import httpx

SYSTEM_PROMPT = (
    "Du bist Hermes, ein hilfreicher KI-Assistent. "
    "Antworte kurz und präzise, da deine Antwort vorgelesen wird. "
    "Maximal 3-4 Sätze, kein Markdown."
)
REQUEST_TIMEOUT = 30


def ask(text: str, screenshot_b64: str | None, settings: dict) -> str:
    """Schickt Text (+ optionalen Screenshot) an MC2 und gibt die Antwort zurück."""
    use_vision = screenshot_b64 is not None
    model = settings.get("vision_model", "Qwen3-VL-2B-Instruct") if use_vision \
        else settings.get("chat_model", "Hermes-4-14B")
    base_url = settings.get("mc2_url", "http://192.168.178.151:9001/v1")
    max_tokens = int(settings.get("max_response_tokens", 200))

    if use_vision:
        user_content = [
            {"type": "text", "text": text},
            {"type": "image_url", "image_url": {
                "url": f"data:image/jpeg;base64,{screenshot_b64}"
            }},
        ]
    else:
        user_content = text

    payload = {
        "model": model,
        "messages": [
            {"role": "system", "content": SYSTEM_PROMPT},
            {"role": "user", "content": user_content},
        ],
        "max_tokens": max_tokens,
        "stream": False,
    }

    try:
        with httpx.Client(timeout=REQUEST_TIMEOUT) as client:
            response = client.post(f"{base_url}/chat/completions", json=payload)
            response.raise_for_status()
            return response.json()["choices"][0]["message"]["content"].strip()
    except httpx.TimeoutException:
        return "Entschuldigung, die Antwort hat zu lange gedauert."
    except Exception as e:
        return f"Verbindungsfehler: {e}"