<Shreyas[m]> "Hi jasje akshaaatt,..." <- Hi Shreyas. Yes, that is what the expected behavior should be! If you want to start working on it, it would be really great. Feel free to ask for help if you feel stuck somewhere :)
s/./:/
nawcom has quit
nawcom joined the channel
minimal has quit
Shreyas[m]
<jasje[m]> "Hi Shreyas. Yes, that is what..." <- Great!, I'll start working on it and let you know if I need any help. Thanks :)
lucifer[m] has quit
tuxpaint
mayhem[m]: by similarity, you mean the similar-recordings labs api?
my naive thought would be to do some sort of recursive similar-recordings, with user configurable depth and breadth for instance, sample N similar recordings for the original track, and recursively generate N similar recordings for those results, performing some random sample over the previous results. this could maybe get you different results each time, since it's an exponential amount of possible similar-tracks you can get with N generations, although i
suspect overlapping similar-recordings would probably decrease your growth rate until your origin-set gained enough like, entropy or whatever.
this is sort of what my app did with the spotify api. it would ask for recs for some song, and then ask for recs for random entries of the currently playing playlist, removing entries as you played them. to be honest, maybe i will just try just using the existing similar-recordings api, and run some statistics to see how the numbers actually work for some different tracks....
it worked pretty well for a few different sorts of tracks between me and a few friends using the spotify api, but no clue how that will compare to the similar recordings api
anyways, im off to bed. i have some meetings tomorrow but maybe can find some time to chat/overlap tomorrow, or can figure a way to brainstorm/communicate async (maybe just on the issue ?)
aerozol[m]
monkey: How’s the YIM coming along? FYI I’ve set aside all of tomorrow to finish off the picture elements. I don’t know if you’re still tweaking stuff but there may be items of interest in LB-1400 btw
mayhem, lucifer, ansh: Hello! Two questions: do you have any PRs to merge? I will publish and deploy a new LB version today.
Second question: Where are we for YIM? We had talked about doing a dry run at this date. I'm working with aerozol on the design, wanted to see where the data is at, and give everyone an alert if work hasn't progressed on it yet.
monkey: I have no other PRs to merge. On YIM, strangely I find myself without tasks, after I fixed the color release table. Happy to help if it is needed.
monkey[m]: perhaps ask aerozol that?
monkey[m]
You were going to generate cover art images from 2024 albums cover art, no?
I believe aerozol was waiting on us to deploy. He wrote a blog post, might need a good reviewe
mayhem[m]
yeah, but if lucifer is going to put that into an endpoint, its kinda moot for me to generate images.
(that we can't share, not at that size)
but, I can still make that happen if we have a reason to do it.
monkey[m]
OK, let's see what lucifer says, if he needs a hand for timing reasons it might be good to help him out
mayhem[m]
I'll surely do that.
ansh[m] joined the channel
ansh[m]
monkey: I don't have any PRs to merge. Same on YIM, I don't have much idea on my tasks. Is there any piece you'd like me to work on?
monkey[m]
ansh: If memory serves, during our team chat we talked about you making a graph to show more genres thn what we did last year (just top 3 in one of the shareable images)
ansh: Although saying that, we're getting awfully close to the date so there might not be enough time for you to work on that. Let me know how long you think that would take, based on the YIM'23 genre data. A basic graph would still be good enough I reckon, as long as it shows more stuff than last year.
For reference, last year's genre work was just these 4 top genres (which excluded topmost genres like 'rock' and 'pop', we would want to show them in this new graph):
monkey[m] uploaded an image: (265KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/chatbrainz.org/oDxdnKbhqgyoGeWttlWmmhhI/image.png >
Let me know if you want to do a quick call to discuss our options
lucifer[m] joined the channel
lucifer[m]
<monkey[m]> "mayhem, lucifer, ansh: Hello..." <- > <@monkey:chatbrainz.org> mayhem, lucifer, ansh: Hello! Two questions: do you have any PRs to merge? I will publish and deploy a new LB version today.
> Second question: Where are we for YIM? We had talked about doing a dry run at this date. I'm working with aerozol on the design, wanted to see where the data is at, and give everyone an alert if work hasn't progressed on it yet.
I can do a dry run today after meeting and stats should be ready by tomorrow.
<mayhem[m]> "yeah, but if lucifer is going to..." <- yup, i can add at an endpoint for it. can you remind me how we wanted to do the images? organize the top releases of this year in some color order or over a background image or something else?
elkiensad[m] joined the channel
elkiensad[m]
<monkey[m]> "For reference, last year's genre..." <- also people was disappointed on Wrapped because of no genres this year, just an extra information
s/was/were/
lucifer[m]
reosarevok, bitmap : when an account is deleted in MB, the editor is marked as deleted. what happens to the email column, is it cleared or left unchanged?
yvanzo[m] joined the channel
yvanzo[m]
lucifer: it is cleared
lucifer[m]
yvanzo: great, thanks
yvanzo[m]
lucifer: What did you try next with Solr 9.7.0/SEARCH-733?
<lucifer[m]> "yup, i can add at an endpoint..." <- are you sure you have enough time to do this? I don't have anything on my plate and can knock this one out fairly easily.
that is probably more important to fix for YIM, that the cover art endpoint.
so, I'm happy to the cover art stuff.
<tuxpaint> "anyways, im off to bed. i have..." <- Interesting ideas -- we'll have to see the data once we generate the large dataset. The set you linked to contains less than 1B datapoints. We have a 26B datapoint dataset we want to process with data back to 2004. That will drastically improve things.
and yes, lets pick a time when we're both around so we can chat.
MyNetAz joined the channel
minimal joined the channel
salaxceitor[m] has quit
lucifer[m]
<mayhem[m]> "are you sure you have enough..." <- sure if you'd like to, feel free
mayhem[m]
yeah, lets do that. please take a look at the stats bug from above -- that needs addressing before yim
lucifer[m]
<yvanzo[m]> "lucifer: What did you try next..." <- yvanzo: i got busy with a couple of other LB important issues last week. I'll get back to it this week.
mayhem: we cap plays per user to 500 for sitewide stats but not in popularity stats, could be possilbly because of that.
mayhem[m]
oh, interesting. could you verify that that is the case?
lucifer[m]
sure
cluster is busy right now so will have to wait a few hours though
yvanzo: mayhem bitmap reosarevok is it possible to setup a meet this week to discuss the second phase of oauth migration?
mayhem[m]
sure, I have a normal work week.
reosarevok[m]
I kinda-sorta do too except Friday
(I'm one hour later again but shouldn't be a big problem)
yvanzo[m]
Same, except Tuesday afternoon as usual
lucifer[m]
how about tomorrow, one hour before regular meeting time?
aerozol: Good morning! I deployed the flairs and published the blog post, would very much appreciate if you could post on social media ❤️
mayhem[m]
sad. I have no flairs. I must donate.
tuxpaint
mayhem[m]: i'll be available from 1PM-5PM CST (in two hours) today if any time in that range works for you. wow, more than 25x the datapoints sounds very cool and exciting
