import os

import numpy as np
import pandas as pd
from keras.models import load_model
import tensorflow as tf

from custom_keras_utils import normalize_timeseries
from preprocess_test_samples import load_data_to_datasets


ROOT_PATH = os.getcwd() + '\\test_samples_0'
CLASS_NAME = 'Расслоение'
CLASS_MAPPING = {
    'ВИ_БезДефекта': 0,
    'ВИ_Дефект': 1,
    'Инородное включение': 2,
    'Непроклей': 3,
    'Непропитка': 4,
    'Пресс_БезДефекта': 5,
    'Пресс_Дефект': 6,
    'Расслоение': 7
}
MODEL_NAME = 'resnet'
MODEL_MAPPING = {
    'lstm_fcn': 1,
    'mlp': 2,
    'resnet': 3
}


with tf.device('/GPU:0'):
    model = load_model(MODEL_NAME + '.h5')
    load_data_to_datasets(ROOT_PATH)

    X_test = pd.read_csv(ROOT_PATH + '\\test_x.csv', header=None, delimiter=' ')
    X_test = normalize_timeseries(X_test).values
    if MODEL_MAPPING[MODEL_NAME] == 3:
        X_test = np.reshape(X_test, (X_test.shape[0], 1, 1, X_test.shape[1]))
    else:
        X_test = np.reshape(X_test, (X_test.shape[0], 1, X_test.shape[1]))

    # Make predictions
    predictions = model.predict(X_test)

    # Convert predictions to class labels
    class_labels = [np.argmax(prediction) for prediction in predictions]

    false_predict = 0.0
    all_predict = 0
    for label in class_labels:
        if label != CLASS_MAPPING[CLASS_NAME]:
            false_predict += 1
        all_predict += 1
    print('Accuracy: {}%'.format((all_predict - false_predict) / all_predict * 100))

    os.remove(ROOT_PATH + '\\test_x.csv')