#metabrainz

Lotheric: that's the spotify player being noisy, are you opted into the Spotify integration? If not, it should not be there.

I linked my account, yeha

ok got it

unlinked myself and relinked using both play and record

the message is gone

:)

thanks

so, is it a standard/convention that tags begin with a lower letter?

and if yes, is it possible to change the name of a tag without changing the value across a selection?

im adding "book" and "series" tags to some mp3s

in the tag list panel picard lists Artist as being capitalized, but then the tag itself is lower case. However if i create a custom tag it appears in the list panel verbatim.

which is what confused me

i figured it out

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

[listenbrainz-server] ishaanshah opened pull request #856 (master…time_range_spark): LB-575: Add support for more time ranges for "Top Artists" (Spark and DB Insert) https://github.com/metabrainz/listenbrainz-serv...

LB-575: Add support for more time ranges for "Top Artists" https://tickets.metabrainz.org/browse/LB-575

iliekcomputers: can have a look once before I update the tests?

[listenbrainz-server] vansika opened pull request #857 (vansika/candidate-recordings…dataframe-doc): update documentation in create_dataframes https://github.com/metabrainz/listenbrainz-serv...

[listenbrainz-server] vansika opened pull request #858 (master…update-recommendation-schema): update recommendation schema to use recording mbid https://github.com/metabrainz/listenbrainz-serv...

Mo''in'!

ruaok: morning!

I have two small pull requests on LB that I want to merge today. Will you have time to take a look?

ishaanshah[m]: hey! I'll look at your PR soon-ish.

iliekcomputers: can do. More urgent then then next couple of hours?

Nah, not that urgent. Just want to deploy them to cron today.

K

Thanks!

done!

[listenbrainz-server] mayhem merged pull request #853 (master…artist-relations): Add artist/artist_credit relations https://github.com/metabrainz/listenbrainz-serv...

[listenbrainz-server] paramsingh merged pull request #854 (master…param/better-notification-emails-for-dumps): Improve dump creation notification email subject https://github.com/metabrainz/listenbrainz-serv...

[listenbrainz-server] paramsingh merged pull request #855 (master…param/delete-old-dumps-from-temp-dir): LB-588: Remove old dumps from the /mnt/dumps as well https://github.com/metabrainz/listenbrainz-serv...

LB-588: Remove old dumps from /mnt/dumps/tmp/archives https://tickets.metabrainz.org/browse/LB-588

Thanks ruaok

Morning!!

Where did the week go?

ishaanshah[m]: i've commented on your PR. I think it'd be a good idea for us to catch up once when you get the time.

Mr_Monkey: my week exactly!

iliekcomputers hey, I am around

ishaanshah[m]: hey

hi

did you see my comments on the PR?

Yes

I agree with all of them, I will make the changes tomorrow

Are the listens in parquet files sorted in ascending order?

i don't think so.

wrt to listened at

ruaok: does meb have an organization google cloud account?

you'll have to do a max operation first, i guess.

Oh, can we do that

_lucifer: yes

ishaanshah[m]: yeah, i'd just load the latest month parquet file and find the max for listened_at

Because where queries will simple because of that

this is scala, but there should be an equivalent.

I was asking for the where queries, min max should be fine

i don't understand the question.

Just a optimisation I though, not necessary to be done right now

I mean if we could tell spark that listens are in ascending order the where queries would just binary search ig

oh

yeah.

the sort would be really expensive though

Yeah, Ig we will have to modify the export to do that

hmm

the dump process is almost insanely inefficient rn

It's an optimisation, maybe we should look at it after we setup a proper pipeline

yeah, true.

I will open a ticket for it

cool, thx!

Other than that you had some things in mind to discuss?

nah, just wanted to make sure we're aligned on the stats calculation comments I had.

Oh, I agree with the seperation part

for the interval of calculation specifically.

Its better to have different commands, because we will have a lot more stats now

we should find the sunday before the latest listen and then calculate stats for that interval

or the 1st of the month before the latest listen and then stats for that interval

But wouldn't that make the graphs static for the whole month

sorry, i'm not clear i guess.

suppose the latest listen is from May 16

then the calculation should be from May 1 to May 16

That wouldn't exactly be last month right

yeah.

I was thinking if 16th may is the latest listen

Last week = 9th-16

Month = 16th april - 16th may

hmm.

Thats what we are doing right now

I mean the PR

i think the stats would be more useful if they're for a particular month vs a more-or-less random range.

like stats for the month of May vs stats for half of April + half of May

i'd say go with what I said for now, we'll change it if people want something different.

should be easy enough to change.

Yep, Yep

What about week and year

same concept for week to week, like Sunday to Sunday, we could show it on the site as stats for the week of May 11.

should week be sunday to sunday too?

and for year to year, stats for 2020

vs stats for may 2019 to may 2020

Hmm

That eliminates the need of where queries :D

So we speed up the pipeline 🎉

Another thing, I think we should look into more columns in db after some time

When we know more about the stats we are gonna calculate

yeah, seeing the changes in the sql in the PR, i think more columns will be needed.

Yeah, the query has become really complex and hard to understand

true

Also does PR #853 mean we will have a MB database on LB now?

we have a connection in prod.

on the dev env we don't

So we can make direct db calls to MB

in LB we can, in spark rn, we can't

Oh, noice

I was gonna need that for some of the stats :D

Anyways, I will make the changes you asked and update the tests :)

awesome, thanks!

reosarevok: thanks, published!

ruaok: can you set up a google cloud project for me with editor access

i want to connect with firebase test lab

it requires the project billing to be enabled but i do not intend to exceed the free quota