app.py

import os
from openai import OpenAI
from gradio import Server
from fastapi.responses import HTMLResponse


client = OpenAI(
    base_url="https://openrouter.ai/api/v1",
    api_key=os.environ.get("OPENROUTER_API_KEY", ""),
)

MODEL = "poolside/laguna-xs.2:free"

app = Server()


@app.api(name="chat")
def chat(message: str, history: list, enable_thinking: bool = True, max_tokens: int = 1024, temperature: float = 0.7) -> dict:
    """Chat with Laguna-XS.2 via OpenRouter. Supports reasoning/thinking mode."""
    messages = []
    for msg in history:
        entry = {"role": msg["role"], "content": msg["content"]}
        # Preserve reasoning_details from prior assistant turns for multi-turn reasoning
        if msg["role"] == "assistant" and msg.get("reasoning_details"):
            entry["reasoning_details"] = msg["reasoning_details"]
        messages.append(entry)
    messages.append({"role": "user", "content": message})

    extra_body = {}
    if enable_thinking:
        extra_body["reasoning"] = {"enabled": True}

    response = client.chat.completions.create(
        model=MODEL,
        messages=messages,
        max_tokens=max_tokens,
        temperature=temperature,
        extra_body=extra_body if extra_body else None,
    )

    choice = response.choices[0].message

    result = {
        "content": choice.content or "",
        "reasoning_details": None,
    }

    # Extract reasoning details if present
    if hasattr(choice, "reasoning_details") and choice.reasoning_details:
        result["reasoning_details"] = choice.reasoning_details

    return result


@app.get("/")
async def homepage():
    html_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "index.html")
    with open(html_path, "r", encoding="utf-8") as f:
        return HTMLResponse(content=f.read())


app.launch(show_error=True)