from fastapi import FastAPI, File, UploadFile, Form
from pydantic import BaseModel
import numpy as np
from tensorflow.keras.models import load_model
import PIL.Image
import io
import pandas as pd
import os

app = FastAPI()


model = load_model("foodnutrition.keras", compile=False)

CLASS_NAMES = [
    'Ayam Bakar', 'Ayam Geprek', 'Ayam Goreng', 'Ayam Tepung', 'Bakso', 'Bakwan', 
    'Batagor', 'Bolu Kukus', 'Burger', 'Cap Cay', 'Chicken Katsu', 'Cilok', 
    'Dimsum', 'Donat', 'Es Teh', 'Gado-Gado', 'Ikan Goreng', 'Jasuke', 
    'Jus Jeruk', 'Kikil', 'Kopi', 'Krepes', 'Lemper', 'Mie Ayam', 
    'Mie Instan', 'Nasi Goreng', 'Nasi Padang', 'Pecel Lele', 'Pizza', 
    'Random Image', 'Rendang', 'Risoles', 'Sate Ayam', 'Seblak', 
    'Siomay', 'Sop', 'Soto', 'Tahu', 'Tangkil', 'Telur Balado', 
    'Telur Dadar', 'Tempe', 'Thai Tea'
]

# Muat file CSV
nutrisi_csv = pd.read_csv('nutrisi_origin.csv')

KOLOM_NAMA = nutrisi_csv.columns[1]

class FoodRequest(BaseModel):
    nama_makanan: str

def get_nutrition_data(food_name: str):
    pencarian = nutrisi_csv[nutrisi_csv[KOLOM_NAMA].str.lower() == food_name.lower()]
    
    kalori, protein, lemak, karbo = 0.0, 0.0, 0.0, 0.0
    kategori = "Tidak Diketahui"

    if not pencarian.empty:
        detail_nutrisi = pencarian.iloc[0]
        
        kal_teks = str(detail_nutrisi.get('Kalori', '0')).lower().replace(',', '.').replace('g', '').strip()
        pro_teks = str(detail_nutrisi.get('Protein', '0')).lower().replace(',', '.').replace('g', '').strip()
        lem_teks = str(detail_nutrisi.get('Lemak', '0')).lower().replace(',', '.').replace('g', '').strip()
        karbo_teks = str(detail_nutrisi.get('Karbohidrat', '0')).lower().replace(',', '.').replace('g', '').strip()
        
        if 'Kategori' in detail_nutrisi:
            kategori = str(detail_nutrisi['Kategori'])

        try: kalori = float(kal_teks)
        except: kalori = 0.0
        
        try: protein = float(pro_teks)
        except: protein = 0.0
        
        try: lemak = float(lem_teks)
        except: lemak = 0.0
        
        try: karbo = float(karbo_teks)
        except: karbo = 0.0

    return {
        "kalori": kalori,
        "protein": protein,
        "lemak": lemak,
        "karbohidrat": karbo,
        "kategori": kategori
    }

@app.get('/')
async def hello():
    return {"message": "Food Nutrition API Ready"}

@app.post('/predict')
async def predict(file: UploadFile = File(...)):
    img_bytes = await file.read()
    img = PIL.Image.open(io.BytesIO(img_bytes)).convert('RGB')
    img = img.resize((244, 244))

    img_array = np.array(img).astype("float32") / 255.0
    img_array = np.expand_dims(img_array, axis=0)

    preds = model.predict(img_array, verbose=0)
    index = np.argmax(preds[0])

    confidence = round(float(np.max(preds[0])) * 100, 2)

    if confidence >= 90.0:
        predicted_label = CLASS_NAMES[index]
    else:
        predicted_label = "Tidak diketahui"

    probability = []
    for i, prob in enumerate(preds[0]):
        probability.append({
            "label": CLASS_NAMES[i],
            "score": round(float(prob) * 100, 2)
        })

    probability = sorted(probability, key=lambda x: x["score"], reverse=True)

    data_nutrisi = get_nutrition_data(predicted_label)

    return {
        "prediction": predicted_label,
        "confidence": confidence,
        "probability": probability,
        "nutrisi": data_nutrisi
    }