]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/youtube.py
Update changelog.
[youtubedl] / youtube_dl / extractor / youtube.py
index 9f748390531b66107f0e2012a0fdee152941d98d..ce2212a7caa13111e1d2835fa65be371a4937847 100644 (file)
@@ -70,6 +70,11 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
 
     _PLAYLIST_ID_RE = r'(?:PL|LL|EC|UU|FL|RD|UL|TL|PU|OLAK5uy_)[0-9A-Za-z-_]{10,}'
 
+    _YOUTUBE_CLIENT_HEADERS = {
+        'x-youtube-client-name': '1',
+        'x-youtube-client-version': '1.20200609.04.02',
+    }
+
     def _set_language(self):
         self._set_cookie(
             '.youtube.com', 'PREF', 'f1=50000000&hl=en',
@@ -301,7 +306,8 @@ class YoutubeEntryListBaseInfoExtractor(YoutubeBaseInfoExtractor):
                         'https://youtube.com/%s' % mobj.group('more'), playlist_id,
                         'Downloading page #%s%s'
                         % (page_num, ' (retry #%d)' % count if count else ''),
-                        transform_source=uppercase_escape)
+                        transform_source=uppercase_escape,
+                        headers=self._YOUTUBE_CLIENT_HEADERS)
                     break
                 except ExtractorError as e:
                     if isinstance(e.cause, compat_HTTPError) and e.cause.code in (500, 503):
@@ -388,8 +394,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                             (?:www\.)?invidious\.drycat\.fr/|
                             (?:www\.)?tube\.poal\.co/|
                             (?:www\.)?vid\.wxzm\.sx/|
+                            (?:www\.)?yewtu\.be/|
                             (?:www\.)?yt\.elukerio\.org/|
                             (?:www\.)?yt\.lelux\.fi/|
+                            (?:www\.)?invidious\.ggc-project\.de/|
+                            (?:www\.)?yt\.maisputain\.ovh/|
+                            (?:www\.)?invidious\.13ad\.de/|
+                            (?:www\.)?invidious\.toot\.koeln/|
+                            (?:www\.)?invidious\.fdn\.fr/|
+                            (?:www\.)?watch\.nettohikari\.com/|
                             (?:www\.)?kgg2m7yk5aybusll\.onion/|
                             (?:www\.)?qklhadlycap4cnod\.onion/|
                             (?:www\.)?axqzx4s6s54s32yentfqojs3x5i7faxza6xo3ehd4bzzsg2ii4fv2iid\.onion/|
@@ -397,6 +410,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                             (?:www\.)?fz253lmuao3strwbfbmx46yu7acac2jz27iwtorgmbqlkurlclmancad\.onion/|
                             (?:www\.)?invidious\.l4qlywnpwqsluw65ts7md3khrivpirse744un3x7mlskqauz5pyuzgqd\.onion/|
                             (?:www\.)?owxfohz4kjyv25fvlqilyxast7inivgiktls3th44jhk3ej3i7ya\.b32\.i2p/|
+                            (?:www\.)?4l2dgddgsrkf2ous66i6seeyi6etzfgrue332grh2n7madpwopotugyd\.onion/|
                             youtube\.googleapis\.com/)                        # the various hostnames, with wildcard subdomains
                          (?:.*?\#/)?                                          # handle anchor (#/) redirect urls
                          (?:                                                  # the various things that can precede the ID:
@@ -1644,8 +1658,63 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         video_id = mobj.group(2)
         return video_id
 
+    def _extract_chapters_from_json(self, webpage, video_id, duration):
+        if not webpage:
+            return
+        player = self._parse_json(
+            self._search_regex(
+                r'RELATED_PLAYER_ARGS["\']\s*:\s*({.+})\s*,?\s*\n', webpage,
+                'player args', default='{}'),
+            video_id, fatal=False)
+        if not player or not isinstance(player, dict):
+            return
+        watch_next_response = player.get('watch_next_response')
+        if not isinstance(watch_next_response, compat_str):
+            return
+        response = self._parse_json(watch_next_response, video_id, fatal=False)
+        if not response or not isinstance(response, dict):
+            return
+        chapters_list = try_get(
+            response,
+            lambda x: x['playerOverlays']
+                       ['playerOverlayRenderer']
+                       ['decoratedPlayerBarRenderer']
+                       ['decoratedPlayerBarRenderer']
+                       ['playerBar']
+                       ['chapteredPlayerBarRenderer']
+                       ['chapters'],
+            list)
+        if not chapters_list:
+            return
+
+        def chapter_time(chapter):
+            return float_or_none(
+                try_get(
+                    chapter,
+                    lambda x: x['chapterRenderer']['timeRangeStartMillis'],
+                    int),
+                scale=1000)
+        chapters = []
+        for next_num, chapter in enumerate(chapters_list, start=1):
+            start_time = chapter_time(chapter)
+            if start_time is None:
+                continue
+            end_time = (chapter_time(chapters_list[next_num])
+                        if next_num < len(chapters_list) else duration)
+            if end_time is None:
+                continue
+            title = try_get(
+                chapter, lambda x: x['chapterRenderer']['title']['simpleText'],
+                compat_str)
+            chapters.append({
+                'start_time': start_time,
+                'end_time': end_time,
+                'title': title,
+            })
+        return chapters
+
     @staticmethod
-    def _extract_chapters(description, duration):
+    def _extract_chapters_from_description(description, duration):
         if not description:
             return None
         chapter_lines = re.findall(
@@ -1679,6 +1748,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             })
         return chapters
 
+    def _extract_chapters(self, webpage, description, video_id, duration):
+        return (self._extract_chapters_from_json(webpage, video_id, duration)
+                or self._extract_chapters_from_description(description, duration))
+
     def _real_extract(self, url):
         url, smuggled_data = unsmuggle_url(url, {})
 
@@ -1825,6 +1898,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         video_details = try_get(
             player_response, lambda x: x['videoDetails'], dict) or {}
 
+        microformat = try_get(
+            player_response, lambda x: x['microformat']['playerMicroformatRenderer'], dict) or {}
+
         video_title = video_info.get('title', [None])[0] or video_details.get('title')
         if not video_title:
             self._downloader.report_warning('Unable to extract video title')
@@ -1902,6 +1978,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             view_count = extract_view_count(video_info)
         if view_count is None and video_details:
             view_count = int_or_none(video_details.get('viewCount'))
+        if view_count is None and microformat:
+            view_count = int_or_none(microformat.get('viewCount'))
 
         if is_live is None:
             is_live = bool_or_none(video_details.get('isLive'))
@@ -2153,7 +2231,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             video_uploader_id = mobj.group('uploader_id')
             video_uploader_url = mobj.group('uploader_url')
         else:
-            self._downloader.report_warning('unable to extract uploader nickname')
+            owner_profile_url = url_or_none(microformat.get('ownerProfileUrl'))
+            if owner_profile_url:
+                video_uploader_id = self._search_regex(
+                    r'(?:user|channel)/([^/]+)', owner_profile_url, 'uploader id',
+                    default=None)
+                video_uploader_url = owner_profile_url
 
         channel_id = (
             str_or_none(video_details.get('channelId'))
@@ -2164,17 +2247,33 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 video_webpage, 'channel id', default=None, group='id'))
         channel_url = 'http://www.youtube.com/channel/%s' % channel_id if channel_id else None
 
-        # thumbnail image
-        # We try first to get a high quality image:
-        m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
-                            video_webpage, re.DOTALL)
-        if m_thumb is not None:
-            video_thumbnail = m_thumb.group(1)
-        elif 'thumbnail_url' not in video_info:
-            self._downloader.report_warning('unable to extract video thumbnail')
+        thumbnails = []
+        thumbnails_list = try_get(
+            video_details, lambda x: x['thumbnail']['thumbnails'], list) or []
+        for t in thumbnails_list:
+            if not isinstance(t, dict):
+                continue
+            thumbnail_url = url_or_none(t.get('url'))
+            if not thumbnail_url:
+                continue
+            thumbnails.append({
+                'url': thumbnail_url,
+                'width': int_or_none(t.get('width')),
+                'height': int_or_none(t.get('height')),
+            })
+
+        if not thumbnails:
             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])
+            # We try first to get a high quality image:
+            m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
+                                video_webpage, re.DOTALL)
+            if m_thumb is not None:
+                video_thumbnail = m_thumb.group(1)
+            thumbnail_url = try_get(video_info, lambda x: x['thumbnail_url'][0], compat_str)
+            if thumbnail_url:
+                video_thumbnail = compat_urllib_parse_unquote_plus(thumbnail_url)
+            if video_thumbnail:
+                thumbnails.append({'url': video_thumbnail})
 
         # upload date
         upload_date = self._html_search_meta(
@@ -2184,6 +2283,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 [r'(?s)id="eow-date.*?>(.*?)</span>',
                  r'(?:id="watch-uploader-info".*?>.*?|["\']simpleText["\']\s*:\s*["\'])(?:Published|Uploaded|Streamed live|Started) on (.+?)[<"\']'],
                 video_webpage, 'upload date', default=None)
+        if not upload_date:
+            upload_date = microformat.get('publishDate') or microformat.get('uploadDate')
         upload_date = unified_strdate(upload_date)
 
         video_license = self._html_search_regex(
@@ -2316,7 +2417,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     errnote='Unable to download video annotations', fatal=False,
                     data=urlencode_postdata({xsrf_field_name: xsrf_token}))
 
-        chapters = self._extract_chapters(description_original, video_duration)
+        chapters = self._extract_chapters(video_webpage, description_original, video_id, video_duration)
 
         # Look for the DASH manifest
         if self._downloader.params.get('youtube_include_dash_manifest', True):
@@ -2407,7 +2508,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             'creator': video_creator or artist,
             'title': video_title,
             'alt_title': video_alt_title or track,
-            'thumbnail': video_thumbnail,
+            'thumbnails': thumbnails,
             'description': video_description,
             'categories': video_categories,
             'tags': video_tags,
@@ -3183,7 +3284,8 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
             more = self._download_json(
                 'https://youtube.com/%s' % mobj.group('more'), self._PLAYLIST_TITLE,
                 'Downloading page #%s' % page_num,
-                transform_source=uppercase_escape)
+                transform_source=uppercase_escape,
+                headers=self._YOUTUBE_CLIENT_HEADERS)
             content_html = more['content_html']
             more_widget_html = more['load_more_widget_html']