#metabrainz

<aerozol[m]> "How did you find those spammers..." <- randomly stumbled across one, then found 30K more with the same spam email domain

you can find hundreds of thousands more with [indexed search](https://musicbrainz.org/search?query=bio%3Ahttp&type=editor&limit=25&method=advanced), but that's far too many to delete by hand, I had to use SQL to handle those

monkey, ansh, lucifer I need some advice so I tried to make a mockup of the genre graph where there is no listen for a specific range, then it is getting populated by the other 6 hour brackets so should I do that or should I keep it as a while hole?

But that wouldn't make sense technically, so how do I go about that

s/a/one/, s/while/sector/, s/hole?/of 0 listens/

Also what if one for a specific time range in top 5 there are more than 1 possible genres

also monkey Thanks for the color sequential suggestion helped a lot

<holycow23[m]> "monkey, ansh, lucifer I need..." <- From a data point of view, I believe it makes more sense to keep one sector of 0 listens

mayhem: two new recording similarity datasets for you to try: https://labs.api.listenbrainz.org/similar-recor... and

i have only generated the two for listens at the moment, not MLHD. artist similarity coming up soon as well.

<holycow23[m]> "Also what if one for a specific..." <- This can only be possible if there are multiple genres with same listen count. So, I don't think its a good solution to just pick any 5. We should ideally use the hierarchy data for resolving this.

Any opinions: monkey reosarevok

<lucifer[m]> "suvid: off the top of my head..." <- yea i think it should be enough only

tbh for the same listen count I would not complicate it in the first implementation, and just pick one

Hi lucifer

I am done with the background tasks, uploading files and creating the import task

I am starting with the actual importer that will process files now :)

Should I make a separate PR for it or continue it in the existing PR only?