llm.py

from __future__ import annotations

import json
from typing import Any, Dict, List, Optional
import requests


class LLMError(RuntimeError):
    pass


def chat_completion(
    api_key: str,
    messages: List[Dict[str, str]],
    model: str = "gpt-4o-mini",
    base_url: str = "https://api.openai.com",
    timeout_s: int = 45,
) -> str:
    """
    Minimal OpenAI-compatible Chat Completions call.
    Expects provider that supports:
      POST {base_url}/v1/chat/completions
    """
    if not api_key or not api_key.strip():
        raise LLMError("LLM API key is empty.")

    url = base_url.rstrip("/") + "/v1/chat/completions"
    headers = {
        "Authorization": f"Bearer {api_key.strip()}",
        "Content-Type": "application/json",
    }
    payload: Dict[str, Any] = {
        "model": model,
        "messages": messages,
        "temperature": 0.3,
    }

    try:
        r = requests.post(url, headers=headers, json=payload, timeout=timeout_s)
    except Exception as e:
        raise LLMError(f"Request failed: {e}") from e

    if r.status_code >= 400:
        raise LLMError(f"LLM HTTP {r.status_code}: {r.text[:400]}")

    data = r.json()
    try:
        return data["choices"][0]["message"]["content"]
    except Exception:
        raise LLMError(f"Unexpected LLM response shape: {json.dumps(data)[:600]}")