import re import base64 from PIL import Image from io import BytesIO import numpy as np from fastapi import FastAPI from pydantic import BaseModel from keras.models import load_model from keras.utils import img_to_array from tensorflow.keras.utils import load_img, img_to_array import uvicorn from keras.layers import BatchNormalization # Memuat model Keras dengan menangani objek kustom try: model = load_model('modelMobileNetNew.h5', custom_objects={'BatchNormalization': BatchNormalization}) except Exception as e: print(f"Error loading model: {e}") app = FastAPI() class ImageData(BaseModel): message: str def base64_to_pil(img_base64): image_data = re.sub('data:image/.+;base64,', '', str(img_base64)) pil_image = Image.open(BytesIO(base64.b64decode(image_data))) return pil_image def model_predict(img, model): img = img.resize((224, 224)) x = img_to_array(img) x = x.reshape(-1, 224, 224, 3) x = x.astype('float32') x = x / 255.0 preds = model.predict(x) return preds target_names = ['australopithecus_aferensis', 'badak_bercula_satu', 'batu_lempung_krisikan', 'fauna_darat_semedo', 'fragmen_tengkorak', 'fragmen_tengkorak_parential', 'fragmen_tengkorak_sambung_macan', 'fragmen_tulang_fermur_bawah', 'gigi_buaya', 'gigi_gajah', 'gigi_hiu', 'homo_erectus_ngawi', 'homo_habilis', 'homo_sapiens_wajak', 'kapak_penetak', 'kapak_primbas'] display_names = ['Australopithecus Aferensis', 'Badak Bercula Satu', 'Batu Lempung Krisikan', 'Fauna Darat Semedo', 'Fragmen Tengkorak', 'Fragmen Tengkorak Parential', 'Fragmen Tengkorak Sambung_macan', 'Fragmen Tulang Fermur Bawah', 'Gigi Buaya', 'Gigi Gajah', 'Gigi Hiu', 'Homo Erectus Ngawi', 'Homo Habilis', 'Homo Sapiens Wajak', 'Kapak Penetak', 'Kapak Primbas'] label_mapping = dict(zip(target_names, display_names)) @app.get("/hello") async def say_hello(): return "hello" @app.post("/predict") async def predict(image_data: ImageData): try: data = image_data.message img = base64_to_pil(data) pred = model_predict(img, model) hasil_label = label_mapping[target_names[np.argmax(pred)]] hasil_prob = "{:.2f}".format(100 * np.max(pred)) return {'message': 'Berhasil melakukan prediksi', 'nama': hasil_label, 'probability': hasil_prob} except Exception as e: print(f"EXCEPTION in Predict POST: {e}") return {'message': str(e), 'nama': '', 'probability': ''} if __name__ == "__main__": uvicorn.run(app, host="0.0.0.0", port=7860)