Merge ea7d7d85d3 into f351440f1d

Closes #8689
Authored by: jshumphrey, bashonly

Co-authored-by: bashonly <88596187+bashonly@users.noreply.github.com>

README.md

@@ -1869,6 +1869,9 @@ The following extractors use this feature:
 #### digitalconcerthall
 * `prefer_combined_hls`: Prefer extracting combined/pre-merged video and audio HLS formats. This will exclude 4K/HEVC video and lossless/FLAC audio formats, which are only available as split video/audio HLS formats
 
+#### bahamut
+* `device_id`: (optional) Device ID got from `https://ani.gamer.com.tw/ajax/getdeviceid.php` (bound to cookies and `User-Agent` HTTP header). The extractor will automatically fetch one if it is not present. E.g. `"bahamut:device_id=1234567890abcdef1234567890abcdef1234567890abcdef1234567890ab"`
+
 **Note**: These options may be changed/removed in the future without concern for backward compatibility
 
 <!-- MANPAGE: MOVE "INSTALLATION" SECTION HERE -->

yt_dlp/extractor/_extractors.py

@@ -196,6 +196,7 @@ from .awaan import (
 )
 from .axs import AxsIE
 from .azmedien import AZMedienIE
+from .bahamut import BahamutIE
 from .baidu import BaiduVideoIE
 from .banbye import (
     BanByeChannelIE,

yt_dlp/extractor/bahamut.py

@@ -0,0 +1,111 @@
+from .common import InfoExtractor
+from ..utils import (
+    ExtractorError,
+    float_or_none,
+    smuggle_url,
+    unified_timestamp,
+    unsmuggle_url,
+)
+from ..utils.traversal import traverse_obj
+
+
+class BahamutIE(InfoExtractor):
+    _VALID_URL = r'https?://ani\.gamer\.com\.tw/animeVideo\.php\?sn=(?P<id>\d+)'
+
+    # see anime_player.js
+    RATING_TO_AGE_LIMIT = {
+        1: 0,
+        2: 6,
+        3: 12,
+        4: 15,
+        5: 18,
+        6: 18,  # age-gated, needs login
+    }
+
+    def _real_extract(self, url):
+        url, unsmuggled_data = unsmuggle_url(url, {})
+        video_id = self._match_id(url)
+        device_id = (
+            self._configuration_arg('device_id', [None], casesense=True)[0]
+            or unsmuggled_data.get('device_id')
+            or self._download_json(
+                'https://ani.gamer.com.tw/ajax/getdeviceid.php', video_id,
+                'Downloading device ID', 'Failed to download device ID',
+                impersonate=True, headers=self.geo_verification_headers())['deviceid'])
+
+        # TODO: extract metadata from webpage
+        metadata = {}
+        if api_result := self._download_json(
+                'https://api.gamer.com.tw/anime/v1/video.php', video_id,
+                'Downloading video info', 'Failed to download video info',
+                impersonate=True, query={'videoSn': video_id}).get('data'):
+
+            metadata.update(traverse_obj(api_result, ('anime', {
+                'description': 'content',
+                'thumbnail': 'cover',
+                'tags': 'tags',
+                'creators': ('director', {lambda x: [x]}),
+                'title': 'title',
+            })))
+            playlist_id = traverse_obj(api_result, ('video', 'animeSn')) or ''
+            if self._yes_playlist(playlist_id, video_id) and unsmuggled_data.get('extract_playlist') is not False:
+                return self.playlist_result(
+                    (self.url_result(
+                        # it may be better to use self.cache for storing device_id
+                        smuggle_url(f'https://ani.gamer.com.tw/animeVideo.php?sn={ep["videoSn"]}', {
+                            'extract_playlist': False,
+                            'device_id': device_id,
+                        }), ie=BahamutIE,
+                        video_id=ep['videoSn'], thumbnail=ep.get('cover')) for ep in traverse_obj(
+                            api_result,
+                            # This (the first ellipsis) extracts episodes of all languages,
+                            # maybe just extract episodes of the current language?
+                            ('anime', 'episodes', ..., ...))),
+                    playlist_id=playlist_id, **metadata)
+
+            # video-specific metadata, extract after returning the playlist result
+            metadata.update(traverse_obj(api_result, ('video', {
+                'thumbnail': 'cover',
+                'title': 'title',
+                'timestamp': ('upTime', {unified_timestamp}),
+                'duration': ('duration', {float_or_none(scale=60)}),
+                'age_limit': ('rating', {lambda x: self.RATING_TO_AGE_LIMIT.get(x)}),
+            })))
+
+        m3u8_info, urlh = self._download_json_handle(
+            'https://ani.gamer.com.tw/ajax/m3u8.php', video_id,
+            note='Downloading m3u8 URL', errnote='Failed to download m3u8 URL', query={
+                'sn': video_id,
+                'device': device_id,
+            }, impersonate=True, headers=self.geo_verification_headers(), expected_status=400)
+
+        formats_fatal = True
+        if urlh.status == 400:
+            # TODO: handle more error codes, search for /case \d+{4}:/g in anime_player.js
+            error_code = traverse_obj(m3u8_info, ('error', 'code'))
+            if error_code == 1011:
+                self.raise_geo_restricted(metadata_available=True)
+                formats_fatal = False
+            elif error_code == 1007:
+                if unsmuggled_data.pop('device_id', None) is not None:
+                    return self.url_result(
+                        smuggle_url(f'https://ani.gamer.com.tw/animeVideo.php?sn={video_id}',
+                                    unsmuggled_data), ie=BahamutIE, video_id=video_id)
+                raise ExtractorError('Invalid device id!')
+            elif error_code == 1017:
+                self.raise_login_required(metadata_available=True)
+                formats_fatal = False
+            else:
+                raise ExtractorError(
+                    traverse_obj(m3u8_info, ('error', 'message')) or 'Failed to download m3u8 URL')
+
+        return {
+            **metadata,
+            'id': video_id,
+            'formats': self._extract_m3u8_formats(
+                m3u8_info.get('src'), video_id, ext='mp4', fatal=formats_fatal, headers={
+                    'Origin': 'https://ani.gamer.com.tw',
+                    **self.geo_verification_headers(),
+                }),
+            'http_headers': {'Origin': 'https://ani.gamer.com.tw'},
+        }

yt_dlp/extractor/ctvnews.py

@@ -1,11 +1,24 @@
+import json
 import re
+import urllib.parse
 
 from .common import InfoExtractor
-from ..utils import orderedSet
+from .ninecninemedia import NineCNineMediaIE
+from ..utils import extract_attributes, orderedSet
+from ..utils.traversal import find_element, traverse_obj
 
 
 class CTVNewsIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:.+?\.)?ctvnews\.ca/(?:video\?(?:clip|playlist|bin)Id=|.*?)(?P<id>[0-9.]+)(?:$|[#?&])'
+    _BASE_REGEX = r'https?://(?:[^.]+\.)?ctvnews\.ca/'
+    _VIDEO_ID_RE = r'(?P<id>\d{5,})'
+    _PLAYLIST_ID_RE = r'(?P<id>\d\.\d{5,})'
+    _VALID_URL = [
+        rf'{_BASE_REGEX}video/c{_VIDEO_ID_RE}',
+        rf'{_BASE_REGEX}video(?:-gallery)?/?\?clipId={_VIDEO_ID_RE}',
+        rf'{_BASE_REGEX}video/?\?(?:playlist|bin)Id={_PLAYLIST_ID_RE}',
+        rf'{_BASE_REGEX}(?!video/)[^?#]*?{_PLAYLIST_ID_RE}/?(?:$|[?#])',
+        rf'{_BASE_REGEX}(?!video/)[^?#]+\?binId={_PLAYLIST_ID_RE}',
+    ]
     _TESTS = [{
         'url': 'http://www.ctvnews.ca/video?clipId=901995',
         'md5': 'b608f466c7fa24b9666c6439d766ab7e',
@@ -17,13 +30,32 @@ class CTVNewsIE(InfoExtractor):
             'timestamp': 1467286284,
             'upload_date': '20160630',
             'categories': [],
-            'tags': [],
-            'season_id': 57981,
-            'duration': 764.631,
-            'series': 'CTV News National story',
-            'thumbnail': r're:^https?://.*\.jpg$',
-            'season': 'Season 0',
             'season_number': 0,
+            'season': 'Season 0',
+            'tags': [],
+            'series': 'CTV News National | Archive | Stories 2',
+            'season_id': '57981',
+            'thumbnail': r're:https?://.*\.jpg$',
+            'duration': 764.631,
+        },
+    }, {
+        'url': 'https://barrie.ctvnews.ca/video/c3030933-here_s-what_s-making-news-for-nov--15?binId=1272429',
+        'md5': '8b8c2b33c5c1803e3c26bc74ff8694d5',
+        'info_dict': {
+            'id': '3030933',
+            'ext': 'flv',
+            'title': 'Here’s what’s making news for Nov. 15',
+            'description': 'Here are the top stories we’re working on for CTV News at 11 for Nov. 15',
+            'thumbnail': 'http://images2.9c9media.com/image_asset/2021_2_22_a602e68e-1514-410e-a67a-e1f7cccbacab_png_2000x1125.jpg',
+            'season_id': '58104',
+            'season_number': 0,
+            'tags': [],
+            'season': 'Season 0',
+            'categories': [],
+            'series': 'CTV News Barrie',
+            'upload_date': '20241116',
+            'duration': 42.943,
+            'timestamp': 1731722452,
         },
     }, {
         'url': 'http://www.ctvnews.ca/video?playlistId=1.2966224',
@@ -46,6 +78,65 @@ class CTVNewsIE(InfoExtractor):
             'id': '1.5736957',
         },
         'playlist_mincount': 6,
+    }, {
+        'url': 'https://www.ctvnews.ca/business/respondents-to-bank-of-canada-questionnaire-largely-oppose-creating-a-digital-loonie-1.6665797',
+        'md5': '24bc4b88cdc17d8c3fc01dfc228ab72c',
+        'info_dict': {
+            'id': '2695026',
+            'ext': 'flv',
+            'season_id': '89852',
+            'series': 'From CTV News Channel',
+            'description': 'md5:796a985a23cacc7e1e2fafefd94afd0a',
+            'season': '2023',
+            'title': 'Bank of Canada asks public about digital currency',
+            'categories': [],
+            'tags': [],
+            'upload_date': '20230526',
+            'season_number': 2023,
+            'thumbnail': 'http://images2.9c9media.com/image_asset/2019_3_28_35f5afc3-10f6-4d92-b194-8b9a86f55c6a_png_1920x1080.jpg',
+            'timestamp': 1685105157,
+            'duration': 253.553,
+        },
+    }, {
+        'url': 'https://stox.ctvnews.ca/video-gallery?clipId=582589',
+        'md5': '135cc592df607d29dddc931f1b756ae2',
+        'info_dict': {
+            'id': '582589',
+            'ext': 'flv',
+            'categories': [],
+            'timestamp': 1427906183,
+            'season_number': 0,
+            'duration': 125.559,
+            'thumbnail': 'http://images2.9c9media.com/image_asset/2019_3_28_35f5afc3-10f6-4d92-b194-8b9a86f55c6a_png_1920x1080.jpg',
+            'series': 'CTV News Stox',
+            'description': 'CTV original footage of the rise and fall of the Berlin Wall.',
+            'title': 'Berlin Wall',
+            'season_id': '63817',
+            'season': 'Season 0',
+            'tags': [],
+            'upload_date': '20150401',
+        },
+    }, {
+        'url': 'https://ottawa.ctvnews.ca/features/regional-contact/regional-contact-archive?binId=1.1164587#3023759',
+        'md5': 'a14c0603557decc6531260791c23cc5e',
+        'info_dict': {
+            'id': '3023759',
+            'ext': 'flv',
+            'season_number': 2024,
+            'timestamp': 1731798000,
+            'season': '2024',
+            'episode': 'Episode 125',
+            'description': 'CTV News Ottawa at Six',
+            'duration': 2712.076,
+            'episode_number': 125,
+            'upload_date': '20241116',
+            'title': 'CTV News Ottawa at Six for Saturday, November 16, 2024',
+            'thumbnail': 'http://images2.9c9media.com/image_asset/2019_3_28_35f5afc3-10f6-4d92-b194-8b9a86f55c6a_png_1920x1080.jpg',
+            'categories': [],
+            'tags': [],
+            'series': 'CTV News Ottawa at Six',
+            'season_id': '92667',
+        },
     }, {
         'url': 'http://www.ctvnews.ca/1.810401',
         'only_matching': True,
@@ -57,29 +148,35 @@ class CTVNewsIE(InfoExtractor):
         'only_matching': True,
     }]
 
+    def _ninecninemedia_url_result(self, clip_id):
+        return self.url_result(f'9c9media:ctvnews_web:{clip_id}', NineCNineMediaIE, clip_id)
+
     def _real_extract(self, url):
         page_id = self._match_id(url)
 
-        def ninecninemedia_url_result(clip_id):
+        if mobj := re.fullmatch(self._VIDEO_ID_RE, urllib.parse.urlparse(url).fragment):
-            return {
+            page_id = mobj.group('id')
-                '_type': 'url_transparent',
-                'id': clip_id,
-                'url': f'9c9media:ctvnews_web:{clip_id}',
-                'ie_key': 'NineCNineMedia',
-            }
 
-        if page_id.isdigit():
+        if re.fullmatch(self._VIDEO_ID_RE, page_id):
-            return ninecninemedia_url_result(page_id)
+            return self._ninecninemedia_url_result(page_id)
-        else:
+
-            webpage = self._download_webpage(f'http://www.ctvnews.ca/{page_id}', page_id, query={
+        webpage = self._download_webpage(f'https://www.ctvnews.ca/{page_id}', page_id, query={
-                'ot': 'example.AjaxPageLayout.ot',
+            'ot': 'example.AjaxPageLayout.ot',
-                'maxItemsPerPage': 1000000,
+            'maxItemsPerPage': 1000000,
-            })
+        })
-            entries = [ninecninemedia_url_result(clip_id) for clip_id in orderedSet(
+        entries = [self._ninecninemedia_url_result(clip_id)
-                re.findall(r'clip\.id\s*=\s*(\d+);', webpage))]
+                   for clip_id in orderedSet(re.findall(r'clip\.id\s*=\s*(\d+);', webpage))]
-            if not entries:
+        if not entries:
-                webpage = self._download_webpage(url, page_id)
+            webpage = self._download_webpage(url, page_id)
-                if 'getAuthStates("' in webpage:
+            if 'getAuthStates("' in webpage:
-                    entries = [ninecninemedia_url_result(clip_id) for clip_id in
+                entries = [self._ninecninemedia_url_result(clip_id) for clip_id in
-                               self._search_regex(r'getAuthStates\("([\d+,]+)"', webpage, 'clip ids').split(',')]
+                           self._search_regex(r'getAuthStates\("([\d+,]+)"', webpage, 'clip ids').split(',')]
-            return self.playlist_result(entries, page_id)
+            else:
+                entries = [
+                    self._ninecninemedia_url_result(clip_id) for clip_id in
+                    traverse_obj(webpage, (
+                        {find_element(tag='jasper-player-container', html=True)},
+                        {extract_attributes}, 'axis-ids', {json.loads}, ..., 'axisId'))
+                ]
+
+        return self.playlist_result(entries, page_id)