#metabrainz

hi there, was trying to import a cover art archive dump in my local mb db, is it possible?

[listenbrainz-android] 1407jasjeet merged pull request #560 (03main…fix-login): Fix: Login https://github.com/metabrainz/listenbrainz-andr...

lucifer: While uploading proguard mappings to sentry I get the following error: https://pastebin.com/r3yU6cnU

I followed an issue here: https://github.com/getsentry/self-hosted/issues...

Proguard file size of our app is about 6.77 mb. Can we do something here?

Also I have some auto-generated auth tokens on android project. If you could remove all the tokens except production one, it would be great!

Ill be skipping on uploading proguard mappings atm until upload body max size is increased.

[listenbrainz-android] 1407jasjeet opened pull request #561 (03main…r8-sentry-enable): Improvement: Enable Sentry & R8 https://github.com/metabrainz/listenbrainz-andr...

kellnerd: Your thoughts? What would something else be?

Last night and this morning I did some more asking around and research on the python thing.

Tulir being the author of maubot and all of the mautrix software, and nexy being a former matrix-nio user

From reading the code, what maubot provides over directly using mautrix-python is handling parsing commands sent into chat, providing the web management interface and easily running many plugins/bots on one Matrix account

And an abstraction over the database that lets you reuse maubot's database, whether that's postgres or sqlite

In terms of data storage, mautrix-python has a pluggable state store that handles storing the state needed to tuse Matrix's sync API, E2EE state and caching things like display names. That can be used with Postgres as the backend

However it doesn't seem to store non-state events, so the actual message timeline, and I'd have to add that functionality re my conversation with lucifer

So we'd end up with all of the Matrix state, the emssage history, and the state for updating the HTML stored in posgres

s/posgres/postgres/

I've not used postges' full text search before, but I have discovered that tantivy offers python bindings, so that part of the plan can stay the same https://github.com/quickwit-oss/tantivy-py

Jade[m]: From what I can see, pg doesn't really provide advanced query syntax, whereas tantivy has that OOTB. Search is secondary though

In terms of tooling, I haven't really done a significant python project before, so I'm not entirely sure what to use for python tooling

Uv and ruff seem popular and easy to use to me. Any opinions on using them?

For handling the JS/CSS, a simple vite config should be good, and the Jinja templates can load the JS entry point

I'm thinking TypeScript, but nothing else fancy, as it's pretty much only the search box that'll need JS

Oh, and ofc sentry and prometheus have official python SDKs, so that should be all good

<Jade[m]> "kellnerd: Your thoughts? What..." <- I mainly voted to be able to see the results, before I realized that it is apparently impossible to withdraw your vote 😅

Ah 😂

As I didn't want to impact the outcome without having read your proposal, I've switched to the most neutral option (for now).

I wouldn't worry too much, one of the votes is from someone I din't recognise

s/din/don/

Can the author of a poll see who voted for which option or did you derive this from the matrix room events? Asking because I only see the tally 😇

Yeah, this is looking at hidden events in the timeline

I din't know why element hides who voted and hides the results before you vote tbh

(A vote event, WjeClptjtsMP0ISH being the ID for the 'something else' option)

Oh, and seems like you can remove your vote on a protocol level, Element just doesn't implement it xD

<Jade[m]> "Uv and ruff seem popular and..." <- these are both great in my experience

If it's the way we're going, i'll write up the python edition of the proposal this evening, and do the timeline and PDF submission tomorrow btw

[listenbrainz-android] 1407jasjeet merged pull request #561 (03main…r8-sentry-enable): Improvement: Enable Sentry & R8 https://github.com/metabrainz/listenbrainz-andr...

[listenbrainz-android] release 032.8.4-beta has been published by 1407jasjeet: https://github.com/metabrainz/listenbrainz-andr...

Hey everyone, just published a 2.8.4-beta release that:... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/...>)

Please try the app out, I really need feedback on whats broken or whats not :)))

s/or/and/

jasje[m]: from about 30 seconds of usage, nothing seems to be broken haha. Listens submitted!

* Hey everyone, just published a 2.8.4-beta release that:... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/...>)

Jade: Thats a relief, I don't want to be in the "works on my device" state haha

Thanks for the feedback, keep testing!!

* Hey everyone, just published a 2.8.4-beta release that:... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/...>)

* Hey everyone, just published a 2.8.4-beta release that:... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/...>)

* Hey everyone, just published a 2.8.4-beta release that:... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/...>)

<Jade[m]> "(loading the top artists in..." <- Ouch, ui bug. Will create a ticket

I recently set up a local, standalone musicbrainz server, installed with mbdump, mbdump-dervied and mbdump-editor. The setup looks successful and I am able to issue simple sql statements to get artists linked to their relationships Example: the Beatles with just their band members with their gender. Now I want to link groups to their main genre and subgenres. I see the genre table and the l_genre_genre table. I see the

artist_tag data but I am at a loss to figure out how to link the Beatles to their genre data. The tag data for the Beatles doesn’t seem to show any meaningful connection to the genre table that I can see. Any help greatly appreciated for a new user in identifying the steps to go from the Beatles to their genre. Do I need more data than the three files I downloaded? Thanks.

rexwilliams1234: https://musicbrainz.org/doc/MusicBrainz_Databas...

have you looked at the schema docs? in theory those three dumps ought to do it you just need to connect all the dots...

<mayhem[m]> "rexwilliams1234: https://..." <- Yes I have looked at the schema docs and couldn't piece the connections together. In the example of the beatles. What would the table connections be. I had thought it would be artist_tag linked to l_genre_genre and then to genre to identify the genre names, expecting rock to be the primary genre. But if artist_tag is the starting point I can't see link to l_genre_genre that

makes sense. Using the beatles as an example - how would you trace their primary genre to the artist group? Do I have the wrong starting point? Or am I missing intermediate steps? Thanks.

ah. I see the confusion.

genres are a superset of tags.

so, fundamentally you really want to be making a tag query, not a genre query.

Thanks. Could that be illustrated a bit further with identifying the tables involved in linking an artist to their primary genre? That might help me figure out genres as a superset of tags .