feat: respect context and add request cancellation

[mudler]

Description

This PR binds the token generation to the request context, and for llama.cpp it implements job cancellation.

It also adds the stop icon now in place of the loading icon, that will abort the request.

Notes for Reviewers

Signed commits

• [ ] Yes, I signed my commits.

[netlify[bot]]

✅ Deploy Preview for localai ready!

Name	Link
🔨 Latest commit	4839d572931cb247fedeabcb41ccffe991bd48f3
🔍 Latest deploy log	https://app.netlify.com/projects/localai/deploys/6910be954e2384000832362a
😎 Deploy Preview	https://deploy-preview-7187--localai.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[mudler]

Seems we can't propagate client disconnection during non-SSE requests due to https://github.com/valyala/fasthttp/issues/468 , also affects go-fiber: https://github.com/gofiber/fiber/issues/1718

[mudler]

just as a note, echo doesn't have such issues: https://github.com/labstack/echo/issues/1581

[mudler]

found an ugly workaround, but works for our case. Would be nice if fasthttp supports this natively, but I guess for now that's the only way we can tackle this.