Merge ea7d7d85d3 into f2a4983df7

Closes #11526
Authored by: jshumphrey

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/archiveorg.py

@@ -205,6 +205,26 @@ class ArchiveOrgIE(InfoExtractor):
                 },
             },
         ],
+    }, {
+        # The reviewbody is None for one of the reviews; just need to extract data without crashing
+        'url': 'https://archive.org/details/gd95-04-02.sbd.11622.sbeok.shnf/gd95-04-02d1t04.shn',
+        'info_dict': {
+            'id': 'gd95-04-02.sbd.11622.sbeok.shnf/gd95-04-02d1t04.shn',
+            'ext': 'mp3',
+            'title': 'Stuck Inside of Mobile with the Memphis Blues Again',
+            'creators': ['Grateful Dead'],
+            'duration': 338.31,
+            'track': 'Stuck Inside of Mobile with the Memphis Blues Again',
+            'description': 'md5:764348a470b986f1217ffd38d6ac7b72',
+            'display_id': 'gd95-04-02d1t04.shn',
+            'location': 'Pyramid Arena',
+            'uploader': 'jon@archive.org',
+            'album': '1995-04-02 - Pyramid Arena',
+            'upload_date': '20040519',
+            'track_number': 4,
+            'release_date': '19950402',
+            'timestamp': 1084927901,
+        },
     }]
 
     @staticmethod
@@ -335,7 +355,7 @@ class ArchiveOrgIE(InfoExtractor):
                 info['comments'].append({
                     'id': review.get('review_id'),
                     'author': review.get('reviewer'),
-                    'text': str_or_none(review.get('reviewtitle'), '') + '\n\n' + review.get('reviewbody'),
+                    'text': join_nonempty('reviewtitle', 'reviewbody', from_dict=review, delim='\n\n'),
                     'timestamp': unified_timestamp(review.get('createdate')),
                     'parent': 'root'})
 

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/facebook.py

@@ -563,13 +563,13 @@ class FacebookIE(InfoExtractor):
                 return extract_video_data(try_get(
                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 
-        def extract_dash_manifest(video, formats):
+        def extract_dash_manifest(vid_data, formats, mpd_url=None):
             dash_manifest = traverse_obj(
-                video, 'dash_manifest', 'playlist', 'dash_manifest_xml_string', expected_type=str)
+                vid_data, 'dash_manifest', 'playlist', 'dash_manifest_xml_string', 'manifest_xml', expected_type=str)
             if dash_manifest:
                 formats.extend(self._parse_mpd_formats(
                     compat_etree_fromstring(urllib.parse.unquote_plus(dash_manifest)),
-                    mpd_url=url_or_none(video.get('dash_manifest_url'))))
+                    mpd_url=url_or_none(video.get('dash_manifest_url')) or mpd_url))
 
         def process_formats(info):
             # Downloads with browser's User-Agent are rate limited. Working around
@@ -619,9 +619,12 @@ class FacebookIE(InfoExtractor):
                         video = video['creation_story']
                         video['owner'] = traverse_obj(video, ('short_form_video_context', 'video_owner'))
                         video.update(reel_info)
-                    fmt_data = traverse_obj(video, ('videoDeliveryLegacyFields', {dict})) or video
+
                     formats = []
                     q = qualities(['sd', 'hd'])
+
+                    # Legacy formats extraction
+                    fmt_data = traverse_obj(video, ('videoDeliveryLegacyFields', {dict})) or video
                     for key, format_id in (('playable_url', 'sd'), ('playable_url_quality_hd', 'hd'),
                                            ('playable_url_dash', ''), ('browser_native_hd_url', 'hd'),
                                            ('browser_native_sd_url', 'sd')):
@@ -629,7 +632,7 @@ class FacebookIE(InfoExtractor):
                         if not playable_url:
                             continue
                         if determine_ext(playable_url) == 'mpd':
-                            formats.extend(self._extract_mpd_formats(playable_url, video_id))
+                            formats.extend(self._extract_mpd_formats(playable_url, video_id, fatal=False))
                         else:
                             formats.append({
                                 'format_id': format_id,
@@ -638,6 +641,28 @@ class FacebookIE(InfoExtractor):
                                 'url': playable_url,
                             })
                     extract_dash_manifest(fmt_data, formats)
+
+                    # New videoDeliveryResponse formats extraction
+                    fmt_data = traverse_obj(video, ('videoDeliveryResponseFragment', 'videoDeliveryResponseResult'))
+                    mpd_urls = traverse_obj(fmt_data, ('dash_manifest_urls', ..., 'manifest_url', {url_or_none}))
+                    dash_manifests = traverse_obj(fmt_data, ('dash_manifests', lambda _, v: v['manifest_xml']))
+                    for idx, dash_manifest in enumerate(dash_manifests):
+                        extract_dash_manifest(dash_manifest, formats, mpd_url=traverse_obj(mpd_urls, idx))
+                    if not dash_manifests:
+                        # Only extract from MPD URLs if the manifests are not already provided
+                        for mpd_url in mpd_urls:
+                            formats.extend(self._extract_mpd_formats(mpd_url, video_id, fatal=False))
+                    for prog_fmt in traverse_obj(fmt_data, ('progressive_urls', lambda _, v: v['progressive_url'])):
+                        format_id = traverse_obj(prog_fmt, ('metadata', 'quality', {str.lower}))
+                        formats.append({
+                            'format_id': format_id,
+                            # sd, hd formats w/o resolution info should be deprioritized below DASH
+                            'quality': q(format_id) - 3,
+                            'url': prog_fmt['progressive_url'],
+                        })
+                    for m3u8_url in traverse_obj(fmt_data, ('hls_playlist_urls', ..., 'hls_playlist_url', {url_or_none})):
+                        formats.extend(self._extract_m3u8_formats(m3u8_url, video_id, 'mp4', fatal=False, m3u8_id='hls'))
+
                     if not formats:
                         # Do not append false positive entry w/o any formats
                         return