X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fyoutube.py;h=323681960e8788a1c9ba735a0e3166f1249a3195;hb=d3f007af183a0f5ed278602128e6bba3cc1350b3;hp=fa1a2e54495ccad9358bbac4418bb458fe8fe5b2;hpb=d8d24a922adb0646f9f78b2fc4a287a72fa7ff73;p=youtube-dl

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index fa1a2e544..323681960 100644
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -17,6 +17,8 @@ from ..compat import (
     compat_chr,
     compat_parse_qs,
     compat_urllib_parse,
+    compat_urllib_parse_unquote,
+    compat_urllib_parse_unquote_plus,
     compat_urllib_request,
     compat_urlparse,
     compat_str,
@@ -29,9 +31,11 @@ from ..utils import (
     get_element_by_id,
     int_or_none,
     orderedSet,
+    str_to_int,
     unescapeHTML,
     unified_strdate,
     uppercase_escape,
+    ISO3166Utils,
 )
 
 
@@ -518,6 +522,38 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 'skip_download': 'requires avconv',
             }
         },
+        # Extraction from multiple DASH manifests (https://github.com/rg3/youtube-dl/pull/6097)
+        {
+            'url': 'https://www.youtube.com/watch?v=FIl7x6_3R5Y',
+            'info_dict': {
+                'id': 'FIl7x6_3R5Y',
+                'ext': 'mp4',
+                'title': 'md5:7b81415841e02ecd4313668cde88737a',
+                'description': 'md5:116377fd2963b81ec4ce64b542173306',
+                'upload_date': '20150625',
+                'uploader_id': 'dorappi2000',
+                'uploader': 'dorappi2000',
+                'formats': 'mincount:33',
+            },
+        },
+        # DASH manifest with segment_list
+        {
+            'url': 'https://www.youtube.com/embed/CsmdDsKjzN8',
+            'md5': '8ce563a1d667b599d21064e982ab9e31',
+            'info_dict': {
+                'id': 'CsmdDsKjzN8',
+                'ext': 'mp4',
+                'upload_date': '20150501',  # According to '<meta itemprop="datePublished"', but in other places it's 20150510
+                'uploader': 'Airtek',
+                'description': 'Retransmisión en directo de la XVIII media maratón de Zaragoza.',
+                'uploader_id': 'UCzTzUmjXxxacNnL8I3m4LnQ',
+                'title': 'Retransmisión XVIII Media maratón Zaragoza 2015',
+            },
+            'params': {
+                'youtube_include_dash_manifest': True,
+                'format': '135',  # bestvideo
+            }
+        },
     ]
 
     def __init__(self, *args, **kwargs):
@@ -782,7 +818,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         return self._download_webpage(url, video_id, note='Searching for annotations.', errnote='Unable to download video annotations.')
 
     def _parse_dash_manifest(
-            self, video_id, dash_manifest_url, player_url, age_gate):
+            self, video_id, dash_manifest_url, player_url, age_gate, fatal=True):
         def decrypt_sig(mobj):
             s = mobj.group(1)
             dec_s = self._decrypt_signature(s, video_id, player_url, age_gate)
@@ -791,7 +827,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         dash_doc = self._download_xml(
             dash_manifest_url, video_id,
             note='Downloading DASH manifest',
-            errnote='Could not download DASH manifest')
+            errnote='Could not download DASH manifest',
+            fatal=fatal)
+
+        if dash_doc is False:
+            return []
 
         formats = []
         for a in dash_doc.findall('.//{urn:mpeg:DASH:schema:MPD:2011}AdaptationSet'):
@@ -804,6 +844,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     # TODO implement WebVTT downloading
                     pass
                 elif mime_type.startswith('audio/') or mime_type.startswith('video/'):
+                    segment_list = r.find('{urn:mpeg:DASH:schema:MPD:2011}SegmentList')
                     format_id = r.attrib['id']
                     video_url = url_el.text
                     filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength'))
@@ -817,6 +858,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                         'filesize': filesize,
                         'fps': int_or_none(r.attrib.get('frameRate')),
                     }
+                    if segment_list is not None:
+                        f.update({
+                            'initialization_url': segment_list.find('{urn:mpeg:DASH:schema:MPD:2011}Initialization').attrib['sourceURL'],
+                            'segment_urls': [segment.attrib.get('media') for segment in segment_list.findall('{urn:mpeg:DASH:schema:MPD:2011}SegmentURL')],
+                            'protocol': 'http_dash_segments',
+                        })
                     try:
                         existing_format = next(
                             fo for fo in formats
@@ -824,6 +871,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     except StopIteration:
                         full_info = self._formats.get(format_id, {}).copy()
                         full_info.update(f)
+                        codecs = r.attrib.get('codecs')
+                        if codecs:
+                            if full_info.get('acodec') == 'none' and 'vcodec' not in full_info:
+                                full_info['vcodec'] = codecs
+                            elif full_info.get('vcodec') == 'none' and 'acodec' not in full_info:
+                                full_info['acodec'] = codecs
                         formats.append(full_info)
                     else:
                         existing_format.update(f)
@@ -839,7 +892,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         # Extract original video URL from URL with redirection, like age verification, using next_url parameter
         mobj = re.search(self._NEXT_URL_RE, url)
         if mobj:
-            url = proto + '://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
+            url = proto + '://www.youtube.com/' + compat_urllib_parse_unquote(mobj.group(1)).lstrip('/')
         video_id = self.extract_id(url)
 
         # Get video webpage
@@ -862,6 +915,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
         # Get video info
         embed_webpage = None
+        is_live = None
         if re.search(r'player-age-gate-content">', video_webpage) is not None:
             age_gate = True
             # We simulate the access to the video from www.youtube.com/v/{video_id}
@@ -883,6 +937,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             add_dash_mpd(video_info)
         else:
             age_gate = False
+            video_info = None
             # Try looking directly into the video webpage
             mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});', video_webpage)
             if mobj:
@@ -893,29 +948,39 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     # Convert to the same format returned by compat_parse_qs
                     video_info = dict((k, [v]) for k, v in args.items())
                     add_dash_mpd(video_info)
-            # We also try looking in get_video_info since it may contain different dashmpd
-            # URL that points to a DASH manifest with possibly different itag set (some itags
-            # are missing from DASH manifest pointed by webpage's dashmpd, some - from DASH
-            # manifest pointed by get_video_info's dashmpd).
-            # The general idea is to take a union of itags of both DASH manifests (for example
-            # video with such 'manifest behavior' see https://github.com/rg3/youtube-dl/issues/6093)
-            self.report_video_info_webpage_download(video_id)
-            for el_type in ['&el=info', '&el=embedded', '&el=detailpage', '&el=vevo', '']:
-                video_info_url = (
-                    '%s://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
-                    % (proto, video_id, el_type))
-                video_info_webpage = self._download_webpage(
-                    video_info_url,
-                    video_id, note=False,
-                    errnote='unable to download video info webpage')
-                get_video_info = compat_parse_qs(video_info_webpage)
-                add_dash_mpd(get_video_info)
-                if not video_info:
-                    video_info = get_video_info
-                if 'token' in get_video_info:
-                    break
+                if args.get('livestream') == '1' or args.get('live_playback') == 1:
+                    is_live = True
+            if not video_info or self._downloader.params.get('youtube_include_dash_manifest', True):
+                # We also try looking in get_video_info since it may contain different dashmpd
+                # URL that points to a DASH manifest with possibly different itag set (some itags
+                # are missing from DASH manifest pointed by webpage's dashmpd, some - from DASH
+                # manifest pointed by get_video_info's dashmpd).
+                # The general idea is to take a union of itags of both DASH manifests (for example
+                # video with such 'manifest behavior' see https://github.com/rg3/youtube-dl/issues/6093)
+                self.report_video_info_webpage_download(video_id)
+                for el_type in ['&el=info', '&el=embedded', '&el=detailpage', '&el=vevo', '']:
+                    video_info_url = (
+                        '%s://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
+                        % (proto, video_id, el_type))
+                    video_info_webpage = self._download_webpage(
+                        video_info_url,
+                        video_id, note=False,
+                        errnote='unable to download video info webpage')
+                    get_video_info = compat_parse_qs(video_info_webpage)
+                    if get_video_info.get('use_cipher_signature') != ['True']:
+                        add_dash_mpd(get_video_info)
+                    if not video_info:
+                        video_info = get_video_info
+                    if 'token' in get_video_info:
+                        break
         if 'token' not in video_info:
             if 'reason' in video_info:
+                if 'The uploader has not made this video available in your country.' in video_info['reason']:
+                    regions_allowed = self._html_search_meta('regionsAllowed', video_webpage, default=None)
+                    if regions_allowed is not None:
+                        raise ExtractorError('YouTube said: This video is available in %s only' % (
+                            ', '.join(map(ISO3166Utils.short2full, regions_allowed.split(',')))),
+                            expected=True)
                 raise ExtractorError(
                     'YouTube said: %s' % video_info['reason'][0],
                     expected=True, video_id=video_id)
@@ -939,7 +1004,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         # uploader
         if 'author' not in video_info:
             raise ExtractorError('Unable to extract uploader name')
-        video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
+        video_uploader = compat_urllib_parse_unquote_plus(video_info['author'][0])
 
         # uploader_id
         video_uploader_id = None
@@ -966,18 +1031,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             self._downloader.report_warning('unable to extract video thumbnail')
             video_thumbnail = None
         else:   # don't panic if we can't find it
-            video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
+            video_thumbnail = compat_urllib_parse_unquote_plus(video_info['thumbnail_url'][0])
 
         # upload date
-        upload_date = None
-        mobj = re.search(r'(?s)id="eow-date.*?>(.*?)</span>', video_webpage)
-        if mobj is None:
-            mobj = re.search(
-                r'(?s)id="watch-uploader-info".*?>.*?(?:Published|Uploaded|Streamed live) on (.*?)</strong>',
-                video_webpage)
-        if mobj is not None:
-            upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
-            upload_date = unified_strdate(upload_date)
+        upload_date = self._html_search_meta(
+            'datePublished', video_webpage, 'upload date', default=None)
+        if not upload_date:
+            upload_date = self._search_regex(
+                [r'(?s)id="eow-date.*?>(.*?)</span>',
+                 r'id="watch-uploader-info".*?>.*?(?:Published|Uploaded|Streamed live|Started) on (.+?)</strong>'],
+                video_webpage, 'upload date', default=None)
+            if upload_date:
+                upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
+        upload_date = unified_strdate(upload_date)
 
         m_cat_container = self._search_regex(
             r'(?s)<h4[^>]*>\s*Category\s*</h4>\s*<ul[^>]*>(.*?)</ul>',
@@ -1011,12 +1077,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 video_description = ''
 
         def _extract_count(count_name):
-            count = self._search_regex(
-                r'id="watch-%s"[^>]*>.*?([\d,]+)\s*</span>' % re.escape(count_name),
-                video_webpage, count_name, default=None)
-            if count is not None:
-                return int(count.replace(',', ''))
-            return None
+            return str_to_int(self._search_regex(
+                r'-%s-button[^>]+><span[^>]+class="yt-uix-button-content"[^>]*>([\d,]+)</span>'
+                % re.escape(count_name),
+                video_webpage, count_name, default=None))
+
         like_count = _extract_count('like')
         dislike_count = _extract_count('dislike')
 
@@ -1028,7 +1093,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             self._downloader.report_warning('unable to extract video duration')
             video_duration = None
         else:
-            video_duration = int(compat_urllib_parse.unquote_plus(video_info['length_seconds'][0]))
+            video_duration = int(compat_urllib_parse_unquote_plus(video_info['length_seconds'][0]))
 
         # annotations
         video_annotations = None
@@ -1131,14 +1196,21 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
         # Look for the DASH manifest
         if self._downloader.params.get('youtube_include_dash_manifest', True):
+            dash_mpd_fatal = True
             for dash_manifest_url in dash_mpds:
                 dash_formats = {}
                 try:
                     for df in self._parse_dash_manifest(
-                            video_id, dash_manifest_url, player_url, age_gate):
+                            video_id, dash_manifest_url, player_url, age_gate, dash_mpd_fatal):
                         # Do not overwrite DASH format found in some previous DASH manifest
                         if df['format_id'] not in dash_formats:
                             dash_formats[df['format_id']] = df
+                        # Additional DASH manifests may end up in HTTP Error 403 therefore
+                        # allow them to fail without bug report message if we already have
+                        # some DASH manifest succeeded. This is temporary workaround to reduce
+                        # burst of bug reports until we figure out the reason and whether it
+                        # can be fixed at all.
+                        dash_mpd_fatal = False
                 except (ExtractorError, KeyError) as e:
                     self.report_warning(
                         'Skipping DASH manifest: %r' % e, video_id)
@@ -1148,8 +1220,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     # fixed values (for example the resolution). See
                     # https://github.com/rg3/youtube-dl/issues/5774 for an
                     # example.
-                    dash_keys = set(df['format_id'] for df in dash_formats.values())
-                    formats = [f for f in formats if f['format_id'] not in dash_keys]
+                    formats = [f for f in formats if f['format_id'] not in dash_formats.keys()]
                     formats.extend(dash_formats.values())
 
         # Check for malformed aspect ratio
@@ -1184,6 +1255,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             'dislike_count': dislike_count,
             'average_rating': float_or_none(video_info.get('avg_rating', [None])[0]),
             'formats': formats,
+            'is_live': is_live,
         }
 
 
@@ -1569,7 +1641,7 @@ class YoutubeSearchURLIE(InfoExtractor):
 
     def _real_extract(self, url):
         mobj = re.match(self._VALID_URL, url)
-        query = compat_urllib_parse.unquote_plus(mobj.group('query'))
+        query = compat_urllib_parse_unquote_plus(mobj.group('query'))
 
         webpage = self._download_webpage(url, query)
         result_code = self._search_regex(