#!/usr/bin/python

################################################################################

##

## Copyright 2016 Udara Karunarathna (IT13021030) and Supun Sudaraka (IT13019914). All rights reserved.

##

################################################################################

import numpy as np

from scipy import optimize

from StringIO import StringIO

import matplotlib.pyplot as plot

from mpl_toolkits.mplot3d import Axes3D

from matplotlib import cm

from neural_network import Neural_Network

from trainer import Trainer

NeuralNetwork = Neural_Network()

#Pass to Trainer Data:

X,Y = NeuralNetwork.readInputFile("input.txt")

Y = np.reshape(Y, (3,1))

print ("Input : ")

print (X)

print ("Results : ")

print (Y)

#normalized training data

X = NeuralNetwork.normalize(X)

Y = Y/100

print ("Input : ")

print (X)

print ("Results :")

print (Y)

TrainObject = Trainer(NeuralNetwork)

x, y, z = TrainObject.train(X,Y)

yHat = NeuralNetwork.forward(X)

print("YHat")

print(yHat)

print("Y")

print(Y)

print(x.shape, y.shape, z.shape)

#to take on 1000 rows of datav(limited no of rows)

x = np.resize(x, 1000)

y = np.resize(y, 1000)

z = np.resize(z, 1000)

#plot wireframe

f = plot.figure()

a = f.add_subplot(111, projection='3d')

a.plot_wireframe(x,y,z)

a.set_xlabel('W1 Values')

a.set_ylabel('W2 Values')

a.set_zlabel('W3 Values')

plot.show()