Other than that, I can say it's not much different from scotty and elbisaur. It just has a better UX and is a nice option for end users.
2025-09-01 24441, 2025
elkiensad[m]
s/exporrt/export/
2025-09-01 24415, 2025
elkiensad[m]
works well, I can send both my LB and Spotify export if anybody is interested
2025-09-01 24451, 2025
elkiensad[m]
It's also nice that we can choose a specific day for the importer's end and start date, but I'd like to be able to choose a specific time as well (like the hour selection in manual addition).
2025-09-01 24415, 2025
elkiensad[m]
<elkiensad[m]> "monkey well, spotify's duplicate..." <- I believe that many of them can be solved with a few logics, even though 100% accuracy seems impossible.
2025-09-01 24457, 2025
elkiensad[m]
* my export is 611.2 mb (zip is only 44 mb though)
2025-09-01 24413, 2025
BlastboomStrice[
<monkey[m]> "elkiensad, Blastboom Strice..." <- > <@monkey:chatbrainz.org> elkiensad, Blastboom Strice : Any feedback on the importer? Everything working as you expect?... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/download/chatbrainz.org/KzPwScknZZxoEaErSwOTDdIy>)
2025-09-01 24429, 2025
BlastboomStrice[
Didnt have to change anything
2025-09-01 24442, 2025
elkiensad[m]
I had to delete ~130 listens for overlapping
2025-09-01 24445, 2025
BlastboomStrice[
Oof
2025-09-01 24457, 2025
BlastboomStrice[
I dont listen much music, mayhe thats why
2025-09-01 24407, 2025
elkiensad[m]
its easy to solve thanks to elbisaur though
2025-09-01 24413, 2025
BlastboomStrice[
That day had like 30songs
2025-09-01 24424, 2025
elkiensad[m]
Spotify's old data is definitely less accurate.
2025-09-01 24439, 2025
BlastboomStrice[
Btw my new extended history got in my email today
2025-09-01 24443, 2025
elkiensad[m]
there's not much ListenBrainz can do
2025-09-01 24404, 2025
BlastboomStrice[
I noticed there are some small differences, somehow for example some tracks on the new history dont have an offline timestamp
2025-09-01 24412, 2025
BlastboomStrice[
And other weird stuff
2025-09-01 24424, 2025
BlastboomStrice[
I suppose it'd work the same though
2025-09-01 24426, 2025
kellnerd[m] joined the channel
2025-09-01 24426, 2025
kellnerd[m]
Hmm, I haven't actually tested it, but having looked at the code I see two potential improvements.
2025-09-01 24450, 2025
elkiensad[m]
of course, we can import general listening patterns quite well now. I can say that many artists came to my top artists tab.
2025-09-01 24408, 2025
BlastboomStrice[
A funny ui thing was that I expected the page to tell me everything was imported and was afraid to press the 🔄 button so I left the page on for like half an hour (the import probably took a minute)
2025-09-01 24456, 2025
BlastboomStrice[
(Didnt read the "its safe to close this page now while listenbrainz imports")
2025-09-01 24457, 2025
kellnerd[m]
1. There is no filtering of listens at all. As already mentioned, it is probably impossible to get this right for everyone, but the [elbisaur docs](https://github.com/kellnerd/elbisaur/#parsi… have a few ideas.
2025-09-01 24423, 2025
BlastboomStrice[
Btw here the developer told me what filters (essentially anything that isnt a track and anything invalid)
Yeah, but no filtering for incomplete listens, that's what I meant.
2025-09-01 24434, 2025
elkiensad[m]
Spotify data needs many filters
2025-09-01 24449, 2025
kellnerd[m]
2. Spotify timestamps are taken as truth. Spotify logs the time when a listen ended, but LB wants to collect start time according to the docs. But I think this is also what the Spotify-LB sync does, so it might work better (i.e. with less duplicates) for users which also use that (not me).
2025-09-01 24419, 2025
BlastboomStrice[ uploaded an image: (137KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/4d2.org/bhsRBuojAjEMQtYTHoZxuOdM/image.png >
2025-09-01 24420, 2025
BlastboomStrice[ uploaded an image: (132KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/4d2.org/aHwwfrcgdKLqSeGZgxYmmkvg/image.png >
2025-09-01 24435, 2025
BlastboomStrice[
Left is from 2025-09-01
2025-09-01 24401, 2025
BlastboomStrice[
Right is from 2024-03-09
2025-09-01 24403, 2025
kellnerd[m]
And then there are the occasional cases when the timestamp is completely off (by hours/days), which can be corrected using the offline timestamp, if present.
2025-09-01 24440, 2025
BlastboomStrice[ uploaded an image: (135KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/4d2.org/xCzwwHlpJPMGALWbhmpRrtTM/image.png >
2025-09-01 24448, 2025
BlastboomStrice[
lol had some private data leak
2025-09-01 24452, 2025
BlastboomStrice[
fixed
2025-09-01 24407, 2025
BlastboomStrice[
Btw I also tried the listenbrainz importer on my alt account
2025-09-01 24423, 2025
BlastboomStrice[
Very quick as well
2025-09-01 24453, 2025
BlastboomStrice[
Which too doesnt have any advanced filters
2025-09-01 24416, 2025
mukti has quit
2025-09-01 24418, 2025
mukti joined the channel
2025-09-01 24423, 2025
monkey[m]
Thank you for the feedback!
2025-09-01 24426, 2025
szepix[m] joined the channel
2025-09-01 24426, 2025
szepix[m]
Is the beta for the iOS app closed? I wanted to join it :/
2025-09-01 24427, 2025
monkey[m]
Pinging jasje for LB iOS questions ^
2025-09-01 24416, 2025
jasje[m] joined the channel
2025-09-01 24417, 2025
jasje[m]
Hi, we’ll renew beta for LB iOS soon, we are facing some developer account issues that need to resolved.
2025-09-01 24450, 2025
monkey[m]
I've compiled the importer feedback into https://tickets.metabrainz.org/browse/LB-1825 , but please feel free to write your comments on the ticket if you run into more issues or inconsistencies
2025-09-01 24450, 2025
BrainzBot
LB-1825: Feedback for history file importer
2025-09-01 24400, 2025
UltimateRiff[m] has quit
2025-09-01 24404, 2025
BlastboomStrice[
<_discord_412> "↑ I'm curious about this too" <- I got an answer!
For Genre-Based Listening Patterns, each listen is joined on its recording_mbid to a precomputed recording → genre table. So the stat effectively pulls genres at the recording granularity (not artist-only like PanoScrobbler’s tag cloud). Any cross-entity propagation (e.g., from artist/release/release-group) would happen upstream when that recording→genre table is built, but this feature itself queries recording genres
2025-09-01 24431, 2025
BlastboomStrice[
and aggregates the results.
2025-09-01 24429, 2025
BlastboomStrice[
If you can help me here, I suppose that when you add a genre to an artist or a release group, they get propagated to the recordings, right?
2025-09-01 24429, 2025
BlastboomStrice[
So essentially this brings genres from all of the entities that are at the recording level and above(?)
