]> jfr.im git - yt-dlp.git/blame - yt_dlp/extractor/twitcasting.py
[extractor] Deprecate `_sort_formats`
[yt-dlp.git] / yt_dlp / extractor / twitcasting.py
CommitLineData
da9a60ca 1import base64
e85a3971 2import itertools
036f9051 3import re
4
29f7c58a 5from .common import InfoExtractor
9b8ee23b 6from ..dependencies import websockets
29f7c58a 7from ..utils import (
8 clean_html,
a4a42602 9 ExtractorError,
29f7c58a 10 float_or_none,
11 get_element_by_class,
12 get_element_by_id,
13 parse_duration,
e6779b94 14 qualities,
29f7c58a 15 str_to_int,
fabb27fc 16 traverse_obj,
e85a3971 17 try_get,
29f7c58a 18 unified_timestamp,
19 urlencode_postdata,
e85a3971 20 urljoin,
29f7c58a 21)
22
036f9051 23
cf0db4d9 24class TwitCastingIE(InfoExtractor):
e85a3971 25 _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<uploader_id>[^/]+)/(?:movie|twplayer)/(?P<id>\d+)'
a4a42602
LTHD
26 _M3U8_HEADERS = {
27 'Origin': 'https://twitcasting.tv',
28 'Referer': 'https://twitcasting.tv/',
29 }
88b54749 30 _TESTS = [{
036f9051 31 'url': 'https://twitcasting.tv/ivetesangalo/movie/2357609',
32 'md5': '745243cad58c4681dc752490f7540d7f',
33 'info_dict': {
34 'id': '2357609',
35 'ext': 'mp4',
00a9a25c 36 'title': 'Live #2357609',
036f9051 37 'uploader_id': 'ivetesangalo',
29f7c58a 38 'description': 'Twitter Oficial da cantora brasileira Ivete Sangalo.',
036f9051 39 'thumbnail': r're:^https?://.*\.jpg$',
29f7c58a 40 'upload_date': '20110822',
41 'timestamp': 1314010824,
42 'duration': 32,
43 'view_count': int,
cf0db4d9
S
44 },
45 'params': {
46 'skip_download': True,
47 },
88b54749
MZ
48 }, {
49 'url': 'https://twitcasting.tv/mttbernardini/movie/3689740',
50 'info_dict': {
51 'id': '3689740',
52 'ext': 'mp4',
53 'title': 'Live playing something #3689740',
54 'uploader_id': 'mttbernardini',
29f7c58a 55 'description': 'Salve, io sono Matto (ma con la e). Questa è la mia presentazione, in quanto sono letteralmente matto (nel senso di strano), con qualcosa in più.',
88b54749 56 'thumbnail': r're:^https?://.*\.jpg$',
29f7c58a 57 'upload_date': '20120212',
58 'timestamp': 1329028024,
59 'duration': 681,
60 'view_count': int,
88b54749
MZ
61 },
62 'params': {
63 'skip_download': True,
64 'videopassword': 'abc',
65 },
a4a42602
LTHD
66 }, {
67 'note': 'archive is split in 2 parts',
68 'url': 'https://twitcasting.tv/loft_heaven/movie/685979292',
69 'info_dict': {
70 'id': '685979292',
71 'ext': 'mp4',
72 'title': '南波一海のhear_here “ナタリー望月哲さんに聞く編集と「渋谷系狂騒曲」”',
73 'duration': 6964.599334,
74 },
75 'playlist_mincount': 2,
88b54749 76 }]
036f9051 77
da9a60ca
L
78 def _parse_data_movie_playlist(self, dmp, video_id):
79 # attempt 1: parse as JSON directly
80 try:
81 return self._parse_json(dmp, video_id)
82 except ExtractorError:
83 pass
84 # attempt 2: decode reversed base64
85 decoded = base64.b64decode(dmp[::-1])
86 return self._parse_json(decoded, video_id)
87
036f9051 88 def _real_extract(self, url):
5ad28e7f 89 uploader_id, video_id = self._match_valid_url(url).groups()
036f9051 90
a06916d9 91 video_password = self.get_param('videopassword')
88b54749
MZ
92 request_data = None
93 if video_password:
94 request_data = urlencode_postdata({
95 'password': video_password,
5a13fdd2 96 }, encoding='utf-8')
df635a09 97 webpage, urlh = self._download_webpage_handle(
d0491a1e 98 url, video_id, data=request_data,
99 headers={'Origin': 'https://twitcasting.tv'})
df635a09
LNO
100 if urlh.geturl() != url and request_data:
101 webpage = self._download_webpage(
102 urlh.geturl(), video_id, data=request_data,
103 headers={'Origin': 'https://twitcasting.tv'},
104 note='Retrying authentication')
f099df14
LNO
105 # has to check here as the first request can contain password input form even if the password is correct
106 if re.search(r'<form\s+method="POST">\s*<input\s+[^>]+?name="password"', webpage):
107 raise ExtractorError('This video is protected by a password, use the --video-password option', expected=True)
036f9051 108
e85a3971 109 title = (clean_html(get_element_by_id('movietitle', webpage))
110 or self._html_search_meta(['og:title', 'twitter:title'], webpage, fatal=True))
cf0db4d9 111
a4a42602
LTHD
112 video_js_data = try_get(
113 webpage,
da9a60ca 114 lambda x: self._parse_data_movie_playlist(self._search_regex(
a4a42602
LTHD
115 r'data-movie-playlist=\'([^\']+?)\'',
116 x, 'movie playlist', default=None), video_id)['2'], list)
fabb27fc 117
a4a42602 118 thumbnail = traverse_obj(video_js_data, (0, 'thumbnailUrl')) or self._og_search_thumbnail(webpage)
29f7c58a 119 description = clean_html(get_element_by_id(
120 'authorcomment', webpage)) or self._html_search_meta(
121 ['description', 'og:description', 'twitter:description'], webpage)
a4a42602
LTHD
122 duration = (try_get(video_js_data, lambda x: sum(float_or_none(y.get('duration')) for y in x) / 1000)
123 or parse_duration(clean_html(get_element_by_class('tw-player-duration-time', webpage))))
29f7c58a 124 view_count = str_to_int(self._search_regex(
a4a42602 125 (r'Total\s*:\s*([\d,]+)\s*Views', r'総視聴者\s*:\s*([\d,]+)\s*</'), webpage, 'views', None))
29f7c58a 126 timestamp = unified_timestamp(self._search_regex(
127 r'data-toggle="true"[^>]+datetime="([^"]+)"',
128 webpage, 'datetime', None))
cf0db4d9 129
a4a42602
LTHD
130 stream_server_data = self._download_json(
131 'https://twitcasting.tv/streamserver.php?target=%s&mode=client' % uploader_id, video_id,
132 'Downloading live info', fatal=False)
e85a3971 133
a4a42602
LTHD
134 is_live = 'data-status="online"' in webpage
135 if not traverse_obj(stream_server_data, 'llfmp4') and is_live:
136 self.raise_login_required(method='cookies')
137
138 base_dict = {
036f9051 139 'title': title,
140 'description': description,
141 'thumbnail': thumbnail,
29f7c58a 142 'timestamp': timestamp,
036f9051 143 'uploader_id': uploader_id,
29f7c58a 144 'duration': duration,
145 'view_count': view_count,
e85a3971 146 'is_live': is_live,
036f9051 147 }
e85a3971 148
a4a42602
LTHD
149 def find_dmu(x):
150 data_movie_url = self._search_regex(
151 r'data-movie-url=(["\'])(?P<url>(?:(?!\1).)+)\1',
152 x, 'm3u8 url', group='url', default=None)
153 if data_movie_url:
154 return [data_movie_url]
155
156 m3u8_urls = (try_get(webpage, find_dmu, list)
157 or traverse_obj(video_js_data, (..., 'source', 'url'))
158 or ([f'https://twitcasting.tv/{uploader_id}/metastream.m3u8'] if is_live else None))
159 if not m3u8_urls:
160 raise ExtractorError('Failed to get m3u8 playlist')
161
162 if is_live:
163 m3u8_url = m3u8_urls[0]
164 formats = self._extract_m3u8_formats(
165 m3u8_url, video_id, ext='mp4', m3u8_id='hls',
166 live=True, headers=self._M3U8_HEADERS)
167
df635a09
LNO
168 if traverse_obj(stream_server_data, ('hls', 'source')):
169 formats.extend(self._extract_m3u8_formats(
170 m3u8_url, video_id, ext='mp4', m3u8_id='source',
171 live=True, query={'mode': 'source'},
172 note='Downloading source quality m3u8',
173 headers=self._M3U8_HEADERS, fatal=False))
a4a42602 174
9b8ee23b 175 if websockets:
a4a42602
LTHD
176 qq = qualities(['base', 'mobilesource', 'main'])
177 streams = traverse_obj(stream_server_data, ('llfmp4', 'streams')) or {}
178 for mode, ws_url in streams.items():
179 formats.append({
180 'url': ws_url,
181 'format_id': 'ws-%s' % mode,
182 'ext': 'mp4',
183 'quality': qq(mode),
df635a09 184 'source_preference': -10,
a4a42602
LTHD
185 # TwitCasting simply sends moof atom directly over WS
186 'protocol': 'websocket_frag',
187 })
188
a4a42602 189 infodict = {
9f14daf2 190 'formats': formats,
191 '_format_sort_fields': ('source', ),
a4a42602 192 }
06b1628d
LNO
193 elif len(m3u8_urls) == 1:
194 formats = self._extract_m3u8_formats(
195 m3u8_urls[0], video_id, 'mp4', headers=self._M3U8_HEADERS)
06b1628d
LNO
196 infodict = {
197 # No problem here since there's only one manifest
198 'formats': formats,
997378f9 199 'http_headers': self._M3U8_HEADERS,
06b1628d 200 }
a4a42602
LTHD
201 else:
202 infodict = {
203 '_type': 'multi_video',
204 'entries': [{
205 'id': f'{video_id}-{num}',
206 'url': m3u8_url,
207 'ext': 'mp4',
208 # Requesting the manifests here will cause download to fail.
209 # So use ffmpeg instead. See: https://github.com/yt-dlp/yt-dlp/issues/382
210 'protocol': 'm3u8',
211 'http_headers': self._M3U8_HEADERS,
212 **base_dict,
213 } for (num, m3u8_url) in enumerate(m3u8_urls)],
214 }
215
216 return {
217 'id': video_id,
218 **base_dict,
219 **infodict,
220 }
221
e85a3971 222
223class TwitCastingLiveIE(InfoExtractor):
224 _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<id>[^/]+)/?(?:[#?]|$)'
225 _TESTS = [{
226 'url': 'https://twitcasting.tv/ivetesangalo',
227 'only_matching': True,
228 }]
229
230 def _real_extract(self, url):
231 uploader_id = self._match_id(url)
232 self.to_screen(
233 'Downloading live video of user {0}. '
234 'Pass "https://twitcasting.tv/{0}/show" to download the history'.format(uploader_id))
235
236 webpage = self._download_webpage(url, uploader_id)
237 current_live = self._search_regex(
238 (r'data-type="movie" data-id="(\d+)">',
239 r'tw-sound-flag-open-link" data-id="(\d+)" style=',),
240 webpage, 'current live ID', default=None)
07ea0014
LNO
241 if not current_live:
242 # fetch unfiltered /show to find running livestreams; we can't get ID of the password-protected livestream above
243 webpage = self._download_webpage(
244 f'https://twitcasting.tv/{uploader_id}/show/', uploader_id,
245 note='Downloading live history')
246 is_live = self._search_regex(r'(?s)(<span\s*class="tw-movie-thumbnail-badge"\s*data-status="live">\s*LIVE)', webpage, 'is live?', default=None)
247 if is_live:
248 # get the first live; running live is always at the first
249 current_live = self._search_regex(
250 r'(?s)<a\s+class="tw-movie-thumbnail"\s*href="/[^/]+/movie/(?P<video_id>\d+)"\s*>.+?</a>',
251 webpage, 'current live ID 2', default=None, group='video_id')
e85a3971 252 if not current_live:
253 raise ExtractorError('The user is not currently live')
254 return self.url_result('https://twitcasting.tv/%s/movie/%s' % (uploader_id, current_live))
255
256
257class TwitCastingUserIE(InfoExtractor):
258 _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<id>[^/]+)/show/?(?:[#?]|$)'
259 _TESTS = [{
260 'url': 'https://twitcasting.tv/noriyukicas/show',
261 'only_matching': True,
262 }]
263
264 def _entries(self, uploader_id):
265 base_url = next_url = 'https://twitcasting.tv/%s/show' % uploader_id
266 for page_num in itertools.count(1):
267 webpage = self._download_webpage(
268 next_url, uploader_id, query={'filter': 'watchable'}, note='Downloading page %d' % page_num)
269 matches = re.finditer(
270 r'''(?isx)<a\s+class="tw-movie-thumbnail"\s*href="(?P<url>/[^/]+/movie/\d+)"\s*>.+?</a>''',
271 webpage)
272 for mobj in matches:
273 yield self.url_result(urljoin(base_url, mobj.group('url')))
274
275 next_url = self._search_regex(
276 r'<a href="(/%s/show/%d-\d+)[?"]' % (re.escape(uploader_id), page_num),
277 webpage, 'next url', default=None)
278 next_url = urljoin(base_url, next_url)
279 if not next_url:
280 return
281
282 def _real_extract(self, url):
283 uploader_id = self._match_id(url)
284 return self.playlist_result(
285 self._entries(uploader_id), uploader_id, '%s - Live History' % uploader_id)