#metabrainz

As consumers are also generic, the present consumer design, validation and db interaction can be used for all the future data sources :)

Apart from this, I designed the basic `queue object`, which is the format the producers of any data source must produce

I also built LB and poked around.

I'd like to work with d3 in future there!

CatQuest: Ask after meeting.

That's for me. Leo__Verto ?

hi Leo__Verto !

Hi!

Spent thursday to sunday at a festival and the rest of the week messing around with pandas and plotly to get my spam data visualizations to a point where I'm happy with them

sorry internet went off

as promised, here's a graph on bio lengths from last wednesday https://github.com/metabrainz/spambrainz/blob/m...

:O

interestingly enough there seems to be almost a gaussian distribution around bio lengths of 3.5k with some outliers to the shorter side

more stuff to come soon!

this is very interesting

dragonzeron, go!

Hi

Leo__Verto: Interesting!

Leo__Verto: Did you compare with length of non-spam editors?

So I have been working on some japanese artist by adding works to the recordings and then I am also merging recordings that I didnt on artist I cleaned up on

Freso, I'm working on how to do that without having access to the prod DB myself but that's the end goal!

I was also wanting to know if someone can possibly look into the Amazon glitch where the image of the album does not show up

I put a request into bugs on metabrainz

dragonzeron: that's mostly solved with "upload ito caa" thoguh

true but even then it kinda becomes a hassle

when I add multiple albums at the time and the amazon thing kinda helps shorten the space and what not

such issues withamazon was one of the reasons caa came ot be

I think it's more likely we'll stop using Amazon for showing covers entirely.

ok

yep

understandable

it's a nice backup so i wouldn't remove it asap. but

I am also planning on putting in a couple of request about updating the famicon of some of the other databases we added so that it would show up on the artist page

fixes on the bugs n it will probably not be prioritised at all

Anyway, reviews is not for extended discussion. :)

yes

if that is what I am supposed to say either way Arigatou Gozaimas

who is left

No one, you were the last. :)

ok

And no one else has spoken up.

So thanks for your reviews, dragonzeron and everyone!

And thanks for your time!

</BANG>

thank you for having me

thank you freso!

Leo__Verto, the missing part of the #musicbrainz log is at https://pastebin.com/GSC55YC5 and the missing part of the #metabrainz log is at https://pastebin.com/UMNV2PUk (times are UCT-6). If you can point me to the desired format, I'll slam together a conversion script for future.

I also forgot to mention I celebrate my 1 year anniversery of working on here today

didn't you mention it earlier?

aw man

dragonzeron: Congrats (again). :)

kartikeyaSh: iliekcomputers tests now pass -- thanks. I'm ok with PR #36.

i really don't want amazon to go away that soon

iliekcomputers : merge when you're happy, please.

rdswift, thank you, I think I still have a modified import script from last time I got logs from you!

Freso: what is an "exit" url?

finally :)

Glad to help.

Leo__Verto: rarg you've been accepting logfro mthers thna me?

/jk

:P

[messybrainz-server] paramsingh closed pull request #40: Don't create database in test.sh (master...tests-fix) https://git.io/vhVXy

Leo__Verto: I just looked at my local db mirror, and it seems that indeed bio is one of the things being sanitised in the dumps. Ah well. :/

well since it would maybe help leo's spambrainx thing, should we prepair special dp dump for him?

[messybrainz-server] paramsingh closed pull request #36: LB-352: Add script to fetch artist MBIDs provided recording MBIDs from MusicBrainz. (master...mba) https://git.io/vproH

db*

Looks like we don't have languages either, so no analysing how much editors stick to languages they know or not. :/

Ah well.

param is on a roll! ᕕ(ᐛ)ᕗ

yvanzo, at this point I'd love to hand you a jupyter notebook to run against the main db but I'm not sure if that'd be GDPR compliant :/

Leo__Verto, not that there's a line or two overlap with the existing logs at either end.

*note*

CatQuest: I'm guessing and Leo__Verto can prepare a query and yvanzo can provide the results of it.

rdswift: doing the good !

rdswift, yeah that's okay, the importer should be able to detect duplicates

oohhh

That way Leo__Verto can get the data in an anonymised form. But I reckon Leo__Verto and yvanzo will figure something out. They're smart peeps. :)

CatQuest: Aaanyway.

the few times I did logs i als odid that, to help find the right spot

"Exit URLs" is a term I kind-of coined (together with yvanzo) for describing URLs that are meant to just redirect you elsewhere.

Freso hi :)

Let me find an example…

Oh yeah, I also removed the [off] lines from the logs posted.

soo. "redirect" urls?

!m rdswift

You're doing good work, rdswift!

Leo__Verto: correct guess ;)

oh I didn't do that, i expected that importer would do that auto

well done rdswift

Freso, I remeber you mentioning that to me last time.

:)

CatQuest: E.g., https://musicbrainz.org/edit/53551784

so that's why "exit" and ont "redirect"

(`exit.sc/?url=` is SoundCloud's "exit URL".)

wait mediafire is still la thing?

ooh "leaving this domain" tyoe urls

yea tohse are hella annoying

!m Freso(Bot) for removing those!

You're doing good work, Freso(Bot) for removing those!!

..don't know how i feel abotu direct download linsk in general though.. I liek better links to pages wher yo ucna click a dl link (not that that's you doing but the original person mind)

Basically, they're addresses that websites/services use between themselves as the end-point, either to keep better stats for where people leave to, for helping to prevent fraud and/or try to make people not leave their site (e.g., Facebook warns people that they're leaving Facebook using `facebook.com/l.php`), or to obscure the HTTP referrer to make it harder for the target site to tell where the visitor comes from (e.g., I think maybe DDG

uses/can use something like this, to prevent sites being able to tell what search words were used on DDG to find the target site).

try to make people not leave their site (e.g., Facebook warns people that they're leaving Facebook

yeai know that ype of url, I find thme seriously nnoying so I am very happy about your bot removing those :D

CatQuest: Like I wrote in the forum, many of the exit URLs will need additional cleaning of various sorts, this is just a "first pass" of cleaning them.

👍 yurp

There's definitely no reason for them to be in MB, but they sometimes sneak in if you for example right click links and "copy URL to clipboard".

exactly

oh god. googledocs does this

i HATE it

it makes copy pasting urls into mb from google doc liner ontes (ype that's a thing) really annoying

CatQuest: https://musicbrainz.org/search/edits?auto_edit_... :)

thnkas. totes voting yes on all the ones that are not wrong :V

(All the "exit URLs" FresoBot has handled so far.)

CatQuest: https://community.metabrainz.org/t/musicbrainz-... :)

lmao, the first one is youtube redirecting to failbook and isntagram, and then there is a instagram redirecting to youtube

>_< XD

!m Freso(Bot) for removing those!

You're doing good work, Freso(Bot) for removing those!!

again, becasue fantastic!

would also be great if we could incorporate the most common ones in our url cleanupscript!

Yeah. I think the issue with that would be that they need two passes, but I haven't really looked at URLCleanup.js much since yvanzo's overhaul of it, so maybe it is (more) feasible now.

(I've also been talking with yvanzo about making FresoBot use URLCleanup.js directly to clean up URLs and fix up their relationships too.)

Freso: It has to be done ahead of URLCleanup.js not inside.

yvanzo: So that part hasn't changed. :)

ruaok:

not a lot of context to evaluate what is doing on...

ah, replayed live logs

I get that.

how many requests? what was the server load?