From ecd14c832de1663a2b4d06827ba584be30167bda Mon Sep 17 00:00:00 2001
From: dirkf <fieldhouse@gmx.net>
Date: Sun, 27 Feb 2022 13:33:28 +0000
Subject: [PATCH] [extractor/facebook] Update extraction improvements and fix
 tests

* avoid crashing in parse_attachment() on invalid attachment
* ignore empty results in <meta> search
---
 youtube_dl/extractor/facebook.py | 40 +++++++++++++++++++++-----------
 1 file changed, 26 insertions(+), 14 deletions(-)

diff --git a/youtube_dl/extractor/facebook.py b/youtube_dl/extractor/facebook.py
index 786a95849..a3f1bd325 100644
--- a/youtube_dl/extractor/facebook.py
+++ b/youtube_dl/extractor/facebook.py
@@ -83,7 +83,7 @@ class FacebookIE(InfoExtractor):
         'info_dict': {
             'id': '274175099429670',
             'ext': 'mp4',
-            'title': 'Asif Nawab Butt',
+            'title': 'Asif',
             'uploader': 'Asif Nawab Butt',
             'uploader_id': '100005115464446',
             'upload_date': '20140506',
@@ -173,6 +173,7 @@ class FacebookIE(InfoExtractor):
         'params': {
             'skip_download': True,
         },
+        'skip': 'This video is only available for registered users',
     }, {
         'url': 'https://www.facebook.com/LaGuiaDelVaron/posts/1072691702860471',
         'info_dict': {
@@ -188,15 +189,16 @@ class FacebookIE(InfoExtractor):
             'skip_download': True,
         },
         # TODO: parse this webpage
-        'skip': 'Cannot parse webpage well',
+        'skip': 'This video is only available for registered users (and cannot parse webpage well)',
     }, {
         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
         'url': 'https://www.facebook.com/groups/1024490957622648/permalink/1396382447100162/',
         'info_dict': {
             'id': '202882990186699',
             'ext': 'mp4',
+
             'title': 'birb (O v O\") | Hello? Yes your uber ride is here',
-            'description': 'md5:963dee8a667a2b49f2059cf7ab54fe55',
+            'description': 'md5:f122681cf504d04be12bb2bdc66c81fd',
             'timestamp': 1486035494,
             'upload_date': '20170202',
             'uploader': 'Elisabeth Ahtn',
@@ -308,15 +310,17 @@ class FacebookIE(InfoExtractor):
         'skip': 'Requires logging in',
     }, {
         # data.node.comet_sections.content.story.attachments[].styles.attachment.media
-        'url': 'https://www.facebook.com/groups/352666925484718/permalink/1112505706167499/',
+        # gone : 'https://www.facebook.com/groups/352666925484718/permalink/1112505706167499/',
+        'url': 'https://www.facebook.com/groups/471738070722101/permalink/471799617382613/',
         'info_dict': {
-            'id': '1080405282779948',
+            'id': '101173872166725',
             'ext': 'mp4',
-            'title': 'Best Chinese Song',
-            'timestamp': 1641225430,
-            'upload_date': '20220103',
-            'uploader_id': '1847991063',
-            'uploader': 'Best Chinese Song',
+            'title': 'md5:0ac174e60ed9d9a5eec588343f6a889c',
+            'description': 'md5:80efdff1371fa5d92758df2fb1b97bfd',
+            'upload_date': '20210517',
+            'timestamp': 1621247381,
+            'uploader': 'md5:34165f897c692f92b13c2cc1eddd8245',
+            'uploader_id': '100957655521680',
         },
     }]
     _SUPPORTED_PAGLETS_REGEX = r'(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_[0-9a-f]+)'
@@ -417,15 +421,23 @@ class FacebookIE(InfoExtractor):
                 video_title = self._html_search_regex(
                     r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
                     webpage, 'alternative title', default=None)
+
+            # _html_search_meta() finds '' ahead of later non-empty results
+            def html_search_meta_non_empty(patterns, html, name, **kwargs):
+                for p in patterns:
+                    result = self._html_search_meta(p, html, name, **kwargs)
+                    if result:
+                        return result
+
             if not video_title:
-                video_title = self._html_search_meta(
+                video_title = html_search_meta_non_empty(
                     ['og:title', 'twitter:title', 'description'],
                     webpage, 'title', default=None)
             if video_title:
                 video_title = limit_length(video_title, 80)
             else:
                 video_title = 'Facebook video #%s' % video_id
-            description = self._html_search_meta(
+            description = html_search_meta_non_empty(
                 ['description', 'og:description', 'twitter:description'],
                 webpage, 'description', default=None)
             uploader = clean_html(get_element_by_id(
@@ -435,7 +447,7 @@ class FacebookIE(InfoExtractor):
             timestamp = int_or_none(self._search_regex(
                 r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
                 'timestamp', default=None))
-            thumbnail = self._html_search_meta(
+            thumbnail = html_search_meta_non_empty(
                 ['og:image', 'twitter:image'], webpage, 'thumbnail', default=None)
             # some webpages contain unretrievable thumbnail urls
             # like https://lookaside.fbsbx.com/lookaside/crawler/media/?media_id=10155168902769113&get_thumbnail=1
@@ -562,7 +574,7 @@ class FacebookIE(InfoExtractor):
                     entries.append(info)
 
                 def parse_attachment(attachment, key='media'):
-                    media = attachment.get(key) or {}
+                    media = try_get(attachment, lambda x: x[key], dict) or {}
                     if media.get('__typename') == 'Video':
                         return parse_graphql_video(media)