mirror of
https://github.com/yt-dlp/yt-dlp.git
synced 2024-11-01 15:40:15 +00:00
Merge branch 'patch-2' of https://github.com/steven7851/youtube-dl into steven7851-patch-2
This commit is contained in:
commit
545ef4f531
@ -3,9 +3,10 @@
|
||||
|
||||
import hashlib
|
||||
import time
|
||||
import uuid
|
||||
from .common import InfoExtractor
|
||||
from ..utils import (ExtractorError, unescapeHTML)
|
||||
from ..compat import (compat_str, compat_basestring)
|
||||
from ..compat import (compat_str, compat_basestring, compat_urllib_parse_urlencode)
|
||||
|
||||
|
||||
class DouyuTVIE(InfoExtractor):
|
||||
@ -21,7 +22,6 @@ class DouyuTVIE(InfoExtractor):
|
||||
'description': 're:.*m7show@163\.com.*',
|
||||
'thumbnail': 're:^https?://.*\.jpg$',
|
||||
'uploader': '7师傅',
|
||||
'uploader_id': '431925',
|
||||
'is_live': True,
|
||||
},
|
||||
'params': {
|
||||
@ -37,7 +37,6 @@ class DouyuTVIE(InfoExtractor):
|
||||
'description': 'md5:746a2f7a253966a06755a912f0acc0d2',
|
||||
'thumbnail': 're:^https?://.*\.jpg$',
|
||||
'uploader': 'douyu小漠',
|
||||
'uploader_id': '3769985',
|
||||
'is_live': True,
|
||||
},
|
||||
'params': {
|
||||
@ -54,7 +53,6 @@ class DouyuTVIE(InfoExtractor):
|
||||
'description': 're:.*m7show@163\.com.*',
|
||||
'thumbnail': 're:^https?://.*\.jpg$',
|
||||
'uploader': '7师傅',
|
||||
'uploader_id': '431925',
|
||||
'is_live': True,
|
||||
},
|
||||
'params': {
|
||||
@ -75,19 +73,39 @@ def _real_extract(self, url):
|
||||
room_id = self._html_search_regex(
|
||||
r'"room_id"\s*:\s*(\d+),', page, 'room id')
|
||||
|
||||
config = None
|
||||
room_url = 'http://m.douyu.com/html5/live?roomId=%s' % room_id
|
||||
room_content = self._download_webpage(room_url, video_id)
|
||||
room_json = self._parse_json(room_content, video_id, fatal=False)
|
||||
|
||||
room = room_json['data']
|
||||
|
||||
show_status = room.get('show_status')
|
||||
# 1 = live, 2 = offline
|
||||
if show_status == '2':
|
||||
raise ExtractorError(
|
||||
'Live stream is offline', expected=True)
|
||||
|
||||
flv_json = None
|
||||
# Douyu API sometimes returns error "Unable to load the requested class: eticket_redis_cache"
|
||||
# Retry with different parameters - same parameters cause same errors
|
||||
for i in range(5):
|
||||
prefix = 'room/%s?aid=android&client_sys=android&time=%d' % (
|
||||
room_id, int(time.time()))
|
||||
auth = hashlib.md5((prefix + '1231').encode('ascii')).hexdigest()
|
||||
tt = int(time.time() / 60)
|
||||
did = uuid.uuid4().hex.upper()
|
||||
|
||||
config_page = self._download_webpage(
|
||||
'http://www.douyutv.com/api/v1/%s&auth=%s' % (prefix, auth),
|
||||
video_id)
|
||||
# Decompile core.swf in webpage by ffdec "Search SWFs in memory"
|
||||
# core.swf is encrypted originally, but ffdec can dump memory to get the decrypted one
|
||||
# If API changes in the future, just use this way to update
|
||||
sign_content = '{room_id}{did}A12Svb&%1UUmf@hC{tt}'.format(room_id = room_id, did = did, tt = tt)
|
||||
sign = hashlib.md5((sign_content).encode('utf-8')).hexdigest()
|
||||
|
||||
payload = {'cdn': 'ws', 'rate': '0', 'tt': tt, 'did': did, 'sign': sign}
|
||||
flv_data = compat_urllib_parse_urlencode(payload)
|
||||
|
||||
flv_request_url = 'http://www.douyu.com/lapi/live/getPlay/%s' % room_id
|
||||
flv_content = self._download_webpage(flv_request_url, video_id, data=flv_data,
|
||||
headers={'Content-Type': 'application/x-www-form-urlencoded'})
|
||||
try:
|
||||
config = self._parse_json(config_page, video_id, fatal=False)
|
||||
flv_json = self._parse_json(flv_content, video_id, fatal=False)
|
||||
except ExtractorError:
|
||||
# Wait some time before retrying to get a different time() value
|
||||
self._sleep(1, video_id, msg_template='%(video_id)s: Error occurs. '
|
||||
@ -95,54 +113,35 @@ def _real_extract(self, url):
|
||||
continue
|
||||
else:
|
||||
break
|
||||
if config is None:
|
||||
if flv_json is None:
|
||||
raise ExtractorError('Unable to fetch API result')
|
||||
|
||||
data = config['data']
|
||||
flv = flv_json['data']
|
||||
|
||||
error_code = config.get('error', 0)
|
||||
error_code = flv_json.get('error', 0)
|
||||
if error_code is not 0:
|
||||
error_desc = 'Server reported error %i' % error_code
|
||||
if isinstance(data, (compat_str, compat_basestring)):
|
||||
error_desc += ': ' + data
|
||||
if isinstance(flv, (compat_str, compat_basestring)):
|
||||
error_desc += ': ' + flv
|
||||
raise ExtractorError(error_desc, expected=True)
|
||||
|
||||
show_status = data.get('show_status')
|
||||
# 1 = live, 2 = offline
|
||||
if show_status == '2':
|
||||
raise ExtractorError(
|
||||
'Live stream is offline', expected=True)
|
||||
base_url = flv['rtmp_url']
|
||||
live_path = flv['rtmp_live']
|
||||
|
||||
base_url = data['rtmp_url']
|
||||
live_path = data['rtmp_live']
|
||||
video_url = '%s/%s' % (base_url, live_path)
|
||||
|
||||
title = self._live_title(unescapeHTML(data['room_name']))
|
||||
description = data.get('show_details')
|
||||
thumbnail = data.get('room_src')
|
||||
|
||||
uploader = data.get('nickname')
|
||||
uploader_id = data.get('owner_uid')
|
||||
|
||||
multi_formats = data.get('rtmp_multi_bitrate')
|
||||
if not isinstance(multi_formats, dict):
|
||||
multi_formats = {}
|
||||
multi_formats['live'] = live_path
|
||||
|
||||
formats = [{
|
||||
'url': '%s/%s' % (base_url, format_path),
|
||||
'format_id': format_id,
|
||||
'preference': 1 if format_id == 'live' else 0,
|
||||
} for format_id, format_path in multi_formats.items()]
|
||||
self._sort_formats(formats)
|
||||
title = self._live_title(unescapeHTML(room['room_name']))
|
||||
description = room.get('notice')
|
||||
thumbnail = room.get('room_src')
|
||||
uploader = room.get('nickname')
|
||||
|
||||
return {
|
||||
'id': room_id,
|
||||
'display_id': video_id,
|
||||
'url': video_url,
|
||||
'title': title,
|
||||
'description': description,
|
||||
'thumbnail': thumbnail,
|
||||
'uploader': uploader,
|
||||
'uploader_id': uploader_id,
|
||||
'formats': formats,
|
||||
'is_live': True,
|
||||
}
|
||||
|
Loading…
Reference in New Issue
Block a user