Litestar (also known as Starlite) is a performant FastAPI-like web framework. I believe using Litestar instead of FastAPI could provide several benefits:
Fast input/output validation via msgspec.
Built-in features like caching, logging, etc.
Native class-based controllers.
Disclaimer: I'm willing to contribute to this feature.
Use case
Most of the ML projects I've worked on with Ray Serve either produce or read large JSON payloads. I think for many ML applications, JSON parsing/serializing efficiency is important. Also, when caching is needed, I always encounter myself having to integrate with libs like aiocache or implement custom functionality.
Description
Litestar (also known as Starlite) is a performant FastAPI-like web framework. I believe using Litestar instead of FastAPI could provide several benefits:
Disclaimer: I'm willing to contribute to this feature.
Use case
Most of the ML projects I've worked on with Ray Serve either produce or read large JSON payloads. I think for many ML applications, JSON parsing/serializing efficiency is important. Also, when caching is needed, I always encounter myself having to integrate with libs like aiocache or implement custom functionality.