import numpy as np
from scipy.special import expit
from utils import random_array
from dataset_utils import DatasetUtils

CYRILLIC_ALPHABET = ['I', 'А', 'Б', 'В', 'Г', 'Д', 'Е', 'Ë', 'Ж', 'З',
                     'И', 'Й', 'К', 'Л', 'М', 'Н', 'О', 'П', 'Р', 'С',
                     'Т', 'У', 'Ф', 'Х', 'Ц', 'Ч', 'Ш', 'Щ', 'Ъ', 'Ы',
                     'Ь', 'Э', 'Ю', 'Я']

"""The neural network class."""
class NeuralNetwork:
    def __init__(self, learning_rate: float, input_resolution: int) -> None:
        self.learning_rate = learning_rate
        self.input_layer_size = input_resolution
        self.hidden_layer_size = len(CYRILLIC_ALPHABET)
        self._hidden_weights = np.random.rand(input_resolution, self.hidden_layer_size)
        self._output_weights = np.random.rand(input_resolution)


    """
    Train the neural network. It loads the dataset contained in ./data,
    converts each image into a numpy array and uses that data for training.
    Algorithm:
    1-. Feedforward the  input matrix.
    2-. Calculate the output layer error.
    3-. Adjust the weights of the output layer via gradient descent.
    3-. Calculate the hidden layer error by backpropagating the output layer
        error.
    4-. Adjust the weights of the hidden layer via gradient descent.
    """
    def train(self):
        pass

    """
    Guess the letter contained in the image file pointed by
    input_image (a path).
    """
    def guess(self, input_image: np.array) -> str:
        output_layer = self._predict(input_image)

        return self.out(output_layer)

    """
    Save the weights to a csv file.
    """
    def save(self):
        pass

    """
    Load the weights from a csv file.
    """
    def load(self, weights_file: str):
        pass

    """
    Feedforwarding.
    """
    def _predict(self, input_layer: np.array):
        hidden_layer_inputs = np.dot(input_layer, self._hidden_weights)
        hidden_layer_outputs = self._get_layer_output(hidden_layer_inputs)
        output_layer_inputs = np.dot(self._output_weights, hidden_layer_outputs)
        
        # The output layer outputs. (Final output of the neural network).
        return self._get_layer_output(output_layer_inputs)

    """
    Apply the sigmoid function to a given layer
    """
    def _get_layer_output(self, layer: np.array):
        return expit(layer)

    """
    Get the result from a sigmoid matrix (the index with the highest chance
    of being the correct answer).
    """
    def out(self, output_layer: np.array):
        return CYRILLIC_ALPHABET[np.argmax(np.transpose(output_layer))]