aboutsummaryrefslogtreecommitdiffstatshomepage
path: root/yt_dlp_plugins/extractor/radiko.py
Commit message (Collapse)AuthorAgeFilesLines
* Fix TOCTOU bug where only one stream was actually downloadablegarret13172 days1-6/+3
| | | | | | | | | | | | | | | | | | | | | | | | | because the playlist_url, domain, first_chunk etc gets re-set by the next items in the loop and the fragments_generator() joyously goes ahead and uses those values, because that's what's there and ^it's a generator so i didnt even need to have the separate fragments_generator func anyway sigh.... well fortunately it failed such that it always downloaded wowza stream which is the best one but still sigh... no wonder both formats looked the same in the spectrum analyser lmao i suppose this is TOCTOU https://en.wikipedia.org/wiki/Time-of-check_to_time-of-use i found this while trying to download 24 hours of silence lmao https://radiko.jp/#!/ts/OC1/20251119050000 "OC1", prob means 'original content 1' ※配信期間外です　LAWSON SEEDS OF CREATION　配信日：2025年11月23日(日)22:00～23:02 ※配信期間外は無音となります。 LAWSON SEEDS OF CREATIONで放送されなかったJAM’S からメンバーへの質問を特別にオンエアー！ あのヒット曲について語る！ wowza stream errors but smartstream is fine so it correctly selected the smartstream stream but the frags it tried to dl were from the wowza stream fun
* move PersonIE from API to mobile web (like RSeasons)garret13178 days1-53/+40
| | | | | | more metadata consistent with whats seen on the site less complicated
* commentgarret13172025-11-071-0/+1
|
* mark chunked formats as from_startgarret13172025-11-071-1/+1
| | | | so that you can interrupt and resume where you left off like normal
* Get the first chunk in advancegarret13172025-11-071-8/+11
| | | | | | | | | so that the returned `url` "works" in mpv etc i can download the first one as part of the normal self._extract_m3u8_formats, and pass the url as the first chunk of my chunking set up then the chunks stuff still works, and i can return an url that will play in mpv ...except that it will be 5 minutes ahead because it's a "live"stream oh well its better than nothing lmao
* Implement lazy chunk downloadinggarret13172025-11-071-4/+7
| | | | | | | | using `http_dash_segments_generator` as seen in youtube live-from-start this way it only goes through the chunks etc one a format has actually been selected so we don't have it downloading eg 20 chunks of every stream only for most to be discarded closes #29
* match "rdk://station-timestring" url for easier debugginggarret13172025-11-071-1/+4
|
* remove stream blacklistgarret13172025-11-061-7/+0
| | | | | it's no longer necessary (hasn't been since 221cefe4810e0bcd5c6faf199ae2589f62edbb77) closes #29
* fully switch to pc_html5 (they killed the radiko.jp ondemand stream)garret13172025-11-061-6/+1
|
* RSeasons: extract description and thumbnailgarret13172025-09-211-1/+8
|
* fix typogarret13172025-09-191-1/+1
|
* Fix search testgarret13172025-09-191-0/+1
|
* oh, absence of a filter means every option doesnt itgarret13172025-09-191-17/+18
|
* Fix search filtering + add support for podcast filteringgarret13172025-09-191-8/+36
|
* disable stream blacklist for livegarret13172025-09-141-1/+3
| | | | | ref https://github.com/garret1317/yt-dlp-rajiko/issues/29#issuecomment-3289577318 github issue #29
* rename "misc" -> "contrib"garret13172025-09-141-1/+1
|
* clarify "indexing (station) regions" textgarret13172025-09-131-1/+1
|
* Don't return any formats when we don't actually have anygarret13172025-09-111-3/+7
|
* update some commentsgarret13172025-09-111-2/+2
|
* add support for "r_seasons" programme pagesgarret13172025-08-291-0/+42
|
* PersonsIE: construct a timefree url_result directlygarret13172025-08-291-11/+8
| | | | | so that --download-archive etc will quickly skip without a full extraction also removes reliance on `radiko_url` always being a share url
* Flag and deprioritise formats with ad insertiongarret13172025-08-261-1/+8
|
* Set ffmpeg args to not send Range headergarret13172025-08-261-3/+10
| | | | github issue #29
* set vcodec none for chunked formatsgarret13172025-08-241-0/+1
| | | | so they get marked as audio only
* move ShareIE tests to test scriptgarret13172025-06-221-31/+0
|
* Move RadikoTimeFreeIE tests to test scriptgarret13172025-06-221-60/+1
|
* set count -> mincount for persons key station testgarret13172025-06-131-1/+1
|
* Extract station logo as thumbnail for live now that banner is gonegarret13172025-06-131-18/+21
| | | | well not technically _gone_, but replaced with a useless placeholder image
* fix search only returning one resultgarret13172025-06-061-1/+1
| | | | | rather embarrassing bug introduced in a659b48a20b6241a68fc046c86c9ad0f71c9bd86 / v1.5 only caught it now because i only listen to weekly programmes so i only get 1 result anyway
* set device to pc_html5 for live as wellgarret13172025-06-011-4/+4
| | | | aSmartPhone7a removed working stream
* switch programme data api to api.radiko.jp as used on sitegarret13172025-05-251-1/+1
| | | | identical content
* make ShareIE return an id in advance as wellgarret13172025-05-251-1/+4
| | | | | | | | | in my excitement i may have prematurely released (har har) v1.5 with the return-id-in-advance trick only added to SearchIE oh well, it can be in ShareIE for next release adding to ShareIE has knock-on effects for PersonsIE as it returns a share url two birds with one stone
* make SearchIE return an id in advance, for download archivesgarret13172025-05-241-3/+13
| | | | | to prevent un-needed extractions when somethings in the archive already absolutely kicking myself that it was this simple and i only just did it
* make tracklist downloading non-fatal so MAJAL worksgarret13172025-05-211-2/+4
| | | | MUSIC AWARDS JAPAN AUDIO LIVE pop-up stream
* chunked: properly fallback to live for ffmpeg etcgarret13172025-05-191-2/+4
|
* cleanup m3u8 request headers into a variablegarret13172025-05-191-2/+4
|
* track real chunk duration to prevent dupes when chunks are longer than expectedgarret13172025-05-181-2/+5
| | | | closes #24 for real this time
* add extractor arg to disable as-live workaroundgarret13172025-05-171-1/+2
|
* add quick-downloading support for as-live formatsgarret13172025-05-171-5/+18
| | | | closes #24
* update "type" comment with new informationgarret13172025-05-171-1/+1
|
* add extractor arg to disable the blacklistgarret13172025-03-301-1/+3
|
* add new alliance-stream-radiko.smartstream.ne.jp format to ffmpeg blacklistgarret13172025-03-271-1/+1
| | | | github issue #25
* switch to different search API url observed on sitegarret13172025-03-061-1/+1
| | | | | | | | | | no change in the response structure*, just the URL afaict (stuff doesnt break at least) interestingly the headers have started including GRPC stuff though *but the search _results_ seem to have changed somewhat had to change my search for ジャズ倶楽部 to specifically ジャズ倶楽部 CRT because loads of results had "ジャズ・クラブ" - not katakana クラブ not kanji 倶楽部 maybe its not changed and i just got unlucky lol
* filter out "cul_area_id" in search params - HTTP Error 415garret13172025-03-061-0/+5
| | | | | | | | site used to use "cul_area_id" in the search url, now it uses "cur_area_id" (with an r) and it outright rejects the old one with HTTP Error 415: Unsupported Media Type probably better to patch in the extractor than to break peoples scripts etc - site does this too idk why they felt the need to break the old one
* missed an importgarret13172025-01-211-0/+1
|
* clearer variable names/comments + fix getting tf30 tokengarret13172024-11-221-11/+11
|
* add debug logs for account sign-in processgarret13172024-11-221-0/+3
|
* Force new token if cached one doesn't have tf30garret13172024-11-221-7/+11
|
* support logging in with accounts for timefree30garret13172024-11-221-5/+39
|
* fix broken time calculation in PersonsIEgarret13172024-11-221-1/+1
| | | | | | | | the .broadcast_day_start() was trying to run on the datetime.timedelta, not the RadikoTime that we calculate so it just ERROR: 'datetime.timedelta' object has no attribute 'broadcast_day_start'
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-11-22 23:45:35 +0000