Introduce a separate backend service for LLM calls (#15831)

This PR introduces a separate backend service for making LLM calls.

It exposes an HTTP interface that can be called by Zed clients. To call
these endpoints, the client must provide a `Bearer` token. These tokens
are issued/refreshed by the collab service over RPC.

We're adding this in a backwards-compatible way. Right now the access
tokens can only be minted for Zed staff, and calling this separate LLM
service is behind the `llm-service` feature flag (which is not
automatically enabled for Zed staff).

Release Notes:

- N/A

---------

Co-authored-by: Marshall <marshall@zed.dev>
Co-authored-by: Marshall Bowers <elliott.codes@gmail.com>

crates/collab/src/main.rs

@@ -83,7 +83,9 @@ async fn main() -> Result<()> {
             if mode.is_llm() {
                 let state = LlmState::new(config.clone(), Executor::Production).await?;
 
-                app = app.layer(Extension(state.clone()));
+                app = app
+                    .merge(collab::llm::routes())
+                    .layer(Extension(state.clone()));
             }
 
             if mode.is_collab() || mode.is_api() {