Cleanup: stale enabled models, refresh_items scheduling, README auto-config docs

backend/workers/ollama_manager.py

@@ -104,6 +104,14 @@ def refresh_models(self):
             self.config.set("llm.available_models", available_models)
             self.log.debug(f"OllamaManager: refreshed model list ({len(available_models)} models)")
 
+            # Reconcile enabled models: remove any that are no longer available
+            enabled_models = self.config.get("llm.enabled_models", [])
+            reconciled = [m for m in enabled_models if m in available_models]
+            if len(reconciled) != len(enabled_models):
+                removed = set(enabled_models) - set(reconciled)
+                self.log.info(f"OllamaManager: removed stale enabled model(s): {', '.join(removed)}")
+                self.config.set("llm.enabled_models", reconciled)
+
         except requests.RequestException as e:
             self.log.warning(f"OllamaManager: could not refresh model list - request error: {e}")
 

backend/workers/refresh_items.py

@@ -13,15 +13,14 @@ class ItemUpdater(BasicWorker):
     type = "refresh-items"
     max_workers = 1
 
-    @classmethod
-    def ensure_job(cls, config=None):
-        """
-        Ensure that the refresher is always running
-
-        :return:  Job parameters for the worker
-        """
-        return {"remote_id": "refresh-items", "interval": 60}
+    # ensure_job is intentionally disabled: this worker currently does nothing
+    # and would only create unnecessary job queue churn. Re-enable when work()
+    # has actual tasks to perform.
+    # @classmethod
+    # def ensure_job(cls, config=None):
+    #     return {"remote_id": "refresh-items", "interval": 60}
 
     def work(self):
+        # Placeholder – no tasks implemented yet.
         self.job.finish()
 

docker/README.md

@@ -63,8 +63,17 @@ the host via the exposed port).
 
 ### Configuring 4CAT to use Ollama
 
+#### Automatic configuration (fresh Docker install with sidecar)
+
+When you start 4CAT for the first time using the Ollama override file, the
+`docker_setup.py` initialisation script automatically detects the `ollama`
+sidecar and sets **LLM Provider Type**, **LLM Server URL**, and **LLM Access**
+for you. You can skip to step 2 below.
+
+#### Manual configuration (or to verify/change settings)
+
 1. Log in as admin and open **Control Panel → Settings**.
-2. Set the following LLM fields:
+2. Confirm or set the following LLM fields:
 
    | Setting | Value |
    |---|---|
@@ -102,3 +111,7 @@ point 4CAT directly at that server:
 - **On the same host**: use `http://host.docker.internal:11434` as the LLM Server URL.
 - **Remote server**: use the server's reachable URL and configure any required
   API key in the *LLM Server API Key* and *LLM Server Authentication Type* settings.
+
+In both cases, configure the LLM settings manually via **Control Panel → Settings**
+(see *Manual configuration* above), using the appropriate server URL instead of
+`http://ollama:11434`.