app.py

"""
Nexus-Core Inference API
13.2M parameter chess engine
"""

from fastapi import FastAPI, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel, Field
import time
import logging
from typing import Optional, List

from engine import NexusCoreEngine

# Logging
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)

# FastAPI
app = FastAPI(
    title="Nexus-Core Inference API",
    description="13.2M parameter Resnet-10 chess engine",
    version="2.0.0"
)

# CORS
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

# Global engine
engine = None


# Models
class MoveRequest(BaseModel):
    fen: str = Field(..., description="FEN notation")
    depth: Optional[int] = Field(5, ge=1, le=8, description="Search depth")
    time_limit: Optional[int] = Field(3000, ge=1000, le=15000, description="Time in ms")


class MoveResponse(BaseModel):
    best_move: str
    evaluation: float
    depth_searched: int
    seldepth: int
    nodes_evaluated: int
    time_taken: int
    nps: int
    pv: List[str]
    tt_hit_rate: Optional[float] = None


class HealthResponse(BaseModel):
    status: str
    model_loaded: bool
    version: str
    model_size_mb: Optional[float] = None


# Startup
@app.on_event("startup")
async def startup_event():
    global engine
    
    logger.info("🚀 Starting Nexus-Core API v2.0...")
    
    try:
        engine = NexusCoreEngine(
            model_path="/app/models/nexus-core.onnx",
            num_threads=2
        )
        logger.info("✅ Engine loaded")
        
    except Exception as e:
        logger.error(f"❌ Failed: {e}")
        raise


# Health
@app.get("/health", response_model=HealthResponse)
async def health_check():
    return {
        "status": "healthy" if engine else "unhealthy",
        "model_loaded": engine is not None,
        "version": "2.0.0",
        "model_size_mb": engine.get_model_size() if engine else None
    }


# Main endpoint
@app.post("/get-move", response_model=MoveResponse)
async def get_move(request: MoveRequest):
    if engine is None:
        raise HTTPException(status_code=503, detail="Engine not loaded")
    
    if not engine.validate_fen(request.fen):
        raise HTTPException(status_code=400, detail="Invalid FEN")
    
    start_time = time.time()
    
    try:
        result = engine.get_best_move(
            fen=request.fen,
            depth=request.depth,
            time_limit=request.time_limit
        )
        
        time_taken = int((time.time() - start_time) * 1000)
        
        logger.info(
            f"Move: {result['best_move']} | "
            f"Eval: {result['evaluation']:+.2f} | "
            f"Depth: {result['depth_searched']}/{result['seldepth']} | "
            f"Nodes: {result['nodes_evaluated']} | "
            f"NPS: {result['nps']}"
        )
        
        return MoveResponse(
            best_move=result['best_move'],
            evaluation=result['evaluation'],
            depth_searched=result['depth_searched'],
            seldepth=result['seldepth'],
            nodes_evaluated=result['nodes_evaluated'],
            time_taken=time_taken,
            nps=result['nps'],
            pv=result['pv'],
            tt_hit_rate=result['tt_stats']['hit_rate']
        )
        
    except Exception as e:
        logger.error(f"Error: {e}")
        raise HTTPException(status_code=500, detail=str(e))


# Root
@app.get("/")
async def root():
    return {
        "name": "Nexus-Core Inference API",
        "version": "2.0.0",
        "model": "13.2M parameters (Resnet-10)",
        "search": "PVS + NMP + LMR",
        "endpoints": {
            "POST /get-move": "Get best move",
            "GET /health": "Health check",
            "GET /docs": "Documentation"
        }
    }


if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=7860, log_level="info")