crawler_xigua/main/demo.py

# -*- coding: utf-8 -*-
# @Author: wangkun
# @Time: 2022/8/23
import ast
import base64
import json
import os
import subprocess
import requests
import urllib3
from main.feishu import Feishu

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


class Demo:
    @classmethod
    def base64_decode(cls):
        s = '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'
        base64_url = base64.b64decode(s).decode('utf8')
        print(base64_url)

    @classmethod
    def ffmpeg(cls):
        video_path = '../videos/陕西第二支预告片，武则天乾陵，气势磅礴，夕阳下拍到神奇的一幕/video1.mp4'
        audio_path = '../videos/陕西第二支预告片，武则天乾陵，气势磅礴，夕阳下拍到神奇的一幕/audio1.mp4'
        out_path = '../videos/陕西第二支预告片，武则天乾陵，气势磅礴，夕阳下拍到神奇的一幕/video.mp4'
        cmd = 'ffmpeg -i ' + video_path + ' -i ' + audio_path + ' -c:v copy -c:a aac -strict experimental -map 0:v:0 -map 1:a:0 ' + out_path
        print(cmd)
        subprocess.call(cmd, shell=True)

        print(os.listdir('../videos/陕西第二支预告片，武则天乾陵，气势磅礴，夕阳下拍到神奇的一幕/'))
        for file in os.listdir('../videos/陕西第二支预告片，武则天乾陵，气势磅礴，夕阳下拍到神奇的一幕/'):
            # print(file.title())
            if file.split('.mp4')[0] == 'video1' or file.split('.mp4')[0] == 'audio1':
                os.remove('../videos/陕西第二支预告片，武则天乾陵，气势磅礴，夕阳下拍到神奇的一幕/' + file)

    @classmethod
    def sheet(cls, log_type):
        sheet = Feishu.get_values_batch(log_type, 'xigua', 'e075e9')
        print(sheet)

    @classmethod
    def get_video_info(cls, gid):
        url = 'https://www.ixigua.com/api/mixVideo/information?'
        headers = {
            "accept-encoding": "gzip, deflate, br",
            "accept-language": "zh-CN,zh-Hans;q=0.9",
            "user-agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) "
                          "AppleWebKit/605.1.15 (KHTML, like Gecko) Version/15.5 Safari/605.1.15",
            "referer": "https://www.ixigua.com/7102614741050196520?logTag=0531c88ac04f38ab2c62",
        }
        params = {
            'mixId': gid,
            'msToken': 'IlG0wd0Pylyw9ghcYiB2YseUmTwrsrqqhXrbIcsSaTcLTJyVlbYJzk20zw3UO-CfrfC'
                       'NVVIOBNjIl7vfBoxnVUwO9ZyzAI3umSKsT5-pef_RRfQCJwmA',
            'X-Bogus': 'DFSzswVupYTANCJOSBk0P53WxM-r',
            '_signature': '_02B4Z6wo0000119LvEwAAIDCuktNZ0y5wkdfS7jAALThuOR8D9yWNZ.EmWHKV0WSn6Px'
                          'fPsH9-BldyxVje0f49ryXgmn7Tzk-swEHNb15TiGqa6YF.cX0jW8Eds1TtJOIZyfc9s5emH7gdWN94',
        }
        cookies = {
            'ixigua-a-s': '1',
            'msToken': 'IlG0wd0Pylyw9ghcYiB2YseUmTwrsrqqhXrbIcsSaTcLTJyVlbYJzk20zw3UO-CfrfCNVVIOB'
                       'NjIl7vfBoxnVUwO9ZyzAI3umSKsT5-pef_RRfQCJwmA',
            'ttwid': '1%7C_yXQeHWwLZgCsgHClOwTCdYSOt_MjdOkgnPIkpi-Sr8%7C1661241238%7Cf57d0c5ef3f1d7'
                     '6e049fccdca1ac54887c34d1f8731c8e51a49780ff0ceab9f8',
            'tt_scid': 'QZ4l8KXDG0YAEaMCSbADdcybdKbUfG4BC6S4OBv9lpRS5VyqYLX2bIR8CTeZeGHR9ee3',
            'MONITOR_WEB_ID': '0a49204a-7af5-4e96-95f0-f4bafb7450ad',
            '__ac_nonce': '06304878000964fdad287',
            '__ac_signature': '_02B4Z6wo00f017Rcr3AAAIDCUVxeW1tOKEu0fKvAAI4cvoYzV-wBhq7B6D8k0no7lb'
                              'FlvYoinmtK6UXjRIYPXnahUlFTvmWVtb77jsMkKAXzAEsLE56m36RlvL7ky.M3Xn52r9t1IEb7IR3ke8',
            'ttcid': 'e56fabf6e85d4adf9e4d91902496a0e882',
            '_tea_utm_cache_1300': 'undefined',
            'support_avif': 'false',
            'support_webp': 'false',
            'xiguavideopcwebid': '7134967546256016900',
            'xiguavideopcwebid.sig': 'xxRww5R1VEMJN_dQepHorEu_eAc',
        }
        urllib3.disable_warnings()
        response = requests.get(url=url, headers=headers, params=params, cookies=cookies, proxies=proxies, verify=False)
        if 'data' not in response.json() or response.json()['data'] == '':
            print(f'response: {response}')
        else:
            video_info = response.json()['data']['gidInformation']['packerData']['video']

            video_url = ''
            audio_url = ''
            video_width = ''
            video_height = ''

            # video_url
            if 'videoResource' not in video_info:
                video_url = 0
                audio_url = 0
                video_width = 0
                video_height = 0
            elif 'dash' in video_info['videoResource']:
                video_url = video_info['videoResource']['dash']['dynamic_video']['dynamic_video_list'][-1]['main_url']
                audio_url = video_info['videoResource']['dash']['dynamic_video']['dynamic_audio_list'][-1]['main_url']
                video_url = base64.b64decode(video_url).decode('utf8')
                audio_url = base64.b64decode(audio_url).decode('utf8')
                video_width = video_info['videoResource']['dash']['dynamic_video']['dynamic_video_list'][-1]['vwidth']
                video_height = video_info['videoResource']['dash']['dynamic_video']['dynamic_video_list'][-1]['vheight']

            elif 'normal' in video_info['videoResource']:
                video_list = video_info['videoResource']['normal']['video_list']
                if 'video_4' in video_list.keys():
                    video_url = video_list['video_4']['main_url']
                    audio_url = video_list['video_4']['main_url']
                    video_url = base64.b64decode(video_url).decode('utf8')
                    audio_url = base64.b64decode(audio_url).decode('utf8')
                    video_width = video_list['video_4']['vwidth']
                    video_height = video_list['video_4']['vheight']
                elif 'video_3' in video_list.keys():
                    video_url = video_list['video_3']['main_url']
                    audio_url = video_list['video_3']['main_url']
                    video_url = base64.b64decode(video_url).decode('utf8')
                    audio_url = base64.b64decode(audio_url).decode('utf8')
                    video_width = video_list['video_3']['vwidth']
                    video_height = video_list['video_3']['vheight']
                elif 'video_2' in video_list.keys():
                    video_url = video_list['video_2']['main_url']
                    audio_url = video_list['video_2']['main_url']
                    video_url = base64.b64decode(video_url).decode('utf8')
                    audio_url = base64.b64decode(audio_url).decode('utf8')
                    video_width = video_list['video_2']['vwidth']
                    video_height = video_list['video_2']['vheight']
                elif 'video_1' in video_list.keys():
                    video_url = video_list['video_1']['main_url']
                    audio_url = video_list['video_1']['main_url']
                    video_url = base64.b64decode(video_url).decode('utf8')
                    audio_url = base64.b64decode(audio_url).decode('utf8')
                    video_width = video_list['video_1']['vwidth']
                    video_height = video_list['video_1']['vheight']
                else:
                    video_url = 0
                    audio_url = 0
                    video_width = 0
                    video_height = 0


            print(f'video_url:{video_url}')
            print(f'audio_url:{audio_url}')
            print(f'video_width:{video_width}')
            print(f'video_height:{video_height}')

    @classmethod
    def get_signature(cls):
        # line = {'level': 'INFO',
        #         'message': '{"message":{"method":"Network.responseReceived","params":{"frameId":"A577824F4AEC626A0E7B1E67515C020B","hasExtraInfo":true,"loaderId":"4C8B8024778C547E73056E298C1199CF","requestId":"20233.355","response":{"connectionId":13,"connectionReused":true,"encodedDataLength":572,"fromDiskCache":false,"fromPrefetchCache":false,"fromServiceWorker":false,"headers":{"content-length":"53","content-type":"application/json; charset=utf-8","date":"Thu, 22 Sep 2022 06:21:39 GMT","eagleid":"7c8486d116638276996804276e","server":"Tengine","server-timing":"inner; dur=40, tt_agw; dur=39, cdn-cache;desc=MISS,edge;dur=10,origin;dur=65","strict-transport-security":"max-age=31536000; includeSubDomains","timing-allow-origin":"*","via":"cache31.l2cm9-6[65,0], cache9.cn2049[75,0]","x-agw-info":"xVgLW9WxHAAG33_khm6L2a8ReWONzX5A1dCyebQN9pV1gLCk8dbQVtZIfVMaO031QpBph1oWkhQ6czSXetfTPHLeDeO-Mct8PB8_S1YfgeNXqqjKQOb9DI-AygaTj86mkA9zn3lILbkE6LDqjB0bu9Ap9JlnN3laAz3Geb9M4xZ759bsHBTFySc=","x-content-type-options":"nosniff","x-download-options":"noopen","x-frame-options":"SAMEORIGIN","x-trans-level":"0","x-tt-logid":"20220922142139010151190080030D8357","x-tt-trace-host":"01d3e0070115308d6e7072e1066a5ff763a5d3a58f00e80e220c124847e6eeb9e70d5462c3b67b716c6b5bf35c4f08d47888aaa9af3610c58aac0d0d0ca7ce5d4cef54fa5205a53e5d7ff2f40cc122d0b128cb2b9fb83f9e6a84965a6ad8c66cb3","x-tt-trace-tag":"id=03;cdn-cache=miss;type=dyn","x-xss-protection":"1; mode=block"},"mimeType":"application/json","protocol":"h2","remoteIPAddress":"124.132.134.227","remotePort":443,"responseTime":1.663827699692927e+12,"securityDetails":{"certificateId":0,"certificateTransparencyCompliance":"unknown","cipher":"AES_256_GCM","encryptedClientHello":false,"issuer":"RapidSSL Global TLS RSA4096 SHA256 2022 CA1","keyExchange":"","keyExchangeGroup":"X25519","protocol":"TLS 1.3","sanList":["*.ixigua.com","ixigua.com"],"serverSignatureAlgorithm":2052,"signedCertificateTimestampList":[],"subjectName":"*.ixigua.com","validFrom":1658102400,"validTo":1689638399},"securityState":"secure","status":200,"statusText":"","timing":{"connectEnd":-1,"connectStart":-1,"dnsEnd":-1,"dnsStart":-1,"proxyEnd":-1,"proxyStart":-1,"pushEnd":0,"pushStart":0,"receiveHeadersEnd":109.989,"requestTime":255819.986241,"sendEnd":0.355,"sendStart":0.177,"sslEnd":-1,"sslStart":-1,"workerFetchStart":-1,"workerReady":-1,"workerRespondWithSettled":-1,"workerStart":-1},"url":"https://www.ixigua.com/at/log/c?msToken=G687B1o_7DTeKdhq-9JMQbDxhK1hzpcdN6DTLUn6SWIq9Xo3ziv-DWrsNqqK8PBrvFzrURV18gBiG82keZcfMdht0qng_bRzQhooy8nwITUATUqn5Uqh0OxC2_XRX9j1vy71&X-Bogus=DFSzKwROEFtBHb-lSsScqggX1TFT&_signature=_02B4Z6wo00001v7TvHwAAIDAjPVk4Z9W.GL-w5jAANysmUCNpvb5Srvg6eExeryhDFeyGu4zoN-PpB3S4tD46UJWh9tF.cXH0ggjdznFn8lC79XMTEkl87e.QyQ80f5TS2h0bl5J77lPHbrN59"},"timestamp":255820.192025,"type":"XHR"}},"webview":"A577824F4AEC626A0E7B1E67515C020B"}',
        #         'timestamp': 1663827699791}
        # message = line['message']
        # message = json.loads(message)
        # url = message['message']['params']['response']['url']
        # print(url)
        # signature = url.split('_signature=')[-1]
        # print(signature)
        s = '_02B4Z6wo000017YitjQAAIDBxARuqbJSv0e2MpKAAI6tf5'
        print(len(s))

    @classmethod
    def get_logs(cls):
        logs = ''


if __name__ == '__main__':
    # Demo.base64_decode()
    # Demo.ffmpeg()
    # Demo.sheet('xigua')
    # Demo.get_video_info('7133153239867851295')
    # Demo.get_signature()
    Demo.get_logs()

    pass