#metabrainz

so for now, we stay put and hope for the best that the US/California doesn't totally melt down.

good week, overall, really!

monkey: go!

Hello everyone!

Last week i worked on the Libre.FM backend importer, specifically looking at the UI portions and replacing the now-empty manual imports page with a placeholder, awaiting the file importers being developed over the course of this GSOC

I looked into improving our error boundaries for thye ListenBrainz player after a suggestion by bitmap (thanks!), but there is more than meets the eye. I found one improvement but looking into it deeper.

I refactored some front-end tests from enzyme to testing-library (ongoing work)

And some PR reviewing to top it all off.

yvanzo: Your turn!

Hi!

Got worse last week, skipped it too.

Fin, go kellnerd?

Hey there!

Last week I had to fix Harmony's Tidal provider again because of a breaking API change.

The change wasn't announced AFAIK, so I had to hotfix this by temporarily losing cover art from Tidal.

Since the API query parameters have changed, I finally had to implement support for legacy URLs in Harmony's caching logic (in order to avoid breaking old permalinks).

Seems to work as expected and Tidal images are back, but I should do some more tests (and write unit tests) before releasing this.

And after more than two years, I've released a new version of my *Guess Unicode punctuation* userscript, which integrates new rules for Hebrew contributed by dvirtz.

Go Hemang Mishra!

Hello!!

Still on my list for today: suvid , m.amanullah7 , Gautam_coder , reosarevok , zas

This week, I mostly worked on implementing Listening Apps Selection Screen and opened a PR for the same.

Also, I fixed a bug in the User Playlist Screen where cover arts were not being fetched due to the large number of parallel Api calls.

That's it for me.

suvid: next?

Hi

Hi! :)

I restarted my GSoC project again last week

Figured out most of the stuff already and kinda back on track

This time it looks more organized and I understand it better :)

Implemented the file upload part and background tasks part

I will make the new schema and test things

after that I will move on to zip file processing for spotify and then make functions that process files for other services as well

That's all for this week

Gautam_coder: next?

Not sure they're active, I can go

Last week I worked on updating my PRs for extra genre data in the API, since now that we have a new search server those things are unblocked and just need to slowly arise from their sleep

Also worked on a couple small improvements, such as MBS-14065

MBS-14065: Show artist credits in release editor duplicates tab https://tickets.metabrainz.org/browse/MBS-14065

And updated some outdated or just wrong flags

Plus a bunch of code review, adding areas, banning evildoers or at least meh-not-cool-doers, etc

m.amanullah7: what about you?

Hey everyone

Still on my list for today: zas

Last week I've added a delete_user function which will earse the user credentials if user disable funkwhale

And also refractor backend

And last week I've finally integrated fw in BrainzPlayer and pushed the changes for review

And also fixed some of the issue like changes refresh method from get to post and worked on fw css as well!

That's it from my side!

Next zas: ?

Does not seem to be around. So let's finish for today!

Thanks everyone :)

More in a week!

</BANG>

Thanks reo

julian45: do different servers have significantly different docker versions?

It seems https://github.com/metabrainz/docker-server-con... fixed warnings on some servers, but others just die because they don't know what timeout is 🫠

(I'm assuming I just need to revert it for now)

auth will have latest docker major, most others will not afaik because defaults for metabrainz-ansible are pretty behind

Thanks everyone!

ok yeah --time was only renamed in docker engine 28 per the docs you linked and i think most MeB servers are pretty far behind that last i checked

Hmm, I see. Some were warning, but I didn't think some might not work at all. I guess I should revert for now, and hope it's backwards-compatible for long enough that the servers get updated in the meantime :D

(and the new ones don't start rejecting it outright)

yeah, i don't imagine it'd get removed outright till docker engine 29 or so, and latest is 28.2 or so

Ok

Pushed a revert commit, let's see

Sorry, I had an emergency to handle, nothing special this week anyway; usual stuff, system maintenance, upgrades & fixes, Picard, MB edits, user support.

Ok, running the prod update again, now without breakages, hopefully :)

Hey lucifer, I tried running the query got a huge trace which ended with... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/...>)

holycow23: need to the entire stack trace to help

* need to see the entire

Yeah how do I see the entire trace of the pyspark shell?

no i mean just paste whatever you do see into a file and upload that to a gist

But it was something related to Mkdirs failed to create file:/stats/user/bookkeeping/genre_trend/last_week/incremental/...

makes sense but even then need to see that complete section of the stack trace

holycow23: did you configure the hdfs_connection properly?

I think I did

in particular, this set config.HDFS_CLUSTER_URI = "hdfs://127.0.0.1:9000"

just print that variable and confirm that is set to that value

yes it is set correctly

Hey lucifer: any update on my pr?

m.amanullah7: yes i reviewed the backend parts of it and one major change is needed. funkwhale has an api to create oauth apps, we should use that instead of harcoding the client id and client secret in the config. take a look at the section on the top of the page: https://docs.funkwhale.audio/swagger/#/

I've pushed the changes and ive integrated fw in bp! U can have a look rn its its priority is lowest but u can change its like other services to top and can test !! I tried to add songs in fw but failed and when i tried to play song I was getting error authentication required!!

i think rest everything should remain the same just instead of reading the client id/secret from config file, you have to call that api to get the data.

there are other smaller changes that i would like and also i am thinking if we should change the existing tables schema to simplify funkwhale implementation, reduce the specific code needed for it at least.

lucifer: updated the [gist](https://gist.github.com/granth23/95232d5ed5c0eff47ad4ab2bb14fc9bb) with the entire trace

as for frontend changes i will ask monkey to do that.

* entire trace which was visible

lucifer[m]: lucifer: I'll update this and let u know!

lucifer[m]: What u can change in schema like removing client id and secret?

m.amanullah7: i am thinking that access token and refresh token's could be stored in external service oauth table. the tokens for spotify/apple music etc. are stored there for example.

What about host url?

that would remain in the funkwhale servers table.

and the client id and client secret would also be added to that table.

so we still have a new table but I am trying to think if it can simplify our implementation.

i'll think more about it and we can see how to move forward on that.

lucifer[m]: lucifer: you're referring to external oauth table?

no the funkwhale servers table.

Okay!!

I'll move access token and refresh token I'm external oauth table and rest all will be in funkwhale servers

holycow23: okay i see, restart pyspark. run the setup commands again. and then import your stats and run it. i think you had imported it before running the setup so the outdated empty value got hardcoded there.

s/I'm/in/

m.amanullah7: you can't do that right now because then you would need to add a new column to store the funkwhale server id to which the token belongs.

lucifer[m]: So re run the shell and then the same commands right?

* commands right? cause I already did that and got the same

yes but ensure that the setup commands are run before any stats imports.

okay

m.amanullah7: i am thinking that we could change the `service` column in external oauth table to store that id.

so its a bigger change if we want to do that. so focus on the remaining parts of your project for now.

i'll discuss with mayhem and monkey this later and if we decide to do it, i'll migrate the existing services first and then you can make the relevant funkwhale changes.

lucifer[m]: should I run this... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/...>)

* should I run this... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/...>)

holycow23: move line 4 to line 12. otherwise yes.

also remove line 1 actually

it shouldn't change anything but is also not needed for the setup anyway.

the key is moving the import on line to after you have patched the config.

* on line 4 to after

lucifer: I've to once look external service table can't we create any kind of relationship with both the tables and connect them!

lucifer[m]: Sure!

yes you can create a foreign key relationship

but i want to avoid it unless necessary

lucifer[m]: I'll work on what you've suggested!

lucifer[m]: made that change got a new [error](https://gist.github.com/granth23/95232d5ed5c0eff47ad4ab2bb14fc9bb)

Going through it right now

holycow23[m]: lucifer: I think we are missing one line in the setup for resolving namenode?

holycow23: like i said before don't import the import at line 4 before configuring.

lucifer: i think we only have host url in extra so can't we create a new column?

s/```/````/, s/listenbrainz_spark.listens.cache import *//, s/```/````/

* `````from... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/...>)

Isn't it the same error, updated the [gist](https://gist.github.com/granth23/95232d5ed5c0eff47ad4ab2bb14fc9bb)

* lucifer: Isn't it

YEs

lucifer[m]: lucifer: my bad! I think only option to connect using foreign key which we dont wants!

Now I'll work on ur reviews and will update u!