Merge pull request #9288 from reyyed/issue#9063fix

[youtube-dl] / youtube_dl / extractor / facebook.py

diff --git a/youtube_dl/extractor/facebook.py b/youtube_dl/extractor/facebook.py
index 6eaa22d894f4d8c2e8415cef17006f590cba7644..0d43acc4ac7ff0162bdcc95471ccdb2376c0f7f2 100644 (file)
--- a/youtube_dl/extractor/facebook.py
+++ b/youtube_dl/extractor/facebook.py
@@ -139,8 +139,8 @@ class FacebookIE(InfoExtractor):
         # Facebook API embed
         # see https://developers.facebook.com/docs/plugins/embedded-video-player
         mobj = re.search(r'''(?x)<div[^>]+
-                class=(?P<q1>[\'"])[^\'"]*\bfb-video\b[^\'"]*(?P=q1)[^>]+
-                data-href=(?P<q2>[\'"])(?P<url>[^\'"]+)(?P=q2)''', webpage)
+                class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
+                data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage)
         if mobj is not None:
             return mobj.group('url')
 
@@ -219,12 +219,23 @@ class FacebookIE(InfoExtractor):
 
         BEFORE = '{swf.addParam(param[0], param[1]);});'
         AFTER = '.forEach(function(variable) {swf.addVariable(variable[0], variable[1]);});'
-        m = re.search(re.escape(BEFORE) + '(?:\n|\\\\n)(.*?)' + re.escape(AFTER), webpage)
-        if m:
-            swf_params = m.group(1).replace('\\\\', '\\').replace('\\"', '"')
+        PATTERN = re.escape(BEFORE) + '(?:\n|\\\\n)(.*?)' + re.escape(AFTER)
+
+        for m in re.findall(PATTERN, webpage):
+            swf_params = m.replace('\\\\', '\\').replace('\\"', '"')
             data = dict(json.loads(swf_params))
             params_raw = compat_urllib_parse_unquote(data['params'])
-            video_data = json.loads(params_raw)['video_data']
+            video_data_candidate = json.loads(params_raw)['video_data']
+            for _, f in video_data_candidate.items():
+                if not f:
+                    continue
+                if isinstance(f, dict):
+                    f = [f]
+                if isinstance(f, list):
+                    continue
+                if f[0].get('video_id') == video_id:
+                    video_data = video_data_candidate
+                    break
 
         def video_data_list2dict(video_data):
             ret = {}