[crunchyroll] parse vilos media data(closes #17343)

youtube_dl/extractor/crunchyroll.py

@@ -7,7 +7,7 @@
 
 from hashlib import sha1
 from math import pow, sqrt, floor
-from .common import InfoExtractor
+from .vrv import VRVIE
 from ..compat import (
     compat_b64decode,
     compat_etree_fromstring,
@@ -18,6 +18,8 @@
 from ..utils import (
     ExtractorError,
     bytes_to_intlist,
+    extract_attributes,
+    float_or_none,
     intlist_to_bytes,
     int_or_none,
     lowercase_escape,
@@ -26,14 +28,13 @@
     unified_strdate,
     urlencode_postdata,
     xpath_text,
-    extract_attributes,
 )
 from ..aes import (
     aes_cbc_decrypt,
 )
 
 
-class CrunchyrollBaseIE(InfoExtractor):
+class CrunchyrollBaseIE(VRVIE):
     _LOGIN_URL = 'https://www.crunchyroll.com/login'
     _LOGIN_FORM = 'login_form'
     _NETRC_MACHINE = 'crunchyroll'
@@ -148,7 +149,7 @@ class CrunchyrollIE(CrunchyrollBaseIE):
             'ext': 'mp4',
             'title': 'Wanna be the Strongest in the World Episode 1 – An Idol-Wrestler is Born!',
             'description': 'md5:2d17137920c64f2f49981a7797d275ef',
-            'thumbnail': 'http://img1.ak.crunchyroll.com/i/spire1-tmb/20c6b5e10f1a47b10516877d3c039cae1380951166_full.jpg',
+            'thumbnail': r're:^https?://.*\.jpg$',
             'uploader': 'Yomiuri Telecasting Corporation (YTV)',
             'upload_date': '20131013',
             'url': 're:(?!.*&amp)',
@@ -221,7 +222,7 @@ class CrunchyrollIE(CrunchyrollBaseIE):
         'info_dict': {
             'id': '535080',
             'ext': 'mp4',
-            'title': '11eyes Episode 1 – Piros éjszaka - Red Night',
+            'title': '11eyes Episode 1 – Red Night ~ Piros éjszaka',
             'description': 'Kakeru and Yuka are thrown into an alternate nightmarish world they call "Red Night".',
             'uploader': 'Marvelous AQL Inc.',
             'upload_date': '20091021',
@@ -437,13 +438,18 @@ def _real_extract(self, url):
         if 'To view this, please log in to verify you are 18 or older.' in webpage:
             self.raise_login_required()
 
+        media = self._parse_json(self._search_regex(
+            r'vilos\.config\.media\s*=\s*({.+?});',
+            webpage, 'vilos media', default='{}'), video_id)
+        media_metadata = media.get('metadata') or {}
+
         video_title = self._html_search_regex(
             r'(?s)<h1[^>]*>((?:(?!<h1).)*?<span[^>]+itemprop=["\']title["\'][^>]*>(?:(?!<h1).)+?)</h1>',
             webpage, 'video_title')
         video_title = re.sub(r' {2,}', ' ', video_title)
-        video_description = self._parse_json(self._html_search_regex(
+        video_description = (self._parse_json(self._html_search_regex(
             r'<script[^>]*>\s*.+?\[media_id=%s\].+?({.+?"description"\s*:.+?})\);' % video_id,
-            webpage, 'description', default='{}'), video_id).get('description')
+            webpage, 'description', default='{}'), video_id) or media_metadata).get('description')
         if video_description:
             video_description = lowercase_escape(video_description.replace(r'\r\n', '\n'))
         video_upload_date = self._html_search_regex(
@@ -456,6 +462,12 @@ def _real_extract(self, url):
             [r'<a[^>]+href="/publisher/[^"]+"[^>]*>([^<]+)</a>', r'<div>\s*Publisher:\s*<span>\s*(.+?)\s*</span>\s*</div>'],
             webpage, 'video_uploader', fatal=False)
 
+        formats = []
+        for stream in media.get('streams', []):
+            formats.extend(self._extract_vrv_formats(
+                stream.get('url'), video_id, stream.get('format'),
+                stream.get('audio_lang'), stream.get('hardsub_lang')))
+        if not formats:
             available_fmts = []
             for a, fmt in re.findall(r'(<a[^>]+token=["\']showmedia\.([0-9]{3,4})p["\'][^>]+>)', webpage):
                 attrs = extract_attributes(a)
@@ -468,8 +480,10 @@ def _real_extract(self, url):
                     available_fmts = re.findall(p, webpage)
                     if available_fmts:
                         break
+            if not available_fmts:
+                available_fmts = self._FORMAT_IDS.keys()
             video_encode_ids = []
-        formats = []
+
             for fmt in available_fmts:
                 stream_quality, stream_format = self._FORMAT_IDS[fmt]
                 video_format = fmt + 'p'
@@ -549,6 +563,16 @@ def _real_extract(self, url):
                 'media_id': video_id,
             })
 
+        subtitles = {}
+        for subtitle in media.get('subtitles', []):
+            subtitle_url = subtitle.get('url')
+            if not subtitle_url:
+                continue
+            subtitles.setdefault(subtitle.get('language', 'enUS'), []).append({
+                'url': subtitle_url,
+                'ext': subtitle.get('format', 'ass'),
+            })
+        if not subtitles:
             subtitles = self.extract_subtitles(video_id, webpage)
 
         # webpage provide more accurate data than series_title from XML
@@ -557,8 +581,8 @@ def _real_extract(self, url):
             webpage, 'series', fatal=False)
         season = xpath_text(metadata, 'series_title')
 
-        episode = xpath_text(metadata, 'episode_title')
+        episode = xpath_text(metadata, 'episode_title') or media_metadata.get('title')
-        episode_number = int_or_none(xpath_text(metadata, 'episode_number'))
+        episode_number = int_or_none(xpath_text(metadata, 'episode_number') or media_metadata.get('episode_number'))
 
         season_number = int_or_none(self._search_regex(
             r'(?s)<h\d[^>]+id=["\']showmedia_about_episode_num[^>]+>.+?</h\d>\s*<h4>\s*Season (\d+)',
@@ -568,7 +592,8 @@ def _real_extract(self, url):
             'id': video_id,
             'title': video_title,
             'description': video_description,
-            'thumbnail': xpath_text(metadata, 'episode_image_url'),
+            'duration': float_or_none(media_metadata.get('duration'), 1000),
+            'thumbnail': xpath_text(metadata, 'episode_image_url') or media_metadata.get('thumbnail', {}).get('url'),
             'uploader': video_uploader,
             'upload_date': video_upload_date,
             'series': series,

youtube_dl/extractor/vrv.py

@@ -72,7 +72,7 @@ def _get_cms_resource(self, resource_key, video_id):
 class VRVIE(VRVBaseIE):
     IE_NAME = 'vrv'
     _VALID_URL = r'https?://(?:www\.)?vrv\.co/watch/(?P<id>[A-Z0-9]+)'
-    _TEST = {
+    _TESTS = [{
         'url': 'https://vrv.co/watch/GR9PNZ396/Hidden-America-with-Jonah-Ray:BOSTON-WHERE-THE-PAST-IS-THE-PRESENT',
         'info_dict': {
             'id': 'GR9PNZ396',
@@ -85,7 +85,28 @@ class VRVIE(VRVBaseIE):
             # m3u8 download
             'skip_download': True,
         },
-    }
+    }]
+
+    def _extract_vrv_formats(self, url, video_id, stream_format, audio_lang, hardsub_lang):
+        if not url or stream_format not in ('hls', 'dash'):
+            return []
+        stream_id = hardsub_lang or audio_lang
+        format_id = '%s-%s' % (stream_format, stream_id)
+        if stream_format == 'hls':
+            adaptive_formats = self._extract_m3u8_formats(
+                url, video_id, 'mp4', m3u8_id=format_id,
+                note='Downloading %s m3u8 information' % stream_id,
+                fatal=False)
+        elif stream_format == 'dash':
+            adaptive_formats = self._extract_mpd_formats(
+                url, video_id, mpd_id=format_id,
+                note='Downloading %s MPD information' % stream_id,
+                fatal=False)
+        if audio_lang:
+            for f in adaptive_formats:
+                if f.get('acodec') != 'none':
+                    f['language'] = audio_lang
+        return adaptive_formats
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
@@ -115,26 +136,9 @@ def _real_extract(self, url):
         for stream_type, streams in streams_json.get('streams', {}).items():
             if stream_type in ('adaptive_hls', 'adaptive_dash'):
                 for stream in streams.values():
-                    stream_url = stream.get('url')
+                    formats.extend(self._extract_vrv_formats(
-                    if not stream_url:
+                        stream.get('url'), video_id, stream_type.split('_')[1],
-                        continue
+                        audio_locale, stream.get('hardsub_locale')))
-                    stream_id = stream.get('hardsub_locale') or audio_locale
-                    format_id = '%s-%s' % (stream_type.split('_')[1], stream_id)
-                    if stream_type == 'adaptive_hls':
-                        adaptive_formats = self._extract_m3u8_formats(
-                            stream_url, video_id, 'mp4', m3u8_id=format_id,
-                            note='Downloading %s m3u8 information' % stream_id,
-                            fatal=False)
-                    else:
-                        adaptive_formats = self._extract_mpd_formats(
-                            stream_url, video_id, mpd_id=format_id,
-                            note='Downloading %s MPD information' % stream_id,
-                            fatal=False)
-                    if audio_locale:
-                        for f in adaptive_formats:
-                            if f.get('acodec') != 'none':
-                                f['language'] = audio_locale
-                    formats.extend(adaptive_formats)
         self._sort_formats(formats)
 
         subtitles = {}