#metabrainz

I've been using this tool called `act` to test workflows locally, quite helpful.

"anythong" - typo of the day?

monkey: lgtm. added a couple of comments where simplifications are probably possible.

Thanks !

[bookbrainz-site] 14MonkeyDo merged pull request #723 (03master…github-workflow-CI): Use Github actions to run CI tests https://github.com/bookbrainz/bookbrainz-site/p...

Thanks !

(again)

[bookbrainz-site] 14dependabot[bot] opened pull request #724 (03master…dependabot/npm_and_yarn/tar-4.4.19): chore(deps): bump tar from 4.4.13 to 4.4.19 https://github.com/bookbrainz/bookbrainz-site/p...

[bookbrainz-site] 14dependabot[bot] opened pull request #725 (03master…dependabot/npm_and_yarn/path-parse-1.0.7): chore(deps): bump path-parse from 1.0.6 to 1.0.7 https://github.com/bookbrainz/bookbrainz-site/p...

no worries :)

ruaok: /home/lucifer/bpm_and_key.csv on michael

thanks!

<ruaok> a lot of ram for stuff that many people won't look at daily. ==> Perhaps allow users to "subscribe" to stats, and then let them know that they will be available starting the next day?

Then if they don't look at them for a length of time (e.g. 6 months), drop their subscription?

an interesting thought for when stats get overwhelming for us, rdswift. right now we just have some optimizations to do... I think it would be a good idea to have stats for everyone as we're trying to get more users to come in.

Perhaps automatically subscribe everybody (including new users) for now, and then change that later when things get overwhelming?

another good thought. I need to ponder all this, because a lot of it doesn't scale well yet

they do -- kinda fodder for my brainstorming about scalability

rough cut of the histograms. the top one is ruaok's and the bottom is monkey's https://usercontent.irccloud-cdn.com/file/ozGMU...

I think that is very cool!

!m lucifer

You're doing good work, lucifer!

:)

thoughts on playing more with this. any variations you think could be interesting to try?

not really. I already felt like we were scraping the barrel. let's be done with it!

👍 sounds good.

huh, we use https://nivo.rocks/components for all other stat graphs and there isn't a histogram here. monkey when you have some time, can you find a suitable histogram component for us to use for bpm histogram ?

Would a bar chart suffice here?

yeah that can probably work too. would be nice if we could eliminate margins between the bars.

ohh i just found a really nice graphic to use for LB's sitewide review. we could show how listens grew each year with this https://nivo.rocks/funnel/

hmm. I'd love to see that with real data.

lucifer: Apparently you should be able to set `padding` to 0 : https://usercontent.irccloud-cdn.com/file/PQnQp...

monkey: fantastic. that should work great then! :D

ruaok: i'll take that was a nerdsnipe ;)

On another bright note, I finally got the Github CI workflow setup working after ironing out all the kinks

lucifer: heh.

I'm certainly susceptible to nerd sniping

is this bpm? https://usercontent.irccloud-cdn.com/file/ozGMU...

yes.

and what's.. y axel?

rose.

its axel rose.

:D

ok so i dpon't know he word in englis and guessed

its listen_cou nt

-_

what is the y thing

axis

axis!

great

what's the y axis measured in? miliseconds?

listen_count

the number of times someone listened to a track with that range of BPM

ah

i'd be interested in seeing mine :o

huh, I wonder why this choses specifically the jp edition https://beta.listenbrainz.org/user/CatCat/?min_...

https://beta.musicbrainz.org/release-group/90ba... it' chooses 2'nd one here

Could be it chooses the first release (by date) of format "CD" rather than any format? ruaok ?

hm

Don't quote me on this, I have no idea.

it makes sense tho

yes, CDs are preferred.

anyway I should probably retag those

Hey, accidentally correct.

eventually I'll be able to tell lb the right one, anyway right

also an interesting thing, that "sanctuary" seems to be taken from a differet edition the the others on that page

CatQuest: You already can, provided you send the release_mbid (if you have that available) in your listen

yea, but I can't fix already sent files, right

erh listens*

And later on, we will have a mechanism to correct mappings that are wrong, but as to determining what is the "right" one…

yea

Tricky.

I can say " all these sets of whitesnake songs with these names are release x"

Which release would you choose, in this particular example?

so for every set of those titles over and over in my listen history,it's *that* album

hm

( <3 Mohama Saz and Heavy Moon)

:D

btw is it possible to include mapping based on song duration, for situations like this

possibly https://beta.musicbrainz.org/release/32cac43a-c... (or wichever was in my "owns" collection)

like suppose, 1 song has multiple versions, they might have different song durations

That's not particularly reliable if I understand correctly

if we could tell lb that so and such releases in collection y on mb is "owned" and to fetch assumed releases for mthat...

CatQuest: Right. And what are you listening to it on?

indeed. for example duration cna be very similar even for very differnt version

riksucks: durations don't work well, because people skip tracks very often. or pause. just not really doable.

monkey: what do you mean?

fair point

Listening to spotify or the like, or the actual CD ?

i ripped the cd (at some point, it's 125 bitrate so i shoul re-reipp it again some how)

and lsitening ot it in itunes

OK, that answers it.

I have to use a different computer to tag files so i'm lazy and it won't get an album tag for quite a while :D

I guess any correction you provide for the mapping of your listens should only apply to you.

For example if I listened to that album it would probably be from Spotify, and I would choose another release (a Digital Media one I guess).

yea

Yeah, not super tricky, in the end I guess. But that requirement to have mapping corrections per user is something to keep in mind

evne g othe other way around if a release is found listeend to but isn't in the user's "owned" collection, suggest adding it :D

hmm interesting.

lucifer: is this?

mine?

CatQuest: yearly listen counts. total number of listens LB has per year.

oh

no this one is not individual, this is sitewide.

yea got that with your latest

woah, I gotta put down that code. I'm seeing double.

double listen counts?

no the graphic you posted. that is what double vision looks like

presumably https://usercontent.irccloud-cdn.com/file/JwILt... and it's lightly coloured edges

oh that, i think its intended to look that way

oh btw, did you figure out anything re timestamps and not showing my profile?

i imported jsut now and there was like 48 pages again but no update on the profile :/

oh! i started looking into the other day you mentioned it by importing your scrobbles on a different user but afterwards something else came up and forgot about it then. i'll look into it again tomorrow.

cheers!

monkey: ^^

lucifer: the last column is BPM and troi can now sort a playlist by slow-fast-slow BPM.

oh, that needs more tuning, not enough tracks present.

awesome!

!m ruaok

You're doing good work, ruaok!

hopefully that will reduce the rollercoaster effect.

thanks for the data!

oh dear. oh dear.

lucifer: I found some problems. 🤣

See "Cantelowes" track near the middle. and the two following it.

yeah, really wrong. what data to you have from AB on this?

the data in AB matches what I got from you. which blows.