>If the WORKDIR doesn’t exist, it will be created even if it’s not used in any subsequent Dockerfile instruction.
nawcom has quit
nawcom joined the channel
sumedh joined the channel
alastairp
_lucifer: mmm, interesting. I guess I always just assumed that you had to create it before switching to it
mrtriad joined the channel
I had a dream last night that I was hosting all participants in the MeB summit, because I had 7 bedrooms in my house, but when we got home I remembered that I already had guests and there wasn't enough space for everyone
mrtriad
Good evening everyone. First of all I want to congrats for the musicbrainz project. Finding such a big and detailed open database was pretty unexpected ahah.
Anyway, I'm working on a small project for my uni. As a starting point I need to place on a nice map artists and albums, but I don't want to annoy with details. The whole thing should work on nodejs. Now, I set up the local dev version of musicbrainz, it works, but now I'm really starting to think if interacting with that server and some of the
community api is more or less convinient that setting up the database on its own.
I'd like to hear what are the advantages of interacting with the server in spite of the direct database, and I guess that this is the best place to ask that. Otherwise tell me simply to go away ;3
ruaok
calendaring is hard. even whle asleep.
reosarevok
mrtriad: depends on what you want to do
ruaok
mrtriad: welcome! and thanks!
reosarevok
The database is always going to be more powerful
(since you don't depend on the specific combinations of data the API offers you)
But on the other hand, you need to maintain a database copy :)
If you don't care about keeping the data up to date and are happy with just a static snapshot, you might as well use the database you already have, I guess :)
mrtriad
That's not a big deal luckily. But yay, thanks for the confirmation.
A few more question since you are so quick ahah. First of all, I installed the dev version during the server installation, though it's not so clear what's the difference from the normal one. (it's not a slave, so it's not up to date, though there are also a few other version to choose and the dev one is much smaller than the others)
ruaok
it has a tiny test database.
which won't be useful for your purposes of collecting data. useful for development and not waiting a long time for data to download.
do you have a list of artists for whom you want releases? how big is it?
sumedh has quit
reosarevok
Yeah, it's basically taking a bunch of big artists and the data connected to them, so that there's enough to play with
So, it'd work for you to test, maybe, but it won't help much otherwise
mrtriad
oh, k. Clear clear.
reosarevok
(also because if you want countries, that's missing some area to country connections)
mrtriad
Also to get the db as a stand alone should I download the big one as I set up the server or should I follow this git https://github.com/lalinsky/mbdata?
(hope that posting link is legal)
reosarevok
Sure, why wouldn't it be :) luks is a friend
Assuming mbdata is up to date, the question is: do you want to be able to use the website too, or do you just need the DB
mrtriad
I just need the DB
reosarevok
If the DB is enough, mbdata is probably all you need. If you want to have a copy of the website locally to view the data with and whatnot, then use the main server
mrtriad
oh, ok. Perfect ^^
Welp, that's it for now. Thanks for your support and the quick replies
reosarevok
ruaok: I'm not missing any recent changes wrt mbdata, right? :)
alastairp
mrtriad: are you talking about using the mbdata db loading/replication functionality?
reosarevok
Just to make sure
ruaok knows nothing
alastairp
reosarevok: it should be up to date. we use it in brainzutils
but it's 2 things: an ORM for accessing the database from python, and some additional tools to set up a database without needing to use musicbrainz-docker or mbserver
mrtriad
Hmm what do you mean by "mbdata db loading/replication functionality"? I still have to take a look at the mbdata functionalities. I guess that at the end of that I'll have a local postgre with the musicbrainz db. Am I right?
reosarevok
I never used the orm, but I remember using it to just get a DB ready back in the day when I didn't want a local MB copy
But this was... well, a long time ago :D
Maybe by now just using the docker files would make your life easier
alastairp
mrtriad: sorry, I may not have explained that clearly. let me read your backlog
mrtriad: I understand that you set up the database with musicbrainz-docker? Does that work well for you?
mrtriad
I set it up without docker. But yay, it works
alastairp
ah, right.
mbdata does the same thing - if you've managed to set it up yourself then you don't need it
yvanzo has quit
reosarevok
Yeah, I guess if you already have all installed, you could drop the sample data, and install the main data, and that would be the same
mrtriad
Oh, ok.
Welp, thanks again :3
Have a nice day and good luck at keeping up the project ^^
Hi _lucifer so apparently since jcenter has served its end, the barcode scanner library that we used to use needed to be switched since its creator had stopped the support and even the repo was no longer available, I have made the switch to the Google ML-Kit Barcode scanning library.
akshaaatt[m] uploaded a video: MusicBrainz (6484KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/WKrNNxNAIENvEzBoBazFqRIl >
yvanzo
Hi nelgin: Sorry for not responding earlier, this is a SEARCH bug (happened in prod too), it is not related to your settings.
akshaaatt[m]: we cannot use google ml library or for that matter any closed source library. otherwise we cannot publish on fdroid.
ROpdebee
reosarevok: RE that tweet, no clue. there's not much to go off, could be due to recent changes, could be similar to MBS-11621, could be something else entirely
searching on pypi for other zip libraries also doesn't turn up much only xtract and reusables seem usable and xtract is GPLv3. reuseables is MIT license. i'll checkout reuseables and see if it serves our purpose.
alastairp
_lucifer: ok, it's a quick fix on 1424, let me do it now. it was related to my question about the new spark image
_lucifer: the reason for using a streaming library is so that we don't have to load all items into memory, then compress it, then send it (that'll make a request take a long time, especially for a large dataset)
another option would be to only allow users to do it for small datasets, but that's annoying too
_lucifer
how about moving the task into background and asking the user to check back later on a page that will get updated once the zip is generated?
we could say retain the zip for a week or so.
alastairp
yeah, that was the other other option, but it has its own set of problems