debian/NEWS: Write news about new behavior of youtube-dl.

[youtubedl] / youtube_dl / extractor / mtv.py

diff --git a/youtube_dl/extractor/mtv.py b/youtube_dl/extractor/mtv.py
index 969db71139b1f81d290ad6549ed3b3d8207da0c7..5447b6c0cab098b895eda0e9f2b3b266fb65a7b0 100644 (file)
--- a/youtube_dl/extractor/mtv.py
+++ b/youtube_dl/extractor/mtv.py
@@ -1,80 +1,195 @@
+from __future__ import unicode_literals
+

 import re
 import re

-import socket
-import xml.etree.ElementTree

 
 from .common import InfoExtractor
 from ..utils import (
 
 from .common import InfoExtractor
 from ..utils import (

-    compat_http_client,
-    compat_str,
-    compat_urllib_error,
-    compat_urllib_request,
-
+    compat_urllib_parse,

     ExtractorError,
     ExtractorError,

+    find_xpath_attr,
+    fix_xml_ampersands,
+    url_basename,
+    RegexNotFoundError,

 )
 
 
 )
 
 

-class MTVIE(InfoExtractor):
-    _VALID_URL = r'^(?P<proto>https?://)?(?:www\.)?mtv\.com/videos/[^/]+/(?P<videoid>[0-9]+)/[^/]+$'
-    _WORKING = False
-
-    def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        if mobj is None:
-            raise ExtractorError(u'Invalid URL: %s' % url)
-        if not mobj.group('proto'):
-            url = 'http://' + url
-        video_id = mobj.group('videoid')
-
-        webpage = self._download_webpage(url, video_id)
-
-        # Some videos come from Vevo.com
-        m_vevo = re.search(r'isVevoVideo = true;.*?vevoVideoId = "(.*?)";',
-                           webpage, re.DOTALL)
-        if m_vevo:
-            vevo_id = m_vevo.group(1);
-            self.to_screen(u'Vevo video detected: %s' % vevo_id)
-            return self.url_result('vevo:%s' % vevo_id, ie='Vevo')
-
-        #song_name = self._html_search_regex(r'<meta name="mtv_vt" content="([^"]+)"/>',
-        #    webpage, u'song name', fatal=False)
-
-        video_title = self._html_search_regex(r'<meta name="mtv_an" content="([^"]+)"/>',
-            webpage, u'title')
-
-        mtvn_uri = self._html_search_regex(r'<meta name="mtvn_uri" content="([^"]+)"/>',
-            webpage, u'mtvn_uri', fatal=False)
-
-        content_id = self._search_regex(r'MTVN.Player.defaultPlaylistId = ([0-9]+);',
-            webpage, u'content id', fatal=False)
-
-        videogen_url = 'http://www.mtv.com/player/includes/mediaGen.jhtml?uri=' + mtvn_uri + '&id=' + content_id + '&vid=' + video_id + '&ref=www.mtvn.com&viewUri=' + mtvn_uri
+def _media_xml_tag(tag):
+    return '{http://search.yahoo.com/mrss/}%s' % tag
+
+
+class MTVServicesInfoExtractor(InfoExtractor):
+    @staticmethod
+    def _id_from_uri(uri):
+        return uri.split(':')[-1]
+
+    # This was originally implemented for ComedyCentral, but it also works here
+    @staticmethod
+    def _transform_rtmp_url(rtmp_video_url):
+        m = re.match(r'^rtmpe?://.*?/(?P<finalid>gsp\..+?/.*)$', rtmp_video_url)
+        if not m:
+            return rtmp_video_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 _get_thumbnail_url(self, uri, itemdoc):
+        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_video_formats(self, mdoc):
+        if re.match(r'.*/error_country_block\.swf$', mdoc.find('.//src').text) is not None:
+            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.')
+        return formats
+
+    def _get_video_info(self, itemdoc):
+        uri = itemdoc.find('guid').text
+        video_id = self._id_from_uri(uri)

         self.report_extraction(video_id)
         self.report_extraction(video_id)

-        request = compat_urllib_request.Request(videogen_url)
-        try:
-            metadataXml = compat_urllib_request.urlopen(request).read()
-        except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
-            raise ExtractorError(u'Unable to download video metadata: %s' % compat_str(err))
+        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'&[^=]*?={.*?}(?=(&|$))', '', mediagen_url)
+        if 'acceptMethods' not in mediagen_url:
+            mediagen_url += '&acceptMethods=fms'
+
+        mediagen_doc = self._download_xml(mediagen_url, video_id,
+            'Downloading video urls')
+
+        description_node = itemdoc.find('description')
+        if description_node is not None:
+            description = description_node.text.strip()
+        else:
+            description = None
+
+        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()
+
+        return {
+            'title': title,
+            'formats': self._extract_video_formats(mediagen_doc),
+            'id': video_id,
+            'thumbnail': self._get_thumbnail_url(uri, itemdoc),
+            'description': description,
+        }

 
 

-        mdoc = xml.etree.ElementTree.fromstring(metadataXml)
-        renditions = mdoc.findall('.//rendition')
+    def _get_videos_info(self, uri):
+        video_id = self._id_from_uri(uri)
+        data = compat_urllib_parse.urlencode({'uri': uri})

 
 

-        # For now, always pick the highest quality.
-        rendition = renditions[-1]
+        idoc = self._download_xml(
+            self._FEED_URL + '?' + data, video_id,
+            'Downloading info', transform_source=fix_xml_ampersands)
+        return [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:
         try:

-            _,_,ext = rendition.attrib['type'].partition('/')
-            format = ext + '-' + rendition.attrib['width'] + 'x' + rendition.attrib['height'] + '_' + rendition.attrib['bitrate']
-            video_url = rendition.find('./src').text
-        except KeyError:
-            raise ExtractorError('Invalid rendition field.')
+            # 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 = self._search_regex(
+                [r'data-mgid="(.*?)"', r'swfobject.embedSWF\(".*?(mgid:.*?)"'],
+                webpage, u'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',
+            'file': '853555.mp4',
+            'md5': '850f3f143316b1e71fa56a4edfd6e0f8',
+            'info_dict': {
+                'title': 'Taylor Swift - "Ours (VH1 Storytellers)"',
+                'description': 'Album: Taylor Swift performs "Ours" for VH1 Storytellers at Harvey Mudd College.',
+            },
+        },
+        {
+            'add_ie': ['Vevo'],
+            'url': 'http://www.mtv.com/videos/taylor-swift/916187/everything-has-changed-ft-ed-sheeran.jhtml',
+            'file': 'USCJY1331283.mp4',
+            'md5': '73b4e7fcadd88929292fe52c3ced8caf',
+            'info_dict': {
+                'title': 'Everything Has Changed',
+                'upload_date': '20130606',
+                'uploader': 'Taylor Swift',
+            },
+            'skip': 'VEVO is only available in some countries',
+        },
+    ]
+
+    def _get_thumbnail_url(self, uri, itemdoc):
+        return 'http://mtv.mtvnimages.com/uri/' + uri

 
 

-        info = {
-            'id': video_id,
-            'url': video_url,
-            'upload_date': None,
-            'title': video_title,
-            'ext': ext,
-            'format': format,
+    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('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 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',

         }
         }

-
-        return [info]
+    }
+    _FEED_URL = 'http://all.mtvworldverticals.com/feed-xml/'