#metabrainz

reosarevok: come over here and we can have a lan party when the gameLan is Game(lan)Over

:D

CatQuest: hey, all this while I realised that you have so many listens in LB

<3

:D

:)

alastairp: nice. I didn't think in that dir. You mean bad strings given to queries leading to crashes?

ruaok: Maybe later this afternoon

[listenbrainz-labs] vansika opened pull request #60 (master…mbid-msid-mapping): Use mbid->msid mapping to get mbid for every listen https://github.com/metabrainz/listenbrainz-labs...

ruaok: ^

wooo! exciting!

Mr_Monkey: http://odyssey.listenbrainz.org/odyssey/

ruaok: I will push changes file by file so that it is not difficult for the reviewers.

Right now create_dataframes has been altered

do you want new JSON base data sets?

I have them ready, I think.

I have the data in HDFS but you can put them on leader. I will host on williams soon

I'm more thinking about having an up to date PR that uses JSON going forward...

yvanzo: I found the issue with the Tags string; my local statistics_event table was incomplete so it wasn't finding the entry for DB:statistics.statistic_event/title:2007-10-17

and normally I revert any DB deletions from the commit since I know I'm missing them

so I ended up keeping the existing entry for DB + stats.js it removed, and also keeping the new one for just stats.js it added

my fault anyway, I'll fix

thank you!

pristine__: the .bz2 files in your dir have been updated to be the JSON version.

+s

hello

long time no see!

pristine__: not just crashes, there's a risk of malicious users being able to delete data or steal private data

SQL client libraries have been aware of this for 20 years, I have no idea why spark decided that it didn't apply to them

possibly because spark is not meant to be used in a live environment?

not that I am defending that argument...

alastairp: https://usercontent.irccloud-cdn.com/file/PbGiw...

the mood_happy is rather quite different for that same radiohead track. :(

the possible argument is that in java there's a defined way of doing sql query arguments which happens outside of the interface to the database

(use a query builder, get a string, pass that to the query method)

but I'm not convinced. no one on the internet sees this as a bad thing

ruaok: right, but this could just as easily be a bad duplicate submission

ah, right.

I'll filter that out anyway. I'll have to listen to the track to see which is a better fit...

we have enough data with this index to actually filter out bad submissions now

how do I go about choosing the right version of that duplicate pair?

simple answer for now: you don't

because we don't know which one is correct

so the filtering needs to be done on the AB level and not ANNOY?

annoy informs the filtration process?

if you have 3 or more submissions for an MBID

alastairp: oooo... I want to read more about it. Do share any resource if you have

you can see if they're all close to each other, or if there is a cluster all the same, with some others with a much larger distance away

pristine__: https://en.wikipedia.org/wiki/SQL_injection

ruaok: sounds good. I will host that on williams and open a PR so that users and get the dumps from FTP and upload to hdfs

pristine__: and of course: https://xkcd.com/327/

alastairp: phew..... Wikipedia is never a resource I look up to....

Isn't it boring kinda?

remember to add the needed license files and such. I should add that to the mapping process.

ruaok: Lord, I cannot stop laughing

Umm.... Did you see the test dump?

I included the copy license file there

The same, no?

I haven't. should I download it?

yea, should be.

You can... It is 1MB

The license in Listenbrainz full dump

Took it from there.

pristine__: the TIMESTAMP file has a second accurate timestamp -- is that what the LB files have?

the live data feed files from MB have subsecond accuracy and timezone.

2011-05-17 17:00:15.816041+00

vs

2019-10-15 10:37:48

Shit. Sorry :(

no need to be sorry. :)

😞

I like it when someone says it is a very good start

ruaok: alastairp <3

hi all

how long does it take the my listens page to update?

i fixed the json error i was facing yesterday

it was due to incorrect content type header

Matthew_: your workaround to the lack of support for pg_amqp looks good, but how are you doing without musicbrainz_collate extension?

pristine__: iliekcomputers ^^

and it works

the vlc plugin is now working. is anyone interested in testing it?

ruaok, rdswift , CatQuest, Pac23 , alastairp : ^

how do we test it? which platforms?

i think it should work on all desktop platforms

for now, i'll build and share the executables

once the plugin is tested properly, we can submit to vlc for inclusion in central repo and future builds

great! do you have a mac build yet?

at the moment i have linux builds

i need to install the windows and mac toolchains to build for those os

once i do that i'll let you know

hm. I have sadly an old mac though

mac os 10.7.11

but I also have windows 7 so i guess i can test it there too

🤔 did i uninstall vlc on android? it was kidna dumb so i think i did , but i can see about re-installing it and trying it there

i am not sure about vlc-android CatQuest

thats why i wrote desktop platforms above

hm

as far as i know vlc android is the only exception

yea i think that makes sence, vlc android is kidna crap :D

is it me or is something odd bout the formatting on this page https://beta.musicbrainz.org/release/35372aaa-0...

cd track lsit ends oddly

Seems normal to me?

hmm

should there be a singleborder at the end?

That's how it is elsewhere too :)

(just that it might look different for one-track releases because you're used to a lot more rows there?)

hm

several row other eleases end with a last border i thoguh..

let me loo at osme releases

they do you know https://beta.musicbrainz.org/release/b6a8463f-2...

But this has a border too?

not at the bottom!

not for me atleast

hm

... it does for me?

huh i refreshed and it's still not there

weird

CatQuest: screenshot maybe?

hmm yea maybe. after dinner and stuff :D

Fwd jonatoni@telegram: Chrome extension using MB API coded at a hacking event: https://twitter.com/codekiwicom/status/11848012...

<3