#metabrainz

Ok, I will.

_lucifer: I can sum it up: "Elastic used a real open source license and lost out on the amazon revenue. In order to 'capture the value' (read stick it to amazon) they changed their license to force amazon to pay. They wrapped the whole thing into "we were forced to do this, we're preserving open source", neither of which are true. Amazon responds by forking elastic search to actually keep it open. This has the potential to kill elastic."

<3

that clears it up. thanks for explanation.

sad state of affairs :(

>This is an Oracle-level move.

very true

companies pretending to be open source and then regretting not getting filthy rich off it have no sympathy from me.

True, but if Amazon really presented the whole thing as a cooperation when it wasn't, that's also a dick move

yep, indeed. Neither party shines here.

true

And there's a history IIRC of Amazon just taking stuff, hosting it and effectively sidestepping the original people

So yeah, shitty situation all around

i think BB uses Elastic search?

and the GPL didn't foresee the cloud and thus we have these problems.

I guess if Amazon keep this fork open *and* take good care of it, it's not necessarily a loss... except that yet another big thing ends up in Amazon.

Which is fucking big enough as it is

it does. but its on Mr_Monkey task list to remove it.

I mean, by now they could probably just buy themselves a continent or two and move operations there...

as long as there is cheap labor to exploit, it would work.

I doubt they would have huge issues just building their workers a Dubai-style "ok we're in the middle of nowhere but you have all you need here" city tbh

Eh, we'll see

"... slum tbh" -- there I fixed it for you.

My point is that if they wanted highly paid experts, they could probably attract them with a non-slum

Maybe they don't even need those, but :)

Like, if they were able to operate just with a slum wouldn't they have moved out of the US ages ago to somewhere with even more corporate-friendly laws?

well, they have highly paid experts. just not in their warehouses.

there is no place in the world more corporate friendly than the US,....

Oh, sure, I was thinking about the tech bits of Google, Amazon etc, not of the Amazon warehouses. Eh, who knows. Let's see what happens

And I dunno, I'm sure some place could be convinced to just make it outright illegal to complain about Google :p

(and charge no taxes too)

yvanzo: what the fuck does the area sortname field search? :D

"Area search terms with no fields specified search the area and sortname fields."

AFAIK areas have no sortnames?

But it's in the table, so it must be a thing...

I guess it's just a copy of the name

(I see the search XML does return it)

(ok, and so does the lookup XML - nvm then, but I should probably still specify it's just equivalent to the name for this)

(... and drop it for ws/3)

Ok, SEARCH-356 confirms that

SEARCH-356: Area Sortname field removed from Database https://tickets.metabrainz.org/browse/SEARCH-356

yvanzo, bitmap: should we have a "drop area and label sortname from WS" ticket for ws/3?

Also, why does artist have primary_alias but not area... jeez, we're so inconsistent :)

reosarevok: there's a script (searchfields.py) in the mbsssss repo that gets the available search fields in a mediawiki table (pipe its output to xclip on linux to have it in the clipboard). doesn't do descriptions, though. you could probably add another namespace at the top of each schema.xml file and add the descriptions as xml attributes ('<field ... mb:description="whatever">') and have it dump that

as well

Yes, I'm using that :)

To figure out what we currently have that we haven't documented

[troi-recommendation-playground] mayhem opened pull request #34 (main…improved-mbid-mapping): Improved mbid mapping, playlist submit improvements and year lookup improvements. https://github.com/metabrainz/troi-recommendati...

reosarevok: probably yes

yvanzo: I'd add a ticket to add primary_alias to all entities, but that would also affect recording, so :/

(and ok, adding the ws/3 one)

reosarevok: added https://tickets.metabrainz.org/browse/SEARCH-639

SEARCH-639: Search artist, label, and place by related area's MBID

Thanks :)

Hello! I am having some trouble in running a local development server for BookBrainz site. I followed the steps mentioned in the github repository. When i point my browser to localhost:9099, it opens up the site but says 'database "bookbrainz" does not exist' . I ran the database-init-docker script and it seemed to work just fine. Can somebody tell

me if I am doing something wrong or maybe I am missing something? Any help will be appreciated. Thanks!

Mr_Monkey: ^

Shivam: that means the BB database is not initialized.

the instructions probably cover how to do that.

Hi Shivam ! It does look like something went wrong when initializing the database

Could you try running the database-init-docker script and seeing if you have any useful output that could shed a light on the issue?

I meant "running the database-init-docker script *again*"

yvanzo: exposing event.ended seems a bit pointless, given it can't be set separately, it's basically just "does it have an end date set". But I documented it as such, anyway

Hopefully it'll avoid people thinking it just means "is the event in the past"

yvanzo: the upgrade to python3 sounds interesting. i found SEARCH-594 and this PR https://github.com/metabrainz/sir/pull/111. but i am unsure what needs to be done.

SEARCH-594: Port Search Index Rebuilder (SIR) to Python 3 https://tickets.metabrainz.org/browse/SEARCH-594

oh! i see the PR mentions to follow the migration guides and upgrade sqlalchemy.

ishaanshah: hi, are you around?

_lucifer: yeah, lots of py2->3 steps for sir. I'm happy to talk you through some items if you want to take a look at it

alastairp: yup, i am interested in working on that. thanks!

yvanzo: also: "If you don't specify a field, the term will be searched for in the area field.". For other entities, it also checks 'alias' - is that not true of area?

(wondering if our docs are wrong or our code is inconsistent)

Sorry, it feels like I'm giving you more work instead of taking it off you D:

ruaok: About the ES/Amazon beef and license change, I must say I find both parties despicable. Shame !

Yep, agreed.

Amazon's ghastly practices I'm very much against, and it's really sad to see the response from Elastic going away from FOSS principles.

On more reason to leave ES behind.

yvanzo: also, tag is not available for places for some reason?

And coordinate search is pretty much useless, since it wants the exact values inc all decimals, so you'd need to know them already

(58.38 doesn't find 58.388226)

Will add a ticket for that last one if not available

_lucifer: the currently blocking bit is upgrading sqlalchemy (because the current version does not support recent python versions)

_lucifer: it is due to some code hacking around sqlalchemy IIRC, alastairp and bitmap may know more about it ^

did anyone copy Mineo's description into a documentation file from the day that we had that meeting?

I don't think so

I'll add a ticket

Seems unintended

reosarevok: maybe keep "terms" plural because a search can have many.

Ok

Added SEARCH-640 / SEARCH-641

SEARCH-640: Don't require *exact* coordinates match for lat / long in places https://tickets.metabrainz.org/browse/SEARCH-640

SEARCH-641: Search for terms on area aliases by default https://tickets.metabrainz.org/browse/SEARCH-641

Thanks!

Hmm, I see release, release group and recording do the same as areas

Re default searches

We could add them to at least RG and release, I guess, but probably not recording then?

We barely have aliases for those, I expect, but

Jeez. I just saw all the fields we index for recording. No wonder it's huge.

We could do it even for recordings, but we better have to estimate size increase first.

Recordings were awfully documented btw

Hopefully I'm figuring out each field properly, but maybe you can take a look to make sure

(almost done)

yvanzo: where do we generate the fallback type for RGs?

(I don't think that's documented anywhere, right? so I should probably link directly to the code)

Currently it says "type of the release group, old type mapping for when we did not have separate primary and secondary types or use standalone for standalone recordings" but "use standalone for standalone recordings" doesn't seem to be true anymore anyway and the first bit is useless

I knew it, will check

i reinstalled docker and started everything from scratch again..when i run database-init script, it tries to build up and create bookbrainz-site but after a while it exits with the following line : 'standard_init_linux.go:219 : exec user process caused: no such file or directory' ..Please help me out

What OS and docker version are you using Shivam ?

I am using docker version 20.10.2 on Windows 10.

yvanzo: https://wiki.musicbrainz.org/index.php?title=Mu... in case you find any errors btw

The two step fix would be:

1. Install linux

2. Run the script in it. Pretty sure the error will be gone. :D

I feel linux provides more comfortable environment for devs. P.S. Just an opinion

yeah i tried it on ubuntu first..but I ran out of storage on that..I guess I'll install ubuntu again with more storage and try it again

Shivam: I'm not sure what's happening there and I'm not on Windows to try and replicate the issue.

From some googling it could potentially be because of line endings (see some solutions here: https://stackoverflow.com/questions/51508150/st...)

You can extend the partition

changing the EOL from crlf to lf

Mr_Monkey: reg. the odd logos in inkscape. Did you figure out some reason behind it. It occurs quite often if you open logos in inkscape but they seem to open perfectly in AI

Oh, found something I dropped in error, fixed :)

Not sure why, no. I made the logos in AI, so it's possible that there are some paths that other vector programs don't like, but… 🤷

zas: please have another look at https://github.com/metabrainz/docker-server-con... -- I'm somewhat blocked by that right now.

It looks better

and the issue about the variables that were not being imported?

we both made mistakes in the names of the variables we wanted to use. no wonder they were not found!