Merge 716972da6b into f2a4983df7

Closes #11526
Authored by: jshumphrey

README.md

@@ -355,6 +355,10 @@ If you fork the project on GitHub, you can run your fork's [build workflow](.git
                                     available. Pass the minimum number of
                                     seconds (or range) to wait between retries
     --no-wait-for-video             Do not wait for scheduled streams (default)
+    --wait-retries RETRIES          Number of retries while waiting for
+                                    scheduled streams to become available
+                                    (default is infinite). --wait-for-video must
+                                    also be set
     --mark-watched                  Mark videos watched (even with --simulate)
     --no-mark-watched               Do not mark videos watched (default)
     --color [STREAM:]POLICY         Whether to emit color codes in output,

yt_dlp/YoutubeDL.py

@@ -1620,17 +1620,26 @@ class YoutubeDL:
     def _handle_extraction_exceptions(func):
         @functools.wraps(func)
         def wrapper(self, *args, **kwargs):
+            wait_retries = 0
+            max_wait_retries = self.params.get('wait_retries')
             while True:
                 try:
                     return func(self, *args, **kwargs)
                 except (CookieLoadError, DownloadCancelled, LazyList.IndexError, PagedList.IndexError):
                     raise
                 except ReExtractInfo as e:
+                    if wait_retries >= max_wait_retries:
+                        if max_wait_retries > 0:
+                            self.report_error(f'Giving up after {wait_retries} {"retries" if wait_retries > 1 else "retry"} while waiting.')
+                        else:
+                            self.report_error('Video is still unavailable after waiting.')
+                        return
                     if e.expected:
                         self.to_screen(f'{e}; Re-extracting data')
                     else:
                         self.to_stderr('\r')
                         self.report_warning(f'{e}; Re-extracting data')
+                    wait_retries += 1
                     continue
                 except GeoRestrictedError as e:
                     msg = e.msg

yt_dlp/__init__.py

@@ -269,6 +269,7 @@ def validate_options(opts):
 
     opts.retries = parse_retries('download', opts.retries)
     opts.fragment_retries = parse_retries('fragment', opts.fragment_retries)
+    opts.wait_retries = parse_retries('waiting', opts.wait_retries)
     opts.extractor_retries = parse_retries('extractor', opts.extractor_retries)
     opts.file_access_retries = parse_retries('file access', opts.file_access_retries)
 
@@ -929,6 +930,7 @@ def parse_options(argv=None):
         'extract_flat': opts.extract_flat,
         'live_from_start': opts.live_from_start,
         'wait_for_video': opts.wait_for_video,
+        'wait_retries': opts.wait_retries,
         'mark_watched': opts.mark_watched,
         'merge_output_format': opts.merge_output_format,
         'final_ext': final_ext,

yt_dlp/extractor/archiveorg.py

@@ -205,6 +205,26 @@ class ArchiveOrgIE(InfoExtractor):
                 },
             },
         ],
+    }, {
+        # The reviewbody is None for one of the reviews; just need to extract data without crashing
+        'url': 'https://archive.org/details/gd95-04-02.sbd.11622.sbeok.shnf/gd95-04-02d1t04.shn',
+        'info_dict': {
+            'id': 'gd95-04-02.sbd.11622.sbeok.shnf/gd95-04-02d1t04.shn',
+            'ext': 'mp3',
+            'title': 'Stuck Inside of Mobile with the Memphis Blues Again',
+            'creators': ['Grateful Dead'],
+            'duration': 338.31,
+            'track': 'Stuck Inside of Mobile with the Memphis Blues Again',
+            'description': 'md5:764348a470b986f1217ffd38d6ac7b72',
+            'display_id': 'gd95-04-02d1t04.shn',
+            'location': 'Pyramid Arena',
+            'uploader': 'jon@archive.org',
+            'album': '1995-04-02 - Pyramid Arena',
+            'upload_date': '20040519',
+            'track_number': 4,
+            'release_date': '19950402',
+            'timestamp': 1084927901,
+        },
     }]
 
     @staticmethod
@@ -335,7 +355,7 @@ class ArchiveOrgIE(InfoExtractor):
                 info['comments'].append({
                     'id': review.get('review_id'),
                     'author': review.get('reviewer'),
-                    'text': str_or_none(review.get('reviewtitle'), '') + '\n\n' + review.get('reviewbody'),
+                    'text': join_nonempty('reviewtitle', 'reviewbody', from_dict=review, delim='\n\n'),
                     'timestamp': unified_timestamp(review.get('createdate')),
                     'parent': 'root'})
 

yt_dlp/extractor/facebook.py

@@ -563,13 +563,13 @@ class FacebookIE(InfoExtractor):
                 return extract_video_data(try_get(
                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 
-        def extract_dash_manifest(video, formats):
+        def extract_dash_manifest(vid_data, formats, mpd_url=None):
             dash_manifest = traverse_obj(
-                video, 'dash_manifest', 'playlist', 'dash_manifest_xml_string', expected_type=str)
+                vid_data, 'dash_manifest', 'playlist', 'dash_manifest_xml_string', 'manifest_xml', expected_type=str)
             if dash_manifest:
                 formats.extend(self._parse_mpd_formats(
                     compat_etree_fromstring(urllib.parse.unquote_plus(dash_manifest)),
-                    mpd_url=url_or_none(video.get('dash_manifest_url'))))
+                    mpd_url=url_or_none(video.get('dash_manifest_url')) or mpd_url))
 
         def process_formats(info):
             # Downloads with browser's User-Agent are rate limited. Working around
@@ -619,9 +619,12 @@ class FacebookIE(InfoExtractor):
                         video = video['creation_story']
                         video['owner'] = traverse_obj(video, ('short_form_video_context', 'video_owner'))
                         video.update(reel_info)
-                    fmt_data = traverse_obj(video, ('videoDeliveryLegacyFields', {dict})) or video
+
                     formats = []
                     q = qualities(['sd', 'hd'])
+
+                    # Legacy formats extraction
+                    fmt_data = traverse_obj(video, ('videoDeliveryLegacyFields', {dict})) or video
                     for key, format_id in (('playable_url', 'sd'), ('playable_url_quality_hd', 'hd'),
                                            ('playable_url_dash', ''), ('browser_native_hd_url', 'hd'),
                                            ('browser_native_sd_url', 'sd')):
@@ -629,7 +632,7 @@ class FacebookIE(InfoExtractor):
                         if not playable_url:
                             continue
                         if determine_ext(playable_url) == 'mpd':
-                            formats.extend(self._extract_mpd_formats(playable_url, video_id))
+                            formats.extend(self._extract_mpd_formats(playable_url, video_id, fatal=False))
                         else:
                             formats.append({
                                 'format_id': format_id,
@@ -638,6 +641,28 @@ class FacebookIE(InfoExtractor):
                                 'url': playable_url,
                             })
                     extract_dash_manifest(fmt_data, formats)
+
+                    # New videoDeliveryResponse formats extraction
+                    fmt_data = traverse_obj(video, ('videoDeliveryResponseFragment', 'videoDeliveryResponseResult'))
+                    mpd_urls = traverse_obj(fmt_data, ('dash_manifest_urls', ..., 'manifest_url', {url_or_none}))
+                    dash_manifests = traverse_obj(fmt_data, ('dash_manifests', lambda _, v: v['manifest_xml']))
+                    for idx, dash_manifest in enumerate(dash_manifests):
+                        extract_dash_manifest(dash_manifest, formats, mpd_url=traverse_obj(mpd_urls, idx))
+                    if not dash_manifests:
+                        # Only extract from MPD URLs if the manifests are not already provided
+                        for mpd_url in mpd_urls:
+                            formats.extend(self._extract_mpd_formats(mpd_url, video_id, fatal=False))
+                    for prog_fmt in traverse_obj(fmt_data, ('progressive_urls', lambda _, v: v['progressive_url'])):
+                        format_id = traverse_obj(prog_fmt, ('metadata', 'quality', {str.lower}))
+                        formats.append({
+                            'format_id': format_id,
+                            # sd, hd formats w/o resolution info should be deprioritized below DASH
+                            'quality': q(format_id) - 3,
+                            'url': prog_fmt['progressive_url'],
+                        })
+                    for m3u8_url in traverse_obj(fmt_data, ('hls_playlist_urls', ..., 'hls_playlist_url', {url_or_none})):
+                        formats.extend(self._extract_m3u8_formats(m3u8_url, video_id, 'mp4', fatal=False, m3u8_id='hls'))
+
                     if not formats:
                         # Do not append false positive entry w/o any formats
                         return

yt_dlp/options.py

@@ -442,6 +442,10 @@ def create_parser():
         '--no-wait-for-video',
         dest='wait_for_video', action='store_const', const=None,
         help='Do not wait for scheduled streams (default)')
+    general.add_option(
+        '--wait-retries',
+        dest='wait_retries', metavar='RETRIES', default='infinite',
+        help='Number of retries while waiting for scheduled streams to become available (default is %default). --wait-for-video must also be set')
     general.add_option(
         '--mark-watched',
         action='store_true', dest='mark_watched', default=False,