import os
from fastapi import FastAPI, File, UploadFile
import tensorflow as tf
import numpy as np
from PIL import Image
import io

app = FastAPI()

ORIGINAL_FILE = "priority_model.bin"
MODEL_H5 = "model_fix.h5"

model = None
error_msg = "Starting..."

@app.on_event("startup")
def load_model():
    global model, error_msg
    try:
        # تحويل bin لـ h5 عشان نهرب من الـ Xet ونرضي TensorFlow 2.15
        if os.path.exists(ORIGINAL_FILE):
            if os.path.exists(MODEL_H5): os.remove(MODEL_H5)
            os.rename(ORIGINAL_FILE, MODEL_H5)
        
        if os.path.exists(MODEL_H5):
            model = tf.keras.models.load_model(MODEL_H5, compile=False)
            error_msg = "None"
        else:
            error_msg = "Model file not found"
    except Exception as e:
        error_msg = str(e)

@app.get("/")
def home():
    return {
        "model_loaded": model is not None,
        "tf_version": tf.__version__,
        "error": error_msg
    }

@app.post("/predict")
async def predict(file: UploadFile = File(...)):
    if model is None: return {"error": "Model not loaded"}
    contents = await file.read()
    image = Image.open(io.BytesIO(contents)).convert("RGB").resize((224, 224))
    img_array = np.array(image) / 255.0
    img_array = np.expand_dims(img_array, axis=0).astype(np.float32)
    predictions = model.predict(img_array)
    classes = ["High Priority", "Low Priority", "Medium Priority"]
    idx = np.argmax(predictions[0])
    return {"prediction": classes[idx], "confidence": f"{float(predictions[0][idx])*100:.2f}%"}