X-Git-Url: https://jfr.im/git/yt-dlp.git/blobdiff_plain/5a13fdd2254d3dde52818e07cd7c5d88a74a6125..f2816634e3be88fe158b342ee33918de3c272a54:/yt_dlp/extractor/twitcasting.py diff --git a/yt_dlp/extractor/twitcasting.py b/yt_dlp/extractor/twitcasting.py index 8c2235a8e..28ea16cc2 100644 --- a/yt_dlp/extractor/twitcasting.py +++ b/yt_dlp/extractor/twitcasting.py @@ -1,14 +1,13 @@ -# coding: utf-8 -from __future__ import unicode_literals - +import base64 import itertools import re from .common import InfoExtractor -from ..downloader.websocket import has_websockets +from ..dependencies import websockets from ..utils import ( - clean_html, ExtractorError, + UserNotLive, + clean_html, float_or_none, get_element_by_class, get_element_by_id, @@ -24,7 +23,7 @@ class TwitCastingIE(InfoExtractor): - _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P[^/]+)/(?:movie|twplayer)/(?P\d+)' + _VALID_URL = r'https?://(?:[^/?#]+\.)?twitcasting\.tv/(?P[^/?#]+)/(?:movie|twplayer)/(?P\d+)' _M3U8_HEADERS = { 'Origin': 'https://twitcasting.tv', 'Referer': 'https://twitcasting.tv/', @@ -40,7 +39,7 @@ class TwitCastingIE(InfoExtractor): 'description': 'Twitter Oficial da cantora brasileira Ivete Sangalo.', 'thumbnail': r're:^https?://.*\.jpg$', 'upload_date': '20110822', - 'timestamp': 1314010824, + 'timestamp': 1313978424, 'duration': 32, 'view_count': int, }, @@ -54,10 +53,10 @@ class TwitCastingIE(InfoExtractor): 'ext': 'mp4', 'title': 'Live playing something #3689740', 'uploader_id': 'mttbernardini', - 'description': 'Salve, io sono Matto (ma con la e). Questa è la mia presentazione, in quanto sono letteralmente matto (nel senso di strano), con qualcosa in più.', + 'description': 'md5:1dc7efa2f1ab932fcd119265cebeec69', 'thumbnail': r're:^https?://.*\.jpg$', - 'upload_date': '20120212', - 'timestamp': 1329028024, + 'upload_date': '20120211', + 'timestamp': 1328995624, 'duration': 681, 'view_count': int, }, @@ -66,36 +65,64 @@ class TwitCastingIE(InfoExtractor): 'videopassword': 'abc', }, }, { - 'note': 'archive is split in 2 parts', 'url': 'https://twitcasting.tv/loft_heaven/movie/685979292', 'info_dict': { 'id': '685979292', 'ext': 'mp4', - '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」”', - 'duration': 6964.599334, + 'title': '【無料配信】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」”', + 'uploader_id': 'loft_heaven', + 'description': 'md5:3a0c7b53019df987ce545c935538bacf', + 'upload_date': '20210604', + 'timestamp': 1622802114, + 'thumbnail': r're:^https?://.*\.jpg$', + 'duration': 6964, + 'view_count': int, + }, + 'params': { + 'skip_download': True, }, - 'playlist_mincount': 2, }] + def _parse_data_movie_playlist(self, dmp, video_id): + # attempt 1: parse as JSON directly + try: + return self._parse_json(dmp, video_id) + except ExtractorError: + pass + # attempt 2: decode reversed base64 + decoded = base64.b64decode(dmp[::-1]) + return self._parse_json(decoded, video_id) + def _real_extract(self, url): uploader_id, video_id = self._match_valid_url(url).groups() + webpage, urlh = self._download_webpage_handle(url, video_id) video_password = self.get_param('videopassword') request_data = None if video_password: request_data = urlencode_postdata({ 'password': video_password, + **self._hidden_inputs(webpage), }, encoding='utf-8') - webpage = self._download_webpage( - url, video_id, data=request_data, - headers={'Origin': 'https://twitcasting.tv'}) + webpage, urlh = self._download_webpage_handle( + url, video_id, data=request_data, + headers={'Origin': 'https://twitcasting.tv'}, + note='Trying video password') + if urlh.url != url and request_data: + webpage = self._download_webpage( + urlh.url, video_id, data=request_data, + headers={'Origin': 'https://twitcasting.tv'}, + note='Retrying authentication') + # has to check here as the first request can contain password input form even if the password is correct + if re.search(r'\s*]+?name="password"', webpage): + raise ExtractorError('This video is protected by a password, use the --video-password option', expected=True) title = (clean_html(get_element_by_id('movietitle', webpage)) or self._html_search_meta(['og:title', 'twitter:title'], webpage, fatal=True)) video_js_data = try_get( webpage, - lambda x: self._parse_json(self._search_regex( + lambda x: self._parse_data_movie_playlist(self._search_regex( r'data-movie-playlist=\'([^\']+?)\'', x, 'movie playlist', default=None), video_id)['2'], list) @@ -106,7 +133,7 @@ def _real_extract(self, url): duration = (try_get(video_js_data, lambda x: sum(float_or_none(y.get('duration')) for y in x) / 1000) or parse_duration(clean_html(get_element_by_class('tw-player-duration-time', webpage)))) view_count = str_to_int(self._search_regex( - (r'Total\s*:\s*([\d,]+)\s*Views', r'総視聴者\s*:\s*([\d,]+)\s*]+datetime="([^"]+)"', webpage, 'datetime', None)) @@ -115,7 +142,7 @@ def _real_extract(self, url): 'https://twitcasting.tv/streamserver.php?target=%s&mode=client' % uploader_id, video_id, 'Downloading live info', fatal=False) - is_live = 'data-status="online"' in webpage + is_live = any(f'data-{x}' in webpage for x in ['is-onlive="true"', 'live-type="live"', 'status="online"']) if not traverse_obj(stream_server_data, 'llfmp4') and is_live: self.raise_login_required(method='cookies') @@ -149,13 +176,14 @@ def find_dmu(x): m3u8_url, video_id, ext='mp4', m3u8_id='hls', live=True, headers=self._M3U8_HEADERS) - formats.extend(self._extract_m3u8_formats( - m3u8_url, video_id, ext='mp4', m3u8_id='source', - live=True, query={'mode': 'source'}, - note='Downloading source quality m3u8', - headers=self._M3U8_HEADERS, fatal=False)) + if traverse_obj(stream_server_data, ('hls', 'source')): + formats.extend(self._extract_m3u8_formats( + m3u8_url, video_id, ext='mp4', m3u8_id='source', + live=True, query={'mode': 'source'}, + note='Downloading source quality m3u8', + headers=self._M3U8_HEADERS, fatal=False)) - if has_websockets: + if websockets: qq = qualities(['base', 'mobilesource', 'main']) streams = traverse_obj(stream_server_data, ('llfmp4', 'streams')) or {} for mode, ws_url in streams.items(): @@ -164,14 +192,22 @@ def find_dmu(x): 'format_id': 'ws-%s' % mode, 'ext': 'mp4', 'quality': qq(mode), + 'source_preference': -10, # TwitCasting simply sends moof atom directly over WS 'protocol': 'websocket_frag', }) - self._sort_formats(formats) - infodict = { - 'formats': formats + 'formats': formats, + '_format_sort_fields': ('source', ), + } + elif len(m3u8_urls) == 1: + formats = self._extract_m3u8_formats( + m3u8_urls[0], video_id, 'mp4', headers=self._M3U8_HEADERS) + infodict = { + # No problem here since there's only one manifest + 'formats': formats, + 'http_headers': self._M3U8_HEADERS, } else: infodict = { @@ -196,10 +232,13 @@ def find_dmu(x): class TwitCastingLiveIE(InfoExtractor): - _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P[^/]+)/?(?:[#?]|$)' + _VALID_URL = r'https?://(?:[^/?#]+\.)?twitcasting\.tv/(?P[^/?#]+)/?(?:[#?]|$)' _TESTS = [{ 'url': 'https://twitcasting.tv/ivetesangalo', 'only_matching': True, + }, { + 'url': 'https://twitcasting.tv/c:unusedlive', + 'expected_exception': 'UserNotLive', }] def _real_extract(self, url): @@ -208,19 +247,38 @@ def _real_extract(self, url): 'Downloading live video of user {0}. ' 'Pass "https://twitcasting.tv/{0}/show" to download the history'.format(uploader_id)) - webpage = self._download_webpage(url, uploader_id) + is_live = traverse_obj(self._download_json( + f'https://frontendapi.twitcasting.tv/watch/user/{uploader_id}', + uploader_id, 'Checking live status', data=b'', fatal=False), ('is_live', {bool})) + if is_live is False: # only raise here if API response was as expected + raise UserNotLive(video_id=uploader_id) + + # Use /show/ page so that password-protected and members-only livestreams can be found + webpage = self._download_webpage( + f'https://twitcasting.tv/{uploader_id}/show/', uploader_id, 'Downloading live history') + is_live = is_live or self._search_regex( + r'(?s)(\s*LIVE)', + webpage, 'is live?', default=False) + # Current live is always the first match current_live = self._search_regex( - (r'data-type="movie" data-id="(\d+)">', - r'tw-sound-flag-open-link" data-id="(\d+)" style=',), - webpage, 'current live ID', default=None) - if not current_live: - raise ExtractorError('The user is not currently live') - return self.url_result('https://twitcasting.tv/%s/movie/%s' % (uploader_id, current_live)) + r'(?s)\d+)"', + webpage, 'current live ID', default=None, group='video_id') + if not is_live or not current_live: + raise UserNotLive(video_id=uploader_id) + + return self.url_result(f'https://twitcasting.tv/{uploader_id}/movie/{current_live}', TwitCastingIE) class TwitCastingUserIE(InfoExtractor): - _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P[^/]+)/show/?(?:[#?]|$)' + _VALID_URL = r'https?://(?:[^/?#]+\.)?twitcasting\.tv/(?P[^/?#]+)/(:?show|archive)/?(?:[#?]|$)' _TESTS = [{ + 'url': 'https://twitcasting.tv/natsuiromatsuri/archive/', + 'info_dict': { + 'id': 'natsuiromatsuri', + 'title': 'natsuiromatsuri - Live History', + }, + 'playlist_mincount': 235, + }, { 'url': 'https://twitcasting.tv/noriyukicas/show', 'only_matching': True, }] @@ -231,8 +289,7 @@ def _entries(self, uploader_id): webpage = self._download_webpage( next_url, uploader_id, query={'filter': 'watchable'}, note='Downloading page %d' % page_num) matches = re.finditer( - r'''(?isx)/[^/]+/movie/\d+)"\s*>.+?''', - webpage) + r'(?s)/[^/"]+/movie/\d+)"', webpage) for mobj in matches: yield self.url_result(urljoin(base_url, mobj.group('url')))