#metabrainz

yep I'm aware of that, but I was afraid that the rate-limits would throttle me a lot. But I guess these URLs will already be included in the database itself, right?

[bookbrainz-site] 14MonkeyDo opened pull request #834 (03master…monkey-fix-redux-slowness): fix(entity-editor): Remove slow deserialization in redux state https://github.com/metabrainz/bookbrainz-site/p...

yvanzo, reosarevok: ^ q3lont 's query

mayhem: alastairp: huh, spotify didn't even send the data in ms, its still s. `"timestamp_utc": "2022-01-06T16:45:41Z"`

Really? fFs

"message_offline_timestamp": 1645968764166,

q3lont: IIRC the urls are not in the database as such, but the database does have info of whether cover art exists for a particular release?

i had earlier seen this and thought they did send in ms but this field is something else.

its 0 for most streams

Let me see how we do it

so the only other time field is timestamp_utc which is still in s resolution. 😞

We generate it in https://github.com/metabrainz/musicbrainz-serve...

We also have CREATE TYPE cover_art_presence AS ENUM ('absent', 'present', 'darkened'); which means you know whether there's something to get before you try to

lucifer: ........

q3lont: Cover art IDs are available through replication according to https://musicbrainz.org/doc/Replication#Tables_...

so basically: you can put together the URL yourself with the replicated data, but it's not directly in the DB as such, AFAICT

so the question is: Can I generate the URLs myself, given the replicated database? Or do I have to query the Covert Art API?

oh ok, thx!

But the release_coverart table does not seem to be replicated, so I’m not sure how these IDs can be used from a mirror…

you beat me to it :)

I'm concerned only about release groups and not releases, fwiw

emailed spotify again.

musicbrainz is a treasure

RG is linked to Cover Art through releases, so…

bitmap: Do you know why the table "release_coverart" is not marked as replicated? It seems that it makes CAA tables less useful to mirrors.

q3lont: FYI, he might not be around for hours because of his timezone. ^

bitmap: Oops, found the answer!

q3lont: yes :)

q3lont: Here is how release and cover art are linked together: https://github.com/metabrainz/musicbrainz-serve...

And to link to release group: https://github.com/metabrainz/musicbrainz-serve...

[bookbrainz-site] 14MonkeyDo opened pull request #835 (03master…monkey-fix-renderauthors): fix(entity-editor): Fix WorkTable HTML structure https://github.com/metabrainz/bookbrainz-site/p...

These tables are available on mirrors by default: https://github.com/metabrainz/musicbrainz-docke...

so it seems that with the replicated db, I can generate a release group's cover art by querying the release_group_cover_art table and generating the covert art url for the linked release, like https://github.com/metabrainz/musicbrainz-serve... does

yes

thanks a lot!

[critiquebrainz] 14amCap1712 opened pull request #416 (03master…missing-moderator-check): Add missing check for moderator CB id https://github.com/metabrainz/critiquebrainz/pu...

[critiquebrainz] 14amCap1712 merged pull request #415 (03master…fix-hide): Fix invalid use of entity_details filter https://github.com/metabrainz/critiquebrainz/pu...

yvanzo: is SEARCH-609 related with the "MB updates too many things" issue we had for entity types? Or is it specifically a separate sir issue??

SEARCH-609: SIR triggers do not work as intended when running a MusicBrainz Slave instance https://tickets.metabrainz.org/browse/SEARCH-609

reosarevok: with "entity types"?

Like when we changed some stuff for RG types and it triggered a reindex because we sent the full change

reosarevok: it is related to "MB updates too many things" at least when it overwrites columns with the same value.

So we really should try and figure out a fix in MB, I guess...

No it doesn’t solve the issue of editing RG types.

reosarevok: It’s more a SEARCH issue really.

The MBS issue is when the server overwrites columns with the same value.

So there's two issues, one is "MBS sometimes sends 'changes' even if nothing has changed" and the other is "MBS sometimes sends legit changes, but they are not relevant to SEARCH and they should get ignored"?

1) yes 2) no they are relevant, but SIR is not able to cope with that many changes at once.

[listenbrainz-server] 14mayhem opened pull request #1943 (03master…migrate-mb-cache-to-bulk-table): Migrate MB metadata cache to bulk table https://github.com/metabrainz/listenbrainz-serv...

lucifer: ^^ for when you get a moment. It looks far worse than it is.

[bookbrainz-site] 14MonkeyDo merged pull request #835 (03master…monkey-fix-renderauthors): fix(entity-editor): Fix WorkTable HTML structure https://github.com/metabrainz/bookbrainz-site/p...

[sir] 14yvanzo opened pull request #125 (03master…fix-image-build): Fix image build and update some dependencies https://github.com/metabrainz/sir/pull/125

[sir] 14yvanzo merged pull request #125 (03master…fix-image-build): Fix image build and update some dependencies https://github.com/metabrainz/sir/pull/125

mayhem: do you really intend this to be a for/else or should this have been an if/else? https://github.com/metabrainz/listenbrainz-serv...

Intended as for/else correct

mayhem: but how could that else ever execute? iirc it executes if the loop terminates with a break. that loop has no break so else would never execute

oh heh. yes, it actually no longer makes sense. will remove.

I really want to use for...else, and I do, but buy the time the code is done, its never needed.

do..while is the way way.

hehe lol

presumably *someone* must have *some* use case when they implemented it.

that second ok = True should also be removed i think.

yeah you could probably rewrite this as a for/else if you want but flags are always the first thing to come to mind

[critiquebrainz] 14alastair merged pull request #416 (03master…missing-moderator-check): Add missing check for moderator CB id https://github.com/metabrainz/critiquebrainz/pu...

hmm the values of the flag look inverted to me ...

I'm just taking that check out. really not that important.

check it gone. lets move on.

change the if to != and i think move these statements inside the if. get rid of the flag

is

ohok

mayhem: did you build a mapping with these changes?

running as we speak

ah ok. nvm, i had confused lines at 366-368 in builk table.

70% complete.

lucifer: I merged https://github.com/metabrainz/critiquebrainz/pu... into master but github doesn't think so. weird

lucifer: of those 3 pending CB PRs, I see you have recent commits on each of them, are they all "done"?

alastairp: maybe a github glitch or maybe some commits differed if you merged locally?

I merged locally and then pushed to the PR branch, so gh should be able to see them

https://github.com/metabrainz/critiquebrainz/pu... is not ready it depends on changes in other PRs t

no problem

others are ready

320 and 328?

yes those are ready

mayhem: PR lgtm thanks. one general comment, can you try moving this line outside if so that it is executed for both full and incremental case. want to test if those config help in the full case as well. https://github.com/metabrainz/listenbrainz-serv...

fine to merge now and try later separately

[sir] 14yvanzo merged pull request #124 (03master…SEARCH-610-selective-live): SEARCH-610: Allow to selectively live update some search indexes only https://github.com/metabrainz/sir/pull/124

how would you judge if it worked or not?

mapping would be built much faster than now.

in particular the execute statement would run in less time?

yes

I was kinda thinking of adding a check to see how long the queries take to run.

let me do that and then run it with both options and we'll see.

actually that isn't even needed. the output already is timestamped, lol.

without the statement

ah nice

Hi monkey!

I had a doubt on how to replace the bootstrap js file like the one here https://github.com/metabrainz/listenbrainz-serv...

Since we have removed the local bootstrap files, how do you suggest we include the bootstrap js now?

Hi !

Wasn't the idea to import it from design-system ?

From DS we get the css

But for things like dropdowns, we require the js

And what of components we will build in DS that need the JS too? (say a custom dropdown for example)

Makes sense to include the bootstrap js also in DS

But how do I make that setup in DS, monkey?

I don't know. Maybe have a peerDependency on https://www.npmjs.com/package/bootstrap ?

And each project then has to install that package and import the JS themselves?

The dependency is already there in DS

Importing the JS part is what I want to do in LB rn

[sir] 14yvanzo opened pull request #126 (03master…SEARCH-609-diff-triggers): SEARCH-609: Triggers only fire message on data change https://github.com/metabrainz/sir/pull/126

akshaaatt: I'm not sure what is best. Importing the file directly from the node_modules BS folder, perhaps?

I tried that but doesn't seem to work for me monkey

We're soon going to move to using React-Bootstrap in LB, and they require to install the bootstrap package and load the JS that way from node_modules: https://react-bootstrap.github.io/getting-start...

Ah.

Can you explain the issues you're having?

The path might be wrong in the context of the docker image. in the built image the node_modules directory is under `/code/listenbrainz/node_modules/…`

Oh okay! Just so I have it right, it should be <script src="/code/listenbrainz/node_modules//bootstrap/dist/js/bootstrap.bundle.min.js"></script> ?

One other solution would be to automatically run a script after npm install to copy the BS file to `js/lib` where the previous one was.

(Or just replacing the file in git)

^that would basically be reverting to what we had before

I think I like the first solution better but no firm opinions

The /code path fix doesn't seem to work still monkey (I did remove the extra '/' I had added)

Yeah, not sure we can access it that way

Right

Just sanity checked with alastairp: probably best to revert to have a copy of bootstrap.min.js in the static folder

Flask won't allow you to access files outside of the statis folder

static*

I did try that as well before but that didn't seem to work well for me monkey

Let me try once again and confirm

Well, it's basically re-creating this file and loading it from there as we did before: https://github.com/metabrainz/listenbrainz-serv...

I understand that. But we'll be replacing it with the 5.1.3 version right

Right

And once we migrate to using react for all our pages (part of the SPA project), we can figure out how to package bootstrap as a dependency in webpack, to be compiled with the rest of our common js code

That sounds really good to me monkey