X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/54477c858fa2504360964e675faf953d78ac57e6..ed6eef637d5115dea4adc4e40f030f0688f29c37:/youtube_dl/extractor/mtv.py?ds=inline

diff --git a/youtube_dl/extractor/mtv.py b/youtube_dl/extractor/mtv.py
index e520e2b..8acea14 100644
--- a/youtube_dl/extractor/mtv.py
+++ b/youtube_dl/extractor/mtv.py
@@ -1,130 +1,468 @@
+from __future__ import unicode_literals
+
 import re
-import xml.etree.ElementTree
 
 from .common import InfoExtractor
+from ..compat import (
+    compat_str,
+    compat_xpath,
+)
 from ..utils import (
-    compat_urllib_parse,
     ExtractorError,
+    find_xpath_attr,
+    fix_xml_ampersands,
+    float_or_none,
+    HEADRequest,
+    RegexNotFoundError,
+    sanitized_Request,
+    strip_or_none,
+    timeconvert,
+    try_get,
+    unescapeHTML,
+    update_url_query,
+    url_basename,
+    xpath_text,
 )
 
+
 def _media_xml_tag(tag):
     return '{http://search.yahoo.com/mrss/}%s' % tag
 
-class MTVIE(InfoExtractor):
-    _VALID_URL = r'^https?://(?:www\.)?mtv\.com/videos/.+?/(?P<videoid>[0-9]+)/[^/]+$'
 
-    _FEED_URL = 'http://www.mtv.com/player/embed/AS3/rss/'
-
-    _TESTS = [
-        {
-            u'url': u'http://www.mtv.com/videos/misc/853555/ours-vh1-storytellers.jhtml',
-            u'file': u'853555.mp4',
-            u'md5': u'850f3f143316b1e71fa56a4edfd6e0f8',
-            u'info_dict': {
-                u'title': u'Taylor Swift - "Ours (VH1 Storytellers)"',
-                u'description': u'Album: Taylor Swift performs "Ours" for VH1 Storytellers at Harvey Mudd College.',
-            },
-        },
-        {
-            u'url': u'http://www.mtv.com/videos/taylor-swift/916187/everything-has-changed-ft-ed-sheeran.jhtml',
-            u'file': u'USCJY1331283.mp4',
-            u'md5': u'73b4e7fcadd88929292fe52c3ced8caf',
-            u'info_dict': {
-                u'title': u'Everything Has Changed',
-                u'upload_date': u'20130606',
-                u'uploader': u'Taylor Swift',
-            },
-            u'skip': u'VEVO is only available in some countries',
-        },
-    ]
+class MTVServicesInfoExtractor(InfoExtractor):
+    _MOBILE_TEMPLATE = None
+    _LANG = None
 
     @staticmethod
     def _id_from_uri(uri):
         return uri.split(':')[-1]
 
-    # This was originally implemented for ComedyCentral, but it also works here
     @staticmethod
-    def _transform_rtmp_url(rtmp_video_url):
-        m = re.match(r'^rtmpe?://.*?/(?P<finalid>gsp\..+?/.*)$', rtmp_video_url)
-        if not m:
-            raise ExtractorError(u'Cannot transform RTMP url')
-        base = 'http://mtvnmobile.vo.llnwd.net/kip0/_pxn=1+_pxI0=Ripod-h264+_pxL0=undefined+_pxM0=+_pxK=18639+_pxE=mp4/44620/mtvnorigin/'
-        return base + m.group('finalid')
+    def _remove_template_parameter(url):
+        # Remove the templates, like &device={device}
+        return re.sub(r'&[^=]*?={.*?}(?=(&|$))', '', url)
+
+    def _get_feed_url(self, uri):
+        return self._FEED_URL
 
     def _get_thumbnail_url(self, uri, itemdoc):
-        return 'http://mtv.mtvnimages.com/uri/' + uri
+        search_path = '%s/%s' % (_media_xml_tag('group'), _media_xml_tag('thumbnail'))
+        thumb_node = itemdoc.find(search_path)
+        if thumb_node is None:
+            return None
+        else:
+            return thumb_node.attrib['url']
+
+    def _extract_mobile_video_formats(self, mtvn_id):
+        webpage_url = self._MOBILE_TEMPLATE % mtvn_id
+        req = sanitized_Request(webpage_url)
+        # Otherwise we get a webpage that would execute some javascript
+        req.add_header('User-Agent', 'curl/7')
+        webpage = self._download_webpage(req, mtvn_id,
+                                         'Downloading mobile page')
+        metrics_url = unescapeHTML(self._search_regex(r'<a href="(http://metrics.+?)"', webpage, 'url'))
+        req = HEADRequest(metrics_url)
+        response = self._request_webpage(req, mtvn_id, 'Resolving url')
+        url = response.geturl()
+        # Transform the url to get the best quality:
+        url = re.sub(r'.+pxE=mp4', 'http://mtvnmobile.vo.llnwd.net/kip0/_pxn=0+_pxK=18639+_pxE=mp4', url, 1)
+        return [{'url': url, 'ext': 'mp4'}]
 
-    def _extract_video_formats(self, metadataXml):
-        if '/error_country_block.swf' in metadataXml:
-            raise ExtractorError(u'This video is not available from your country.', expected=True)
-        mdoc = xml.etree.ElementTree.fromstring(metadataXml.encode('utf-8'))
-        renditions = mdoc.findall('.//rendition')
+    def _extract_video_formats(self, mdoc, mtvn_id, video_id):
+        if re.match(r'.*/(error_country_block\.swf|geoblock\.mp4|copyright_error\.flv(?:\?geo\b.+?)?)$', mdoc.find('.//src').text) is not None:
+            if mtvn_id is not None and self._MOBILE_TEMPLATE is not None:
+                self.to_screen('The normal version is not available from your '
+                               'country, trying with the mobile version')
+                return self._extract_mobile_video_formats(mtvn_id)
+            raise ExtractorError('This video is not available from your country.',
+                                 expected=True)
 
         formats = []
         for rendition in mdoc.findall('.//rendition'):
-            try:
-                _, _, ext = rendition.attrib['type'].partition('/')
-                rtmp_video_url = rendition.find('./src').text
-                formats.append({'ext': ext,
-                                'url': self._transform_rtmp_url(rtmp_video_url),
-                                'format_id': rendition.get('bitrate'),
-                                'width': int(rendition.get('width')),
-                                'height': int(rendition.get('height')),
-                                })
-            except (KeyError, TypeError):
-                raise ExtractorError('Invalid rendition field.')
+            if rendition.get('method') == 'hls':
+                hls_url = rendition.find('./src').text
+                formats.extend(self._extract_m3u8_formats(
+                    hls_url, video_id, ext='mp4', entry_protocol='m3u8_native',
+                    m3u8_id='hls'))
+            else:
+                # fms
+                try:
+                    _, _, ext = rendition.attrib['type'].partition('/')
+                    rtmp_video_url = rendition.find('./src').text
+                    if 'error_not_available.swf' in rtmp_video_url:
+                        raise ExtractorError(
+                            '%s said: video is not available' % self.IE_NAME,
+                            expected=True)
+                    if rtmp_video_url.endswith('siteunavail.png'):
+                        continue
+                    formats.extend([{
+                        'ext': 'flv' if rtmp_video_url.startswith('rtmp') else ext,
+                        'url': rtmp_video_url,
+                        'format_id': '-'.join(filter(None, [
+                            'rtmp' if rtmp_video_url.startswith('rtmp') else None,
+                            rendition.get('bitrate')])),
+                        'width': int(rendition.get('width')),
+                        'height': int(rendition.get('height')),
+                    }])
+                except (KeyError, TypeError):
+                    raise ExtractorError('Invalid rendition field.')
+        self._sort_formats(formats)
         return formats
 
-    def _get_video_info(self, itemdoc):
+    def _extract_subtitles(self, mdoc, mtvn_id):
+        subtitles = {}
+        for transcript in mdoc.findall('.//transcript'):
+            if transcript.get('kind') != 'captions':
+                continue
+            lang = transcript.get('srclang')
+            subtitles[lang] = [{
+                'url': compat_str(typographic.get('src')),
+                'ext': typographic.get('format')
+            } for typographic in transcript.findall('./typographic')]
+        return subtitles
+
+    def _get_video_info(self, itemdoc, use_hls=True):
         uri = itemdoc.find('guid').text
         video_id = self._id_from_uri(uri)
         self.report_extraction(video_id)
-        mediagen_url = itemdoc.find('%s/%s' % (_media_xml_tag('group'), _media_xml_tag('content'))).attrib['url']
+        content_el = itemdoc.find('%s/%s' % (_media_xml_tag('group'), _media_xml_tag('content')))
+        mediagen_url = self._remove_template_parameter(content_el.attrib['url'])
+        mediagen_url = mediagen_url.replace('device={device}', '')
         if 'acceptMethods' not in mediagen_url:
-            mediagen_url += '&acceptMethods=fms'
-        mediagen_page = self._download_webpage(mediagen_url, video_id,
-                                               u'Downloading video urls')
+            mediagen_url += '&' if '?' in mediagen_url else '?'
+            mediagen_url += 'acceptMethods='
+            mediagen_url += 'hls' if use_hls else 'fms'
 
-        description_node = itemdoc.find('description')
-        if description_node is not None:
-            description = description_node.text.strip()
-        else:
-            description = None
+        mediagen_doc = self._download_xml(mediagen_url, video_id,
+                                          'Downloading video urls')
+
+        item = mediagen_doc.find('./video/item')
+        if item is not None and item.get('type') == 'text':
+            message = '%s returned error: ' % self.IE_NAME
+            if item.get('code') is not None:
+                message += '%s - ' % item.get('code')
+            message += item.text
+            raise ExtractorError(message, expected=True)
+
+        description = strip_or_none(xpath_text(itemdoc, 'description'))
+
+        timestamp = timeconvert(xpath_text(itemdoc, 'pubDate'))
+
+        title_el = None
+        if title_el is None:
+            title_el = find_xpath_attr(
+                itemdoc, './/{http://search.yahoo.com/mrss/}category',
+                'scheme', 'urn:mtvn:video_title')
+        if title_el is None:
+            title_el = itemdoc.find(compat_xpath('.//{http://search.yahoo.com/mrss/}title'))
+        if title_el is None:
+            title_el = itemdoc.find(compat_xpath('.//title'))
+            if title_el.text is None:
+                title_el = None
 
-        info = {
-            'title': itemdoc.find('title').text,
-            'formats': self._extract_video_formats(mediagen_page),
+        title = title_el.text
+        if title is None:
+            raise ExtractorError('Could not find video title')
+        title = title.strip()
+
+        # This a short id that's used in the webpage urls
+        mtvn_id = None
+        mtvn_id_node = find_xpath_attr(itemdoc, './/{http://search.yahoo.com/mrss/}category',
+                                       'scheme', 'urn:mtvn:id')
+        if mtvn_id_node is not None:
+            mtvn_id = mtvn_id_node.text
+
+        formats = self._extract_video_formats(mediagen_doc, mtvn_id, video_id)
+
+        return {
+            'title': title,
+            'formats': formats,
+            'subtitles': self._extract_subtitles(mediagen_doc, mtvn_id),
             'id': video_id,
             'thumbnail': self._get_thumbnail_url(uri, itemdoc),
             'description': description,
+            'duration': float_or_none(content_el.attrib.get('duration')),
+            'timestamp': timestamp,
         }
 
-        # TODO: Remove when #980 has been merged
-        info.update(info['formats'][-1])
+    def _get_feed_query(self, uri):
+        data = {'uri': uri}
+        if self._LANG:
+            data['lang'] = self._LANG
+        return data
+
+    def _get_videos_info(self, uri, use_hls=True):
+        video_id = self._id_from_uri(uri)
+        feed_url = self._get_feed_url(uri)
+        info_url = update_url_query(feed_url, self._get_feed_query(uri))
+        return self._get_videos_info_from_url(info_url, video_id, use_hls)
+
+    def _get_videos_info_from_url(self, url, video_id, use_hls=True):
+        idoc = self._download_xml(
+            url, video_id,
+            'Downloading info', transform_source=fix_xml_ampersands)
+
+        title = xpath_text(idoc, './channel/title')
+        description = xpath_text(idoc, './channel/description')
+
+        return self.playlist_result(
+            [self._get_video_info(item, use_hls) for item in idoc.findall('.//item')],
+            playlist_title=title, playlist_description=description)
+
+    def _extract_triforce_mgid(self, webpage, data_zone=None, video_id=None):
+        triforce_feed = self._parse_json(self._search_regex(
+            r'triforceManifestFeed\s*=\s*({.+?})\s*;\s*\n', webpage,
+            'triforce feed', default='{}'), video_id, fatal=False)
+
+        data_zone = self._search_regex(
+            r'data-zone=(["\'])(?P<zone>.+?_lc_promo.*?)\1', webpage,
+            'data zone', default=data_zone, group='zone')
+
+        feed_url = try_get(
+            triforce_feed, lambda x: x['manifest']['zones'][data_zone]['feed'],
+            compat_str)
+        if not feed_url:
+            return
+
+        feed = self._download_json(feed_url, video_id, fatal=False)
+        if not feed:
+            return
+
+        return try_get(feed, lambda x: x['result']['data']['id'], compat_str)
+
+    def _extract_mgid(self, webpage):
+        try:
+            # the url can be http://media.mtvnservices.com/fb/{mgid}.swf
+            # or http://media.mtvnservices.com/{mgid}
+            og_url = self._og_search_video_url(webpage)
+            mgid = url_basename(og_url)
+            if mgid.endswith('.swf'):
+                mgid = mgid[:-4]
+        except RegexNotFoundError:
+            mgid = None
+
+        if mgid is None or ':' not in mgid:
+            mgid = self._search_regex(
+                [r'data-mgid="(.*?)"', r'swfobject.embedSWF\(".*?(mgid:.*?)"'],
+                webpage, 'mgid', default=None)
+
+        if not mgid:
+            sm4_embed = self._html_search_meta(
+                'sm4:video:embed', webpage, 'sm4 embed', default='')
+            mgid = self._search_regex(
+                r'embed/(mgid:.+?)["\'&?/]', sm4_embed, 'mgid', default=None)
+
+        if not mgid:
+            mgid = self._extract_triforce_mgid(webpage)
+
+        return mgid
+
+    def _real_extract(self, url):
+        title = url_basename(url)
+        webpage = self._download_webpage(url, title)
+        mgid = self._extract_mgid(webpage)
+        videos_info = self._get_videos_info(mgid)
+        return videos_info
+
+
+class MTVServicesEmbeddedIE(MTVServicesInfoExtractor):
+    IE_NAME = 'mtvservices:embedded'
+    _VALID_URL = r'https?://media\.mtvnservices\.com/embed/(?P<mgid>.+?)(\?|/|$)'
+
+    _TEST = {
+        # From http://www.thewrap.com/peter-dinklage-sums-up-game-of-thrones-in-45-seconds-video/
+        'url': 'http://media.mtvnservices.com/embed/mgid:uma:video:mtv.com:1043906/cp~vid%3D1043906%26uri%3Dmgid%3Auma%3Avideo%3Amtv.com%3A1043906',
+        'md5': 'cb349b21a7897164cede95bd7bf3fbb9',
+        'info_dict': {
+            'id': '1043906',
+            'ext': 'mp4',
+            'title': 'Peter Dinklage Sums Up \'Game Of Thrones\' In 45 Seconds',
+            'description': '"Sexy sexy sexy, stabby stabby stabby, beautiful language," says Peter Dinklage as he tries summarizing "Game of Thrones" in under a minute.',
+            'timestamp': 1400126400,
+            'upload_date': '20140515',
+        },
+    }
 
-        return info
+    @staticmethod
+    def _extract_url(webpage):
+        mobj = re.search(
+            r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//media.mtvnservices.com/embed/.+?)\1', webpage)
+        if mobj:
+            return mobj.group('url')
 
-    def _get_videos_info(self, uri):
+    def _get_feed_url(self, uri):
         video_id = self._id_from_uri(uri)
-        data = compat_urllib_parse.urlencode({'uri': uri})
-        infoXml = self._download_webpage(self._FEED_URL +'?' + data, video_id,
-                                         u'Downloading info')
-        idoc = xml.etree.ElementTree.fromstring(infoXml.encode('utf-8'))
-        return [self._get_video_info(item) for item in idoc.findall('.//item')]
+        config = self._download_json(
+            'http://media.mtvnservices.com/pmt/e1/access/index.html?uri=%s&configtype=edge' % uri, video_id)
+        return self._remove_template_parameter(config['feedWithQueryParams'])
 
     def _real_extract(self, url):
         mobj = re.match(self._VALID_URL, url)
-        video_id = mobj.group('videoid')
+        mgid = mobj.group('mgid')
+        return self._get_videos_info(mgid)
+
+
+class MTVIE(MTVServicesInfoExtractor):
+    IE_NAME = 'mtv'
+    _VALID_URL = r'https?://(?:www\.)?mtv\.com/(?:video-clips|(?:full-)?episodes)/(?P<id>[^/?#.]+)'
+    _FEED_URL = 'http://www.mtv.com/feeds/mrss/'
+
+    _TESTS = [{
+        'url': 'http://www.mtv.com/video-clips/vl8qof/unlocking-the-truth-trailer',
+        'md5': '1edbcdf1e7628e414a8c5dcebca3d32b',
+        'info_dict': {
+            'id': '5e14040d-18a4-47c4-a582-43ff602de88e',
+            'ext': 'mp4',
+            'title': 'Unlocking The Truth|July 18, 2016|1|101|Trailer',
+            'description': '"Unlocking the Truth" premieres August 17th at 11/10c.',
+            'timestamp': 1468846800,
+            'upload_date': '20160718',
+        },
+    }, {
+        'url': 'http://www.mtv.com/full-episodes/94tujl/unlocking-the-truth-gates-of-hell-season-1-ep-101',
+        'only_matching': True,
+    }, {
+        'url': 'http://www.mtv.com/episodes/g8xu7q/teen-mom-2-breaking-the-wall-season-7-ep-713',
+        'only_matching': True,
+    }]
+
+
+class MTV81IE(InfoExtractor):
+    IE_NAME = 'mtv81'
+    _VALID_URL = r'https?://(?:www\.)?mtv81\.com/videos/(?P<id>[^/?#.]+)'
+
+    _TEST = {
+        'url': 'http://www.mtv81.com/videos/artist-to-watch/the-godfather-of-japanese-hip-hop-segment-1/',
+        'md5': '1edbcdf1e7628e414a8c5dcebca3d32b',
+        'info_dict': {
+            'id': '5e14040d-18a4-47c4-a582-43ff602de88e',
+            'ext': 'mp4',
+            'title': 'Unlocking The Truth|July 18, 2016|1|101|Trailer',
+            'description': '"Unlocking the Truth" premieres August 17th at 11/10c.',
+            'timestamp': 1468846800,
+            'upload_date': '20160718',
+        },
+    }
+
+    def _extract_mgid(self, webpage):
+        return self._search_regex(
+            r'getTheVideo\((["\'])(?P<id>mgid:.+?)\1', webpage,
+            'mgid', group='id')
 
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
         webpage = self._download_webpage(url, video_id)
+        mgid = self._extract_mgid(webpage)
+        return self.url_result('http://media.mtvnservices.com/embed/%s' % mgid)
+
+
+class MTVVideoIE(MTVServicesInfoExtractor):
+    IE_NAME = 'mtv:video'
+    _VALID_URL = r'''(?x)^https?://
+        (?:(?:www\.)?mtv\.com/videos/.+?/(?P<videoid>[0-9]+)/[^/]+$|
+           m\.mtv\.com/videos/video\.rbml\?.*?id=(?P<mgid>[^&]+))'''
+
+    _FEED_URL = 'http://www.mtv.com/player/embed/AS3/rss/'
+
+    _TESTS = [
+        {
+            'url': 'http://www.mtv.com/videos/misc/853555/ours-vh1-storytellers.jhtml',
+            'md5': '850f3f143316b1e71fa56a4edfd6e0f8',
+            'info_dict': {
+                'id': '853555',
+                'ext': 'mp4',
+                'title': 'Taylor Swift - "Ours (VH1 Storytellers)"',
+                'description': 'Album: Taylor Swift performs "Ours" for VH1 Storytellers at Harvey Mudd College.',
+                'timestamp': 1352610000,
+                'upload_date': '20121111',
+            },
+        },
+    ]
 
-        # Some videos come from Vevo.com
-        m_vevo = re.search(r'isVevoVideo = true;.*?vevoVideoId = "(.*?)";',
-                           webpage, re.DOTALL)
-        if m_vevo:
-            vevo_id = m_vevo.group(1);
-            self.to_screen(u'Vevo video detected: %s' % vevo_id)
-            return self.url_result('vevo:%s' % vevo_id, ie='Vevo')
+    def _get_thumbnail_url(self, uri, itemdoc):
+        return 'http://mtv.mtvnimages.com/uri/' + uri
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+        video_id = mobj.group('videoid')
+        uri = mobj.groupdict().get('mgid')
+        if uri is None:
+            webpage = self._download_webpage(url, video_id)
 
-        uri = self._html_search_regex(r'/uri/(.*?)\?', webpage, u'uri')
+            # Some videos come from Vevo.com
+            m_vevo = re.search(
+                r'(?s)isVevoVideo = true;.*?vevoVideoId = "(.*?)";', webpage)
+            if m_vevo:
+                vevo_id = m_vevo.group(1)
+                self.to_screen('Vevo video detected: %s' % vevo_id)
+                return self.url_result('vevo:%s' % vevo_id, ie='Vevo')
+
+            uri = self._html_search_regex(r'/uri/(.*?)\?', webpage, 'uri')
         return self._get_videos_info(uri)
+
+
+class MTVDEIE(MTVServicesInfoExtractor):
+    IE_NAME = 'mtv.de'
+    _VALID_URL = r'https?://(?:www\.)?mtv\.de/(?:artists|shows|news)/(?:[^/]+/)*(?P<id>\d+)-[^/#?]+/*(?:[#?].*)?$'
+    _TESTS = [{
+        'url': 'http://www.mtv.de/artists/10571-cro/videos/61131-traum',
+        'info_dict': {
+            'id': 'music_video-a50bc5f0b3aa4b3190aa',
+            'ext': 'flv',
+            'title': 'MusicVideo_cro-traum',
+            'description': 'Cro - Traum',
+        },
+        'params': {
+            # rtmp download
+            'skip_download': True,
+        },
+        'skip': 'Blocked at Travis CI',
+    }, {
+        # mediagen URL without query (e.g. http://videos.mtvnn.com/mediagen/e865da714c166d18d6f80893195fcb97)
+        'url': 'http://www.mtv.de/shows/933-teen-mom-2/staffeln/5353/folgen/63565-enthullungen',
+        'info_dict': {
+            'id': 'local_playlist-f5ae778b9832cc837189',
+            'ext': 'flv',
+            'title': 'Episode_teen-mom-2_shows_season-5_episode-1_full-episode_part1',
+        },
+        'params': {
+            # rtmp download
+            'skip_download': True,
+        },
+        'skip': 'Blocked at Travis CI',
+    }, {
+        'url': 'http://www.mtv.de/news/77491-mtv-movies-spotlight-pixels-teil-3',
+        'info_dict': {
+            'id': 'local_playlist-4e760566473c4c8c5344',
+            'ext': 'mp4',
+            'title': 'Article_mtv-movies-spotlight-pixels-teil-3_short-clips_part1',
+            'description': 'MTV Movies Supercut',
+        },
+        'params': {
+            # rtmp download
+            'skip_download': True,
+        },
+        'skip': 'Das Video kann zur Zeit nicht abgespielt werden.',
+    }]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+
+        webpage = self._download_webpage(url, video_id)
+
+        playlist = self._parse_json(
+            self._search_regex(
+                r'window\.pagePlaylist\s*=\s*(\[.+?\]);\n', webpage, 'page playlist'),
+            video_id)
+
+        def _mrss_url(item):
+            return item['mrss'] + item.get('mrssvars', '')
+
+        # news pages contain single video in playlist with different id
+        if len(playlist) == 1:
+            return self._get_videos_info_from_url(_mrss_url(playlist[0]), video_id)
+
+        for item in playlist:
+            item_id = item.get('id')
+            if item_id and compat_str(item_id) == video_id:
+                return self._get_videos_info_from_url(_mrss_url(item), video_id)