[generic] Extract subtitles from video.js (#3156)

[yt-dlp.git] / yt_dlp / extractor / generic.py

diff --git a/yt_dlp/extractor/generic.py b/yt_dlp/extractor/generic.py
index 97e34808f105d708860f30a95af98e79e1f0a1a1..4a2e3015803b5731d212a5b3f7cb6dca85598755 100644 (file)
--- a/yt_dlp/extractor/generic.py
+++ b/yt_dlp/extractor/generic.py
@@ -17,6 +17,7 @@
 )
 from ..utils import (
     determine_ext,
+    dict_get,
     ExtractorError,
     float_or_none,
     HEADRequest,
@@ -31,6 +32,7 @@
     parse_resolution,
     sanitized_Request,
     smuggle_url,
+    str_or_none,
     unescapeHTML,
     unified_timestamp,
     unsmuggle_url,
@@ -3778,11 +3780,12 @@ def _real_extract(self, url):
 
         # Video.js embed
         mobj = re.search(
-            r'(?s)\bvideojs\s*\(.+?\.src\s*\(\s*((?:\[.+?\]|{.+?}))\s*\)\s*;',
+            r'(?s)\bvideojs\s*\(.+?([a-zA-Z0-9_$]+)\.src\s*\(\s*((?:\[.+?\]|{.+?}))\s*\)\s*;',
             webpage)
         if mobj is not None:
+            varname = mobj.group(1)
             sources = self._parse_json(
-                mobj.group(1), video_id, transform_source=js_to_json,
+                mobj.group(2), video_id, transform_source=js_to_json,
                 fatal=False) or []
             if not isinstance(sources, list):
                 sources = [sources]
@@ -3819,6 +3822,21 @@ def _real_extract(self, url):
                             'Referer': full_response.geturl(),
                         },
                     })
+            # https://docs.videojs.com/player#addRemoteTextTrack
+            # https://html.spec.whatwg.org/multipage/media.html#htmltrackelement
+            for sub_match in re.finditer(rf'(?s){re.escape(varname)}' r'\.addRemoteTextTrack\(({.+?})\s*,\s*(?:true|false)\)', webpage):
+                sub = self._parse_json(
+                    sub_match.group(1), video_id, transform_source=js_to_json, fatal=False) or {}
+                src = str_or_none(sub.get('src'))
+                if not src:
+                    continue
+                subtitles.setdefault(dict_get(sub, ('language', 'srclang')) or 'und', []).append({
+                    'url': compat_urlparse.urljoin(url, src),
+                    'name': sub.get('label'),
+                    'http_headers': {
+                        'Referer': full_response.geturl(),
+                    },
+                })
             if formats or subtitles:
                 self.report_detected('video.js embed')
                 self._sort_formats(formats)