#metabrainz

Morning!!

GoodMorningg!

mo’’in’

Morning!!!

Moin !

moin

ishaanshah[m]: i won't be able to make our meeting normal time today either, i pushed it an hour earlier.

Argh

I hate the "authority control" name

People keep thinking we're the "owners" of some Wikipedia pages because they have authority control -> musicbrainz

Do they want to pay us taxes or something?

huh.

Mr_Monkey: https://tickets.metabrainz.org/browse/BB-472

BB-472: The Neverending Identifierslist

(yes my ticekt names are refrences to some books.)

Mr_Monkey: on the playlist page, which is behaving well for me (woo!), the timestamps for tracks are still shown.

I'd love to also show some metadata for the playlist.

I've thought more about your suggestion of giving playlists UUIDs and then saving them locally. I think that sounds like a good first step, since that allows me to go back to the playlist in question and look at the comments to find relevant data.

what should we do with the timestamps? can we put something else useful there?

Mr_Monkey: this is probably a schema change. no? https://tickets.metabrainz.org/browse/BB-473

BB-473: Create a comprehensive relationship-type editing interface

(also sorry for focusing on the bb thask things now that you're 50% on LB. it's jsut that i was planning on doing all these tickets before we heard of he priority change somewhat :D I hope lots of bb tickets now does not stress you)

Thanks CatQuest ! I'll have a look later. More tickets is always good !

ruaok: Right, timestamps.

I'm not sure we need to put anything there.

and playlist metadaa. lets noodle on that too.

But definitely removing them

ok. I'm sure we'll find other stuff to put there in time.

Playlist metadata we should add to the html flask template. Since we're reusing the existing react components as much as possible that makes the most sense

hmm. ok.

then I'll change the structure of the data that is posted and do the parsing for you, then pass you nothing but the list of listens and then add new fields to template.

👍

Let me know where you're pushing that, I'll do the timestamp change there too

k, let me tuck away a couple of things I'm working on first.

oh ho that reminds me. i actually have some listenbrainz ticket ( ^ิ ॢᗨ^ิﾟ) ohohoho

Oh no.

Moiiin

alastairp: I have studied your papers. I have to mentions that it is an incredible work. Could we please arrange a meeting e.g. in the middle of the next week? I have some questions about the implementation process, especially on the training phase, as the evaluation on the validation sets is obvious I think.

*mention

pristine__: is #846 ready to merge?

Yeah. Once you approve. I will test on the cluster then.

I have tested on locally

ok.

jmp_music: sure. all days are good for me. How about Tuesday/Wednesday, the same time that we talked the other week? (3:30 CEST, 4:30 EEST)

jmp_music: by the way, I didn't send you a confirmation, but I also uploaded audio for the datasets to the same drive folder

pristine__: Hi. I was willing to jump into the Spark stuff for recommendations. Can you help get into it? I am quite much interested in recommendations :)

shivam-kapila: let's talk about it tomorrow?

Sure thing. Thanks.

shivam-kapila: when will your endsems take place?

i'm wondering if it's worth it to get a start on the gsoc project instead

iliekcomputers: No notification yet. The earlier notification said to open college on july 1 and then exams in July end. But I think they are gonna change things.

* july 1 -> June 1

pristine__: gah. *another* giant PR. :(

when you do cleanup AND code development in on PR it makes the PRs a nightmare to go through.

can you please please please not do this again?

> i'm wondering if it's worth it to get a start on the gsoc project instead.

I also had that thought. I had some points in mind. I will discuss it with the team and if its okay I will get it started.

we can get started anytime you want.

Hm. I guess then I should start from the next week itself.

I will see and plan the early step things this weekend.

alastairp: Cool! thanks! yeah, Wednesday at 3:30 CEST sounds good

👍

ruaok: haven't done cleanup. The scripts are inter connected, so the tests fail so I had to do it in one PR. Clean up is a result of PEP-8 . I have pushed a commit for every script to make it a lil easy to review.

[listenbrainz-server] mayhem closed pull request #846 (vansika/candidate-recordings…candidate-recordings-tar): generate candidate recordings for recommendation toolkit https://github.com/metabrainz/listenbrainz-serv...

then break that PR into at least 2 PRs, please.

with iliekcomputers' lead we're all working to submit smaller PRs, so please do that as well.

Okay

it might be best to do the PEP-8 cleanup first, then the other changes.

it will make the changes easier to read, since the code overall is easier to read due to the PEP-8 cleanup.

my general thinking these days is that if you're doing changes to existing code and the PR is more than 250 lines, it can probably be broken up.

Right.

iliekcomputers: We will be calculating all the stats at once right?

Like, all_time, last_year, last_month. last_week

ishaanshah[m]: make a seperate command for the month thing for now. I'll have to see how long it takes.

I have done that

I was thinking we should load all the listens and go on filtering

step by step

we'll have to experiment, doing that for month, week makes sense because they're in the same file

for the others, i'm not so sure.

Right now time taken to load a months data is almost instatanous

Hmm, I have created seperate functions for all right now

Each functions loads the data it needs and filters it out accordingly

I will let it be for now then

We can figure the optimisations later

sounds reasonable to me

Mr_Monkey: ping

prabal: pong

Hi

Do you know how to mock res and req ? I need them for writing tests

I found a `mock-req-res` package

Will try that

iliekcomputers: ishaanshah[m]: Hi, can I download artist stats?

alastairp: Yes you can

Just send a request to the api endpint

api.listenbrainz.org/1/stats/user/<user-name>;

is there a dump for all users?

no

not yet

ah. I have a great idea that I'd like to try out today

obviously the data is being copied from spark to listenbrainz though. I'll wait for iliekcomputers to jump in and tell me where it is, and I'll get my own copy

its store in a pg database

stored*

alastairp: Does SQL query work for you for the time being?

statistics.user

yes that ^^

alastairp: this dump should have the table ideally: http://ftp.musicbrainz.org/pub/musicbrainz/list...

it's a bit old, but new dump should be coming up next week, i fixed cron.

the table is statistics.user

iliekcomputers: perfect, that's plenty for now

[listenbrainz-server] vansika opened pull request #852 (vansika/candidate-recordings…candidate-set-for-all): Generate candidate sets for all users at once. https://github.com/metabrainz/listenbrainz-serv...

Any way to clear RMQ queue

I sent a wrong message and my RAM is full 😬

ishaanshah[m]: Try to bash into rabbitmq container

docker exec -it listenbrainz_rabbitmq_1 bash

then `rabbitmqctl purge_queue <queue_name>`

iliekcomputers: is that a postgres dump format. or tsv? I tried to just read it manually and parse the data, but json module in python fails to load the data column

postgres

Mr_Monkey: I found a way. mock-req-res wasn't required :)