#musicbrainz-devel

so I think I'm ok to continue on my own, thanks

cool, good luck :)

yeah, I suspect all of what I've done will be easier than this last part ;)

I've found a demo of something that should be what I want, just need to implement it

yuck. https://github.com/jreese/multiprocessing-keybo... uses message passing to send a "I was cancelled" message

ruaok: When did you want 100k by?

there wasn't a milesone

Aw.

"Soon"

hey all, someone here able to help me with running a server ?

I get some errors thrown at me and can't figure out the cause

when I set sub DEVELOPMENT_SERVER { 0 }

gizah: try running `gulp` in /usr/local/musicbrainzz

command not found ;)

or ./node_modules/.bin/gulp

this is assuming you've followed run npm install first

it seems I have no gulp

I followed the install.md guide

though it is a bitch as it seems quite incompatible with freebsd

I've come far enough to be able to run with development 1 but without it it still dies ;)

ah, for production setups you'll need node.js, and then need to run npm install to get gulp

I did install phantomjs

is that an equivalent ? :P

nope

installing node.js now from ports :)

bitmap:

does that look like correct output off npm install ? :)

yeah, looks sane

:)

thx! :)

I got it up now :)

great :)

does the server need to be stopped for replication ?

if it is only a slave that is

otherwise start/stopping is another challange ;)

from my rather limited experience, I've activated auto replication and it's never gone down

auto replication using cron ?

or another way ?

>75k unique.

Yes, applying replication packets shouldn’t disturb the server.

good :)

now to modify all the scripts that have hardcoded paths :(

kepstin-laptop: did you have an exheres for the acousticbrainz stuff?

kloeri: it's not really ready for that yet; it's all unfinished build stuff off unmerged git branches right now.

I think I did do one of the python dependencies tho, I should post that.

k

just wanted to contribute some data from my music collection if it doesn't require much work from me :)

kloeri: wait until thursday

and it'll require less work

but that's two days away!

:)

alastairp: any news on the analyzer hanging issue that I was hitting?

kloeri: IKR!

kepstin-laptop: Do you have access to `timeout`?

It should be part of the GNU coreutils.

Freso: hmm, yeah. In theory that should work, if I e.g. run the scanner from a find command separately on each file.

Combine it with find/ls to parallel and hanging files will get skipped over.

parallel has issues with the log file not having synchronized accesses

so it loses files :/

kepstin-laptop: I thought that was fixed with SQLite?

... when did that happen?

Yesterday.

well then.

Man. Keep up already. ;)

does it import the old logfile into the sqlite?

I'm using `(somehow seed the files) | parallel --eta timeout 5m python2 ./abzsubmit`

Yep.

cool

do I need an essentia update?

Nope.

I'm not sure there is one.

I'm using the latest audiocodec-something last I checked.

huh, apparently I never built python with sqlite support on my server box

>60k unique lossless.

hmm. does abzsubmit work with python3?

alastairp: ^

I think there's some work done towards that end, but I'm not sure it's ready yet.

(I noticed that it's importing the print() function as of yesterday, so...)

it's using print_function, so maybe

haven't tested it past that though

alright, I've built a command with find, parallel, and timeout. my log has been migrated to sqlite already

lets see what happens :)

hmm. I have a couple of those misdetected silent files in here

whee, looks like this combination of command-line tools is somehow managing to do what I want :)

kepstin-laptop: :)

find appears to have decided to list files in no particular order

kepstin-laptop: Sure it isn't parallel? Or abzsubmit?

parallel starts jobs in the order that it receives them by default

and abzsubmit is only being run on individual files

Are you using find with full/absolute paths?

yeah

Weird.

i guess find just returns stuff in the order that it gets from readdir without sorting,

(rsync and ls both have explicit sorting steps when listing files)

Less than 1500 files to go now...

whee, broke 80k unique tracks

er, I guess that's recordings, not tracks?

Yeah.

100k unique is beginning to be within reach.

how many people do we have submitting?

has to be less than 20 or so

Probably less than, but close to, 10.

Unless there are a bunch of people submitting who doesn't say anything in this channel.

hmm, only need about 10k unique tracks per person then :)

Actually, there might be some MTG people.

kepstin-laptop: I think there will be some publicity about it next week - alastairp says the blocking bugs in essentia should be ready by Thursday.

I suppose I'll run this stuff over the songs that I don't have lossless copies of, too

force all those students to submit their music in order to pass their courses :D

I can't remember when I last edited as much as I've edited today, just because I "needed" to submit AB data. :|

kepstin-laptop: That's what I'm doing. I reckon it's better to have analyzed a lossy recording than not to have it analysed at all.

what a romantic

ah, such song gems as http://musicbrainz.org/recording/5544367d-7d2e-...

apparently that has a danceability of ~1.3

and is in the key of A# major? :)

CallerNo6: As I always say, it's better to have encoded lossily, than to never have encoded at all.

(even if 14 minutes later, it was still groan worthy)

Freso: did you ever receive the 300 DKK?

ruaok: I did.

ah, good to know.

then I will leave the woman a great review.

:)

that house worked really well for a summit location.

Yep.

that was a fairly nice house indeed

surprisingly ok, even.

:)

:o apparently I've submitted or skipped my whole flac folder!