHCoAxmZq2pauV1cuw9kUeLkR38iqHO8RtCgUo0-geLgwtyzvSLFDUfWwgMIhVBcCXOIgxfLBcMawLG1TNqBrpCeUNYEwFfEhwO6vV_HCXsCqfz5DeBS6S9xUw9urO__Dg1cqHByqX_k9g0iUw1urN0Qicf6qFGFqc8Qi9GQqgxWDLm3UgfiJHBwww8DBeLyvO1ucS1n92mO9gmxgFmbuN1LGcpWhXBBBLC'
                }
                # headers = {
                #     'pragma': 'no-cache',
                #     'cache-control': 'no-cache',
                #     'x-zse-93': '101_3_3.0',
                #     'accept-language': 'zh-CN,zh;q=0.9,en;q=0.8',
                #     'x-zst-81': '3_2.0aR_sn77yn6O92wOB8hPZnQr0EMYxc4f18wNBUgpTQ6nxERFZMRY0-4Lm-h3_tufIwJS8gcxTgJS_'
                #                 'AuPZNcXCTwxI78YxEM20s4PGDwN8gGcYAupMWufIoLVqr4gxrRPOI0cY7HL8qun9g93mFukyigcmebS_'
                #                 'FwOYPRP0E4rZUrN9DDom3hnynAUMnAVPF_PhaueTF7C89Uw8_w2YiDNBOcC_ggeL2HN1fTN0WvUq_'
                #                 'UY9aCLBMQO0VvUfoTp9y9eTV4xC8b3YCGNs2bSfXwCZpUw06hcMJ_w0B8NmhvwMcGe9iqCZlcSftv3qTug'
                #                 '_-BVqgwL1r7H99CeV24C9_Up9HqHmj9CsqUH8tBNLUbxm1gOprAO_6CL0Vg_z9DVs9rS1R9x08wxG2Cxme9'
                #                 'FLlUc9jJu027gB3UVBJqpBbqUCrwo1nqHmggO_Nr9_BqcxfC2GnvSqNcxyIDOfiuw9dgCC27emaUpMxuVqq'
                #                 'uFYwUSL2BXs',
                #     'accept': '*/*',
                #     'referer': referer,
                #     'sec-ch-ua': '"Chromium";v="106", "Google Chrome";v="106", "Not;A=Brand";v="99"',
                #     'sec-ch-ua-mobile': '?0',
                #     'sec-ch-ua-platform': '"macOS"',
                #     'sec-fetch-dest': 'empty',
                #     'sec-fetch-mode': 'cors',
                #     'sec-fetch-site': 'same-origin',
                #     'user-agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 '
                #                   '(KHTML, like Gecko) Chrome/106.0.0.0 Safari/537.36',
                #     'x-ab-pb': 'CsIBCAAbAD8ARwC0AGkBagF0ATsCzALXAtgCTwNQA6ADoQOiA7cD8wP0AzMEjASNBKYE1gQR'
                #                'BVEFiwWMBZ4FMAYxBusGJwd3B3gH2AfcB90HZwh0CHYIeQjaCD8JQglgCY0JwwnECcUJxgnH'
                #                'CcgJyQnKCcsJzAnRCfQJBApJCmUKawqYCqUKqQq+CsQK1ArdCu0K/Qr+CjsLPAtDC0YLcQt2'
                #                'C4ULhwuNC8AL1wvgC+UL5gssDDgMcQyPDKwMuQzDDMkM+AwSYQEAAAAAAAAAAAMAAAAAAAAA'
                #                'AAAAAAQEAAQAAAAAAQAAAQAAAAACAgQAAAYAAAEBAAAAAAAAAAAAAAAAAwAAAAABAAAAAQEAA'
                #                'AAAAQABAAAAAAAFAAIBAAAGAgYAAAECAAA=',
                #     'x-requested-with': 'fetch',
                #     # 'accept-encoding': 'gzip, deflate, br',
                #     'accept-encoding': 'gzip, deflate',
                #     'x-zse-96': '2.0_wSyMHDQ7MZYfDZWRCM=Il3PaIuFuaTZf=spPXrAiJo83rlG65AU+b6IIBIr5nJ5L',
                # }
                # params = {
                #     'offset': str(cls.offset),
                #     'limit': '20',
                #     'similar_aggregation': True,
                #     'include': 'similar_zvideo,creation_relationship',
                # }
                # cookies = {
                #     '_zap': 'a17e5882-775f-49dd-b541-e854c0299999',
                #     'd_c0': '"AFAcR4hr0RKPTn2-RL935cm51KQZn9-Xlb0=|1616071586"',
                #     '_9755xjdesxxd_': '32',
                #     'YD00517437729195%3AWM_TID': 'yR6Kt%2FwMI2xEFEQRVBYrou2J%2F14yfdzt',
                #     'q_c1': '947ece3f1ccd4ccd83b718c113fe3935|1658911548000|1658911548000',
                #     '_xsrf': 'cAM8ovVh43f0oJE2erl6ho8Y4CTZhFKF',
                #     '__snaker__id': 'mqTPtP3IwbuoHA2K',
                #     'gdxidpyhxdE': 'a2d%2FeD%2BEUNxnnTaLDjQfd28sI39g3pnKAqS%2BmodWMqYcMa3AMzV5eom2APR2z5rDHAVvfubp'
                #                    '%2BWEzkcxdb9UjWNT0gbshTpBmHzU1%2BMB%2BfW5sAAUz1%5CTWIWbP34z7SV4TCB%2BfbQtoAdRA'
                #                    'ZANzEbcollrLlPmyAG%2FtPICs5ewAvswneclw%3A1666061422009',
                #     'YD00517437729195%3AWM_NI': 'bTzKLUbqXTrTGVFS91056cj27L%2F9PQ8MG8ofPHnJc98014ktS5xa09N8yJZho7j'
                #                                 'EeQfeueH0sJLDH5YmaA2nCP1kQ1jaB4FChGGNA7HbM%2B5yRCuYOmGjfIs5LOdoxm'
                #                                 '1WSHY%3D',
                #     'YD00517437729195%3AWM_NIKE': '9ca17ae2e6ffcda170e2e6eeb2f5628de7a5b2db5cfc8e8eb6d85f929e8aacc'
                #                                   '150b6ed9bd8b33c93afc089cf2af0fea7c3b92aa28cb6accf5fa2af98b5f1608'
                #                                   '58a8ab4bc3bf4ebfa84e93af594e1dad35bfc9389d9ee6e9498a2abe233fcf0f'
                #                                   'fd5ea8091ebb882b3338ee8b7abd964ade9a6a6e85ba6bbbb86cb3cb19684a5c'
                #                                   'b3f9bb3afa2b1668190ff8cd43baabf8ed5c549ac9989afc43d8cbe83ccb8739'
                #                                   '29dbbbace6785b6aaaaf23f9091ac90fb67f3899da9dc37e2a3',
                #     'captcha_ticket_v2': '2|1:0|10:1666060537|17:captcha_ticket_v2|704:eyJ2YWxpZGF0ZSI6IkNOMzFfVkk0'
                #                          'TmhfRVJ1SGxhQW4yV0ZHWFlXTjdIWXM3TnZQSUFMY19MRGdVV1dMX3Vody0wMFAtYnA0UXBFV'
                #                          'nI5akI1NFNVajhZVEpLeXVLODJWajJoVU83MnBya2xyLi5TalExekwyZ3hQU3JjZTI4TkJweT'
                #                          'h2NEg5QjFyVUlBME1zSGtfUlRqQ3lDd1BkQ2Y3c3J3WWl0V0Qyci5FTWV6eEJRMUdQcGlWS2d'
                #                          'CXzRkQzhSWmNMT1ZmWnlxUDZWbTh6ZEdDdmVnVkhFTTFGVWNpNVJRV2NWLjZ6V2V4UUpJOWJi'
                #                          'bUdRZHpJQ0VQUFlZZU5ObVBKOTlKMUw1OWFxUmRxUzVtY1hDbXBsRzVBTnM0LmEyS1BDRGNTY'
                #                          'WVTWUlpMFdfc0JlTE44QlZCXzFkWjFLV2xoUmx3bkR6QjJMdUdOcm1IRXRaejQ2T0ViVzhVS3'
                #                          'ZSNjU2YUJEUk9tYUNscVM5cDk2T2cwY1JadlhEZFZOSlFSeG0wNkRUMEs5cHRLLS1hajFWaWt'
                #                          'EMnp4UU50MEpFZE9PRGs2RWtqdklnWnM0S0s0bUVBd0tGS2FIMklJaXVod1dMck9WOGdKLUpw'
                #                          'NjlYNU9hWWxNcEtCbk9pU2JYNUJfUnBmeU1hLlAwbUlBYXYwbnBvY0xRWmdMMXJpelN5OFVaV'
                #                          'l9QMXduWTByR1BtMyJ9|557dafa54cac0a5cd55b3e8ebbb626e6817010e985c2a3c60a82f'
                #                          'b00bcb8b72a',
                #     'captcha_session_v2': '2|1:0|10:1666060537|18:captcha_session_v2|88:SUlnbE5uR29rUC83Z0VvWVpjTS9'
                #                           'JV3VKSFFYTUl4T1hFbVdZZnB2Wms1Z1cxYkZFQndYWHhqVDRGWEE0NzVaTQ==|a40c62b74a'
                #                           '94ad618205342e3f87091b525a50fe9c8f153af02b08a3569fa7c0',
                #     'z_c0': '2|1:0|10:1666767387|4:z_c0|92:Mi4xVFdsTlB3QUFBQUFBVUJ4SGlHdlJFaVlBQUFCZ0FsVk5CVjg3WkFB'
                #             'OTBXb185c2wyUjJ6ZmQ1OTYtamliYWJWTUp3|abd71d40da50ae308430e26ce358d34cd4bccdfe575bf544d'
                #             '212019c7189ecdb',
                #     'Hm_lvt_98beee57fd2ef70ccdd5ca52b9740c49': '1665574748,1666678329,1667194593',
                #     'SESSIONID': 'hyrUgL9IzongRfviGSTmGR2sDyVliSwHuWy0fXouVk7',
                #     'JOID': 'W14WBkoyfRA9faU0WD7qwUz2wuJAVQFkbBLmCBdAHXtcK8tNNNogTVJ6rzdabb3cImOxRThXGFkflrdU1CD0UWI=',
                #     'osd': 'U10VAU46fhM6ea03WznuyU_1xeZIVgJjaBrlCxBEFXhfLM9FN9knSVp5rDBeZb7fJWe5RjtQHFEclbBQ3CP3VmY=',
                #     'tst': 'v',
                #     'NOT_UNREGISTER_WAITING': '1',
                #     'Hm_lpvt_98beee57fd2ef70ccdd5ca52b9740c49': '1667373061',
                #     'unlock_ticket': 'AGBWwkpQuhUmAAAAYAJVTQ0fYmOTbjeHqRp3XiDPD3ZVB5KkeaEeYw==',
                #     'KLBRSID': '031b5396d5ab406499e2ac6fe1bb1a43|1667373074|1667370552',
                # }

                urllib3.disable_warnings()
                response = requests.get(url=url, headers=headers, data=payload, verify=False, proxies=proxies)
                cls.offset += 20
                if 'data' not in response.json():
                    Common.logger(log_type).error('response:{}\n', response.text)
                    cls.offset = 0
                    return
                elif len(response.json()['data']) == 0:
                    Common.logger(log_type).error('response:{}\n', response.text)
                    cls.offset = 0
                    return
                else:
                    feeds = response.json()['data']
                    is_end = response.json()['paging']['is_end']
                    for i in range(len(feeds)):
                        # video_title
                        if 'title' not in feeds[i]:
                            video_title = 0
                        else:
                            video_title = feeds[i]['title']

                        # video_id
                        if 'video' not in feeds[i]:
                            video_id = 0
                        elif 'video_id' not in feeds[i]['video']:
                            video_id = 0
                        else:
                            video_id = feeds[i]['video']['video_id']

                        # play_cnt
                        if 'play_count' not in feeds[i]:
                            play_cnt = 0
                        else:
                            play_cnt = feeds[i]['play_count']

                        # comment_cnt
                        if 'comment_count' not in feeds[i]:
                            comment_cnt = 0
                        else:
                            comment_cnt = feeds[i]['comment_count']

                        # like_cnt
                        if 'liked_count' not in feeds[i]:
                            like_cnt = 0
                        else:
                            like_cnt = feeds[i]['liked_count']

                        # share_cnt
                        if 'share_count' not in feeds[i]:
                            share_cnt = 0
                        else:
                            share_cnt = feeds[i]['share_count']

                        # voteup_cnt 赞同数
                        if 'voteup_count' not in feeds[i]:
                            voteup_cnt = 0
                        else:
                            voteup_cnt = feeds[i]['voteup_count']

                        # publish_time
                        if 'published_at' not in feeds[i]:
                            publish_time = 0
                        else:
                            publish_time = feeds[i]['published_at']

                        # duration
                        if 'video' not in feeds[i]:
                            duration = 0
                        elif 'duration' not in feeds[i]['video']:
                            duration = 0
                        else:
                            duration = feeds[i]['video']['duration']

                        # width / height / video_url
                        if 'video' not in feeds[i]:
                            video_width = 0
                            video_height = 0
                            video_url = 0
                        elif 'playlist' in feeds[i]['video'] and 'fhd' in feeds[i]['video']['playlist'] \
                                and 'width' in feeds[i]['video']['playlist']['fhd'] \
                                and 'height' in feeds[i]['video']['playlist']['fhd'] \
                                and 'play_url' in feeds[i]['video']['playlist']['fhd']:
                            video_width = feeds[i]['video']['playlist']['fhd']['width']
                            video_height = feeds[i]['video']['playlist']['fhd']['height']
                            video_url = feeds[i]['video']['playlist']['fhd']['play_url']
                        elif 'playlist' in feeds[i]['video'] and 'fhd' in feeds[i]['video']['playlist'] \
                                and 'width' in feeds[i]['video']['playlist']['fhd'] \
                                and 'height' in feeds[i]['video']['playlist']['fhd'] \
                                and 'url' in feeds[i]['video']['playlist']['fhd']:
                            video_width = feeds[i]['video']['playlist']['fhd']['width']
                            video_height = feeds[i]['video']['playlist']['fhd']['height']
                            video_url = feeds[i]['video']['playlist']['fhd']['url']
                        elif 'playlist' in feeds[i]['video'] and 'hd' in feeds[i]['video']['playlist'] \
                                and 'width' in feeds[i]['video']['playlist']['hd'] \
                                and 'height' in feeds[i]['video']['playlist']['hd'] \
                                and 'play_url' in feeds[i]['video']['playlist']['hd']:
                            video_width = feeds[i]['video']['playlist']['hd']['width']
                            video_height = feeds[i]['video']['playlist']['hd']['height']
                            video_url = feeds[i]['video']['playlist']['hd']['play_url']
                        elif 'playlist' in feeds[i]['video'] and 'hd' in feeds[i]['video']['playlist'] \
                                and 'width' in feeds[i]['video']['playlist']['hd'] \
                                and 'height' in feeds[i]['video']['playlist']['hd'] \
                                and 'url' in feeds[i]['video']['playlist']['hd']:
                            video_width = feeds[i]['video']['playlist']['hd']['width']
                            video_height = feeds[i]['video']['playlist']['hd']['height']
                            video_url = feeds[i]['video']['playlist']['hd']['url']
                        elif 'playlist' in feeds[i]['video'] and 'ld' in feeds[i]['video']['playlist'] \
                                and 'width' in feeds[i]['video']['playlist']['ld'] \
                                and 'height' in feeds[i]['video']['playlist']['ld'] \
                                and 'play_url' in feeds[i]['video']['playlist']['ld']:
                            video_width = feeds[i]['video']['playlist']['ld']['width']
                            video_height = feeds[i]['video']['playlist']['ld']['height']
                            video_url = feeds[i]['video']['playlist']['ld']['play_url']
                        elif 'playlist' in feeds[i]['video'] and 'ld' in feeds[i]['video']['playlist'] \
                                and 'width' in feeds[i]['video']['playlist']['ld'] \
                                and 'height' in feeds[i]['video']['playlist']['ld'] \
                                and 'url' in feeds[i]['video']['playlist']['ld']:
                            video_width = feeds[i]['video']['playlist']['ld']['width']
                            video_height = feeds[i]['video']['playlist']['ld']['height']
                            video_url = feeds[i]['video']['playlist']['ld']['url']
                        elif 'playlist' in feeds[i]['video'] and 'sd' in feeds[i]['video']['playlist'] \
                                and 'width' in feeds[i]['video']['playlist']['sd'] \
                                and 'height' in feeds[i]['video']['playlist']['sd'] \
                                and 'play_url' in feeds[i]['video']['playlist']['sd']:
                            video_width = feeds[i]['video']['playlist']['sd']['width']
                            video_height = feeds[i]['video']['playlist']['sd']['height']
                            video_url = feeds[i]['video']['playlist']['sd']['play_url']
                        elif 'playlist' in feeds[i]['video'] and 'sd' in feeds[i]['video']['playlist'] \
                                and 'width' in feeds[i]['video']['playlist']['sd'] \
                                and 'height' in feeds[i]['video']['playlist']['sd'] \
                                and 'url' in feeds[i]['video']['playlist']['sd']:
                            video_width = feeds[i]['video']['playlist']['sd']['width']
                            video_height = feeds[i]['video']['playlist']['sd']['height']
                            video_url = feeds[i]['video']['playlist']['sd']['url']

                        elif 'playlist_v2' in feeds[i]['video'] and 'fhd' in feeds[i]['video']['playlist_v2'] \
                                and 'width' in feeds[i]['video']['playlist_v2']['fhd'] \
                                and 'height' in feeds[i]['video']['playlist_v2']['fhd'] \
                                and 'play_url' in feeds[i]['video']['playlist_v2']['fhd']:
                            video_width = feeds[i]['video']['playlist_v2']['fhd']['width']
                            video_height = feeds[i]['video']['playlist_v2']['fhd']['height']
                            video_url = feeds[i]['video']['playlist_v2']['fhd']['play_url']
                        elif 'playlist_v2' in feeds[i]['video'] and 'fhd' in feeds[i]['video']['playlist_v2'] \
                                and 'width' in feeds[i]['video']['playlist_v2']['fhd'] \
                                and 'height' in feeds[i]['video']['playlist_v2']['fhd'] \
                                and 'url' in feeds[i]['video']['playlist_v2']['fhd']:
                            video_width = feeds[i]['video']['playlist_v2']['fhd']['width']
                            video_height = feeds[i]['video']['playlist_v2']['fhd']['height']
                            video_url = feeds[i]['video']['playlist_v2']['fhd']['url']
                        elif 'playlist_v2' in feeds[i]['video'] and 'hd' in feeds[i]['video']['playlist_v2'] \
                                and 'width' in feeds[i]['video']['playlist_v2']['hd'] \
                                and 'height' in feeds[i]['video']['playlist_v2']['hd'] \
                                and 'play_url' in feeds[i]['video']['playlist_v2']['hd']:
                            video_width = feeds[i]['video']['playlist_v2']['hd']['width']
                            video_height = feeds[i]['video']['playlist_v2']['hd']['height']
                            video_url = feeds[i]['video']['playlist_v2']['hd']['play_url']
                        elif 'playlist_v2' in feeds[i]['video'] and 'hd' in feeds[i]['video']['playlist_v2'] \
                                and 'width' in feeds[i]['video']['playlist_v2']['hd'] \
                                and 'height' in feeds[i]['video']['playlist_v2']['hd'] \
                                and 'url' in feeds[i]['video']['playlist_v2']['hd']:
                            video_width = feeds[i]['video']['playlist_v2']['hd']['width']
                            video_height = feeds[i]['video']['playlist_v2']['hd']['height']
                            video_url = feeds[i]['video']['playlist_v2']['hd']['url']
                        elif 'playlist_v2' in feeds[i]['video'] and 'ld' in feeds[i]['video']['playlist_v2'] \
                                and 'width' in feeds[i]['video']['playlist_v2']['ld'] \
                                and 'height' in feeds[i]['video']['playlist_v2']['ld'] \
                                and 'play_url' in feeds[i]['video']['playlist_v2']['ld']:
                            video_width = feeds[i]['video']['playlist_v2']['ld']['width']
                            video_height = feeds[i]['video']['playlist_v2']['ld']['height']
                            video_url = feeds[i]['video']['playlist_v2']['ld']['play_url']
                        elif 'playlist_v2' in feeds[i]['video'] and 'ld' in feeds[i]['video']['playlist_v2'] \
                                and 'width' in feeds[i]['video']['playlist_v2']['ld'] \
                                and 'height' in feeds[i]['video']['playlist_v2']['ld'] \
                                and 'url' in feeds[i]['video']['playlist_v2']['ld']:
                            video_width = feeds[i]['video']['playlist_v2']['ld']['width']
                            video_height = feeds[i]['video']['playlist_v2']['ld']['height']
                            video_url = feeds[i]['video']['playlist_v2']['ld']['url']
                        elif 'playlist_v2' in feeds[i]['video'] and 'sd' in feeds[i]['video']['playlist_v2'] \
                                and 'width' in feeds[i]['video']['playlist_v2']['sd'] \
                                and 'height' in feeds[i]['video']['playlist_v2']['sd'] \
                                and 'play_url' in feeds[i]['video']['playlist_v2']['sd']:
                            video_width = feeds[i]['video']['playlist_v2']['sd']['width']
                            video_height = feeds[i]['video']['playlist_v2']['sd']['height']
                            video_url = feeds[i]['video']['playlist_v2']['sd']['play_url']
                        elif 'playlist_v2' in feeds[i]['video'] and 'sd' in feeds[i]['video']['playlist_v2'] \
                                and 'width' in feeds[i]['video']['playlist_v2']['sd'] \
                                and 'height' in feeds[i]['video']['playlist_v2']['sd'] \
                                and 'url' in feeds[i]['video']['playlist_v2']['sd']:
                            video_width = feeds[i]['video']['playlist_v2']['sd']['width']
                            video_height = feeds[i]['video']['playlist_v2']['sd']['height']
                            video_url = feeds[i]['video']['playlist_v2']['sd']['url']
                        else:
                            video_width = 0
                            video_height = 0
                            video_url = 0

                        # cover_url
                        if 'video' not in feeds[i]:
                            cover_url = 0
                        elif 'thumbnail' not in feeds[i]['video']:
                            cover_url = 0
                        else:
                            cover_url = feeds[i]['video']['thumbnail']

                        # user_name / uid / user_type / url_token / avatar_url
                        if 'author' not in feeds[i]:
                            user_name = 0
                            uid = 0
                            user_type = 0
                            url_token = 0
                            avatar_url = 0
                        elif 'author' in feeds[i] \
                                and 'name' in feeds[i]['author'] \
                                and 'uid' in feeds[i]['author'] \
                                and 'user_type' in feeds[i]['author'] \
                                and 'url_token' in feeds[i]['author'] \
                                and 'avatar_url_template' in feeds[i]['author']:
                            user_name = feeds[i]['author']['name']
                            uid = feeds[i]['author']['uid']
                            user_type = feeds[i]['author']['user_type']
                            url_token = feeds[i]['author']['url_token']
                            avatar_url = feeds[i]['author']['avatar_url_template']
                        elif 'author' in feeds[i] \
                                and 'name' in feeds[i]['author'] \
                                and 'uid' in feeds[i]['author'] \
                                and 'user_type' in feeds[i]['author'] \
                                and 'url_token' in feeds[i]['author'] \
                                and 'avatar_url' in feeds[i]['author']:
                            user_name = feeds[i]['author']['name']
                            uid = feeds[i]['author']['uid']
                            user_type = feeds[i]['author']['user_type']
                            url_token = feeds[i]['author']['url_token']
                            avatar_url = feeds[i]['author']['avatar_url']
                        else:
                            user_name = 0
                            uid = 0
                            user_type = 0
                            url_token = 0
                            avatar_url = 0

                        Common.logger(log_type).info('video_title:{}', video_title)
                        Common.logger(log_type).info('duration:{}秒', int(duration))
                        Common.logger(log_type).info(
                            'publish_time:{}', time.strftime("%Y/%m/%d %H:%M:%S", time.localtime(publish_time)))

                        if video_title == 0 or cover_url == 0 or video_url == 0:
                            Common.logger(log_type).info('无效视频\n')
                        elif int(time.time()) - int(publish_time) >= 3600 * 24 * 30:
                            Common.logger(log_type).info('发布时间超过30天\n')
                            cls.offset = 0
                            return
                        elif int(duration) < 60:
                            Common.logger(log_type).info('时长{}<60秒\n', int(duration))
                        elif str(video_id) in [x for y in Feishu.get_values_batch(log_type, 'zhihu', '8871e3') for x in
                                               y]:
                            Common.logger(log_type).info('视频已下载\n')
                        elif str(video_id) in [x for y in Feishu.get_values_batch(log_type, 'zhihu', '4MGuux') for x in
                                               y]:
                            Common.logger(log_type).info('视频已下载\n')
                        else:
                            Common.download_method(log_type, 'cover', video_title, cover_url)
                            Common.download_method(log_type, 'video', video_title, video_url)
                            # 保存视频信息至 "./videos/{download_video_title}/info.txt"
                            with open("./videos/" + video_title + "/" + "info.txt", "a", encoding="UTF-8") as f_a:
                                f_a.write(str(video_id) + "\n" +
                                          str(video_title) + "\n" +
                                          str(int(duration)) + "\n" +
                                          str(play_cnt) + "\n" +
                                          str(comment_cnt) + "\n" +
                                          str(like_cnt) + "\n" +
                                          str(share_cnt) + "\n" +
                                          str(video_width) + '*' + str(video_height) + "\n" +
                                          str(publish_time) + "\n" +
                                          str(user_name) + "\n" +
                                          str(avatar_url) + "\n" +
                                          str(video_url) + "\n" +
                                          str(cover_url) + "\n" +
                                          "zhihu_follow" + str(int(time.time())))
                            Common.logger(log_type).info("==========视频信息已保存至info.txt==========")

                            our_video_id = Publish.upload_and_publish(log_type, env, our_uid)
                            if env == 'dev':
                                our_video_link = "https://testadmin.piaoquantv.com/cms/post-detail/" + str(
                                    our_video_id) + "/info"
                            else:
                                our_video_link = "https://admin.piaoquantv.com/cms/post-detail/" + str(
                                    our_video_id) + "/info"
                            Common.logger(log_type).info("视频上传完成:{}", video_title)

                            Feishu.insert_columns(log_type, 'zhihu', '4MGuux', 'ROWS', 1, 2)
                            time.sleep(1)
                            upload_time = int(time.time())
                            values = [[
                                time.strftime("%Y/%m/%d %H:%M:%S", time.localtime(upload_time)),
                                "定向榜",
                                video_title,
                                str(video_id),
                                our_video_link,
                                play_cnt,
                                comment_cnt,
                                like_cnt,
                                share_cnt,
                                voteup_cnt,
                                int(duration),
                                str(video_width) + '*' + str(video_height),
                                time.strftime("%Y/%m/%d %H:%M:%S", time.localtime(int(publish_time))),
                                user_name,
                                uid,
                                'https://www.zhihu.com/' + user_type + '/' + url_token,
                                user_type,
                                url_token,
                                avatar_url,
                                cover_url,
                                video_url]]
                            Feishu.update_values(log_type, 'zhihu', "4MGuux", "F2:Z2", values)
                            Common.logger(log_type).info("视频已保存至云文档:{}\n", video_title)

                    if is_end is True:
                        Common.logger(log_type).info('到底了\n')
                        cls.offset = 0
                        return
            except Exception as e:
                Common.logger(log_type).error('get_follow_feeds异常:{}', e)

    @classmethod
    def get_all_users_feeds(cls, log_type, env):
        try:
            user_list = cls.get_users_from_feishu(log_type)
            if len(user_list) == 0:
                Common.logger(log_type).info('定向用户列表为空\n')
            else:
                for k, v in user_list.items():
                    user_name = k
                    url_token = v.split(',')[0]
                    referer = v.split(',')[1]
                    our_uid = v.split(',')[2]
                    Common.logger(log_type).info('开始抓取 {} 主页视频\n', user_name)
                    cls.get_follow_feeds(log_type, url_token, referer, our_uid, env)
                    Common.logger(log_type).info('{} 主页视频抓取完毕，休眠 60 秒\n', user_name)
                    cls.offset = 0
                    time.sleep(60)
        except Exception as e:
            Common.logger(log_type).error('get_all_users_feeds异常:{}', e)


if __name__ == '__main__':
    ZhihuFollow.get_all_users_feeds('follow', 'dev')

    pass