[facebook] Add support for Relay based pages(closes #26823)

youtube_dl/extractor/facebook.py

@@ -16,11 +16,13 @@ from ..utils import (
     clean_html,
     error_to_compat_str,
     ExtractorError,
+    float_or_none,
     get_element_by_id,
     int_or_none,
     js_to_json,
     limit_length,
     parse_count,
+    qualities,
     sanitized_Request,
     try_get,
     urlencode_postdata,
@@ -327,6 +329,14 @@ class FacebookIE(InfoExtractor):
                 return extract_video_data(try_get(
                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 
+        formats = []
+
+        def extract_dash_manifest(video):
+            dash_manifest = video.get('dash_manifest')
+            if dash_manifest:
+                formats.extend(self._parse_mpd_formats(
+                    compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
+
         if not video_data:
             server_js_data = self._parse_json(self._search_regex([
                 r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_\d+)',
@@ -334,6 +344,61 @@ class FacebookIE(InfoExtractor):
             ], webpage, 'js data', default='{}'), video_id, js_to_json, False)
             video_data = extract_from_jsmods_instances(server_js_data)
 
+        if not video_data:
+            graphql_data = self._parse_json(self._search_regex(
+                r'handleWithCustomApplyEach\([^,]+,\s*({.*?"(?:dash_manifest|playable_url(?:_quality_hd)?)"\s*:\s*"[^"]+".*?})\);',
+                webpage, 'graphql data', default='{}'), video_id, fatal=False) or {}
+            for require in (graphql_data.get('require') or []):
+                if require[0] == 'RelayPrefetchedStreamCache':
+                    def parse_graphql_video(video):
+                        q = qualities(['sd', 'hd'])
+                        for (suffix, format_id) in [('', 'sd'), ('_quality_hd', 'hd')]:
+                            playable_url = video.get('playable_url' + suffix)
+                            if not playable_url:
+                                continue
+                            formats.append({
+                                'format_id': format_id,
+                                'quality': q(format_id),
+                                'url': playable_url,
+                            })
+                        extract_dash_manifest(video)
+                        self._sort_formats(formats)
+                        v_id = video.get('videoId') or video.get('id') or video_id
+                        info = {
+                            'id': v_id,
+                            'formats': formats,
+                            'thumbnail': try_get(video, lambda x: x['thumbnailImage']['uri']),
+                            'uploader_id': try_get(video, lambda x: x['owner']['id']),
+                            'timestamp': int_or_none(video.get('publish_time')),
+                            'duration': float_or_none(video.get('playable_duration_in_ms'), 1000),
+                        }
+                        description = try_get(video, lambda x: x['savable_description']['text'])
+                        title = video.get('name')
+                        if title:
+                            info.update({
+                                'title': title,
+                                'description': description,
+                            })
+                        else:
+                            info['title'] = description or 'Facebook video #%s' % v_id
+                        return webpage, info
+
+                    data = try_get(require, lambda x: x[3][1]['__bbox']['result']['data'], dict) or {}
+
+                    attachments = try_get(data, [
+                        lambda x: x['video']['story']['attachments'],
+                        lambda x: x['video']['creation_story']['attachments'],
+                        lambda x: x['node']['comet_sections']['content']['story']['attachments']
+                    ], list) or []
+                    for attachment in attachments:
+                        media = attachment.get('media') or try_get(attachment, lambda x: x['style_type_renderer']['attachment']['media'], dict) or {}
+                        if media.get('__typename') == 'Video':
+                            return parse_graphql_video(media)
+
+                    video = data.get('video') or {}
+                    if video:
+                        return parse_graphql_video(video)
+
         if not video_data:
             if not fatal_if_no_video:
                 return webpage, False
@@ -375,7 +440,6 @@ class FacebookIE(InfoExtractor):
             raise ExtractorError('Cannot parse data')
 
         subtitles = {}
-        formats = []
         for f in video_data:
             format_id = f['stream_type']
             if f and isinstance(f, dict):
@@ -394,10 +458,7 @@ class FacebookIE(InfoExtractor):
                             'url': src,
                             'preference': preference,
                         })
-            dash_manifest = f[0].get('dash_manifest')
+            extract_dash_manifest(f[0])
-            if dash_manifest:
-                formats.extend(self._parse_mpd_formats(
-                    compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
             subtitles_src = f[0].get('subtitles_src')
             if subtitles_src:
                 subtitles.setdefault('en', []).append({'url': subtitles_src})