[extractor/common] Extract HLS subtitle tracks

_extract_m3u8_formats is renamed to _extract_m3u8_formats_and_subtitles
and extended to handle subtitle tracks instead of skipping them;
a wrapper with the old name is provided for compatibility.

_parse_m3u8_formats is likewise renamed and extended, but without adding
the compatibility wrapper; the test suite is adjusted to test the enhanced
method instead.

test/test_InfoExtractor.py

@@ -684,17 +684,19 @@ def test_parse_m3u8_formats(self):
                     'width': 1920,
                     'height': 1080,
                     'vcodec': 'avc1.64002a',
-                }]
+                }],
+                {}
             ),
         ]
 
-        for m3u8_file, m3u8_url, expected_formats in _TEST_CASES:
+        for m3u8_file, m3u8_url, expected_formats, expected_subs in _TEST_CASES:
             with io.open('./test/testdata/m3u8/%s.m3u8' % m3u8_file,
                          mode='r', encoding='utf-8') as f:
-                formats = self.ie._parse_m3u8_formats(
+                formats, subs = self.ie._parse_m3u8_formats_and_subtitles(
                     f.read(), m3u8_url, ext='mp4')
                 self.ie._sort_formats(formats)
                 expect_value(self, formats, expected_formats, None)
+                expect_value(self, subs, expected_subs, None)
 
     def test_parse_mpd_formats(self):
         _TEST_CASES = [

yt_dlp/extractor/common.py

@@ -1879,11 +1879,21 @@ def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m
             'format_note': 'Quality selection URL',
         }
 
-    def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
+    def _extract_m3u8_formats(self, *args, **kwargs):
-                              entry_protocol='m3u8', preference=None, quality=None,
+        fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
-                              m3u8_id=None, note=None, errnote=None,
+        if subs:
-                              fatal=True, live=False, data=None, headers={},
+            self.report_warning(bug_reports_message(
-                              query={}):
+                "Ignoring subtitle tracks found in the HLS manifest; "
+                "if any subtitle tracks are missing,"
+            ))
+        return fmts
+
+    def _extract_m3u8_formats_and_subtitles(
+            self, m3u8_url, video_id, ext=None, entry_protocol='m3u8',
+            preference=None, quality=None, m3u8_id=None, note=None,
+            errnote=None, fatal=True, live=False, data=None, headers={},
+            query={}):
+
         res = self._download_webpage_handle(
             m3u8_url, video_id,
             note=note or 'Downloading m3u8 information',
@@ -1891,30 +1901,34 @@ def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
             fatal=fatal, data=data, headers=headers, query=query)
 
         if res is False:
-            return []
+            return [], {}
 
         m3u8_doc, urlh = res
         m3u8_url = urlh.geturl()
 
-        return self._parse_m3u8_formats(
+        return self._parse_m3u8_formats_and_subtitles(
             m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
             preference=preference, quality=quality, m3u8_id=m3u8_id,
             note=note, errnote=errnote, fatal=fatal, live=live, data=data,
             headers=headers, query=query, video_id=video_id)
 
-    def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
+    def _parse_m3u8_formats_and_subtitles(
-                            entry_protocol='m3u8', preference=None, quality=None,
+            self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8',
-                            m3u8_id=None, live=False, note=None, errnote=None,
+            preference=None, quality=None, m3u8_id=None, live=False, note=None,
-                            fatal=True, data=None, headers={}, query={}, video_id=None):
+            errnote=None, fatal=True, data=None, headers={}, query={},
+            video_id=None):
+
         if '#EXT-X-FAXS-CM:' in m3u8_doc:  # Adobe Flash Access
-            return []
+            return [], {}
 
         if (not self._downloader.params.get('allow_unplayable_formats')
                 and re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc)):  # Apple FairPlay
-            return []
+            return [], {}
 
         formats = []
 
+        subtitles = {}
+
         format_url = lambda u: (
             u
             if re.match(r'^https?://', u)
@@ -2001,7 +2015,7 @@ def _extract_m3u8_playlist_formats(format_url=None, m3u8_doc=None, video_id=None
                 }
                 formats.append(f)
 
-            return formats
+            return formats, subtitles
 
         groups = {}
         last_stream_inf = {}
@@ -2013,6 +2027,15 @@ def extract_media(x_media_line):
             if not (media_type and group_id and name):
                 return
             groups.setdefault(group_id, []).append(media)
+            # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
+            if media_type == 'SUBTITLES':
+                lang = media['LANGUAGE']  # XXX: normalise?
+                url = format_url(media['URI'])
+                sub_info = {
+                    'url': url,
+                    'ext': determine_ext(url),
+                }
+                subtitles.setdefault(lang, []).append(sub_info)
             if media_type not in ('VIDEO', 'AUDIO'):
                 return
             media_url = media.get('URI')
@@ -2160,7 +2183,7 @@ def build_stream_name():
                         formats.append(http_f)
 
                 last_stream_inf = {}
-        return formats
+        return formats, subtitles
 
     @staticmethod
     def _xpath_ns(path, namespace=None):