import tensorflow as tf
import pandas as pd
import numpy as np

mnist = pd.read_csv('../input/train.csv')
mnist_test = pd.read_csv('../input/test.csv')

x = tf.placeholder(tf.float32, [None, 784])
y_ = tf.placeholder(tf.float32, [None, 10])
#output_y = tf.placeholder(tf.int32, [None, 1])

def weight_variable(shape):
    initial = tf.truncated_normal(shape, stddev = 0.1)
    return tf.Variable(initial)

def bias_variable(shape):
    initial = tf.constant(0.1, shape=shape)
    return tf.Variable(initial)
    
def conv2d(x, W):
    return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')

def max_pool(x):
    return tf.nn.max_pool(x, ksize=[1,2,2,1], strides=[1,2,2,1], padding='SAME')
    
#initailizing the weight and bias variables
W_conv1 = weight_variable([5, 5, 1, 32])
b_conv1 = bias_variable([32])
#reshaping the input variable to a 4d tensor
x_image = tf.reshape(x, [-1, 28, 28, 1])
#convolution and max pooling
h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)
h_pool1 = max_pool(h_conv1)

#initializing the weight and bias variables
W_conv2 = weight_variable([5, 5, 32, 64])
b_conv2 = bias_variable([64])

#convolution and max pooling
h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2 = max_pool(h_conv2)

W_fc1 = weight_variable([7*7*64, 1024])
b_fc1 = bias_variable([1024])

h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)

#prob that neuron's output should be kept during dropout
keep_prob = tf.placeholder(tf.float32)
h_fc1_dropout = tf.nn.dropout(h_fc1, keep_prob)

W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])

y_conv = tf.matmul(h_fc1_dropout, W_fc2) + b_fc2

#Function to randomly select batches of training data
def next_batch(num):
    l = np.random.randint(0, 42000, num)
    return (mnist.iloc[l, 1:], np.eye(10)[mnist.iloc[l, 0]])
    
cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y_conv))
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)
predictions = tf.equal(tf.arg_max(y_, 1), tf.arg_max(y_conv, 1))
accuracy = tf.reduce_mean(tf.cast(predictions, tf.float32))
result = tf.arg_max(y_conv, 1)
sess = tf.Session()
sess.run(tf.global_variables_initializer())
with sess.as_default():
    for i in range(3500):
        batch = next_batch(50)
        if i%100 == 0:
            train_accuracy = accuracy.eval(feed_dict={x:batch[0], y_:batch[1], keep_prob:0.5})
            print ("step %d has accuracy %g"%(i, train_accuracy))
        train_step.run(feed_dict={x:batch[0], y_:batch[1], keep_prob:0.5})
    low = 0
    arr = []
    while (low <= (28000-50)):
        arr.extend(result.eval(feed_dict={x:mnist_test.iloc[low:low+50, :], keep_prob:1.0}).tolist())
        #print("Step: \n" + str(low))
        low += 50
    
    y_test_labels = pd.DataFrame({"ImageId": list(range(1,len(arr)+1)),'Label': arr}).to_csv(index=False, header=True)
    
    print (y_test_labels)
    
    #y_output = tf.constant([tf.range(0, len(arr)+1, dtype=tf.int32), tf.constant(arr)], shape = [None, 2])
    
    tf.write_file(filename='output.csv', contents=y_test_labels)