Merge 2eaf303b63 into f2a4983df7

Closes #11526
Authored by: jshumphrey

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/extractor/zoom.py

@@ -7,6 +7,7 @@ from ..utils import (
     parse_resolution,
     str_or_none,
     traverse_obj,
+    update_url,
     url_basename,
     urlencode_postdata,
     urljoin,
@@ -34,6 +35,7 @@ class ZoomIE(InfoExtractor):
             'ext': 'mp4',
             'title': 'Prépa AF2023 - Séance 5 du 11 avril - R20/VM/GO',
         },
+        'skip': 'This recording has expired',
     }, {
         # share URL
         'url': 'https://us02web.zoom.us/rec/share/hkUk5Zxcga0nkyNGhVCRfzkA2gX_mzgS3LpTxEEWJz9Y_QpIQ4mZFOUx7KZRZDQA.9LGQBdqmDAYgiZ_8',
@@ -61,41 +63,59 @@ class ZoomIE(InfoExtractor):
         return self._search_json(
             r'window\.__data__\s*=', webpage, 'data', video_id, transform_source=js_to_json)
 
-    def _get_real_webpage(self, url, base_url, video_id, url_type):
+    def _try_login(self, url, base_url, video_id, form):
-        webpage = self._download_webpage(url, video_id, note=f'Downloading {url_type} webpage')
+        # This will most likely only work for password-protected meetings
-        try:
-            form = self._form_hidden_inputs('password_form', webpage)
-        except ExtractorError:
-            return webpage
-
         password = self.get_param('videopassword')
         if not password:
             raise ExtractorError(
                 'This video is protected by a passcode, use the --video-password option', expected=True)
+
         is_meeting = form.get('useWhichPasswd') == 'meeting'
         validation = self._download_json(
-            base_url + 'rec/validate%s_passwd' % ('_meet' if is_meeting else ''),
+            base_url + 'nws/recording/1.0/validate%s-passwd' % ('-meeting' if is_meeting else ''),
             video_id, 'Validating passcode', 'Wrong passcode', data=urlencode_postdata({
-                'id': form[('meet' if is_meeting else 'file') + 'Id'],
+                'id': form[('meeting' if is_meeting else 'file') + '_id'],
                 'passwd': password,
                 'action': form.get('action'),
             }))
+
         if not validation.get('status'):
             raise ExtractorError(validation['errorMessage'], expected=True)
-        return self._download_webpage(url, video_id, note=f'Re-downloading {url_type} webpage')
+
+    def _get_real_webpage(self, url, base_url, video_id, url_type):
+        webpage = self._download_webpage(url, video_id, note=f'Downloading {url_type} webpage')
+
+        data = self._get_page_data(webpage, video_id)
+        if data.get('componentName') != 'need-password':  # not password protected
+            return webpage
+
+        # Password-protected:
+        self._try_login(url, base_url, video_id, form=data)
+        # Return the new HTML document
+        new_url = f"{base_url}rec/share/{data['meeting_id']}"
+        return self._download_webpage(new_url, video_id, note=f'Re-downloading {url_type} webpage')
+
+    def _get_share_redirect_url(self, url, base_url, video_id):
+        """Converts a `/rec/share` url to the corresponding `/rec/play` url, performs login if necessary"""
+        webpage = self._get_real_webpage(url, base_url, video_id, 'share')
+        meeting_id = self._get_page_data(webpage, video_id)['meetingId']
+        redirect_dict = self._download_json(
+            f'{base_url}nws/recording/1.0/play/share-info/{meeting_id}',
+            video_id, note='Downloading share info JSON')['result']
+        redirect_path = redirect_dict.pop('redirectUrl')
+        url = update_url(urljoin(base_url, redirect_path), query_update=redirect_dict)
+
+        if redirect_dict.get('componentName') == 'need-password':
+            # First login, then return redirection URL
+            return self._get_share_redirect_url(url, base_url, video_id)
+
+        return url
 
     def _real_extract(self, url):
         base_url, url_type, video_id = self._match_valid_url(url).group('base_url', 'type', 'id')
-        query = {}
 
         if url_type == 'share':
-            webpage = self._get_real_webpage(url, base_url, video_id, 'share')
+            url = self._get_share_redirect_url(url, base_url, video_id)
-            meeting_id = self._get_page_data(webpage, video_id)['meetingId']
-            redirect_path = self._download_json(
-                f'{base_url}nws/recording/1.0/play/share-info/{meeting_id}',
-                video_id, note='Downloading share info JSON')['result']['redirectUrl']
-            url = urljoin(base_url, redirect_path)
-            query['continueMode'] = 'true'
 
         webpage = self._get_real_webpage(url, base_url, video_id, 'play')
         file_id = self._get_page_data(webpage, video_id)['fileId']
@@ -104,10 +124,12 @@ class ZoomIE(InfoExtractor):
             raise ExtractorError('Unable to extract file ID')
 
         data = self._download_json(
-            f'{base_url}nws/recording/1.0/play/info/{file_id}', video_id, query=query,
+            f'{base_url}nws/recording/1.0/play/info/{file_id}', video_id, query={
+                'continueMode': 'true',  # Makes this return value include interpreter audio information
+            },
             note='Downloading play info JSON')['result']
-
         subtitles = {}
+        # XXX: Would be more appropriate to parse chapters separate from subtitles
         for _type in ('transcript', 'cc', 'chapter'):
             if data.get(f'{_type}Url'):
                 subtitles[_type] = [{
@@ -117,6 +139,19 @@ class ZoomIE(InfoExtractor):
 
         formats = []
 
+        if data.get('interpreterAudioList'):
+            for audio in data.get('interpreterAudioList'):
+                formats.append({
+                    'format_note': f'Intepreter: {audio["languageText"]}',
+                    'url': audio['audioUrl'],
+                    'format_id': f'interpreter-{ audio["icon"].lower()}',
+                    'ext': 'm4a',
+                    # There doesn't seem to be an explicit field for a standardized language code,
+                    # sometimes the `language` field may be more accurate than `icon`
+                    'language': audio['icon'].lower(),
+                    'vcodec': 'none',
+                })
+
         if data.get('viewMp4Url'):
             formats.append({
                 'format_note': 'Camera stream',