projects / youtube-dl / blobdiff
? search:
summary | shortlog | log | commit | commitdiff | tree
raw | inline | side by side
[extractor/common] Support multiple properties in _og_search_property
[youtube-dl] / youtube_dl / extractor / common.py
diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py
index 40ddf175c37419395cae02edb5d1fdf5118a60bf..3b6a5491dbd1f17ef0d056b385955354c82df156 100644 (file)
--- a/youtube_dl/extractor/common.py
+++ b/youtube_dl/extractor/common.py
@@ -22,8 +22,10 @@ from ..compat import (
     compat_str,
     compat_urllib_error,
     compat_urllib_parse_urlencode,
     compat_str,
     compat_urllib_error,
     compat_urllib_parse_urlencode,
+    compat_urllib_request,
     compat_urlparse,
 )
     compat_urlparse,
 )
+from ..downloader.f4m import remove_encrypted_media
 from ..utils import (
     NO_DEFAULT,
     age_restricted,
 from ..utils import (
     NO_DEFAULT,
     age_restricted,
@@ -42,13 +44,19 @@ from ..utils import (
     sanitized_Request,
     unescapeHTML,
     unified_strdate,
     sanitized_Request,
     unescapeHTML,
     unified_strdate,
+    unified_timestamp,
     url_basename,
     url_basename,
+    xpath_element,
     xpath_text,
     xpath_with_ns,
     determine_protocol,
     parse_duration,
     mimetype2ext,
     xpath_text,
     xpath_with_ns,
     determine_protocol,
     parse_duration,
     mimetype2ext,
+    update_Request,
     update_url_query,
     update_url_query,
+    parse_m3u8_attributes,
+    extract_attributes,
+    parse_codecs,
 )
 
 
 )
 
 
@@ -156,11 +164,12 @@ class InfoExtractor(object):
                         * "height" (optional, int)
                         * "resolution" (optional, string "{width}x{height"},
                                         deprecated)
                         * "height" (optional, int)
                         * "resolution" (optional, string "{width}x{height"},
                                         deprecated)
+                        * "filesize" (optional, int)
     thumbnail:      Full URL to a video thumbnail image.
     description:    Full video description.
     uploader:       Full name of the video uploader.
     license:        License name the video is licensed under.
     thumbnail:      Full URL to a video thumbnail image.
     description:    Full video description.
     uploader:       Full name of the video uploader.
     license:        License name the video is licensed under.
-    creator:        The main artist who created the video.
+    creator:        The creator of the video.
     release_date:   The date (YYYYMMDD) when the video was released.
     timestamp:      UNIX timestamp of the moment the video became available.
     upload_date:    Video upload date (YYYYMMDD).
     release_date:   The date (YYYYMMDD) when the video was released.
     timestamp:      UNIX timestamp of the moment the video became available.
     upload_date:    Video upload date (YYYYMMDD).
@@ -229,6 +238,24 @@ class InfoExtractor(object):
     episode_number: Number of the video episode within a season, as an integer.
     episode_id:     Id of the video episode, as a unicode string.
 
     episode_number: Number of the video episode within a season, as an integer.
     episode_id:     Id of the video episode, as a unicode string.
 
+    The following fields should only be used when the media is a track or a part of
+    a music album:
+
+    track:          Title of the track.
+    track_number:   Number of the track within an album or a disc, as an integer.
+    track_id:       Id of the track (useful in case of custom indexing, e.g. 6.iii),
+                    as a unicode string.
+    artist:         Artist(s) of the track.
+    genre:          Genre(s) of the track.
+    album:          Title of the album the track belongs to.
+    album_type:     Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
+    album_artist:   List of all artists appeared on the album (e.g.
+                    "Ash Borer / Fell Voices" or "Various Artists", useful for splits
+                    and compilations).
+    disc_number:    Number of the disc or other physical medium the track belongs to,
+                    as an integer.
+    release_year:   Year (YYYY) when the album was released.
+
     Unless mentioned otherwise, the fields should be Unicode strings.
 
     Unless mentioned otherwise, None is equivalent to absence of information.
     Unless mentioned otherwise, the fields should be Unicode strings.
 
     Unless mentioned otherwise, None is equivalent to absence of information.
@@ -346,7 +373,7 @@ class InfoExtractor(object):
     def IE_NAME(self):
         return compat_str(type(self).__name__[:-2])
 
     def IE_NAME(self):
         return compat_str(type(self).__name__[:-2])
 
-    def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers=None, query=None):
+    def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
         """ Returns the response handle """
         if note is None:
             self.report_download_webpage(video_id)
         """ Returns the response handle """
         if note is None:
             self.report_download_webpage(video_id)
@@ -355,12 +382,14 @@ class InfoExtractor(object):
                 self.to_screen('%s' % (note,))
             else:
                 self.to_screen('%s: %s' % (video_id, note))
                 self.to_screen('%s' % (note,))
             else:
                 self.to_screen('%s: %s' % (video_id, note))
-        # data, headers and query params will be ignored for `Request` objects
-        if isinstance(url_or_request, compat_str):
+        if isinstance(url_or_request, compat_urllib_request.Request):
+            url_or_request = update_Request(
+                url_or_request, data=data, headers=headers, query=query)
+        else:
             if query:
                 url_or_request = update_url_query(url_or_request, query)
             if query:
                 url_or_request = update_url_query(url_or_request, query)
-            if data or headers:
-                url_or_request = sanitized_Request(url_or_request, data, headers or {})
+            if data is not None or headers:
+                url_or_request = sanitized_Request(url_or_request, data, headers)
         try:
             return self._downloader.urlopen(url_or_request)
         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
         try:
             return self._downloader.urlopen(url_or_request)
         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
@@ -376,7 +405,7 @@ class InfoExtractor(object):
                 self._downloader.report_warning(errmsg)
                 return False
 
                 self._downloader.report_warning(errmsg)
                 return False
 
-    def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers=None, query=None):
+    def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
         """ Returns a tuple (page content as string, URL handle) """
         # Strip hashes from the URL (#1038)
         if isinstance(url_or_request, (compat_str, str)):
         """ Returns a tuple (page content as string, URL handle) """
         # Strip hashes from the URL (#1038)
         if isinstance(url_or_request, (compat_str, str)):
@@ -469,7 +498,7 @@ class InfoExtractor(object):
 
         return content
 
 
         return content
 
-    def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None, data=None, headers=None, query=None):
+    def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None, data=None, headers={}, query={}):
         """ Returns the data of the page as a string """
         success = False
         try_count = 0
         """ Returns the data of the page as a string """
         success = False
         try_count = 0
@@ -490,7 +519,7 @@ class InfoExtractor(object):
 
     def _download_xml(self, url_or_request, video_id,
                       note='Downloading XML', errnote='Unable to download XML',
 
     def _download_xml(self, url_or_request, video_id,
                       note='Downloading XML', errnote='Unable to download XML',
-                      transform_source=None, fatal=True, encoding=None, data=None, headers=None, query=None):
+                      transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
         """Return the xml as an xml.etree.ElementTree.Element"""
         xml_string = self._download_webpage(
             url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
         """Return the xml as an xml.etree.ElementTree.Element"""
         xml_string = self._download_webpage(
             url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
@@ -504,7 +533,7 @@ class InfoExtractor(object):
                        note='Downloading JSON metadata',
                        errnote='Unable to download JSON metadata',
                        transform_source=None,
                        note='Downloading JSON metadata',
                        errnote='Unable to download JSON metadata',
                        transform_source=None,
-                       fatal=True, encoding=None, data=None, headers=None, query=None):
+                       fatal=True, encoding=None, data=None, headers={}, query={}):
         json_string = self._download_webpage(
             url_or_request, video_id, note, errnote, fatal=fatal,
             encoding=encoding, data=data, headers=headers, query=query)
         json_string = self._download_webpage(
             url_or_request, video_id, note, errnote, fatal=fatal,
             encoding=encoding, data=data, headers=headers, query=query)
@@ -698,9 +727,14 @@ class InfoExtractor(object):
                     [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
 
     def _og_search_property(self, prop, html, name=None, **kargs):
                     [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
 
     def _og_search_property(self, prop, html, name=None, **kargs):
+        if not isinstance(prop, (list, tuple)):
+            prop = [prop]
         if name is None:
         if name is None:
-            name = 'OpenGraph %s' % prop
-        escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
+            name = 'OpenGraph %s' % prop[0]
+        og_regexes = []
+        for p in prop:
+            og_regexes.extend(self._og_regexes(p))
+        escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
         if escaped is None:
             return None
         return unescapeHTML(escaped)
         if escaped is None:
             return None
         return unescapeHTML(escaped)
@@ -724,10 +758,12 @@ class InfoExtractor(object):
         return self._og_search_property('url', html, **kargs)
 
     def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
         return self._og_search_property('url', html, **kargs)
 
     def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
+        if not isinstance(name, (list, tuple)):
+            name = [name]
         if display_name is None:
         if display_name is None:
-            display_name = name
+            display_name = name[0]
         return self._html_search_regex(
         return self._html_search_regex(
-            self._meta_regex(name),
+            [self._meta_regex(n) for n in name],
             html, display_name, fatal=fatal, group='content', **kwargs)
 
     def _dc_search_uploader(self, html):
             html, display_name, fatal=fatal, group='content', **kwargs)
 
     def _dc_search_uploader(self, html):
@@ -776,15 +812,17 @@ class InfoExtractor(object):
         return self._html_search_meta('twitter:player', html,
                                       'twitter card player')
 
         return self._html_search_meta('twitter:player', html,
                                       'twitter card player')
 
-    def _search_json_ld(self, html, video_id, **kwargs):
+    def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
         json_ld = self._search_regex(
             r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
             html, 'JSON-LD', group='json_ld', **kwargs)
         if not json_ld:
             return {}
         json_ld = self._search_regex(
             r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
             html, 'JSON-LD', group='json_ld', **kwargs)
         if not json_ld:
             return {}
-        return self._json_ld(json_ld, video_id, fatal=kwargs.get('fatal', True))
+        return self._json_ld(
+            json_ld, video_id, fatal=kwargs.get('fatal', True),
+            expected_type=expected_type)
 
 
-    def _json_ld(self, json_ld, video_id, fatal=True):
+    def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
         if isinstance(json_ld, compat_str):
             json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
         if not json_ld:
         if isinstance(json_ld, compat_str):
             json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
         if not json_ld:
@@ -792,6 +830,8 @@ class InfoExtractor(object):
         info = {}
         if json_ld.get('@context') == 'http://schema.org':
             item_type = json_ld.get('@type')
         info = {}
         if json_ld.get('@context') == 'http://schema.org':
             item_type = json_ld.get('@type')
+            if expected_type is not None and expected_type != item_type:
+                return info
             if item_type == 'TVEpisode':
                 info.update({
                     'episode': unescapeHTML(json_ld.get('name')),
             if item_type == 'TVEpisode':
                 info.update({
                     'episode': unescapeHTML(json_ld.get('name')),
@@ -810,6 +850,19 @@ class InfoExtractor(object):
                     'title': unescapeHTML(json_ld.get('headline')),
                     'description': unescapeHTML(json_ld.get('articleBody')),
                 })
                     'title': unescapeHTML(json_ld.get('headline')),
                     'description': unescapeHTML(json_ld.get('articleBody')),
                 })
+            elif item_type == 'VideoObject':
+                info.update({
+                    'url': json_ld.get('contentUrl'),
+                    'title': unescapeHTML(json_ld.get('name')),
+                    'description': unescapeHTML(json_ld.get('description')),
+                    'thumbnail': json_ld.get('thumbnailUrl'),
+                    'duration': parse_duration(json_ld.get('duration')),
+                    'timestamp': unified_timestamp(json_ld.get('uploadDate')),
+                    'filesize': float_or_none(json_ld.get('contentSize')),
+                    'tbr': int_or_none(json_ld.get('bitrate')),
+                    'width': int_or_none(json_ld.get('width')),
+                    'height': int_or_none(json_ld.get('height')),
+                })
         return dict((k, v) for k, v in info.items() if v is not None)
 
     @staticmethod
         return dict((k, v) for k, v in info.items() if v is not None)
 
     @staticmethod
@@ -819,7 +872,7 @@ class InfoExtractor(object):
         for input in re.findall(r'(?i)<input([^>]+)>', html):
             if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
                 continue
         for input in re.findall(r'(?i)<input([^>]+)>', html):
             if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
                 continue
-            name = re.search(r'name=(["\'])(?P<value>.+?)\1', input)
+            name = re.search(r'(?:name|id)=(["\'])(?P<value>.+?)\1', input)
             if not name:
                 continue
             value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
             if not name:
                 continue
             value = re.search(r'value=(["\'])(?P<value>.*?)\1', input)
@@ -851,7 +904,11 @@ class InfoExtractor(object):
                 f['ext'] = determine_ext(f['url'])
 
             if isinstance(field_preference, (list, tuple)):
                 f['ext'] = determine_ext(f['url'])
 
             if isinstance(field_preference, (list, tuple)):
-                return tuple(f.get(field) if f.get(field) is not None else -1 for field in field_preference)
+                return tuple(
+                    f.get(field)
+                    if f.get(field) is not None
+                    else ('' if field == 'format_id' else -1)
+                    for field in field_preference)
 
             preference = f.get('preference')
             if preference is None:
 
             preference = f.get('preference')
             if preference is None:
@@ -964,7 +1021,7 @@ class InfoExtractor(object):
 
     def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
                              transform_source=lambda s: fix_xml_ampersands(s).strip(),
 
     def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
                              transform_source=lambda s: fix_xml_ampersands(s).strip(),
-                             fatal=True):
+                             fatal=True, m3u8_id=None):
         manifest = self._download_xml(
             manifest_url, video_id, 'Downloading f4m manifest',
             'Unable to download f4m manifest',
         manifest = self._download_xml(
             manifest_url, video_id, 'Downloading f4m manifest',
             'Unable to download f4m manifest',
@@ -978,25 +1035,54 @@ class InfoExtractor(object):
 
         return self._parse_f4m_formats(
             manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
 
         return self._parse_f4m_formats(
             manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
-            transform_source=transform_source, fatal=fatal)
+            transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
 
     def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
                            transform_source=lambda s: fix_xml_ampersands(s).strip(),
 
     def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
                            transform_source=lambda s: fix_xml_ampersands(s).strip(),
-                           fatal=True):
+                           fatal=True, m3u8_id=None):
+        # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
+        akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
+        if akamai_pv is not None and ';' in akamai_pv.text:
+            playerVerificationChallenge = akamai_pv.text.split(';')[0]
+            if playerVerificationChallenge.strip() != '':
+                return []
+
         formats = []
         manifest_version = '1.0'
         media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
         if not media_nodes:
             manifest_version = '2.0'
             media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
         formats = []
         manifest_version = '1.0'
         media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
         if not media_nodes:
             manifest_version = '2.0'
             media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
+        # Remove unsupported DRM protected media from final formats
+        # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
+        media_nodes = remove_encrypted_media(media_nodes)
+        if not media_nodes:
+            return formats
         base_url = xpath_text(
             manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
             'base URL', default=None)
         if base_url:
             base_url = base_url.strip()
         base_url = xpath_text(
             manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
             'base URL', default=None)
         if base_url:
             base_url = base_url.strip()
+
+        bootstrap_info = xpath_element(
+            manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
+            'bootstrap info', default=None)
+
         for i, media_el in enumerate(media_nodes):
         for i, media_el in enumerate(media_nodes):
-            if manifest_version == '2.0':
-                media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
+            tbr = int_or_none(media_el.attrib.get('bitrate'))
+            width = int_or_none(media_el.attrib.get('width'))
+            height = int_or_none(media_el.attrib.get('height'))
+            format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
+            # If <bootstrapInfo> is present, the specified f4m is a
+            # stream-level manifest, and only set-level manifests may refer to
+            # external resources.  See section 11.4 and section 4 of F4M spec
+            if bootstrap_info is None:
+                media_url = None
+                # @href is introduced in 2.0, see section 11.6 of F4M spec
+                if manifest_version == '2.0':
+                    media_url = media_el.attrib.get('href')
+                if media_url is None:
+                    media_url = media_el.attrib.get('url')
                 if not media_url:
                     continue
                 manifest_url = (
                 if not media_url:
                     continue
                 manifest_url = (
@@ -1006,29 +1092,43 @@ class InfoExtractor(object):
                 # since bitrates in parent manifest (this one) and media_url manifest
                 # may differ leading to inability to resolve the format by requested
                 # bitrate in f4m downloader
                 # since bitrates in parent manifest (this one) and media_url manifest
                 # may differ leading to inability to resolve the format by requested
                 # bitrate in f4m downloader
-                if determine_ext(manifest_url) == 'f4m':
-                    formats.extend(self._extract_f4m_formats(
+                ext = determine_ext(manifest_url)
+                if ext == 'f4m':
+                    f4m_formats = self._extract_f4m_formats(
                         manifest_url, video_id, preference=preference, f4m_id=f4m_id,
                         manifest_url, video_id, preference=preference, f4m_id=f4m_id,
-                        transform_source=transform_source, fatal=fatal))
+                        transform_source=transform_source, fatal=fatal)
+                    # Sometimes stream-level manifest contains single media entry that
+                    # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
+                    # At the same time parent's media entry in set-level manifest may
+                    # contain it. We will copy it from parent in such cases.
+                    if len(f4m_formats) == 1:
+                        f = f4m_formats[0]
+                        f.update({
+                            'tbr': f.get('tbr') or tbr,
+                            'width': f.get('width') or width,
+                            'height': f.get('height') or height,
+                            'format_id': f.get('format_id') if not tbr else format_id,
+                        })
+                    formats.extend(f4m_formats)
+                    continue
+                elif ext == 'm3u8':
+                    formats.extend(self._extract_m3u8_formats(
+                        manifest_url, video_id, 'mp4', preference=preference,
+                        m3u8_id=m3u8_id, fatal=fatal))
                     continue
                     continue
-            tbr = int_or_none(media_el.attrib.get('bitrate'))
             formats.append({
             formats.append({
-                'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),
+                'format_id': format_id,
                 'url': manifest_url,
                 'url': manifest_url,
-                'ext': 'flv',
+                'ext': 'flv' if bootstrap_info is not None else None,
                 'tbr': tbr,
                 'tbr': tbr,
-                'width': int_or_none(media_el.attrib.get('width')),
-                'height': int_or_none(media_el.attrib.get('height')),
+                'width': width,
+                'height': height,
                 'preference': preference,
             })
         return formats
 
                 'preference': preference,
             })
         return formats
 
-    def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
-                              entry_protocol='m3u8', preference=None,
-                              m3u8_id=None, note=None, errnote=None,
-                              fatal=True):
-
-        formats = [{
+    def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
+        return {
             'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
             'url': m3u8_url,
             'ext': ext,
             'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
             'url': m3u8_url,
             'ext': ext,
@@ -1036,7 +1136,14 @@ class InfoExtractor(object):
             'preference': preference - 1 if preference else -1,
             'resolution': 'multiple',
             'format_note': 'Quality selection URL',
             'preference': preference - 1 if preference else -1,
             'resolution': 'multiple',
             'format_note': 'Quality selection URL',
-        }]
+        }
+
+    def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
+                              entry_protocol='m3u8', preference=None,
+                              m3u8_id=None, note=None, errnote=None,
+                              fatal=True, live=False):
+
+        formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
 
         format_url = lambda u: (
             u
 
         format_url = lambda u: (
             u
@@ -1076,23 +1183,11 @@ class InfoExtractor(object):
             }]
         last_info = None
         last_media = None
             }]
         last_info = None
         last_media = None
-        kv_rex = re.compile(
-            r'(?P<key>[a-zA-Z_-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)')
         for line in m3u8_doc.splitlines():
             if line.startswith('#EXT-X-STREAM-INF:'):
         for line in m3u8_doc.splitlines():
             if line.startswith('#EXT-X-STREAM-INF:'):
-                last_info = {}
-                for m in kv_rex.finditer(line):
-                    v = m.group('val')
-                    if v.startswith('"'):
-                        v = v[1:-1]
-                    last_info[m.group('key')] = v
+                last_info = parse_m3u8_attributes(line)
             elif line.startswith('#EXT-X-MEDIA:'):
             elif line.startswith('#EXT-X-MEDIA:'):
-                last_media = {}
-                for m in kv_rex.finditer(line):
-                    v = m.group('val')
-                    if v.startswith('"'):
-                        v = v[1:-1]
-                    last_media[m.group('key')] = v
+                last_media = parse_m3u8_attributes(line)
             elif line.startswith('#') or not line.strip():
                 continue
             else:
             elif line.startswith('#') or not line.strip():
                 continue
             else:
@@ -1103,13 +1198,21 @@ class InfoExtractor(object):
                 format_id = []
                 if m3u8_id:
                     format_id.append(m3u8_id)
                 format_id = []
                 if m3u8_id:
                     format_id.append(m3u8_id)
-                last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') != 'SUBTITLES' else None
-                format_id.append(last_media_name if last_media_name else '%d' % (tbr if tbr else len(formats)))
+                last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') not in ('SUBTITLES', 'CLOSED-CAPTIONS') else None
+                # Despite specification does not mention NAME attribute for
+                # EXT-X-STREAM-INF it still sometimes may be present
+                stream_name = last_info.get('NAME') or last_media_name
+                # Bandwidth of live streams may differ over time thus making
+                # format_id unpredictable. So it's better to keep provided
+                # format_id intact.
+                if not live:
+                    format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
                 f = {
                     'format_id': '-'.join(format_id),
                     'url': format_url(line.strip()),
                     'tbr': tbr,
                     'ext': ext,
                 f = {
                     'format_id': '-'.join(format_id),
                     'url': format_url(line.strip()),
                     'tbr': tbr,
                     'ext': ext,
+                    'fps': float_or_none(last_info.get('FRAME-RATE')),
                     'protocol': entry_protocol,
                     'preference': preference,
                 }
                     'protocol': entry_protocol,
                     'preference': preference,
                 }
@@ -1118,24 +1221,17 @@ class InfoExtractor(object):
                     width_str, height_str = resolution.split('x')
                     f['width'] = int(width_str)
                     f['height'] = int(height_str)
                     width_str, height_str = resolution.split('x')
                     f['width'] = int(width_str)
                     f['height'] = int(height_str)
-                codecs = last_info.get('CODECS')
-                if codecs:
-                    vcodec, acodec = [None] * 2
-                    va_codecs = codecs.split(',')
-                    if len(va_codecs) == 1:
-                        # Audio only entries usually come with single codec and
-                        # no resolution. For more robustness we also check it to
-                        # be mp4 audio.
-                        if not resolution and va_codecs[0].startswith('mp4a'):
-                            vcodec, acodec = 'none', va_codecs[0]
-                        else:
-                            vcodec = va_codecs[0]
-                    else:
-                        vcodec, acodec = va_codecs[:2]
+                # Unified Streaming Platform
+                mobj = re.search(
+                    r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
+                if mobj:
+                    abr, vbr = mobj.groups()
+                    abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
                     f.update({
                     f.update({
-                        'acodec': acodec,
-                        'vcodec': vcodec,
+                        'vbr': vbr,
+                        'abr': abr,
                     })
                     })
+                f.update(parse_codecs(last_info.get('CODECS')))
                 if last_media is not None:
                     f['m3u8_media'] = last_media
                     last_media = None
                 if last_media is not None:
                     f['m3u8_media'] = last_media
                     last_media = None
@@ -1236,21 +1332,21 @@ class InfoExtractor(object):
         m3u8_count = 0
 
         srcs = []
         m3u8_count = 0
 
         srcs = []
-        videos = smil.findall(self._xpath_ns('.//video', namespace))
-        for video in videos:
-            src = video.get('src')
+        media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
+        for medium in media:
+            src = medium.get('src')
             if not src or src in srcs:
                 continue
             srcs.append(src)
 
             if not src or src in srcs:
                 continue
             srcs.append(src)
 
-            bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
-            filesize = int_or_none(video.get('size') or video.get('fileSize'))
-            width = int_or_none(video.get('width'))
-            height = int_or_none(video.get('height'))
-            proto = video.get('proto')
-            ext = video.get('ext')
+            bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
+            filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
+            width = int_or_none(medium.get('width'))
+            height = int_or_none(medium.get('height'))
+            proto = medium.get('proto')
+            ext = medium.get('ext')
             src_ext = determine_ext(src)
             src_ext = determine_ext(src)
-            streamer = video.get('streamer') or base
+            streamer = medium.get('streamer') or base
 
             if proto == 'rtmp' or streamer.startswith('rtmp'):
                 rtmp_count += 1
 
             if proto == 'rtmp' or streamer.startswith('rtmp'):
                 rtmp_count += 1
@@ -1324,7 +1420,7 @@ class InfoExtractor(object):
             if not src or src in urls:
                 continue
             urls.append(src)
             if not src or src in urls:
                 continue
             urls.append(src)
-            ext = textstream.get('ext') or determine_ext(src) or mimetype2ext(textstream.get('type'))
+            ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
             lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
             subtitles.setdefault(lang, []).append({
                 'url': src,
             lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
             subtitles.setdefault(lang, []).append({
                 'url': src,
@@ -1390,6 +1486,13 @@ class InfoExtractor(object):
             compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, formats_dict=formats_dict)
 
     def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}):
             compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, formats_dict=formats_dict)
 
     def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}):
+        """
+        Parse formats from MPD manifest.
+        References:
+         1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
+            http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
+         2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
+        """
         if mpd_doc.get('type') == 'dynamic':
             return []
 
         if mpd_doc.get('type') == 'dynamic':
             return []
 
@@ -1422,8 +1525,16 @@ class InfoExtractor(object):
                         s_e = segment_timeline.findall(_add_ns('S'))
                         if s_e:
                             ms_info['total_number'] = 0
                         s_e = segment_timeline.findall(_add_ns('S'))
                         if s_e:
                             ms_info['total_number'] = 0
+                            ms_info['s'] = []
                             for s in s_e:
                             for s in s_e:
-                                ms_info['total_number'] += 1 + int(s.get('r', '0'))
+                                r = int(s.get('r', 0))
+                                ms_info['total_number'] += 1 + r
+                                ms_info['s'].append({
+                                    't': int(s.get('t', 0)),
+                                    # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
+                                    'd': int(s.attrib['d']),
+                                    'r': r,
+                                })
                     else:
                         timescale = segment_template.get('timescale')
                         if timescale:
                     else:
                         timescale = segment_template.get('timescale')
                         if timescale:
@@ -1460,7 +1571,7 @@ class InfoExtractor(object):
                         continue
                     representation_attrib = adaptation_set.attrib.copy()
                     representation_attrib.update(representation.attrib)
                         continue
                     representation_attrib = adaptation_set.attrib.copy()
                     representation_attrib.update(representation.attrib)
-                    # According to page 41 of ISO/IEC 29001-1:2014, @mimeType is mandatory
+                    # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
                     mime_type = representation_attrib['mimeType']
                     content_type = mime_type.split('/')[0]
                     if content_type == 'text':
                     mime_type = representation_attrib['mimeType']
                     content_type = mime_type.split('/')[0]
                     if content_type == 'text':
@@ -1504,9 +1615,40 @@ class InfoExtractor(object):
                                 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
                             media_template = representation_ms_info['media_template']
                             media_template = media_template.replace('$RepresentationID$', representation_id)
                                 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
                             media_template = representation_ms_info['media_template']
                             media_template = media_template.replace('$RepresentationID$', representation_id)
-                            media_template = re.sub(r'\$(Number|Bandwidth)(?:%(0\d+)d)?\$', r'%(\1)\2d', media_template)
+                            media_template = re.sub(r'\$(Number|Bandwidth|Time)\$', r'%(\1)d', media_template)
+                            media_template = re.sub(r'\$(Number|Bandwidth|Time)%([^$]+)\$', r'%(\1)\2', media_template)
                             media_template.replace('$$', '$')
                             media_template.replace('$$', '$')
-                            representation_ms_info['segment_urls'] = [media_template % {'Number': segment_number, 'Bandwidth': representation_attrib.get('bandwidth')} for segment_number in range(representation_ms_info['start_number'], representation_ms_info['total_number'] + representation_ms_info['start_number'])]
+
+                            # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
+                            # can't be used at the same time
+                            if '%(Number' in media_template:
+                                representation_ms_info['segment_urls'] = [
+                                    media_template % {
+                                        'Number': segment_number,
+                                        'Bandwidth': representation_attrib.get('bandwidth'),
+                                    }
+                                    for segment_number in range(
+                                        representation_ms_info['start_number'],
+                                        representation_ms_info['total_number'] + representation_ms_info['start_number'])]
+                            else:
+                                representation_ms_info['segment_urls'] = []
+                                segment_time = 0
+
+                                def add_segment_url():
+                                    representation_ms_info['segment_urls'].append(
+                                        media_template % {
+                                            'Time': segment_time,
+                                            'Bandwidth': representation_attrib.get('bandwidth'),
+                                        }
+                                    )
+
+                                for num, s in enumerate(representation_ms_info['s']):
+                                    segment_time = s.get('t') or segment_time
+                                    add_segment_url()
+                                    for r in range(s.get('r', 0)):
+                                        segment_time += s['d']
+                                        add_segment_url()
+                                    segment_time += s['d']
                         if 'segment_urls' in representation_ms_info:
                             f.update({
                                 'segment_urls': representation_ms_info['segment_urls'],
                         if 'segment_urls' in representation_ms_info:
                             f.update({
                                 'segment_urls': representation_ms_info['segment_urls'],
@@ -1533,6 +1675,62 @@ class InfoExtractor(object):
                         self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
         return formats
 
                         self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
         return formats
 
+    def _parse_html5_media_entries(self, base_url, webpage):
+        def absolute_url(video_url):
+            return compat_urlparse.urljoin(base_url, video_url)
+
+        def parse_content_type(content_type):
+            if not content_type:
+                return {}
+            ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
+            if ctr:
+                mimetype, codecs = ctr.groups()
+                f = parse_codecs(codecs)
+                f['ext'] = mimetype2ext(mimetype)
+                return f
+            return {}
+
+        entries = []
+        for media_tag, media_type, media_content in re.findall(r'(?s)(<(?P<tag>video|audio)[^>]*>)(.*?)</(?P=tag)>', webpage):
+            media_info = {
+                'formats': [],
+                'subtitles': {},
+            }
+            media_attributes = extract_attributes(media_tag)
+            src = media_attributes.get('src')
+            if src:
+                media_info['formats'].append({
+                    'url': absolute_url(src),
+                    'vcodec': 'none' if media_type == 'audio' else None,
+                })
+            media_info['thumbnail'] = media_attributes.get('poster')
+            if media_content:
+                for source_tag in re.findall(r'<source[^>]+>', media_content):
+                    source_attributes = extract_attributes(source_tag)
+                    src = source_attributes.get('src')
+                    if not src:
+                        continue
+                    f = parse_content_type(source_attributes.get('type'))
+                    f.update({
+                        'url': absolute_url(src),
+                        'vcodec': 'none' if media_type == 'audio' else None,
+                    })
+                    media_info['formats'].append(f)
+                for track_tag in re.findall(r'<track[^>]+>', media_content):
+                    track_attributes = extract_attributes(track_tag)
+                    kind = track_attributes.get('kind')
+                    if not kind or kind == 'subtitles':
+                        src = track_attributes.get('src')
+                        if not src:
+                            continue
+                        lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
+                        media_info['subtitles'].setdefault(lang, []).append({
+                            'url': absolute_url(src),
+                        })
+            if media_info['formats']:
+                entries.append(media_info)
+        return entries
+
     def _live_title(self, name):
         """ Generate the title for a live video """
         now = datetime.datetime.now()
     def _live_title(self, name):
         """ Generate the title for a live video """
         now = datetime.datetime.now()
@@ -1593,7 +1791,7 @@ class InfoExtractor(object):
 
         any_restricted = False
         for tc in self.get_testcases(include_onlymatching=False):
 
         any_restricted = False
         for tc in self.get_testcases(include_onlymatching=False):
-            if 'playlist' in tc:
+            if tc.get('playlist', []):
                 tc = tc['playlist'][0]
             is_restricted = age_restricted(
                 tc.get('info_dict', {}).get('age_limit'), age_limit)
                 tc = tc['playlist'][0]
             is_restricted = age_restricted(
                 tc.get('info_dict', {}).get('age_limit'), age_limit)
@@ -1646,6 +1844,13 @@ class InfoExtractor(object):
     def _mark_watched(self, *args, **kwargs):
         raise NotImplementedError('This method must be implemented by subclasses')
 
     def _mark_watched(self, *args, **kwargs):
         raise NotImplementedError('This method must be implemented by subclasses')
 
+    def geo_verification_headers(self):
+        headers = {}
+        geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
+        if geo_verification_proxy:
+            headers['Ytdl-request-proxy'] = geo_verification_proxy
+        return headers
+
 
 class SearchInfoExtractor(InfoExtractor):
     """
 
 class SearchInfoExtractor(InfoExtractor):
     """
Totally unverified clone of youtube-dl