X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/90cd207d52d77e075e62cc256cc39dceee05bbf3..382a868cee069f08aacf0b89c9d689ec420d6b2c:/youtube_dl/extractor/mtv.py

diff --git a/youtube_dl/extractor/mtv.py b/youtube_dl/extractor/mtv.py
index 24a79ae..bc7f49e 100644
--- a/youtube_dl/extractor/mtv.py
+++ b/youtube_dl/extractor/mtv.py
@@ -1,43 +1,30 @@
+from __future__ import unicode_literals
+
 import re
-import xml.etree.ElementTree
 
-from .common import InfoExtractor
-from ..utils import (
+from .subtitles import SubtitlesInfoExtractor
+from ..compat import (
     compat_urllib_parse,
+    compat_urllib_request,
+    compat_str,
+)
+from ..utils import (
     ExtractorError,
+    find_xpath_attr,
+    fix_xml_ampersands,
+    HEADRequest,
+    unescapeHTML,
+    url_basename,
+    RegexNotFoundError,
 )
 
+
 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'add_ie': ['Vevo'],
-            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(SubtitlesInfoExtractor):
+    _MOBILE_TEMPLATE = None
 
     @staticmethod
     def _id_from_uri(uri):
@@ -48,45 +35,98 @@ class MTVIE(InfoExtractor):
     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 rtmp_video_url
+        base = 'http://viacommtvstrmfs.fplive.net/'
         return base + m.group('finalid')
 
+    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 = compat_urllib_request.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):
+        if re.match(r'.*/(error_country_block\.swf|geoblock\.mp4)$', 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')),
-                                })
+                if rtmp_video_url.endswith('siteunavail.png'):
+                    continue
+                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.')
+        self._sort_formats(formats)
         return formats
 
+    def _extract_subtitles(self, mdoc, mtvn_id):
+        subtitles = {}
+        FORMATS = {
+            'scc': 'cea-608',
+            'eia-608': 'cea-608',
+            'xml': 'ttml',
+        }
+        subtitles_format = FORMATS.get(
+            self._downloader.params.get('subtitlesformat'), 'ttml')
+        for transcript in mdoc.findall('.//transcript'):
+            if transcript.get('kind') != 'captions':
+                continue
+            lang = transcript.get('srclang')
+            for typographic in transcript.findall('./typographic'):
+                captions_format = typographic.get('format')
+                if captions_format == subtitles_format:
+                    subtitles[lang] = compat_str(typographic.get('src'))
+                    break
+        if self._downloader.params.get('listsubtitles', False):
+            self._list_available_subtitles(mtvn_id, subtitles)
+        return self.extract_subtitles(mtvn_id, subtitles)
+
     def _get_video_info(self, itemdoc):
         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']
         # Remove the templates, like &device={device}
-        mediagen_url = re.sub(r'&[^=]*?={.*?}(?=(&|$))', u'', mediagen_url)
+        mediagen_url = re.sub(r'&[^=]*?={.*?}(?=(&|$))', '', mediagen_url)
         if 'acceptMethods' not in mediagen_url:
             mediagen_url += '&acceptMethods=fms'
-        mediagen_page = self._download_webpage(mediagen_url, video_id,
-                                               u'Downloading video urls')
+
+        mediagen_doc = self._download_xml(mediagen_url, video_id,
+                                          'Downloading video urls')
 
         description_node = itemdoc.find('description')
         if description_node is not None:
@@ -94,40 +134,156 @@ class MTVIE(InfoExtractor):
         else:
             description = None
 
-        info = {
-            'title': itemdoc.find('title').text,
-            'formats': self._extract_video_formats(mediagen_page),
+        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('.//{http://search.yahoo.com/mrss/}title')
+        if title_el is None:
+            title_el = itemdoc.find('.//title')
+            if title_el.text is None:
+                title_el = None
+
+        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
+
+        return {
+            'title': title,
+            'formats': self._extract_video_formats(mediagen_doc, mtvn_id),
+            'subtitles': self._extract_subtitles(mediagen_doc, mtvn_id),
             'id': video_id,
             'thumbnail': self._get_thumbnail_url(uri, itemdoc),
             'description': description,
         }
 
-        # TODO: Remove when #980 has been merged
-        info.update(info['formats'][-1])
-
-        return info
-
     def _get_videos_info(self, uri):
         video_id = self._id_from_uri(uri)
+        feed_url = self._get_feed_url(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')]
+        idoc = self._download_xml(
+            feed_url + '?' + data, video_id,
+            'Downloading info', transform_source=fix_xml_ampersands)
+        return self.playlist_result(
+            [self._get_video_info(item) for item in idoc.findall('.//item')])
+
+    def _real_extract(self, url):
+        title = url_basename(url)
+        webpage = self._download_webpage(url, title)
+        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')
+
+        videos_info = self._get_videos_info(mgid)
+        if self._downloader.params.get('listsubtitles', False):
+            return
+        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.',
+        },
+    }
+
+    def _get_feed_url(self, uri):
+        video_id = self._id_from_uri(uri)
+        site_id = uri.replace(video_id, '')
+        config_url = ('http://media.mtvnservices.com/pmt/e1/players/{0}/'
+                      'context4/context5/config.xml'.format(site_id))
+        config_doc = self._download_xml(config_url, video_id)
+        feed_node = config_doc.find('.//feed')
+        feed_url = feed_node.text.strip().split('?')[0]
+        return feed_url
 
     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):
+    _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.',
+            },
+        },
+    ]
+
+    def _get_thumbnail_url(self, uri, itemdoc):
+        return 'http://mtv.mtvnimages.com/uri/' + uri
 
-        webpage = self._download_webpage(url, video_id)
+    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)
 
-        # 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')
+            # 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, u'uri')
+            uri = self._html_search_regex(r'/uri/(.*?)\?', webpage, 'uri')
         return self._get_videos_info(uri)
+
+
+class MTVIggyIE(MTVServicesInfoExtractor):
+    IE_NAME = 'mtviggy.com'
+    _VALID_URL = r'https?://www\.mtviggy\.com/videos/.+'
+    _TEST = {
+        'url': 'http://www.mtviggy.com/videos/arcade-fire-behind-the-scenes-at-the-biggest-music-experiment-yet/',
+        'info_dict': {
+            'id': '984696',
+            'ext': 'mp4',
+            'title': 'Arcade Fire: Behind the Scenes at the Biggest Music Experiment Yet',
+        }
+    }
+    _FEED_URL = 'http://all.mtvworldverticals.com/feed-xml/'