yt-dlp/yt_dlp/extractor/huya.py

# coding: utf-8
from __future__ import unicode_literals

import hashlib
import random

from ..compat import compat_urlparse, compat_b64decode

from ..utils import (
    ExtractorError,
    int_or_none,
    js_to_json,
    str_or_none,
    try_get,
    unescapeHTML,
    update_url_query,
)

from .common import InfoExtractor


class HuyaLiveIE(InfoExtractor):
    _VALID_URL = r'https?://(?:www\.|m\.)?huya\.com/(?P<id>[^/#?&]+)(?:\D|$)'
    IE_NAME = 'huya:live'
    IE_DESC = 'huya.com'
    TESTS = [{
        'url': 'https://www.huya.com/572329',
        'info_dict': {
            'id': '572329',
            'title': str,
            'description': str,
            'is_live': True,
            'view_count': int,
        },
        'params': {
            'skip_download': True,
        },
    }, {
        'url': 'https://www.huya.com/xiaoyugame',
        'only_matching': True
    }]

    _RESOLUTION = {
        '蓝光4M': {
            'width': 1920,
            'height': 1080,
        },
        '超清': {
            'width': 1280,
            'height': 720,
        },
        '流畅': {
            'width': 800,
            'height': 480
        }
    }

    def _real_extract(self, url):
        video_id = self._match_id(url)
        webpage = self._download_webpage(url, video_id=video_id)
        json_stream = self._search_regex(r'"stream":\s+"([a-zA-Z0-9+=/]+)"', webpage, 'stream', default=None)
        if not json_stream:
            raise ExtractorError('Video is offline', expected=True)
        stream_data = self._parse_json(compat_b64decode(json_stream).decode(), video_id=video_id,
                                       transform_source=js_to_json)
        room_info = try_get(stream_data, lambda x: x['data'][0]['gameLiveInfo'])
        if not room_info:
            raise ExtractorError('Can not extract the room info', expected=True)
        title = room_info.get('roomName') or room_info.get('introduction') or self._html_extract_title(webpage)
        screen_type = room_info.get('screenType')
        live_source_type = room_info.get('liveSourceType')
        stream_info_list = stream_data['data'][0]['gameStreamInfoList']
        formats = []
        for stream_info in stream_info_list:
            stream_url = stream_info.get('sFlvUrl')
            if not stream_url:
                continue
            stream_name = stream_info.get('sStreamName')
            re_secret = not screen_type and live_source_type in (0, 8, 13)
            params = dict(compat_urlparse.parse_qsl(unescapeHTML(stream_info['sFlvAntiCode'])))
            fm, ss = '', ''
            if re_secret:
                fm, ss = self.encrypt(params, stream_info, stream_name)
            for si in stream_data.get('vMultiStreamInfo'):
                rate = si.get('iBitRate')
                if rate:
                    params['ratio'] = rate
                else:
                    params.pop('ratio', None)
                if re_secret:
                    params['wsSecret'] = hashlib.md5(
                        '_'.join([fm, params['u'], stream_name, ss, params['wsTime']]))
                formats.append({
                    'ext': stream_info.get('sFlvUrlSuffix'),
                    'format_id': str_or_none(stream_info.get('iLineIndex')),
                    'tbr': rate,
                    'url': update_url_query(f'{stream_url}/{stream_name}.{stream_info.get("sFlvUrlSuffix")}',
                                            query=params),
                    **self._RESOLUTION.get(si.get('sDisplayName'), {}),
                })

        self._sort_formats(formats)

        return {
            'id': video_id,
            'title': title,
            'formats': formats,
            'view_count': room_info.get('totalCount'),
            'thumbnail': room_info.get('screenshot'),
            'description': room_info.get('contentIntro'),
            'http_headers': {
                'Origin': 'https://www.huya.com',
                'Referer': 'https://www.huya.com/',
            },
        }

    def encrypt(self, params, stream_info, stream_name):
        ct = int_or_none(params.get('wsTime'), 16) + random.random()
        presenter_uid = stream_info['lPresenterUid']
        if not stream_name.startswith(str(presenter_uid)):
            uid = presenter_uid
        else:
            uid = int_or_none(ct % 1e7 * 1e6 % 0xffffffff)
        u1 = uid & 0xffffffff00000000
        u2 = uid & 0xffffffff
        u3 = uid & 0xffffff
        u = u1 | u2 >> 24 | u3 << 8
        params.update({
            'u': str_or_none(u),
            'seqid': str_or_none(int_or_none(ct * 1000) + uid),
            'ver': '1',
            'uuid': int_or_none(ct % 1e7 * 1e6 % 0xffffffff),
            't': '100',
        })
        fm = compat_b64decode(params['fm']).decode().split('_', 1)[0]
        ss = hashlib.md5('|'.join([params['seqid'], params['ctype'], params['t']]))
        return fm, ss