unum-cloud / usearch

Fast Open-Source Search & Clustering engine × for Vectors & 🔜 Strings × in C++, C, Python, JavaScript, Rust, Java, Objective-C, Swift, C#, GoLang, and Wolfram 🔍
https://unum-cloud.github.io/usearch/
Apache License 2.0
2.27k stars 142 forks source link

Feature: possibility of a streaming API? #203

Closed philbooth closed 1 year ago

philbooth commented 1 year ago

Describe what you are looking for

Caveat, this question may not make sense: I'm coming at it from a node.js perspective and have little understanding of the underlying C++ implementation.

When handling large datasets in node, it's customary to pass streams around instead of loading it all into memory as a string. Given that usearch can already serialize indices to disk, it would be really neat if I could open a stream when fetching/reading a document and connect it to a usearch index for streamed serialization.

Would something like even that be possible? I'd be happy to try and implement it if so, with guidance.

Can you contribute to the implementation?

Is your feature request specific to a certain interface?

C++ implementation

Contact Details

pmbooth@gmail.com

Is there an existing issue for this?

Code of Conduct

ashvardanian commented 1 year ago

Hey, @philbooth! Thanks for reaching out!

I am not sure I understand correctly. We currently have a view() function, that allows viewing an index from disk, without loading it into RAM. Is that what you are looking for?

philbooth commented 1 year ago

Do you know what, I just realised this ticket makes no sense at all, sorry. Been looking at lots of different dependencies this morning and got myself mixed up. Ignore this, I'm closing it! 😊

ashvardanian commented 1 year ago

No worries, @philbooth! Don't hesitate to ping us if you have any cool ideas on your mind, and please share a link to what you build using USearch on our Discord - contributors will be very excited 🤗