[ie/crunchyroll] Fix stream extraction (#10005)

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

diff --git a/yt_dlp/extractor/theplatform.py b/yt_dlp/extractor/theplatform.py
index 537f6f6cd0479ba61b26242de43f2883572c65d0..9160f5ec6b8e95932c86383f09dcee9aa784fba2 100644 (file)
--- a/yt_dlp/extractor/theplatform.py
+++ b/yt_dlp/extractor/theplatform.py
@@ -19,7 +19,11 @@
     xpath_with_ns,
     mimetype2ext,
     find_xpath_attr,
+    traverse_obj,
+    update_url,
+    urlhandle_detect_ext,
 )
+from ..networking import HEADRequest
 
 default_ns = 'http://www.w3.org/2005/SMIL21/Language'
 _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
@@ -45,7 +49,7 @@ def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL d
                     raise ExtractorError(
                         error_element.attrib['abstract'], expected=True)
 
-        smil_formats = self._parse_smil_formats(
+        smil_formats, subtitles = self._parse_smil_formats_and_subtitles(
             meta, smil_url, video_id, namespace=default_ns,
             # the parameters are from syfy.com, other sites may use others,
             # they also work for nbc.com
@@ -65,8 +69,6 @@ def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL d
 
                 formats.append(_format)
 
-        subtitles = self._parse_smil_subtitles(meta, default_ns)
-
         return formats, subtitles
 
     def _download_theplatform_metadata(self, path, video_id):
@@ -102,6 +104,10 @@ def _add_chapter(start_time, end_time):
                 _add_chapter(chapter.get('startTime'), chapter.get('endTime'))
             _add_chapter(tp_chapters[-1].get('startTime'), tp_chapters[-1].get('endTime') or duration)
 
+        def extract_site_specific_field(field):
+            # A number of sites have custom-prefixed keys, e.g. 'cbc$seasonNumber'
+            return traverse_obj(info, lambda k, v: v and k.endswith(f'${field}'), get_all=False)
+
         return {
             'title': info['title'],
             'subtitles': subtitles,
@@ -111,6 +117,14 @@ def _add_chapter(start_time, end_time):
             'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
             'uploader': info.get('billingCode'),
             'chapters': chapters,
+            'creator': traverse_obj(info, ('author', {str})) or None,
+            'categories': traverse_obj(info, (
+                'categories', lambda _, v: v.get('label') in ('category', None), 'name', {str})) or None,
+            'tags': traverse_obj(info, ('keywords', {lambda x: re.split(r'[;,]\s?', x) if x else None})),
+            'location': extract_site_specific_field('region'),
+            'series': extract_site_specific_field('show'),
+            'season_number': int_or_none(extract_site_specific_field('seasonNumber')),
+            'media_type': extract_site_specific_field('programmingType') or extract_site_specific_field('type'),
         }
 
     def _extract_theplatform_metadata(self, path, video_id):
@@ -164,7 +178,8 @@ class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
         'params': {
             # rtmp download
             'skip_download': True,
-        }
+        },
+        'skip': 'CNet no longer uses ThePlatform',
     }, {
         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
         'info_dict': {
@@ -173,7 +188,8 @@ class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
             'uploader': 'EGSM',
-        }
+        },
+        'skip': 'Dead link',
     }, {
         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
         'only_matching': True,
@@ -191,6 +207,7 @@ class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
             'upload_date': '20150701',
             'uploader': 'NBCU-NEWS',
         },
+        'skip': 'Error: Player PID "nbcNewsOffsite" is disabled',
     }, {
         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
         # geo-restricted (US), HLS encrypted with AES-128
@@ -297,6 +314,17 @@ def _real_extract(self, url):
 
         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 
+        # With some sites, manifest URL must be forced to extract HLS formats
+        if not traverse_obj(formats, lambda _, v: v['format_id'].startswith('hls')):
+            m3u8_url = update_url(url, query='mbr=true&manifest=m3u', fragment=None)
+            urlh = self._request_webpage(
+                HEADRequest(m3u8_url), video_id, 'Checking for HLS formats', 'No HLS formats found', fatal=False)
+            if urlh and urlhandle_detect_ext(urlh) == 'm3u8':
+                m3u8_fmts, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
+                    m3u8_url, video_id, m3u8_id='hls', fatal=False)
+                formats.extend(m3u8_fmts)
+                self._merge_subtitles(m3u8_subs, target=subtitles)
+
         ret = self._extract_theplatform_metadata(path, video_id)
         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
         ret.update({