but I just deleted bunch of listens so it could be one of them
2025-01-13 01329, 2025
vardhan_ joined the channel
2025-01-13 01339, 2025
rustynova[m]
I know that it broke when fetching a parent work. So I think it may be an instrument or work relation that isn't set in the schema.
2025-01-13 01339, 2025
rustynova[m]
I hope I catch it because that it's actually a pain to find those
2025-01-13 01358, 2025
mayhem[m] joined the channel
2025-01-13 01359, 2025
mayhem[m]
<ruebeckscube> "hi folks! new here, super..." <- Hi! Sadly restricting the DB isn't possible right now. You might consider using the API to fetch the data you need -- that depends on how much data you really need to fetch.
2025-01-13 01351, 2025
elkiensad[m]
<rustynova[m]> "I know that it broke when..." <- might be this one?
I was asking. I couldn't find anything related to taste in the docs you linked
2025-01-13 01330, 2025
mayhem[m]
feedback = love/hate, pins = pins. not sure what else to tell you.
2025-01-13 01321, 2025
hash[m]
oh alright. thanks
2025-01-13 01334, 2025
vardhan_ has quit
2025-01-13 01320, 2025
rustynova[m]
I did a sample of a year of listens and i'm still not past the recording fetching phase.
2025-01-13 01320, 2025
rustynova[m]
Would love to improve the speed of the recording redirect fetching, but when it comes to works, I'd get like 1% of speed Improvement
2025-01-13 01338, 2025
rustynova[m]
Or I could try adding a multi mirror fetcher to parallel the fetching
2025-01-13 01355, 2025
ruebeckscube[m] joined the channel
2025-01-13 01356, 2025
ruebeckscube[m]
<mayhem[m]> "Hi! Sadly restricting the DB isn..." <- the similar-artists API is def an option, but I would love to be able to access the raw similarity scores for two mbids that are stored in the database—considered trying to contribute an API for that but I'd be calling it like 1000 times for each search on my website so not really feasible even in a batched setup
2025-01-13 01330, 2025
ruebeckscube[m]
i think I might just have to suck it up and pay for more server storage! the docs mention building indices locally, I'll see if I can limit that to the ones I need to save space