import numpy
import matplotlib.pyplot
import scipy.special
%matplotlib inline

class neuralNetwork:
    def __init__(self, inputnodes, hiddennodes, outputnodes, learningrate):  # 初始化神经网络
        self.inodes = inputnodes
        self.hnodes = hiddennodes
        self.onodes = outputnodes
        self.lr = learningrate  # 学习率
        # from node i to node j in the next layer
        self.wih = numpy.random.normal(0.0, pow(self.hnodes, -0.5), (self.hnodes, self.inodes))  # 权重矩阵
        self.who = numpy.random.normal(0.0, pow(self.onodes, -0.5), (self.onodes, self.hnodes))
        self.activation_function = lambda x: scipy.special.expit(x)  # 应用激活函数

    def train(self, inputs_list, targets_list):
        inputs = numpy.array(inputs_list, ndmin=2).T
        targets = numpy.array(targets_list, ndmin=2).T
        hidden_inputs = numpy.dot(self.wih, inputs)
        hidden_outputs = self.activation_function(hidden_inputs)
        final_inputs = numpy.dot(self.who, hidden_outputs)
        final_outputs = self.activation_function(final_inputs)

        output_errors = targets - final_outputs
        hidden_errors = numpy.dot(self.who.T, output_errors)

        self.who += self.lr*numpy.dot((output_errors*final_outputs*(1.0-final_outputs)), numpy.transpose(hidden_outputs))  # 调整隐藏层与输出层之间的权重
        self.wih += self.lr*numpy.dot((hidden_errors*hidden_outputs*(1.0-hidden_outputs)), numpy.transpose(inputs))  # 调整输入层与隐藏层之间的权重

    def query(self, inputs_list):
        inputs = numpy.array(inputs_list, ndmin=2).T  # 输入矩阵的转置矩阵
        hidden_inputs = numpy.dot(self.wih, inputs)
        hidden_outputs = self.activation_function(hidden_inputs)
        final_inputs = numpy.dot(self.who, hidden_outputs)
        final_outputs = self.activation_function(final_inputs)
        return final_outputs

input_nodes = 784  # 28*28，输入层节点数
hidden_nodes = 100  # 隐藏层节点数
output_nodes = 10  # 输出层节点

learning_rate = 0.2  # 学习率

# 生成一个神经网络实例
n = neuralNetwork(input_nodes,hidden_nodes,output_nodes,learning_rate)

# 加载训练数据
training_data_file = open("mnist_dataset/mnist_train.csv", 'r')
training_data_list = training_data_file.readlines()
training_data_file.close()

# 训练神经网络
epochs = 5  # 训练5个世代(本篇的学习率为0.2，训练数据为六万组，训练五个世代，测试数据为一万组)
for e in range(epochs):
    for record in training_data_list:
        all_values = record.split(',')
        inputs = (numpy.asfarray(all_values[1:])/255.0*0.99)+0.01
        targets = numpy.zeros(output_nodes) + 0.01
        targets[int(all_values[0])] = 0.99
        n.train(inputs,targets)

# 加载测试数据(本篇使用了一万组测试数据，正确率为95.11%，结果可能不大一致，但大概都在94%~95%)
test_data_file = open("mnist_dataset/mnist_test.csv", 'r')
test_data_list = test_data_file.readlines()
test_data_file.close()

# 测试神经网络
scorecard = []
for record in test_data_list:
    all_values = record.split(',')
    correct_label = int(all_values[0])
    # print(correct_label, "correct lable")
    inputs = (numpy.asfarray(all_values[1:])/255.0*0.99)+0.01
    outputs = n.query(inputs)
    label = numpy.argmax(outputs)
    # print(label, "networks's answer")
    if(label == correct_label):
        scorecard.append(1)
    else:
        scorecard.append(0)
        pass
    pass
# print(scorecard)  # 输出训练成绩
scorecard_array = numpy.asarray(scorecard)
print("performances = ", scorecard_array.sum()*1.0/scorecard_array.size)