feat: add doc endpoint

src/routes/v1/models.rs

@@ -6,6 +6,25 @@ use axum::{
 use crate::dto::api;
 use crate::{errors::OllamaError, state::app_state::AppState};
 
+#[utoipa::path(
+    get,
+    path = "/models",
+    tag = "models",
+    responses(
+        (
+            status = 200,
+            description = "List of locally available Ollama models",
+            body = api::ModelsResponse,
+            content_type = "application/json",
+        ),
+        (
+            status = 500,
+            description = "Internal server error (Ollama or network failure)",
+            body = api::ErrorResponse,
+            example = json!({ "error": "connection refused" })
+        )
+    )
+)]
 pub async fn list_models(
     State(state): State<AppState>,
 ) -> Result<Json<api::ModelsResponse>, (axum::http::StatusCode, String)> {
@@ -15,6 +34,49 @@ pub async fn list_models(
     }
 }
 
+#[utoipa::path(
+    post,
+    path = "/models/{model}/load",
+    tag = "models",
+    params(
+        ("model" = String, Path, description = "Name of the model to load into memory (e.g. 'llama3')")
+    ),
+    request_body(
+        content = api::LoadModelBody,
+        description = "Load model request",
+        content_type = "application/json",
+        example = json!({ "keep_alive": "10m" })
+    ),
+    responses(
+        (
+            status = 200,
+            description = "Model successfully loaded into memory",
+            body = api::LoadModelResponse,
+            content_type = "application/json",
+        ),
+        (
+            status = 400,
+            description = "Invalid or missing keep_alive format",
+            body = api::ErrorResponse,
+            examples(
+                ("Missing" = (value = json!({ "error": "keep alive is required and cannot be empty" }))),
+                ("Invalid" = (value = json!({ "error": "invalid keep_alive '10x' — use 30s / 10m / 2h, a plain integer, or -1" })))
+            )
+        ),
+        (
+            status = 404,
+            description = "Model not found locally",
+            body = api::ErrorResponse,
+            example = json!({ "error": "model 'llama3' not found — run `ollama pull llama3`" })
+        ),
+        (
+            status = 500,
+            description = "Internal server error (Ollama or network failure)",
+            body = api::ErrorResponse,
+            example = json!({ "error": "connection refused" })
+        )
+    )
+)]
 pub async fn load_model(
     State(state): State<AppState>,
     Path(model): Path<String>,
@@ -29,6 +91,34 @@ pub async fn load_model(
     Ok(Json(response))
 }
 
+#[utoipa::path(
+    delete,
+    path = "/models/{model}/load",
+    tag = "models",
+    params(
+        ("model" = String, Path, description = "Name of the model to unload from memory (e.g. 'llama3')")
+    ),
+    responses(
+        (
+            status = 200,
+            description = "Model successfully unloaded from memory",
+            body = api::UnloadModelResponse,
+            content_type = "application/json",
+        ),
+        (
+            status = 404,
+            description = "Model not found locally",
+            body = api::ErrorResponse,
+            example = json!({ "error": "model 'llama3' not found — run `ollama pull llama3`" })
+        ),
+        (
+            status = 500,
+            description = "Internal server error (Ollama or network failure)",
+            body = api::ErrorResponse,
+            example = json!({ "error": "connection refused" })
+        )
+    )
+)]
 pub async fn unload_model(
     State(state): State<AppState>,
     Path(model): Path<String>,