#metabrainz

Freso/legoktm: Note that Discourse also does something when reading, or when having pages open for a while. So not posting alone is not enough.

alastairp: https://community.metabrainz.org/t/acousticbrai...

alastairp: I've posted the pre-proposal to the forum, please take a look and comment

cetko: perfect, thanks

alastairp: that memory leak PR is interesting.

I wonder how that manifests itself.

ruaok: yeah, really odd that no browser seems to solve this

it was good sleuthing, or just good luck that we found someone who understands how js works :-P

now we need to get the rate-limiting patch finally integrated and debugged.

that stupid PR has been blocking things for so long. :(

we should go have lunch sometime this week. you in at MOB at all this week?

armalcolite: hi. did you solve your problems with listenstore and cassandra version?

how did you install the software? We have a vagrant system with fixed dependencies for a reason. If you can't use the vagrant image you will need to make sure you install the exact software versions that we support

if you're installing a newer version and you have problems, we would accept any patches that you can submit to make the new version work

ruaok: yeah, I think I'll be around

any idea which day?

tomorrow is no good.

I've normally been doing thursday afternoon. I think that's still good for this week

armalcolite: you asked about LB-84, I think it still has problems. I think we would have closed the ticket if it was fixed.

You can look at the source code for this part of the server and see if we have a check for it

ruaok: oh, no. thursday is no good

perhaps wednesday afternoon then

ok.

alastairp: Just had a look at the code review. Making the suggested changes

cool. as I said, the patch is great, just a few style things to make it fit with the rest of the code

yeah

alastairp: yeah. my cassandra and listenstore are working fine.

alastairp: also, regarding the PR, we still dont get the spotify details, so i wasn't sure what to do with the code. My guess was to delete them, but i commented them for review. will work on the review.

alastairp: Also i was working on the ticket, LB-13 where people could export the personal listen history.

armalcolite: OK. You should assign the ticket to you, so others know that you are working on it

alastairp: I have made the necessary changes in the user.py and listenstore.py to gain. I have to now add a button on the frontend

alastairp: any specific suggestion?

also, it would be a good idea if you started a discussion on the tickets that you are working on, so that we can discuss any changes that are needed

alastairp: i cant comment on the tickets !!

it's better to talk about what we want to do before you make the changes, because if we have different ideas then you would have to re-do it

right. that's unusual. perhaps you can talk to Freso when he is up, to see if he can work out why your account can't make comments

alastairp: ok. will sort this out.

and do we have to export json from cassandra database?

bcoz currently i m doing it.

what do you mean?

for exporting the listens, we can just dump all the listens of user in json format.

yes, that sounds like a good idea

so, i have made changes in the fetch_listens() to achieve json output as well.

or something more compact like tsv

Gentlecat: hmm. I don't think I'm too worried either way

compress it and it'll be small

what to export?

last.fm archive is a zip with a bunch of stuff in it

just track_name, artist and listened at?

or everything, like mdib's

everything

i think, json would be good.

we can the import it in future as well.

yes. we can start with json, and if we want to change it in the future we can change

yes! I would like some more import functions too

for me the most important is to be able to import from a lastfm archive file.

but it's a good idea to be able to import a listenbrainz archive too

this should be a separate ticket

oh. i searched but now they have remobed the feature.

yes, when they released the new website it went away

I hope it will return

can u share the archive?

but we should be able to import archives from the old website

yes, I will when I get home

but we should finish these tickets first. You should not work on too many at the same time, so that other people have the chance to work on them too

sure. will complete those first

Freso: can you check my access to tickets.musicbrainz.org, i cant comment on any issue.

you should also tell him your username, so that he doesn't have to ask again

give as much information as you can

Freso: username is the same, "armalcolite", email-pinkeshbadjatiya@gmail.com, group: jira-users

:)

armalcolite - it is a rock which was first discovered on moon. (Shows my love with astronomy.)

alastairp: that evaluation is still running :(

hi all I just found out about the brainz community and different brainz projects. anyone know what the current status of listenbrainz is?

alastair@spike:/tmp/tmpVmS77z/results$ ls -l|wc -l

1435

for each combination -> params/results file, 720 combinations = 1440 total files

almost finished!

don't know if you did, but might be worth noting to others to not make them very large

I said it in the document: http://www.dtic.upf.edu/~aporter/amplab/

"The training process takes 2 hours per 1,000 instances in your dataset. If other students are doing the same task you will not have time in the last day before the task is due. Start early!"

petepy: it had been stalled for a bit, but with summer of code and some other organizational guff out of the way, more attention is being paid to it again.

though that seems to be a lie

it's weird - sergio did some examples witha few thousand items and it only took about 4 or 5 hours

perhaps he had less than you

you used 2k right?

maybe depends on the load