[yt-dlp.git] / yt_dlp / extractor / twitcasting.py

# coding: utf-8
from __future__ import unicode_literals

import itertools
import re

from .common import InfoExtractor
from ..downloader.websocket import has_websockets
from ..utils import (
    clean_html,
    ExtractorError,
    float_or_none,
    get_element_by_class,
    get_element_by_id,
    parse_duration,
    qualities,
    str_to_int,
    traverse_obj,
    try_get,
    unified_timestamp,
    urlencode_postdata,
    urljoin,
)


class TwitCastingIE(InfoExtractor):
    _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<uploader_id>[^/]+)/(?:movie|twplayer)/(?P<id>\d+)'
    _M3U8_HEADERS = {
        'Origin': 'https://twitcasting.tv',
        'Referer': 'https://twitcasting.tv/',
    }
    _TESTS = [{
        'url': 'https://twitcasting.tv/ivetesangalo/movie/2357609',
        'md5': '745243cad58c4681dc752490f7540d7f',
        'info_dict': {
            'id': '2357609',
            'ext': 'mp4',
            'title': 'Live #2357609',
            'uploader_id': 'ivetesangalo',
            'description': 'Twitter Oficial da cantora brasileira Ivete Sangalo.',
            'thumbnail': r're:^https?://.*\.jpg$',
            'upload_date': '20110822',
            'timestamp': 1314010824,
            'duration': 32,
            'view_count': int,
        },
        'params': {
            'skip_download': True,
        },
    }, {
        'url': 'https://twitcasting.tv/mttbernardini/movie/3689740',
        'info_dict': {
            'id': '3689740',
            'ext': 'mp4',
            'title': 'Live playing something #3689740',
            'uploader_id': 'mttbernardini',
            'description': 'Salve, io sono Matto (ma con la e). Questa è la mia presentazione, in quanto sono letteralmente matto (nel senso di strano), con qualcosa in più.',
            'thumbnail': r're:^https?://.*\.jpg$',
            'upload_date': '20120212',
            'timestamp': 1329028024,
            'duration': 681,
            'view_count': int,
        },
        'params': {
            'skip_download': True,
            '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,
        },
        'playlist_mincount': 2,
    }]

    def _real_extract(self, url):
        uploader_id, video_id = self._match_valid_url(url).groups()

        video_password = self.get_param('videopassword')
        request_data = None
        if video_password:
            request_data = urlencode_postdata({
                'password': video_password,
            }, encoding='utf-8')
        webpage = self._download_webpage(
            url, video_id, data=request_data,
            headers={'Origin': 'https://twitcasting.tv'})

        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(
                r'data-movie-playlist=\'([^\']+?)\'',
                x, 'movie playlist', default=None), video_id)['2'], list)

        thumbnail = traverse_obj(video_js_data, (0, 'thumbnailUrl')) or self._og_search_thumbnail(webpage)
        description = clean_html(get_element_by_id(
            'authorcomment', webpage)) or self._html_search_meta(
            ['description', 'og:description', 'twitter:description'], webpage)
        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*</'), webpage, 'views', None))
        timestamp = unified_timestamp(self._search_regex(
            r'data-toggle="true"[^>]+datetime="([^"]+)"',
            webpage, 'datetime', None))

        stream_server_data = self._download_json(
            '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
        if not traverse_obj(stream_server_data, 'llfmp4') and is_live:
            self.raise_login_required(method='cookies')

        base_dict = {
            'title': title,
            'description': description,
            'thumbnail': thumbnail,
            'timestamp': timestamp,
            'uploader_id': uploader_id,
            'duration': duration,
            'view_count': view_count,
            'is_live': is_live,
        }

        def find_dmu(x):
            data_movie_url = self._search_regex(
                r'data-movie-url=(["\'])(?P<url>(?:(?!\1).)+)\1',
                x, 'm3u8 url', group='url', default=None)
            if data_movie_url:
                return [data_movie_url]

        m3u8_urls = (try_get(webpage, find_dmu, list)
                     or traverse_obj(video_js_data, (..., 'source', 'url'))
                     or ([f'https://twitcasting.tv/{uploader_id}/metastream.m3u8'] if is_live else None))
        if not m3u8_urls:
            raise ExtractorError('Failed to get m3u8 playlist')

        if is_live:
            m3u8_url = m3u8_urls[0]
            formats = self._extract_m3u8_formats(
                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 has_websockets:
                qq = qualities(['base', 'mobilesource', 'main'])
                streams = traverse_obj(stream_server_data, ('llfmp4', 'streams')) or {}
                for mode, ws_url in streams.items():
                    formats.append({
                        'url': ws_url,
                        'format_id': 'ws-%s' % mode,
                        'ext': 'mp4',
                        'quality': qq(mode),
                        # TwitCasting simply sends moof atom directly over WS
                        'protocol': 'websocket_frag',
                    })

            self._sort_formats(formats)

            infodict = {
                'formats': formats
            }
        else:
            infodict = {
                '_type': 'multi_video',
                'entries': [{
                    'id': f'{video_id}-{num}',
                    'url': m3u8_url,
                    'ext': 'mp4',
                    # Requesting the manifests here will cause download to fail.
                    # So use ffmpeg instead. See: https://github.com/yt-dlp/yt-dlp/issues/382
                    'protocol': 'm3u8',
                    'http_headers': self._M3U8_HEADERS,
                    **base_dict,
                } for (num, m3u8_url) in enumerate(m3u8_urls)],
            }

        return {
            'id': video_id,
            **base_dict,
            **infodict,
        }


class TwitCastingLiveIE(InfoExtractor):
    _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<id>[^/]+)/?(?:[#?]|$)'
    _TESTS = [{
        'url': 'https://twitcasting.tv/ivetesangalo',
        'only_matching': True,
    }]

    def _real_extract(self, url):
        uploader_id = self._match_id(url)
        self.to_screen(
            '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)
        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))


class TwitCastingUserIE(InfoExtractor):
    _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<id>[^/]+)/show/?(?:[#?]|$)'
    _TESTS = [{
        'url': 'https://twitcasting.tv/noriyukicas/show',
        'only_matching': True,
    }]

    def _entries(self, uploader_id):
        base_url = next_url = 'https://twitcasting.tv/%s/show' % uploader_id
        for page_num in itertools.count(1):
            webpage = self._download_webpage(
                next_url, uploader_id, query={'filter': 'watchable'}, note='Downloading page %d' % page_num)
            matches = re.finditer(
                r'''(?isx)<a\s+class="tw-movie-thumbnail"\s*href="(?P<url>/[^/]+/movie/\d+)"\s*>.+?</a>''',
                webpage)
            for mobj in matches:
                yield self.url_result(urljoin(base_url, mobj.group('url')))

            next_url = self._search_regex(
                r'<a href="(/%s/show/%d-\d+)[?"]' % (re.escape(uploader_id), page_num),
                webpage, 'next url', default=None)
            next_url = urljoin(base_url, next_url)
            if not next_url:
                return

    def _real_extract(self, url):
        uploader_id = self._match_id(url)
        return self.playlist_result(
            self._entries(uploader_id), uploader_id, '%s - Live History' % uploader_id)
Commit	Line	Data
036f9051	1	# coding: utf-8
	2	from __future__ import unicode_literals
	3
e85a3971	4	import itertools
036f9051	5	import re
036f9051	6
29f7c58a	7	from .common import InfoExtractor
e6779b94	8	from ..downloader.websocket import has_websockets
29f7c58a	9	from ..utils import (
29f7c58a	10	clean_html,
a4a42602	11	ExtractorError,
29f7c58a	12	float_or_none,
	13	get_element_by_class,
	14	get_element_by_id,
	15	parse_duration,
e6779b94	16	qualities,
29f7c58a	17	str_to_int,
fabb27fc	18	traverse_obj,
e85a3971	19	try_get,
29f7c58a	20	unified_timestamp,
29f7c58a	21	urlencode_postdata,
e85a3971	22	urljoin,
29f7c58a	23	)
29f7c58a	24
036f9051	25
cf0db4d9	26	class TwitCastingIE(InfoExtractor):
e85a3971	27	_VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<uploader_id>[^/]+)/(?:movie\|twplayer)/(?P<id>\d+)'
a4a42602 LTHD	28	_M3U8_HEADERS = {
	29	'Origin': 'https://twitcasting.tv',
	30	'Referer': 'https://twitcasting.tv/',
	31	}
88b54749	32	_TESTS = [{
036f9051	33	'url': 'https://twitcasting.tv/ivetesangalo/movie/2357609',
	34	'md5': '745243cad58c4681dc752490f7540d7f',
	35	'info_dict': {
	36	'id': '2357609',
	37	'ext': 'mp4',
00a9a25c	38	'title': 'Live #2357609',
036f9051	39	'uploader_id': 'ivetesangalo',
29f7c58a	40	'description': 'Twitter Oficial da cantora brasileira Ivete Sangalo.',
036f9051	41	'thumbnail': r're:^https?://.*\.jpg$',
29f7c58a	42	'upload_date': '20110822',
	43	'timestamp': 1314010824,
	44	'duration': 32,
	45	'view_count': int,
cf0db4d9 S	46	},
	47	'params': {
	48	'skip_download': True,
	49	},
88b54749 MZ	50	}, {
	51	'url': 'https://twitcasting.tv/mttbernardini/movie/3689740',
	52	'info_dict': {
	53	'id': '3689740',
	54	'ext': 'mp4',
	55	'title': 'Live playing something #3689740',
	56	'uploader_id': 'mttbernardini',
29f7c58a	57	'description': 'Salve, io sono Matto (ma con la e). Questa è la mia presentazione, in quanto sono letteralmente matto (nel senso di strano), con qualcosa in più.',
88b54749	58	'thumbnail': r're:^https?://.*\.jpg$',
29f7c58a	59	'upload_date': '20120212',
	60	'timestamp': 1329028024,
	61	'duration': 681,
	62	'view_count': int,
88b54749 MZ	63	},
	64	'params': {
	65	'skip_download': True,
	66	'videopassword': 'abc',
	67	},
a4a42602 LTHD	68	}, {
	69	'note': 'archive is split in 2 parts',
	70	'url': 'https://twitcasting.tv/loft_heaven/movie/685979292',
	71	'info_dict': {
	72	'id': '685979292',
	73	'ext': 'mp4',
	74	'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」”',
	75	'duration': 6964.599334,
	76	},
	77	'playlist_mincount': 2,
88b54749	78	}]
036f9051	79
036f9051	80	def _real_extract(self, url):
5ad28e7f	81	uploader_id, video_id = self._match_valid_url(url).groups()
036f9051	82
a06916d9	83	video_password = self.get_param('videopassword')
88b54749 MZ	84	request_data = None
	85	if video_password:
	86	request_data = urlencode_postdata({
	87	'password': video_password,
5a13fdd2	88	}, encoding='utf-8')
d0491a1e	89	webpage = self._download_webpage(
	90	url, video_id, data=request_data,
	91	headers={'Origin': 'https://twitcasting.tv'})
036f9051	92
e85a3971	93	title = (clean_html(get_element_by_id('movietitle', webpage))
e85a3971	94	or self._html_search_meta(['og:title', 'twitter:title'], webpage, fatal=True))
cf0db4d9	95
a4a42602 LTHD	96	video_js_data = try_get(
	97	webpage,
	98	lambda x: self._parse_json(self._search_regex(
	99	r'data-movie-playlist=\'([^\']+?)\'',
	100	x, 'movie playlist', default=None), video_id)['2'], list)
fabb27fc	101
a4a42602	102	thumbnail = traverse_obj(video_js_data, (0, 'thumbnailUrl')) or self._og_search_thumbnail(webpage)
29f7c58a	103	description = clean_html(get_element_by_id(
	104	'authorcomment', webpage)) or self._html_search_meta(
	105	['description', 'og:description', 'twitter:description'], webpage)
a4a42602 LTHD	106	duration = (try_get(video_js_data, lambda x: sum(float_or_none(y.get('duration')) for y in x) / 1000)
a4a42602 LTHD	107	or parse_duration(clean_html(get_element_by_class('tw-player-duration-time', webpage))))
29f7c58a	108	view_count = str_to_int(self._search_regex(
a4a42602	109	(r'Total\s:\s([\d,]+)\sViews', r'総視聴者\s:\s([\d,]+)\s</'), webpage, 'views', None))
29f7c58a	110	timestamp = unified_timestamp(self._search_regex(
	111	r'data-toggle="true"[^>]+datetime="([^"]+)"',
	112	webpage, 'datetime', None))
cf0db4d9	113
a4a42602 LTHD	114	stream_server_data = self._download_json(
	115	'https://twitcasting.tv/streamserver.php?target=%s&mode=client' % uploader_id, video_id,
	116	'Downloading live info', fatal=False)
e85a3971	117
a4a42602 LTHD	118	is_live = 'data-status="online"' in webpage
	119	if not traverse_obj(stream_server_data, 'llfmp4') and is_live:
	120	self.raise_login_required(method='cookies')
	121
	122	base_dict = {
036f9051	123	'title': title,
	124	'description': description,
	125	'thumbnail': thumbnail,
29f7c58a	126	'timestamp': timestamp,
036f9051	127	'uploader_id': uploader_id,
29f7c58a	128	'duration': duration,
29f7c58a	129	'view_count': view_count,
e85a3971	130	'is_live': is_live,
036f9051	131	}
e85a3971	132
a4a42602 LTHD	133	def find_dmu(x):
	134	data_movie_url = self._search_regex(
	135	r'data-movie-url=(["\'])(?P<url>(?:(?!\1).)+)\1',
	136	x, 'm3u8 url', group='url', default=None)
	137	if data_movie_url:
	138	return [data_movie_url]
	139
	140	m3u8_urls = (try_get(webpage, find_dmu, list)
	141	or traverse_obj(video_js_data, (..., 'source', 'url'))
	142	or ([f'https://twitcasting.tv/{uploader_id}/metastream.m3u8'] if is_live else None))
	143	if not m3u8_urls:
	144	raise ExtractorError('Failed to get m3u8 playlist')
	145
	146	if is_live:
	147	m3u8_url = m3u8_urls[0]
	148	formats = self._extract_m3u8_formats(
	149	m3u8_url, video_id, ext='mp4', m3u8_id='hls',
	150	live=True, headers=self._M3U8_HEADERS)
	151
	152	formats.extend(self._extract_m3u8_formats(
	153	m3u8_url, video_id, ext='mp4', m3u8_id='source',
	154	live=True, query={'mode': 'source'},
	155	note='Downloading source quality m3u8',
	156	headers=self._M3U8_HEADERS, fatal=False))
	157
	158	if has_websockets:
	159	qq = qualities(['base', 'mobilesource', 'main'])
	160	streams = traverse_obj(stream_server_data, ('llfmp4', 'streams')) or {}
	161	for mode, ws_url in streams.items():
	162	formats.append({
	163	'url': ws_url,
	164	'format_id': 'ws-%s' % mode,
	165	'ext': 'mp4',
	166	'quality': qq(mode),
	167	# TwitCasting simply sends moof atom directly over WS
	168	'protocol': 'websocket_frag',
	169	})
	170
	171	self._sort_formats(formats)
	172
	173	infodict = {
	174	'formats': formats
	175	}
	176	else:
	177	infodict = {
	178	'_type': 'multi_video',
	179	'entries': [{
	180	'id': f'{video_id}-{num}',
	181	'url': m3u8_url,
	182	'ext': 'mp4',
	183	# Requesting the manifests here will cause download to fail.
	184	# So use ffmpeg instead. See: https://github.com/yt-dlp/yt-dlp/issues/382
	185	'protocol': 'm3u8',
	186	'http_headers': self._M3U8_HEADERS,
	187	**base_dict,
	188	} for (num, m3u8_url) in enumerate(m3u8_urls)],
	189	}
	190
	191	return {
	192	'id': video_id,
	193	**base_dict,
	194	**infodict,
	195	}
	196
e85a3971	197
	198	class TwitCastingLiveIE(InfoExtractor):
	199	_VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<id>[^/]+)/?(?:[#?]\|$)'
	200	_TESTS = [{
	201	'url': 'https://twitcasting.tv/ivetesangalo',
	202	'only_matching': True,
	203	}]
	204
	205	def _real_extract(self, url):
	206	uploader_id = self._match_id(url)
	207	self.to_screen(
	208	'Downloading live video of user {0}. '
	209	'Pass "https://twitcasting.tv/{0}/show" to download the history'.format(uploader_id))
	210
	211	webpage = self._download_webpage(url, uploader_id)
	212	current_live = self._search_regex(
	213	(r'data-type="movie" data-id="(\d+)">',
	214	r'tw-sound-flag-open-link" data-id="(\d+)" style=',),
	215	webpage, 'current live ID', default=None)
	216	if not current_live:
	217	raise ExtractorError('The user is not currently live')
	218	return self.url_result('https://twitcasting.tv/%s/movie/%s' % (uploader_id, current_live))
	219
	220
	221	class TwitCastingUserIE(InfoExtractor):
	222	_VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<id>[^/]+)/show/?(?:[#?]\|$)'
	223	_TESTS = [{
	224	'url': 'https://twitcasting.tv/noriyukicas/show',
	225	'only_matching': True,
	226	}]
	227
	228	def _entries(self, uploader_id):
	229	base_url = next_url = 'https://twitcasting.tv/%s/show' % uploader_id
	230	for page_num in itertools.count(1):
	231	webpage = self._download_webpage(
	232	next_url, uploader_id, query={'filter': 'watchable'}, note='Downloading page %d' % page_num)
	233	matches = re.finditer(
	234	r'''(?isx)<a\s+class="tw-movie-thumbnail"\shref="(?P<url>/[^/]+/movie/\d+)"\s>.+?</a>''',
	235	webpage)
	236	for mobj in matches:
	237	yield self.url_result(urljoin(base_url, mobj.group('url')))
	238
	239	next_url = self._search_regex(
	240	r'<a href="(/%s/show/%d-\d+)[?"]' % (re.escape(uploader_id), page_num),
	241	webpage, 'next url', default=None)
	242	next_url = urljoin(base_url, next_url)
	243	if not next_url:
	244	return
	245
	246	def _real_extract(self, url):
	247	uploader_id = self._match_id(url)
	248	return self.playlist_result(
	249	self._entries(uploader_id), uploader_id, '%s - Live History' % uploader_id)