server.py

"""
server.py — Phase 2 backend service for the CURT inventory assistant.

FastAPI server exposing:
    POST /chat       → {session_id, message} → {session_id, reply}
    GET  /inventory  → current inventory as JSON
    GET  /health     → liveness check

The LLM (Groq gpt-oss-20b) decides which tools to call; tools are executed
against the real SQLite database. Conversation memory is kept per session.

Run with:
    .venv/Scripts/python.exe -m uvicorn server:app --reload --port 8000
"""

from contextlib import asynccontextmanager

from fastapi import FastAPI, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel, Field

import database as db
import llm_assistant


@asynccontextmanager
async def lifespan(app: FastAPI):
    # Make sure the database exists and is seeded before serving requests
    db.init_db()
    db.seed_db()
    yield


app = FastAPI(
    title="CURT Inventory Assistant API",
    description="LLM-powered inventory assistant with function calling (Phase 2).",
    version="2.0.0",
    lifespan=lifespan,
)

# Allow the Streamlit app (and local tools) to call the API
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_methods=["*"],
    allow_headers=["*"],
)


# --------------------------------------------------------------------------
# Schemas
# --------------------------------------------------------------------------

class ChatRequest(BaseModel):
    session_id: str = Field(..., min_length=1, description="Conversation/session id")
    message: str = Field(..., min_length=1, description="The user's question")


class ChatResponse(BaseModel):
    session_id: str
    reply: str


# --------------------------------------------------------------------------
# Endpoints
# --------------------------------------------------------------------------

@app.post("/chat", response_model=ChatResponse)
def chat(req: ChatRequest) -> ChatResponse:
    """Send a user message; get the assistant's reply. Memory is per session_id."""
    try:
        reply = llm_assistant.chat(req.session_id, req.message)
    except Exception as exc:
        raise HTTPException(status_code=502, detail=f"LLM backend error: {exc}") from exc
    return ChatResponse(session_id=req.session_id, reply=reply)


@app.get("/inventory")
def inventory() -> dict:
    """Return the current inventory as JSON."""
    parts = db.get_all_parts()
    return {
        "count": len(parts),
        "total_units": sum(p["quantity"] for p in parts),
        "parts": parts,
    }


@app.get("/health")
def health() -> dict:
    return {"status": "ok", "model": llm_assistant.MODEL}


@app.post("/reset/{session_id}")
def reset(session_id: str) -> dict:
    """Forget a session's conversation history (handy while demoing)."""
    llm_assistant.reset_session(session_id)
    return {"session_id": session_id, "reset": True}


if __name__ == "__main__":
    import uvicorn

    uvicorn.run(app, host="127.0.0.1", port=8000)