[yt-dlp.git] / youtube_dl / extractor / arte.py

# encoding: utf-8
from __future__ import unicode_literals

import re

from .common import InfoExtractor
from ..compat import (
    compat_parse_qs,
    compat_urllib_parse_urlparse,
)
from ..utils import (
    find_xpath_attr,
    unified_strdate,
    get_element_by_attribute,
    int_or_none,
    qualities,
)

# There are different sources of video in arte.tv, the extraction process
# is different for each one. The videos usually expire in 7 days, so we can't
# add tests.


class ArteTvIE(InfoExtractor):
    _VALID_URL = r'http://videos\.arte\.tv/(?P<lang>fr|de)/.*-(?P<id>.*?)\.html'
    IE_NAME = 'arte.tv'

    def _real_extract(self, url):
        mobj = re.match(self._VALID_URL, url)
        lang = mobj.group('lang')
        video_id = mobj.group('id')

        ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
        ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
        ref_xml_doc = self._download_xml(
            ref_xml_url, video_id, note='Downloading metadata')
        config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
        config_xml_url = config_node.attrib['ref']
        config = self._download_xml(
            config_xml_url, video_id, note='Downloading configuration')

        formats = [{
            'format_id': q.attrib['quality'],
            # The playpath starts at 'mp4:', if we don't manually
            # split the url, rtmpdump will incorrectly parse them
            'url': q.text.split('mp4:', 1)[0],
            'play_path': 'mp4:' + q.text.split('mp4:', 1)[1],
            'ext': 'flv',
            'quality': 2 if q.attrib['quality'] == 'hd' else 1,
        } for q in config.findall('./urls/url')]
        self._sort_formats(formats)

        title = config.find('.//name').text
        thumbnail = config.find('.//firstThumbnailUrl').text
        return {
            'id': video_id,
            'title': title,
            'thumbnail': thumbnail,
            'formats': formats,
        }


class ArteTVPlus7IE(InfoExtractor):
    IE_NAME = 'arte.tv:+7'
    _VALID_URL = r'https?://(?:www\.)?arte\.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'

    @classmethod
    def _extract_url_info(cls, url):
        mobj = re.match(cls._VALID_URL, url)
        lang = mobj.group('lang')
        query = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
        if 'vid' in query:
            video_id = query['vid'][0]
        else:
            # This is not a real id, it can be for example AJT for the news
            # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
            video_id = mobj.group('id')
        return video_id, lang

    def _real_extract(self, url):
        video_id, lang = self._extract_url_info(url)
        webpage = self._download_webpage(url, video_id)
        return self._extract_from_webpage(webpage, video_id, lang)

    def _extract_from_webpage(self, webpage, video_id, lang):
        patterns_templates = (r'arte_vp_url=["\'](.*?%s.*?)["\']', r'data-url=["\']([^"]+%s[^"]+)["\']')
        ids = (video_id, '')
        # some pages contain multiple videos (like
        # http://www.arte.tv/guide/de/sendungen/XEN/xenius/?vid=055918-015_PLUS7-D),
        # so we first try to look for json URLs that contain the video id from
        # the 'vid' parameter.
        patterns = [t % re.escape(_id) for _id in ids for t in patterns_templates]
        json_url = self._html_search_regex(
            patterns, webpage, 'json vp url', default=None)
        if not json_url:
            iframe_url = self._html_search_regex(
                r'<iframe[^>]+src=(["\'])(?P<url>.+\bjson_url=.+?)\1',
                webpage, 'iframe url', group='url')
            json_url = compat_parse_qs(
                compat_urllib_parse_urlparse(iframe_url).query)['json_url'][0]
        return self._extract_from_json_url(json_url, video_id, lang)

    def _extract_from_json_url(self, json_url, video_id, lang):
        info = self._download_json(json_url, video_id)
        player_info = info['videoJsonPlayer']

        upload_date_str = player_info.get('shootingDate')
        if not upload_date_str:
            upload_date_str = player_info.get('VDA', '').split(' ')[0]

        title = player_info['VTI'].strip()
        subtitle = player_info.get('VSU', '').strip()
        if subtitle:
            title += ' - %s' % subtitle

        info_dict = {
            'id': player_info['VID'],
            'title': title,
            'description': player_info.get('VDE'),
            'upload_date': unified_strdate(upload_date_str),
            'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
        }
        qfunc = qualities(['HQ', 'MQ', 'EQ', 'SQ'])

        formats = []
        for format_id, format_dict in player_info['VSR'].items():
            f = dict(format_dict)
            versionCode = f.get('versionCode')

            langcode = {
                'fr': 'F',
                'de': 'A',
            }.get(lang, lang)
            lang_rexs = [r'VO?%s' % langcode, r'VO?.-ST%s' % langcode]
            lang_pref = (
                None if versionCode is None else (
                    10 if any(re.match(r, versionCode) for r in lang_rexs)
                    else -10))
            source_pref = 0
            if versionCode is not None:
                # The original version with subtitles has lower relevance
                if re.match(r'VO-ST(F|A)', versionCode):
                    source_pref -= 10
                # The version with sourds/mal subtitles has also lower relevance
                elif re.match(r'VO?(F|A)-STM\1', versionCode):
                    source_pref -= 9
            format = {
                'format_id': format_id,
                'preference': -10 if f.get('videoFormat') == 'M3U8' else None,
                'language_preference': lang_pref,
                'format_note': '%s, %s' % (f.get('versionCode'), f.get('versionLibelle')),
                'width': int_or_none(f.get('width')),
                'height': int_or_none(f.get('height')),
                'tbr': int_or_none(f.get('bitrate')),
                'quality': qfunc(f.get('quality')),
                'source_preference': source_pref,
            }

            if f.get('mediaType') == 'rtmp':
                format['url'] = f['streamer']
                format['play_path'] = 'mp4:' + f['url']
                format['ext'] = 'flv'
            else:
                format['url'] = f['url']

            formats.append(format)

        self._check_formats(formats, video_id)
        self._sort_formats(formats)

        info_dict['formats'] = formats
        return info_dict


# It also uses the arte_vp_url url from the webpage to extract the information
class ArteTVCreativeIE(ArteTVPlus7IE):
    IE_NAME = 'arte.tv:creative'
    _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/(?:magazine?/)?(?P<id>[^?#]+)'

    _TESTS = [{
        'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
        'info_dict': {
            'id': '72176',
            'ext': 'mp4',
            'title': 'Folge 2 - Corporate Design',
            'upload_date': '20131004',
        },
    }, {
        'url': 'http://creative.arte.tv/fr/Monty-Python-Reunion',
        'info_dict': {
            'id': '160676',
            'ext': 'mp4',
            'title': 'Monty Python live (mostly)',
            'description': 'Événement ! Quarante-cinq ans après leurs premiers succès, les légendaires Monty Python remontent sur scène.\n',
            'upload_date': '20140805',
        }
    }]


class ArteTVFutureIE(ArteTVPlus7IE):
    IE_NAME = 'arte.tv:future'
    _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(?P<id>.+)'

    _TESTS = [
        {
            'url': 'http://future.arte.tv/fr/info-sciences/les-ecrevisses-aussi-sont-anxieuses',
            'info_dict': {
                'id': '050940-028-A',
                'ext': 'mp4',
                'title': 'Les écrevisses aussi peuvent être anxieuses',
            },
        },
        {
            'url': 'http://future.arte.tv/fr/la-science-est-elle-responsable',
            'info_dict': {
                'id': '061982-002-A',
                'ext': 'mp4',
                'title': 'Brian P. Schmidt - Prix Nobel de physique 2011',
            },
        }
    ]


class ArteTVDDCIE(ArteTVPlus7IE):
    IE_NAME = 'arte.tv:ddc'
    _VALID_URL = r'https?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>.+)'

    def _real_extract(self, url):
        video_id, lang = self._extract_url_info(url)
        if lang == 'folge':
            lang = 'de'
        elif lang == 'emission':
            lang = 'fr'
        webpage = self._download_webpage(url, video_id)
        scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
        script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
        javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
        json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url')
        return self._extract_from_json_url(json_url, video_id, lang)


class ArteTVConcertIE(ArteTVPlus7IE):
    IE_NAME = 'arte.tv:concert'
    _VALID_URL = r'https?://concert\.arte\.tv/(?P<lang>de|fr)/(?P<id>.+)'

    _TEST = {
        'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde',
        'md5': '9ea035b7bd69696b67aa2ccaaa218161',
        'info_dict': {
            'id': '186',
            'ext': 'mp4',
            'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"',
            'upload_date': '20140128',
            'description': 'md5:486eb08f991552ade77439fe6d82c305',
        },
    }


class ArteTVEmbedIE(ArteTVPlus7IE):
    IE_NAME = 'arte.tv:embed'
    _VALID_URL = r'''(?x)
        http://www\.arte\.tv
        /playerv2/embed\.php\?json_url=
        (?P<json_url>
            http://arte\.tv/papi/tvguide/videos/stream/player/
            (?P<lang>[^/]+)/(?P<id>[^/]+)[^&]*
        )
    '''

    def _real_extract(self, url):
        mobj = re.match(self._VALID_URL, url)
        video_id = mobj.group('id')
        lang = mobj.group('lang')
        json_url = mobj.group('json_url')
        return self._extract_from_json_url(json_url, video_id, lang)
Commit	Line	Data
69a0c470	1	# encoding: utf-8
3798eadc PH	2	from __future__ import unicode_literals
3798eadc PH	3
d5822b96	4	import re
d5822b96 PH	5
d5822b96 PH	6	from .common import InfoExtractor
393ca8c9 S	7	from ..compat import (
	8	compat_parse_qs,
	9	compat_urllib_parse_urlparse,
	10	)
d5822b96	11	from ..utils import (
df50a412	12	find_xpath_attr,
d5822b96	13	unified_strdate,
56a8ab7d	14	get_element_by_attribute,
d24a2b20	15	int_or_none,
aff2f4f4	16	qualities,
d5822b96 PH	17	)
d5822b96 PH	18
5f6a1245	19	# There are different sources of video in arte.tv, the extraction process
c40f5cf4 JMF	20	# is different for each one. The videos usually expire in 7 days, so we can't
	21	# add tests.
	22
d5822b96	23
515bbe4b	24	class ArteTvIE(InfoExtractor):
878d11ec	25	_VALID_URL = r'http://videos\.arte\.tv/(?P<lang>fr\|de)/.-(?P<id>.?)\.html'
3798eadc	26	IE_NAME = 'arte.tv'
d5822b96	27
d5822b96	28	def _real_extract(self, url):
7a249480 PH	29	mobj = re.match(self._VALID_URL, url)
	30	lang = mobj.group('lang')
	31	video_id = mobj.group('id')
	32
8de64cac PH	33	ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
8de64cac PH	34	ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
a4ff6c47 PH	35	ref_xml_doc = self._download_xml(
a4ff6c47 PH	36	ref_xml_url, video_id, note='Downloading metadata')
df50a412	37	config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
8de64cac	38	config_xml_url = config_node.attrib['ref']
878d11ec	39	config = self._download_xml(
a4ff6c47	40	config_xml_url, video_id, note='Downloading configuration')
37b6a661	41
878d11ec	42	formats = [{
5a184030	43	'format_id': q.attrib['quality'],
459af434 JMF	44	# The playpath starts at 'mp4:', if we don't manually
	45	# split the url, rtmpdump will incorrectly parse them
	46	'url': q.text.split('mp4:', 1)[0],
	47	'play_path': 'mp4:' + q.text.split('mp4:', 1)[1],
b2799ff9	48	'ext': 'flv',
878d11ec	49	'quality': 2 if q.attrib['quality'] == 'hd' else 1,
b2799ff9	50	} for q in config.findall('./urls/url')]
878d11ec PH	51	self._sort_formats(formats)
	52
	53	title = config.find('.//name').text
	54	thumbnail = config.find('.//firstThumbnailUrl').text
	55	return {
	56	'id': video_id,
	57	'title': title,
	58	'thumbnail': thumbnail,
7a249480	59	'formats': formats,
515bbe4b	60	}
c40f5cf4 JMF	61
	62
	63	class ArteTVPlus7IE(InfoExtractor):
3798eadc	64	IE_NAME = 'arte.tv:+7'
f66ede43	65	_VALID_URL = r'https?://(?:www\.)?arte\.tv/guide/(?P<lang>fr\|de)/(?:(?:sendungen\|emissions)/)?(?P<id>.?)/(?P<name>.?)(\?.*)?'
c40f5cf4	66
69a0c470 JMF	67	@classmethod
	68	def _extract_url_info(cls, url):
	69	mobj = re.match(cls._VALID_URL, url)
c40f5cf4	70	lang = mobj.group('lang')
a8f1d167 JMF	71	query = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
	72	if 'vid' in query:
	73	video_id = query['vid'][0]
	74	else:
	75	# This is not a real id, it can be for example AJT for the news
	76	# http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
	77	video_id = mobj.group('id')
69a0c470	78	return video_id, lang
c40f5cf4	79
69a0c470 JMF	80	def _real_extract(self, url):
69a0c470 JMF	81	video_id, lang = self._extract_url_info(url)
c40f5cf4	82	webpage = self._download_webpage(url, video_id)
69a0c470 JMF	83	return self._extract_from_webpage(webpage, video_id, lang)
	84
	85	def _extract_from_webpage(self, webpage, video_id, lang):
a8f1d167 JMF	86	patterns_templates = (r'arte_vp_url=["\'](.?%s.?)["\']', r'data-url=["\']([^"]+%s[^"]+)["\']')
	87	ids = (video_id, '')
	88	# some pages contain multiple videos (like
	89	# http://www.arte.tv/guide/de/sendungen/XEN/xenius/?vid=055918-015_PLUS7-D),
	90	# so we first try to look for json URLs that contain the video id from
	91	# the 'vid' parameter.
	92	patterns = [t % re.escape(_id) for _id in ids for t in patterns_templates]
88ce273d	93	json_url = self._html_search_regex(
a8f1d167	94	patterns, webpage, 'json vp url', default=None)
393ca8c9 S	95	if not json_url:
	96	iframe_url = self._html_search_regex(
	97	r'<iframe[^>]+src=(["\'])(?P<url>.+\bjson_url=.+?)\1',
5e39123b	98	webpage, 'iframe url', group='url')
393ca8c9 S	99	json_url = compat_parse_qs(
393ca8c9 S	100	compat_urllib_parse_urlparse(iframe_url).query)['json_url'][0]
56a8ab7d	101	return self._extract_from_json_url(json_url, video_id, lang)
c40f5cf4	102
56a8ab7d	103	def _extract_from_json_url(self, json_url, video_id, lang):
893f8832	104	info = self._download_json(json_url, video_id)
c40f5cf4 JMF	105	player_info = info['videoJsonPlayer']
c40f5cf4 JMF	106
99b67fec PH	107	upload_date_str = player_info.get('shootingDate')
	108	if not upload_date_str:
	109	upload_date_str = player_info.get('VDA', '').split(' ')[0]
	110
74214d35 S	111	title = player_info['VTI'].strip()
	112	subtitle = player_info.get('VSU', '').strip()
	113	if subtitle:
	114	title += ' - %s' % subtitle
	115
c40f5cf4 JMF	116	info_dict = {
c40f5cf4 JMF	117	'id': player_info['VID'],
74214d35	118	'title': title,
c40f5cf4	119	'description': player_info.get('VDE'),
99b67fec	120	'upload_date': unified_strdate(upload_date_str),
c40f5cf4 JMF	121	'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
c40f5cf4 JMF	122	}
aff2f4f4	123	qfunc = qualities(['HQ', 'MQ', 'EQ', 'SQ'])
c40f5cf4	124
aff2f4f4	125	formats = []
d24a2b20	126	for format_id, format_dict in player_info['VSR'].items():
aff2f4f4 PH	127	f = dict(format_dict)
	128	versionCode = f.get('versionCode')
	129
	130	langcode = {
	131	'fr': 'F',
	132	'de': 'A',
	133	}.get(lang, lang)
	134	lang_rexs = [r'VO?%s' % langcode, r'VO?.-ST%s' % langcode]
	135	lang_pref = (
	136	None if versionCode is None else (
	137	10 if any(re.match(r, versionCode) for r in lang_rexs)
	138	else -10))
	139	source_pref = 0
	140	if versionCode is not None:
	141	# The original version with subtitles has lower relevance
	142	if re.match(r'VO-ST(F\|A)', versionCode):
	143	source_pref -= 10
	144	# The version with sourds/mal subtitles has also lower relevance
	145	elif re.match(r'VO?(F\|A)-STM\1', versionCode):
	146	source_pref -= 9
	147	format = {
	148	'format_id': format_id,
	149	'preference': -10 if f.get('videoFormat') == 'M3U8' else None,
	150	'language_preference': lang_pref,
	151	'format_note': '%s, %s' % (f.get('versionCode'), f.get('versionLibelle')),
	152	'width': int_or_none(f.get('width')),
	153	'height': int_or_none(f.get('height')),
	154	'tbr': int_or_none(f.get('bitrate')),
1b7b1d6e	155	'quality': qfunc(f.get('quality')),
aff2f4f4	156	'source_preference': source_pref,
c40f5cf4	157	}
aff2f4f4 PH	158
	159	if f.get('mediaType') == 'rtmp':
	160	format['url'] = f['streamer']
	161	format['play_path'] = 'mp4:' + f['url']
	162	format['ext'] = 'flv'
c40f5cf4	163	else:
aff2f4f4 PH	164	format['url'] = f['url']
	165
	166	formats.append(format)
	167
c06a9f87	168	self._check_formats(formats, video_id)
aff2f4f4	169	self._sort_formats(formats)
c40f5cf4	170
aff2f4f4	171	info_dict['formats'] = formats
c40f5cf4 JMF	172	return info_dict
	173
	174
	175	# It also uses the arte_vp_url url from the webpage to extract the information
	176	class ArteTVCreativeIE(ArteTVPlus7IE):
3798eadc	177	IE_NAME = 'arte.tv:creative'
01d906ff	178	_VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr\|de)/(?:magazine?/)?(?P<id>[^?#]+)'
c40f5cf4	179
01d906ff	180	_TESTS = [{
3798eadc	181	'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
3798eadc	182	'info_dict': {
01d906ff	183	'id': '72176',
39a743fb	184	'ext': 'mp4',
01d906ff PH	185	'title': 'Folge 2 - Corporate Design',
01d906ff PH	186	'upload_date': '20131004',
c40f5cf4	187	},
01d906ff PH	188	}, {
	189	'url': 'http://creative.arte.tv/fr/Monty-Python-Reunion',
	190	'info_dict': {
	191	'id': '160676',
	192	'ext': 'mp4',
	193	'title': 'Monty Python live (mostly)',
b5f4775b PH	194	'description': 'Événement ! Quarante-cinq ans après leurs premiers succès, les légendaires Monty Python remontent sur scène.\n',
b5f4775b PH	195	'upload_date': '20140805',
01d906ff PH	196	}
01d906ff PH	197	}]
c40f5cf4	198
69a0c470 JMF	199
69a0c470 JMF	200	class ArteTVFutureIE(ArteTVPlus7IE):
3798eadc	201	IE_NAME = 'arte.tv:future'
24114fee FC	202	_VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr\|de)/(?P<id>.+)'
	203
	204	_TESTS = [
	205	{
	206	'url': 'http://future.arte.tv/fr/info-sciences/les-ecrevisses-aussi-sont-anxieuses',
	207	'info_dict': {
	208	'id': '050940-028-A',
	209	'ext': 'mp4',
	210	'title': 'Les écrevisses aussi peuvent être anxieuses',
	211	},
69a0c470	212	},
24114fee FC	213	{
	214	'url': 'http://future.arte.tv/fr/la-science-est-elle-responsable',
	215	'info_dict': {
	216	'id': '061982-002-A',
	217	'ext': 'mp4',
	218	'title': 'Brian P. Schmidt - Prix Nobel de physique 2011',
	219	},
	220	}
	221	]
56a8ab7d	222
ac5118bc	223
56a8ab7d	224	class ArteTVDDCIE(ArteTVPlus7IE):
3798eadc	225	IE_NAME = 'arte.tv:ddc'
39a743fb	226	_VALID_URL = r'https?://ddc\.arte\.tv/(?P<lang>emission\|folge)/(?P<id>.+)'
56a8ab7d	227
56a8ab7d CD	228	def _real_extract(self, url):
	229	video_id, lang = self._extract_url_info(url)
	230	if lang == 'folge':
	231	lang = 'de'
	232	elif lang == 'emission':
	233	lang = 'fr'
	234	webpage = self._download_webpage(url, video_id)
	235	scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
	236	script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
	237	javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
	238	json_url = self._search_regex(r"json_url=(.)&rendering_place.", javascriptPlayerGenerator, 'json url')
	239	return self._extract_from_json_url(json_url, video_id, lang)
4966a0b2 JMF	240
	241
	242	class ArteTVConcertIE(ArteTVPlus7IE):
	243	IE_NAME = 'arte.tv:concert'
	244	_VALID_URL = r'https?://concert\.arte\.tv/(?P<lang>de\|fr)/(?P<id>.+)'
	245
	246	_TEST = {
	247	'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde',
	248	'md5': '9ea035b7bd69696b67aa2ccaaa218161',
	249	'info_dict': {
	250	'id': '186',
	251	'ext': 'mp4',
	252	'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"',
	253	'upload_date': '20140128',
a9c2896e	254	'description': 'md5:486eb08f991552ade77439fe6d82c305',
4966a0b2 JMF	255	},
4966a0b2 JMF	256	}
893f8832 PH	257
	258
	259	class ArteTVEmbedIE(ArteTVPlus7IE):
	260	IE_NAME = 'arte.tv:embed'
	261	_VALID_URL = r'''(?x)
	262	http://www\.arte\.tv
	263	/playerv2/embed\.php\?json_url=
	264	(?P<json_url>
	265	http://arte\.tv/papi/tvguide/videos/stream/player/
	266	(?P<lang>[^/]+)/(?P<id>[^/]+)[^&]*
	267	)
	268	'''
	269
	270	def _real_extract(self, url):
	271	mobj = re.match(self._VALID_URL, url)
	272	video_id = mobj.group('id')
	273	lang = mobj.group('lang')
	274	json_url = mobj.group('json_url')
	275	return self._extract_from_json_url(json_url, video_id, lang)