from fastapi import FastAPI, File, UploadFile, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from PIL import Image
import io
import torch
from transformers import AutoImageProcessor, AutoModelForImageClassification
import numpy as np

app = FastAPI()

# CORS ayarları - Tüm kaynaklardan gelen isteklere izin ver
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

# Model ve processor'ı yükle
MODEL_NAME = "Pavarissy/ConvNextV2-large-DogBreed"
print(f"🔄 Loading model: {MODEL_NAME}")

try:
    processor = AutoImageProcessor.from_pretrained(MODEL_NAME)
    model = AutoModelForImageClassification.from_pretrained(MODEL_NAME)
    model.eval()
    print("✅ Model loaded successfully!")
except Exception as e:
    print(f"❌ Error loading model: {e}")
    raise e

@app.get("/")
async def root():
    """Ana endpoint - API durumunu gösterir"""
    return {
        "message": "Dog Breed Classification API",
        "model": MODEL_NAME,
        "status": "ready",
        "endpoints": {
            "predict": "/predict_pet (POST)",
            "health": "/health (GET)"
        }
    }

@app.post("/predict_pet")
async def predict_pet(image: UploadFile = File(...)):
    """
    Pet (köpek ırkı) tahmini endpoint'i

    Expected response format:
    {
        "predicted_label": "n02085620-Chihuahua",
        "confidence": 0.95,
        "detection": {
            "box": {"x": 50, "y": 50, "width": 400, "height": 400}
        }
    }
    """
    try:
        # Resmi oku ve RGB'ye çevir
        image_bytes = await image.read()
        img = Image.open(io.BytesIO(image_bytes)).convert('RGB')

        # Orijinal görüntü boyutları
        width, height = img.size
        print(f"📸 Image received: {width}x{height}")

        # Model için preprocessing
        inputs = processor(images=img, return_tensors="pt")

        # Tahmin yap
        with torch.no_grad():
            outputs = model(**inputs)
            logits = outputs.logits

        # Softmax ile olasılıkları hesapla
        probabilities = torch.nn.functional.softmax(logits, dim=-1)
        confidence, predicted_idx = torch.max(probabilities, dim=-1)

        # Tahmin edilen sınıfı al
        predicted_label = model.config.id2label[predicted_idx.item()]
        confidence_score = confidence.item()

        # Basit bir detection box oluştur (görüntünün %80'i merkeze yerleştirilmiş)
        box_margin = 0.1
        detection_box = {
            "x": float(width * box_margin),
            "y": float(height * box_margin),
            "width": float(width * (1 - 2 * box_margin)),
            "height": float(height * (1 - 2 * box_margin))
        }

        # Yanıt hazırla (React Native app'in beklediği formatta)
        response = {
            "predicted_label": predicted_label,
            "confidence": float(confidence_score),
            "detection": {
                "box": detection_box
            },
            "imageDimensions": {
                "width": width,
                "height": height
            }
        }

        print(f"✅ Prediction: {predicted_label} (confidence: {confidence_score:.4f})")
        return response

    except Exception as e:
        print(f"❌ Error during prediction: {str(e)}")
        raise HTTPException(status_code=500, detail=f"Prediction error: {str(e)}")

@app.get("/health")
async def health_check():
    """Sağlık kontrolü endpoint'i"""
    return {
        "status": "healthy",
        "model_loaded": model is not None,
        "model_name": MODEL_NAME
    }

# Hugging Face Space'te uvicorn otomatik çalışır
if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=7860)