#metabrainz

Morning!

Morning!!

Morning!!

moin moin!

hi ferbncode!

yvanzo: bitmap: security notifications on jenkins. I'll release a new version today

hi alastairp o/

how are you? :)

very good, surviving from day to day

how about you?

good too :)

Hi Suyash! Long time no see !

Hi Mr_Monkey o/, yes 😅 how are you? hope you're well and staying safe <3

Doing all right! Safe and sound at least.

alastairp: Hi! I saw your comment on my PR. i wanted to ask which of the above approach should I follow. https://www.irccloud.com/pastebin/KMPR4goL/

abhinavohri: hi! this is a long-standing question in python, and I'm sure that everyone has their own opinions

personally, I think it's best to use the second approach, especially if the file uses multiple methods/types from a single module

ok....Thanks

e.g. imagine if we use time.monotonic, and time.time, and time.strptime. personally I think that it's cleaner to just have a single `import time`

there are some cases where the first approach is OK, for example if you _only_ use monotonic, and it's clear where that import comes from, but I don't often use it

I will make the changes in the PR

ok

often when I have these kinds of questions I look at the google python style guide: https://google.github.io/styleguide/pyguide.htm...

Use import statements for packages and modules only, not for individual classes or functions. The namespace management convention is simple. The source of each identifier is indicated in a consistent way; x.Obj says that object Obj is defined in module x.

OK ...I would keep that in mind

ferbncode: we had a small discussion about CB yesterday: https://chatlogs.metabrainz.org/brainzbot/metab...

you might have some feedback

pristine___: see if my comment on the PR makes sense

alastairp: got it, thanks!, I'll read through

ruaok: Hi o/, how are you? how is it there in Hamburg? :)

ruaok: Hi, is bono.metabrainz.org:8000 down?

sunny and nice. with lots of nautical shit. :) doing well, you?

ishaanshah: something weird is currently happening. hang on.

zas: you about?

even LB looks down

sorry its working

everything appears down.

ruaok: yeah cool. Let's merge this PR, I will open a seperate PR for scaling and will update the comment the, cool?

ruaok: not as sunny as usual here, but doing good :)

alastairp: > I'm going to start by putting out a call for volunteers on the forums, and we'll see what comes out of that and go from there

I agree, I haven't been able to much time in CB the last year, and I am sure we can benefit from the community for volunteers and contributions

zas: https://status.gandi.net/

pristine___: 👍

ishaanshah: I suspect that the service on bono is up, but we can't resolve the IP addresses to the servers right now.

ruaok: its back up again

as u said dns issues

yeah, I suspect it never went down.

Hi, may I how many approx editors accounts does MB have rn?

Know*

2,094,046

see section editors for many more details.

Thanks!

ruaok, iliekcomputers ^

iliekcomputers: The PR is ready for review

ruaok: is this error familiar to you?

this one too

I wonder if it happens when two CI instances are running simultaneously

ishaanshah: ruaok: https://dougturnbull.org/index.php/2019/08/28/i...

next step: geographically-constrained recommendations

(or geographically-diverse if that's your jam)

Yep, should be possible to do with troi. Especially with the new country endpoint.

Perhaps we need to tune that a bit for better effectiveness.

ishaanshah: I've seen it before, but have no idea what it might be

what a great motivation for having a platform-agnostic recommendation system

Exactly!

jenkins going down

up. thank you for flying with BrainzAirlines

@alastairp: 🤣

reosarevok: hi, are you here today?

I am at the moment at least :)

I'm looking at MBS-10388

MBS-10388: Add some genre tags to MBS https://tickets.metabrainz.org/browse/MBS-10388

_lucifer hey lucifer! what about the app launch?

just looking at some methods of matching genres to mb genres

alastairp: oh, neat

this is just a basic Levenshtein distance

note that genres from our datasets are sometimes 'genre/subgenre'

mhm

I think throwing this together first, then asking someone to hand check it makes sense

in this case, the match rate (80-100) is matched just on the subgenre part

*70-100

Seems that anything under 90 seems to generally be wrong

There are some other matching heuristics that I'm going to try, e.g. some that insist that all words must be there, will fix issues such as alternative punk -> alternative pop

right, most of the time it seems we should just go with an exact match

although bebop/bop was a nice catch

Exact match automatic, show stuff over maybe 85 to a human to confirm

white_shadow: a bit behind schedule. some ui bugs are to be fixed. as soon as that is done will make th release

ok np

sounds good to me

Are we writing test cases in the app?

194/347 had an exact match anyway, which is pretty nice

what do you think about the genre/subgenre situation?

a single tag to represent the whole thing, or one for genre, one for subgenre? (since MB has many tags that we consider 'subgenres'

Hmm