noita-tools
noita-tools copied to clipboard
Seed Info API call?
Would be pretty cool if we could do a http request to get seedinfo for a specific seed and get a json with the raw data
Or a feature on the website to get a list of all seed numbers matching the criteria I just want to generate a list of seed numbers to go through, without clicking Find Next all the time ;)
Hey! I can add a “find all” button that will search for all the seeds and give you a json with some data. What do you think it should contain?
Some things are generated lazily, so I'd have to add those in as well as provide some data structure API to use.
I'm not sure when the next update will be, though (I'm working on a big update with maps)
A "find all" button would be really useful too yea :) Just a list with seed numbers is fine, or even the way it is right now if that works for a big list of seeds (doesn't have to be json, I just figured it'd be easy to dump all the raw data for a single seed into json format as the output, but maybe not so easy in typescript)
API call is probably gonna be too much data... and you'd have to do the processing server side I presume
Not exactly on topic, but maybe also a way to "copy link" for search criteria like on https://cr4xy.dev ? (Redoing all the filters from scratch every time can be really annoying hehe)
@Dextrome Find all is now an option that returns a list of seeds. I'll think about how to implement "copy link"
Very nice :) thanks
@Dextrome is this still something that you would like? I got wasm working server-side, so now the generation can happen there!
Sounds pretty interesting :)
Does that mean you can continue running the search even if you lose connection?
Yes, that's right!
Hey @Dextrome, turns out, search is very impractical to implement server-side. For reference, for $20/month for a compute-oriented server (aws c7g), you'll get a search speed of (of a very heavy search) ~50 seeds/s/core, while my laptop gets around 80 seeds/s/core. For this to be viable, there needs to be another way, and I do have one in mind, but not in an API sense.