Support protocol thread configuration

src/meeting_lab/llm/ollama.py

@@ -62,6 +62,7 @@ def generate_once(
     timeout: int,
     num_ctx: int,
     num_predict: int,
+    num_thread: int | None = None,
 ) -> OllamaGeneration:
     payload = {
         "model": model,
@@ -74,6 +75,10 @@ def generate_once(
             "num_predict": num_predict,
         },
     }
+    if num_thread is not None:
+        if type(num_thread) is not int or num_thread <= 0:
+            raise ValueError("Protocol thread count must be a positive integer.")
+        payload["options"]["num_thread"] = num_thread
     started = time.perf_counter()
     try:
         response = requests.post(generate_url(endpoint), json=payload, timeout=timeout)