feat: move /completions under /v1/completions

[tomasmcm]

This PR updates the server endpoints in local.ai to match the endpoints from llama-cpp-python.
Specifically it changes /completions to /v1/completions as in llama-cpp-python/llama_cpp/server/app.py:599
And /model to /v1/models as in llama-cpp-python/llama_cpp/server/app.py:805

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
local-ai-web	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Sep 20, 2023 8:18pm

[tomasmcm]

Can confirm this works 👌 with Continue.dev using GGML(server_url="http://localhost:8080") which is amazing!

[louisgv]

LGTM! 👍 - this is somewhat related to #42 , but we can ponder on that and just get this more compatible with other API.

We will need to fix the wiki next: https://github.com/louisgv/local.ai/wiki

Side-note: I wonder if we can add both v1 and also the base API? (So that when we update to v2, we can point base to the latest version :-?...

@tomasmcm can you investigate if that's possible?

[louisgv]

Actually... would be much better if we exposes the baseURL as a config instead - so that folks can put /v1 or /? (Similar to how we have the config for port etc...)

EDIT: scratch this idea - it's basically #42

[tomasmcm]

@louisgv agreed, one step at a time. Having a compatible API is essential for many use cases. Making it customisable is even better but that can be done next 😊
Feel free to merge this (I don't have write access to do it)