i'll rebase this on latest master once, regenerate all the models and update the PR.
2024-01-13 01353, 2024
lucifer
if we merge this PR, then i shouldn't need to update the `.gitmodules` file?
2024-01-13 01324, 2024
kellnerd
Where in the LB codebase does the cover art selection happen? I have a listen here where the displayed cover art is literally the worst choice for the whole release group: An image with types other+Raw/unedited from the only release which does not have a front cover 😂
2024-01-13 01323, 2024
lucifer
kellnerd: link to the listen?
2024-01-13 01308, 2024
kellnerd
Sent you a PM
2024-01-13 01348, 2024
mayhem
clearly we need to document all this cover art black magic. I honestly never thought cover art could get so complicated.
2024-01-13 01357, 2024
kellnerd
It was submitted manually via the website, which seems to think that is the canonical release.
2024-01-13 01342, 2024
kellnerd
"mbid_mapping" suggests a better cover but that is apparently ignored in favor of the canonical release, although that one has no front cover.
2024-01-13 01335, 2024
kellnerd
The problem is that the listens contains a release MBID although the "Add listens" popup gives me no choice to select a release, only the recording.
2024-01-13 01334, 2024
lucifer
kellnerd: i took a look at the data, this seems like a bug in frontend or the backend to me. only front cover arts should be used.
2024-01-13 01353, 2024
lucifer
the cover art that appears in your case is not marked front so it should never show up in any case.
2024-01-13 01300, 2024
lucifer
i'll try to fix it.
2024-01-13 01306, 2024
kellnerd
I thought so. Falling back to something else if there is no front cover across the whole RG might make sense, but not if other releases have a front cover. Thanks!
no worries. let me know if we can merge these, i would prefer to rebase the existing solr9 PRs on the master before moving ahead
2024-01-13 01350, 2024
yvanzo
lucifer: I see, I will test it further as we don’t have enough test coverage for it.
2024-01-13 01306, 2024
lucifer
yvanzo: makes sense
2024-01-13 01312, 2024
lucifer
let me know if i can help with the testing.
2024-01-13 01303, 2024
GauravGupta993
Hello, I am interested in contributing to listenbrainz-server. I have worked on Flask and MERN tech stacks before but this would be my first time contributing to a big project. I have already set up the development environment on my PC. It would be really helpful if I could get some guidance on which issue to take first and how to start
I probably won't write a direct importer, but something that imports these into a sqlite DB so I can diff them with last.fm
2024-01-13 01328, 2024
bttf has quit
2024-01-13 01303, 2024
bitmap
to fix all the dupes/incorrect timestamps caused by spotify's crap official scrobbler
2024-01-13 01323, 2024
kellnerd
Yeah, I have already disbaled the LB Spotify integration after seeing the many errors in the first 50 automatic imports already :D
2024-01-13 01348, 2024
kellnerd
I think it also treats streams as listens when you have only listened a single second.
2024-01-13 01333, 2024
kellnerd
So, e.g. play until the start of track X, pause, go away, come back and hit play, and you have a duplicate listen for track X.
2024-01-13 01315, 2024
bitmap
I have the exact same issue on last.fm, so yeah, def. a spotify bug :\
2024-01-13 01353, 2024
bitmap
but it also submits dupes for entire albums at staggered times, I guess related to me using clients on multiple devices/machines
2024-01-13 01347, 2024
bitmap
the data in the JSON they gave me looks fine though
2024-01-13 01358, 2024
kellnerd
Is it still the same format as in the linked PR's description?
2024-01-13 01301, 2024
kellnerd
ms_played, reason_start and reason_end might be useful to get rid of these bad short listens if they are included
2024-01-13 01319, 2024
bitmap
the format is the same, and those are indeed useful. reason_end could be, e.g., "trackdone", "endplay", and "unexpected-exit-while-paused". but an abnormally short ms_played is also a giveaway
2024-01-13 01349, 2024
bitmap
I can send you the PDF documentation they provided with the dumps if you're interested, but it should come with yours later
2024-01-13 01346, 2024
kellnerd
I guess I will be patient and focus on my .scrobbler.log files and "full release listens noted on paper" first :)
2024-01-13 01306, 2024
bitmap
ok! just curious, what is the source of your .scrobbler.log files?
2024-01-13 01338, 2024
kellnerd
My portable Sansa MP3 player with Rockbox custom firmware
2024-01-13 01322, 2024
bitmap
neat
2024-01-13 01348, 2024
bitmap
I used to have a cowon player with rockbox a long time ago, wish it still worked
2024-01-13 01342, 2024
kellnerd
Mine is 16 years old and still working, only the clip on the back has died long ago :D
2024-01-13 01355, 2024
bitmap
wow! mine only lasted a few years sadly
2024-01-13 01313, 2024
bitmap
I was probably 16 when I got it, 'cause it could play flac and ogg files XD
2024-01-13 01346, 2024
RetroPunk has quit
2024-01-13 01328, 2024
RetroPunk joined the channel
2024-01-13 01351, 2024
kellnerd
Hmm, how long does it take the LB mapper to notice upstream data changes on MB? A changed recording title, for example. (Changing it had no effect when I listened to it again shortly after.)
