crawler_kuaishou/main/demo.py

# -*- coding: utf-8 -*-
# @Author: wangkun
# @Time: 2022/3/31
from datetime import date, timedelta
import datetime
import json
import re
import time

import requests
import urllib3


class Demo:
    @classmethod
    def demo1(cls):
        download_video_resolution = "720*1280"
        download_video_width = download_video_resolution.split("*")[0]
        download_video_height = download_video_resolution.split("*")[-1]
        print(download_video_resolution)
        print(download_video_width)
        print(download_video_height)

    @classmethod
    def time(cls):
        # 推荐
        time1 = int(time.time()*1000)
        print(time1)
        # 不推荐
        time2 = round(time.time())*1000
        print(time2)

        # 统一获取当前时间
        now = datetime.datetime.now()
        print(type(now))
        print(f"now:{now}")
        # 昨天
        yesterday = (date.today() + timedelta(days=-1)).strftime("%Y-%m-%d")
        print(type(yesterday))
        print(f"昨天:{yesterday}")
        # 今天
        today = date.today()
        print(type(today))
        print(f"今天:{today}")
        # 明天
        tomorrow = (date.today() + timedelta(days=1)).strftime("%Y-%m-%d")
        print(type(tomorrow))
        print(f"明天:{tomorrow}")

    @classmethod
    def get_douyin_feeds(cls):
        """
        获取抖音feed流视频 https://www.douyin.com
        """
        url = "https://www.douyin.com/aweme/v1/web/tab/feed/?"
        params = {
            "device_platform": "webapp",
            "aid": "6383",
            "channel": "channel_pc_web",
            "count": "10",
            "refresh_index": "4",
            "video_type_select": "0",
            "version_code": "170400",
            "version_name": "17.4.0",
            "cookie_enabled": "true",
            "screen_width": "1920",
            "screen_height": "1080",
            "browser_language": "zh-CN",
            "browser_platform": "MacIntel",
            "browser_name": "Chrome",
            "browser_version": "99.0.4844.84",
            "browser_online": "true",
            "engine_name": "Blink",
            "engine_version": "99.0.4844.84",
            "os_name": "Mac OS",
            "os_version": "10.15.7",
            "platform": "PC",
            "cpu_core_num": "8",
            "device_memory": "8",
            "downlink": "10",
            "effective_type": "4g",
            "round_trip_time": "50",
            "msToken": "304uY1lV7HmHkR1G1QUaFqg0yrL5_WqrFOR8qCbl3hOsl8aSNI_18vIfpTGNhNRVZx7ysRiCHpcBKhpujTsbbC"
                       "ZEDbG7pllZzlO3tlrBOs2TFYUgJdsvbw==",
            "X-Bogus": "DFSzswVYPVsANat/Sl8eGc3WxM23",
            "_signature": "qaJgTwAAy.aVqLslyfC7aKmiYF"
        }
        cookies = {
            "_tea_utm_cache_6383": "undefined",
            "ttwid": "1%7CETZk6sDMDSBgewWhKJXghFN4cwXTz0fLuhsLEngD_Nk%7C1648812136%7Cfa66fa81ccfe3f552f4"
                     "e8b8327e72cbbc5e897141c25a5fcd32defaed1466d3e",
            "passport_csrf_token": "e2d0f1ed9fd22463be9f389137a781ce",
            "passport_csrf_token_default": "e2d0f1ed9fd22463be9f389137a781ce",
            "s_v_web_id": "verify_l1h7nzwr_ABN0FA2f_BTrM_4zSH_8WPN_2KY2iZFmbhE2",
            "_tea_utm_cache_1300": "undefined",
            "_tea_utm_cache_2285": "undefined",
            "ttcid": "3220eeda36a244beadd32a4b44d2044b31",
            "douyin.com": "",
            "__ac_nonce": "06247fb0f00f050ccc9b2",
            "__ac_signature": "_02B4Z6wo00f01AN7DoAAAIDB5nv.qI7xGZQDWwoAAGKfo4rd5YCAYF8o5PyppIpsdKxV0k2NerO"
                              "f1VEQr3eJftkpgon9tcveDVpmfY555vzTTvRznegS1ax3KJXnoav2ZdEoYzwR3wDszPCk5d",
            "strategyABtestKey": "1648865029.449",
            "AB_LOGIN_GUIDE_TIMESTAMP": "1648865029279",
            "THEME_STAY_TIME": "299621",
            "IS_HIDE_THEME_CHANGE": "1",
            "home_can_add_dy_2_desktop": "0",
            "tt_scid": "vUl8CBW1SMQp2l5GmUIja5A6ziY1LByrsoN.P-wvKuutiB8ftvlfK.9ZEeehNC5u821d",
            "pwa_guide_count": "2",
            "msToken": "EHCmp9Qw7PAChI3do-MQPjOR29hf4ZFLYNrGl89HkFKdO5Iwb8n7z5fpETrgim2zFTIkGT"
                       "ObOxH7HCrHCLVEX5eAuwAS1A2sjKH4MHEfjfPqA06Lo4v9Pw==",
        }
        try:
            urllib3.disable_warnings()
            r = requests.get(url=url, params=params, cookies=cookies, verify=False)
            # response = json.loads(r.content.decode("utf8"))
            print(r)
            print(type(r.text))
            print(r.text)
        except Exception as e:
            print(e)

    @classmethod
    def demo2(cls):
        s = "0"
        print(int(int(s) / 10))

    @classmethod
    def get_weishi_feeds(cls):
        url = "https://api.weishi.qq.com/trpc.weishi.weishi_h5_proxy.weishi_h5_proxy/WxminiGetFeedList"
        cookies = {
            "wesee_authtype": "3",
            "wesee_openid":	"oWGa05FrwkuUvT-4n1qGeQuhVsc8",
            "wesee_openkey": "8c3ec202f5d679fb5ee6d9f643640d9a2580ba504612e2d979a881d3169caf189e2a5c1d532eeff172bc21cf2"
                             "6230941ccbc10243a7879e8165ca608c17060de606a6d08afe0a3abd5250629314f9a99e9d1003b201bf5ec",
            "wesee_personid": "1593522421826902",
            "wesee_refresh_token": "",
            "wesee_access_token": "8c3ec202f5d679fb5ee6d9f643640d9a2580ba504612e2d979a881d3169caf18"
                                  "9e2a5c1d532eeff172bc21cf26230941ccbc10243a7879e8165ca608c17060de6"
                                  "06a6d08afe0a3abd5250629314f9a99e9d1003b201bf5ec",
            "wesee_thr_appid": "wx75ee9f19b93e5c46",
            "wesee_ichid": "8"
        }
        json_data = {
            "req_body": {
                "requestType": 16,
                "isrefresh": 0,
                "isfirst": 0,
                "attachInfo": "",
                "scene_id": 22,
                "requestExt": {
                    "mini_openid": "oWGa05FrwkuUvT-4n1qGeQuhVsc8",
                    "notLogin-personid": "1593522421826902"
                }
            },
            "req_header": {
                "mapExt": "{\"imageSize\":\"480\",\"adaptScene\":\"PicHDWebpLimitScene\"}"
            }
        }
        try:
            urllib3.disable_warnings()
            r = requests.post(url=url, cookies=cookies, json=json_data, verify=False)
            response = json.loads(r.content.decode("utf8"))
            feeds = response["rsp_body"]["feeds"]
            for feed in feeds:
                print(feed)
        except Exception as e:
            print(e)

    @classmethod
    def edit_str(cls):
        title_list = ["#上海战疫 上海累计感染超20万！这条被淹没的热搜，令全网泪目… 疫情一定要攻克，但所有人都不该遗忘这些弱者。#上海累计报告本土阳性感染者超20万例 #农民工",
                      "#重庆地火村 #地火村 #旅行",
                      "第79集 | 湖南最值得去的六个景区，每一个都是绝色…… #快手带你去旅行 #旅游胜地 #旅游",
                      "霸王条款不废除，断供有多可怕。 #涨知识 #生活小常识 # 生活常识",
                      "秦始皇还活着？地宫中有不明物体缓缓移动 #历史 #秦始皇 #新春寄语  @快手热点(O40300129)",
                      "#夏日荷花  #国花牡丹 #昙花一现",
                      "国内最良心的8个景区，这才是景区最该有的样子，看看你去过几个？ #旅行  #旅游 ",
                      "狗子呆在水里三天三夜，终于练成捕鱼神功，一口一个大鲶鱼 #狗狗  #神奇动物  #快手放映室  @快手热点(O40300129) ",
                      "#集结吧光合创作者  养鸡小伙：喂鸡摆出各种造型，被称为鸡司令。",
                      "89岁农民老艺人自食其力，街头卖艺表演“捏碎碗片”绝技，现场听到咔吱咔吱响，人狠功夫硬！这功夫已失传，以后再看不到了！#集结吧光合创作者 #农民 #街头表演  @快手光合作者助手(O40300118)  @快手热点(O40300129)  @我要上热门(O1907752910)",
                      "我国最贵最有名的三棵树，你知道哪三棵吗？#旅游 #旅行攻略 #黄山迎客松",
                      "潘长江带来热舞，蔡明 郭达也来了！太嗨了！歌词太棒了！ @快手涨粉助手(O1815060199)  @快手热点(O40300129)  @快手平台帐号(O90041) #潘长江 #搞笑 #集结吧光合创作者",
                      "#带你看世界 给大家带来一期烟花盛宴，希望大家能够喜欢，带上你的那个她一起来看吧 #烟花 #视觉震撼"
                      ]
        for title in title_list:
            title_split1 = title.split(" #")
            if title_split1[0] != "":
                title1 = title_split1[0]
            else:
                title1 = title_split1[0]

            title_split2 = title1.split(" #")
            if title_split2[0] != "":
                title2 = title_split2[0]
            else:
                title2 = title_split2[-1]

            title_split3 = title2.split("@")
            if title_split3[0] != "":
                title3 = title_split3[0]
            else:
                title3 = title_split3[-1]

            print(title3)
            title = title3.replace("\n", "").replace("#", "").replace("/", "").replace("\r", "")
            print(title)

        # new_title = re.compile(r'(#)(.*)(#)')
        # print(new_title.sub(r'', title))

    @classmethod
    def kuaishou_sensitive_words(cls):
        sensitive_words = [
            "汽车",
            "电影解说",
            "放映室",
            "解说电影",
            "断供",
        ]
        return sensitive_words

    @classmethod
    def sensitive_words(cls):
        title_list = ["#上海战疫 上海累计感染超20万！这条被淹没的热搜，令全网泪目… 疫情一定要攻克，但所有人都不该遗忘这些弱者。#上海累计报告本土阳性感染者超20万例 #农民工",
                      "#重庆地火村 #地火村 #旅行",
                      "第79集 | 湖南最值得去的六个景区，每一个都是绝色…… #快手带你去旅行 #旅游胜地 #旅游",
                      "霸王条款不废除，断供有多可怕。 #涨知识 #生活小常识 # 生活常识",
                      "秦始皇还活着？地宫中有不明物体缓缓移动 #历史 #秦始皇 #新春寄语  @快手热点(O40300129)",
                      "#夏日荷花  #国花牡丹 #昙花一现",
                      "国内最良心的8个景区，这才是景区最该有的样子，看看你去过几个？ #旅行  #旅游 ",
                      "狗子呆在水里三天三夜，终于练成捕鱼神功，一口一个大鲶鱼 #狗狗  #神奇动物  #快手放映室  @快手热点(O40300129) ",
                      "#集结吧光合创作者  养鸡小伙：喂鸡摆出各种造型，被称为鸡司令。",
                      "89岁农民老艺人自食其力，街头卖艺表演“捏碎碗片”绝技，现场听到咔吱咔吱响，人狠功夫硬！这功夫已失传，以后再看不到了！#集结吧光合创作者 #农民 #街头表演  @快手光合作者助手(O40300118)  @快手热点(O40300129)  @我要上热门(O1907752910)",
                      "我国最贵最有名的三棵树，你知道哪三棵吗？#旅游 #旅行攻略 #黄山迎客松",
                      "潘长江带来热舞，蔡明 郭达也来了！太嗨了！歌词太棒了！ @快手涨粉助手(O1815060199)  @快手热点(O40300129)  @快手平台帐号(O90041) #潘长江 #搞笑 #集结吧光合创作者",
                      "#带你看世界 给大家带来一期烟花盛宴，希望大家能够喜欢，带上你的那个她一起来看吧 #烟花 #视觉震撼"
                      ]
        print(cls.kuaishou_sensitive_words())
        for title in title_list:
            for word in cls.kuaishou_sensitive_words():
                if word in title:
                    print(f"敏感词：{word}")
                    print(f"敏感词视频：{title}")
                    cls.kuaishou_sensitive_words().remove(word)
                else:
                    print(f"正常视频：{title}")
                    cls.kuaishou_sensitive_words().remove(word)


if __name__ == "__main__":
    demo = Demo()
    # demo.demo1()
    demo.time()
    # demo.get_douyin_feeds()
    # demo.demo2()
    # demo.get_weishi_feeds()
    # demo.edit_str()
    # demo.sensitive_words()

    pass