#metabrainz

/

      • antlarr2 has quit
      • antlarr joined the channel
      • Thuraabtech has quit
      • agatzk has quit
      • serial-ata has quit
      • jasje joined the channel
      • lucifer
        mayhem: i took a look in the listens in spark but couldn't find listens with this mbid or even any having `a forest` in the track name. its possible the mapper matched this to some other mbid causing your submitted mbid to be overidden.
      • regarding dupes in recommendations, that is a known issue no? and we have filters in daily jams to tackle this.
      • i'll have to check to confirm but i don't think spark has any specification against recommending the same thing multiple times. could do a distinct items query in post processing but at least for daily jams purposes troi handles it already i thik.
      • vibhoo_24: best to ask monkey when he is around later today.
      • mayhem: oh i found the entries now all are `|2022-12-23 10:14:56|1 |A Forest - 2006 Remaster|null |`, so mapping issue i think.
      • vibhoo_24 joined the channel
      • jivte joined the channel
      • Thuraabtech joined the channel
      • jivte has quit
      • jasje has quit
      • jivte joined the channel
      • jivte
        lucifer: Hey can u tell what to do when a module doesn't contain ts module declaration file
      • lucifer
        jivte: module as in some external library?
      • jivte
        yeah i installed a third party module for datetime
      • as soon I start using it the deadly white component appears
      • :(
      • lucifer
        first option would be to check if there is a corresponding types module for it. what's the library you want to use?
      • jivte
        react-datetime-picker
      • I tried the types approach but no luck
      • lucifer
      • try installing `@types/react-datetime-picker` as well
      • jivte
        actually I tried it but got no result
      • same WSOD
      • lucifer
        what error do you get?
      • jivte
        WHITE SCREEN RENDERS
      • lucifer
        that can be due to many reasons. do you get any specific errors in console or static builder container logs
      • jivte
        let me check :|
      • reosarevok
        lucifer, mayhem: we do not separate remastered recordings in MB, so maybe the mapper should explicitly ignore "remaster" in titles?
      • (I bet there's an artist who named their tracks "Remaster", but)
      • lucifer
        reosarevok: in this case it seems the track name submitted (probably by spotify) had remaster in the track name.
      • reosarevok
        Yes, Spotify does that often
      • I'm just wondering if we can get rid of that before trying to match
      • lucifer
        should be possible i think.
      • in this particular case however, the mapper already prefers the original track over remastered.
      • reosarevok
        Most often in my experience those are either "- Remastered" or "- YYYY Remaster", so easy to filter out
      • lucifer
        there is no match because `a` and `the` are removed from the track and artist name as stop words.
      • reosarevok
        :D
      • Shouldn't "Forest Cure" still match?
      • lucifer
        i thought the same but doesn't happen due to some reason: https://labs.api.listenbrainz.org/mbid-mapping?...
      • reosarevok
        Well, that seems to be your bug :D
      • (I'd still suggest dropping remaster more generally if you don't already, but that's I guess a different issue then)
      • reosarevok plays https://www.youtube.com/watch?v=JRtFHNen20M
      • jivte
        lucifer: thanks for the help i found out my bugs :)
      • mayhem
        lucifer: moin!
      • I think it would be good to remove the dups from the recs -- this will become a problem for other things and I would very much like to remove them just to be sure. please do that.
      • lucifer
        morning!
      • mayhem
        and maybe next week or so, I think you and I should tackle the mapping and move a couple of things forward.
      • Freso
        alastairp: Pong
      • mayhem
        namely, improving our detunings, so that that case above and lots of other cases are improved.
      • vibhoo_24 has quit
      • and releases in mapper -- a couple of days of concerted hacking could be good.
      • lucifer
        mayhem: fix for dupes already deployed currently generating new recs to confirm it.
      • mayhem
        <3
      • lucifer
        improving mapper makes sense
      • jasje joined the channel
      • same recording mbid dupes should be gone now
      • mayhem
        Great, thank you. Will dig in when I get to the office.
      • BrainzGit
        [musicbrainz-server] 14reosarevok opened pull request #2809 (03master…MBS-12169): MBS-12169: OpenAPI docs for WS/2 [WIP] https://github.com/metabrainz/musicbrainz-serve...
      • reosarevok
        monkey, alastairp ^ so I put the basics into the codebase.
      • monkey: how did you set up the actual docs page for the BB API?
      • Would be nice to do that for MB so that as I work on this people can actually try it and give suggestions
      • monkey
        Using two neat packages: one takes JSDocs and transforms them into swagger definition, the other serves the Swagger live docs website using expressJS (the web server framework we use)
      • Maxr1998_ joined the channel
      • So it integrates automagically into it
      • Maxr1998 has quit
      • alastairp
        morning
      • reosarevok: amazing! thanks
      • reosarevok: lucifer was looking into compiling swagger to html and serving it as static files, he might be able to help
      • monkey
      • reosarevok
        I mean, that's still half the artist docs :D
      • But playing with it would be neat
      • I split them into separate files because it would be ridiculous otherwise, but maybe the process to turn them into a static file needs to combine them, dunno
      • lucifer
        monkey, reosarevok, alastairp: its possible to embed swagger docs in the readthedocs if wanted. (pending some issues with theme)
      • monkey
        That's pretty neat
      • reosarevok
        I mean, I'd be happy to just have them show under musicbrainz.org/doc/API or whatever
      • Since we don't use readthedocs at the moment :)
      • lucifer
        reosarevok: yup, i mean you can embed in a html page. just serve the openapi json schema file from some publicly available url.
      • alastairp
        reosarevok: I think that multiple files is a good idea, especially if there are entities which are 90% the same but with just a different entity name. it could be possible to automatically generate these included files from a template
      • reosarevok
        I expect descriptions and whatnot are going to be different for every entity (one of the best things about this is literally that you can customize the descriptions very nicely and show them as appropriate for every entity, so we can document how each include plays with each specific entity type for example)
      • lucifer
      • just copy the html file from here https://github.com/swagger-api/swagger-ui/blob/... and embed it in the page you want to
      • reosarevok
        So I guess first it'd need a tool to merge the files into one, then to publish that one (in staticbrainz I guess), then access that for the display
      • mayhem
        lucifer: the raw recs look a lot better now, thanks.
      • vibhoo_24 joined the channel
      • vibhoo_24 has quit
      • BrainzGit
        [bookbrainz-site] 14MonkeyDo merged pull request #919 (03master…vscode-docker): Docker config for vscode debugging console https://github.com/metabrainz/bookbrainz-site/p...
      • vibhoo_24 joined the channel
      • Thuraabtech has quit
      • vibhoo_24 has quit
      • jasje
        akshaaatt: facing this issue thats described here : debugImplementation("androidx.fragment:fragment-testing-manifest:1.6.0-alpha04") androidTestImplementation("androidx.fragment:fragment-testing:1.6.0-alpha04")
      • oops
      • here
      • jasje has quit
      • smach has quit
      • BrainzGit
        [listenbrainz-server] 14amCap1712 merged pull request #2297 (03master…load-listens): Load all listen parquet files at once https://github.com/metabrainz/listenbrainz-serv...
      • [listenbrainz-server] 14amCap1712 opened pull request #2341 (03master…dedup-recs): Fix duplicates in CF recommendations https://github.com/metabrainz/listenbrainz-serv...
      • jivte
        lucifer: monkey: Hey how to add a type declaration file if it doesn't exist
      • monkey
        Is that still for the datepicker library?
      • jivte
        yeah that library is a headache
      • it shows when we import dist.nostyle
      • dist/nostyle
      • monkey
        Here's an example of how we declare dummy types for libraries that don't have existing types: https://github.com/metabrainz/listenbrainz-serv...
      • But if the library has types let's use that and figure out what your issues are
      • What is your issue exactly? Any error message we could look at?
      • jivte
        I dived quite deep into library it has only one type and i.e. index.d.ts
      • no type exists for one which we are looking even in @types/react-datetime-picker
      • BrainzGit
        [listenbrainz-server] 14amCap1712 merged pull request #2341 (03master…dedup-recs): Fix duplicates in CF recommendations https://github.com/metabrainz/listenbrainz-serv...
      • vibhoo_24 joined the channel
      • monkey
        I still don't understand what the issue is jivte. What is the result you see? VSCode warnings? Compilation failure with webpack? Something else?
      • A single index.d.ts file for typings of a library is the most common pattern.
      • jivte
        yeah but like in date-picker library nostyles.d.ts file exists it is not there
      • but no worries by declaring it in utils we got it covered.
      • monkey
        By declaring it in utils you're declaring a dummy typing, which means no typing help in your editor or at compile time, so definitely not the ideal result :)
      • I'm happy to try to help you resolve the typing issue properly, but you haven't given me enough information to do that at the moment. What's the actual error or warning you're getting?
      • jivte_ joined the channel
      • jivte
      • d4rk joined the channel
      • d4rkie has quit
      • monkey
        Thanks jivte that's helpful. And so you get the same message after installing the @types/react-datetime-picker library? I think the issue could be simpler and have to do with webpack not being aware of the new package. Can you try rebuilding the docker image for the static_builder container and trying again?
      • vibhoo_24 has quit
      • jivte
        monkey: already tried rebuilding
      • monkey
        The command would be `./develop.sh up --build static_builder`
      • Ah
      • jivte
        lets try one more time
      • jasje joined the channel
      • monkey
        Looks like you're not the only one with this issue jivte : https://github.com/wojtekmaj/react-datetime-pic...
      • If all else fails, do add a dummy type definition as discussed above, with a comment explaning why we're doing it and pointing to that github issue
      • jivte
        monkey: Okk thanks and by the way both pr's work is complete
      • and
      • need to discuss abt add album feature
      • we can discuss when ur free :)
      • monkey
      • Sure, will review in a few days, not going to be available much this week.
      • jivte
        monkey: In the link mentioned the fix is done how can we implement it
      • can u plz help?
      • monkey
        In that PR I don't think there should be anything we need to do, the types should be recognized when importing the entry/nostyle file
      • jivte
        oh I see it is done in react-date-picker not react-datetime-picker
      • monkey
        So we're back to "let's see if rebuilding the docker image will work", and barring that I'll have to do some testing myself when I have a bit of time
      • jivte
        rebuilding failed :(
      • same error
      • monkey
        Ah, that could be it. In which case, you can open a PR for that other repo, citing the PR above