"""This tutorial introduces the LeNet5 neural network architecture
using Theano.  LeNet5 is a convolutional neural network, good for
classifying images. This tutorial shows how to build the architecture,
and comes with all the hyper-parameters you need to reproduce the
paper's MNIST results.


This implementation simplifies the model in the following ways:

 - LeNetConvPool doesn't implement location-specific gain and bias parameters
 - LeNetConvPool doesn't implement pooling by average, it implements pooling
   by max.
 - Digit classification is implemented with a logistic regression rather than
   an RBF network
 - LeNet5 was not fully-connected convolutions at second layer

References:
 - Y. LeCun, L. Bottou, Y. Bengio and P. Haffner:
   Gradient-Based Learning Applied to Document
   Recognition, Proceedings of the IEEE, 86(11):2278-2324, November 1998.
   http://yann.lecun.com/exdb/publis/pdf/lecun-98.pdf

"""
import cPickle
import gzip
import os
import sys
import time

import numpy

import theano
import theano.tensor as T
from theano.tensor.signal import downsample
from  theano.tensor.nnet.Conv3D import conv3D
from theano.tensor.nnet import conv
from theano.tensor.nnet import conv3d2d
from theano.sandbox.cuda import fftconv
from logistic_sgd import LogisticRegression, load_data
from mlp import HiddenLayer
mode_default = theano.compile.mode.get_default_mode()
mode_fft = mode_default.including('gpu', 'conv3d_fft', 'convgrad3d_fft','convtransp3d_fft')

class LeNetConvPoolLayer(object):
    """Pool Layer of a convolutional network """

    def __init__(self, rng, input, filter_shape, image_shape, poolsize,pad):
        """
        Allocate a LeNetConvPoolLayer with shared variable internal parameters.

        :type rng: numpy.random.RandomState
        :param rng: a random number generator used to initialize weights

        :type input: theano.tensor.dtensor4
        :param input: symbolic image tensor, of shape image_shape

        :type filter_shape: tuple or list of length 4
        :param filter_shape: (number of filters, num input feature maps,
                              filter height,filter width)

        :type image_shape: tuple or list of length 4
        :param image_shape: (batch size, num input feature maps,
                             image height, image width,)

        :type poolsize: tuple or list of length 2
        :param poolsize: the downsampling (pooling) factor (#rows,#cols)
        """

        assert image_shape[2] == filter_shape[2]
        self.input = input

        # there are "num input feature maps * filter height * filter width * filter depth"
        # inputs to each hidden unit
        fan_in = numpy.prod(filter_shape[2:])
        # each unit in the lower layer receives a gradient from:
        # "num output feature maps * filter height * filter width" /
        #   pooling size
        fan_out = (filter_shape[0] * numpy.prod(filter_shape[3:]) /
                   numpy.prod(poolsize))
        # initialize weights with random weights
        W_bound = numpy.sqrt(6. / (fan_in + fan_out))
        self.W = theano.shared(numpy.asarray(
            rng.uniform(low=-W_bound, high=W_bound, size=filter_shape),
            dtype=theano.config.floatX),
                               borrow=True)
    #print self.W
        # the bias is a 1D tensor -- one bias per output feature map
        b_values = numpy.zeros((filter_shape[0],), dtype=theano.config.floatX)
        self.b = theano.shared(value=b_values, borrow=True)

        # convolve input feature maps with filters
        d = theano.shared(numpy.ndarray(shape=(3,),dtype=int))
    d.set_value([1, 1, 1])
    #w1=numpy.transpose(self.W,(0,3,4,2,1))
    #self.b.dimshuffle('x', 0,'x', 'x', 'x')
    #bias2=self.b.transpose(0,2,3,4,1)
        conv_out = conv3d2d.conv3d(signals=input, filters=self.W, signals_shape=image_shape,filters_shape=filter_shape)
        #conv_out = conv.conv2d(input=input, filters=self.W,
                #filter_shape=filter_shape, image_shape=image_shape)

        # downsample each feature map individually, using maxpooling
        #conv_out_m=numpy.transpose(conv_out, (0, 4, 3, 1, 2))
        pooled_out = downsample.max_pool_2d(input=conv_out,ds=poolsize, ignore_border=True)
        #pooled_out_m = numpy.transpose(pooled_out, (0, 3, 4, 2, 1))

        # add the bias term. Since the bias is a vector (1D array), we first
        # reshape it to a tensor of shape (1,n_filters,1,1). Each bias will
        # thus be broadcasted across mini-batches and feature map
        # width & height
        self.output = T.tanh(pooled_out+self.b.dimshuffle('x','x',0,'x','x'))

        #print 'hi'


        # store parameters of this layer
        self.params = [self.W, self.b]


def evaluate_lenet5(learning_rate=0.01, n_epochs=10,
                    dataset='./kth_thesis.pkl',
                    nkerns=[5, 10, 128], batch_size=10):
    """ Demonstrates lenet on MNIST dataset

    :type learning_rate: float
    :param learning_rate: learning rate used (factor for the stochastic
                          gradient)

    :type n_epochs: int
    :param n_epochs: maximal number of epochs to run the optimizer

    :type dataset: string
    :param dataset: path to the dataset used for training /testing (MNIST here)

    :type nkerns: list of ints
    :param nkerns: number of kernels on each layer
    """

    rng = numpy.random.RandomState(23455)

    datasets = load_data(dataset)

    #print datasets[0]
    #print datasets[1]
    #print datasets[2]

    train_set_x, train_set_y = datasets[0]
    test_set_x, test_set_y = datasets[1]

    #print train_set_x
    #print train_set_y
    #print valid_set_x
    #print valid_set_y
    #print test_set_x
    #print test_set_y
    # compute number of minibatches for training, validation and testing
    n_train_batches = train_set_x.get_value(borrow=True).shape[0]
    n_test_batches = test_set_x.get_value(borrow=True).shape[0]
    n_train_batches /= batch_size
    n_test_batches /= batch_size

    # allocate symbolic variables for the data
    index = T.lscalar()  # index to a [mini]batch
    x = T.matrix('x')   # the data is presented as rasterized images
    y = T.ivector('y')  # the labels are presented as 1D vector of
                        # [int] labels

    ishape = (80, 60)  # this is the size of MNIST images

    ######################
    # BUILD ACTUAL MODEL #
    ######################
    print '... building the model'

    # Reshape matrix of rasterized images of shape (batch_size,28*28)
    # to a 4D tensor, compatible with our LeNetConvPoolLayer
    layer0_input = x.reshape((batch_size, 9, 1, 60, 80))
    #print layer0_input

    # Construct the first convolutional pooling layer:
    # filtering reduces the image size to (28-5+1,28-5+1)=(24,24)
    # maxpooling reduces this further to (24/2,24/2) = (12,12)
    # 4D output tensor is thus of shape (batch_size,nkerns[0],12,12)
    layer0 = LeNetConvPoolLayer(rng, input=layer0_input,
            image_shape=(batch_size, 9, 1, 60, 80),
            filter_shape=(nkerns[0], 4, 1, 7, 9), poolsize=(3, 3),pad=False)

    #print layer0.output


    # Construct the second convolutional pooling layer
    # filtering reduces the image size to (12-5+1,12-5+1)=(8,8)
    # maxpooling reduces this further to (8/2,8/2) = (4,4)
    # 4D output tensor is thus of shape (nkerns[0],nkerns[1],4,4)
    layer1 = LeNetConvPoolLayer(rng, input=layer0.output,
            image_shape=(batch_size, 6, nkerns[0], 18, 24),
            filter_shape=(nkerns[1], 4, nkerns[0], 7, 7), poolsize=(3, 3),pad=False)

    #print layer1.output

    # the TanhLayer being fully-connected, it operates on 2D matrices of
    # shape (batch_size,num_pixels) (i.e matrix of rasterized images).
    # This will generate a matrix of shape (20,32*4*4) = (20,512)
    layer2 = LeNetConvPoolLayer(rng, input=layer1.output,
            image_shape=(batch_size, 3, nkerns[1] , 4, 6 ),
            filter_shape=(nkerns[2], 3, nkerns[1] , 4, 6 ), poolsize=(1, 1),pad=False)

    #print layer2.output
    layer3_input = layer2.output.flatten(2)


    # classify the values of the fully-connected sigmoidal layer
    layer3 = HiddenLayer(rng,input=layer3_input,n_in=128,n_out=20,activation=T.tanh)
    layer4 = LogisticRegression(input=layer3.output, n_in=20, n_out=6)

    # the cost we minimize during training is the NLL of the model
    cost = layer4.negative_log_likelihood(y)

    # create a function to compute the mistakes that are made by the model
    test_model = theano.function([index], layer4.errors(y),
             givens={
                x: test_set_x[index * batch_size: (index + 1) * batch_size],
                y: test_set_y[index * batch_size: (index + 1) * batch_size]},mode=mode_fft)


    # create a list of all model parameters to be fit by gradient descent
    params = layer4.params+layer3.params + layer2.params + layer1.params + layer0.params

    # create a list of gradients for all model parameters
    grads = T.grad(cost, params)

    # train_model is a function that updates the model parameters by
    # SGD Since this model has many parameters, it would be tedious to
    # manually create an update rule for each model parameter. We thus
    # create the updates list by automatically looping over all
    # (params[i],grads[i]) pairs.
    updates = []
    for param_i, grad_i in zip(params, grads):
        updates.append((param_i, param_i - learning_rate * grad_i))

    train_model = theano.function([index], cost, updates=updates,
          givens={
            x: train_set_x[index * batch_size: (index + 1) * batch_size],
            y: train_set_y[index * batch_size: (index + 1) * batch_size]},mode=mode_fft)

    ###############
    # TRAIN MODEL #
    ###############
    print '... training'
    # early-stopping parameters
    patience = 100000  # look as this many examples regardless
    patience_increase = 2  # wait this much longer when a new best is
                           # found
    improvement_threshold = 0.995  # a relative improvement of this much is
                                   # considered significant
    test_frequency = n_train_batches
                                  # go through this many
                                  # minibatche before checking the network
                                  # on the validation set; in this case we
                                  # check every epoch

    best_params = None
    best_validation_loss = numpy.inf
    best_iter = 0
    test_score = 0.
    start_time = time.clock()

    epoch = 0
    done_looping = False

    while (epoch < n_epochs) and (not done_looping):
        epoch = epoch + 1
        for minibatch_index in xrange(n_train_batches):

            iter = (epoch - 1) * n_train_batches + minibatch_index

            #if iter % 100 == 0:
                #print 'training @ iter = ', iter
            #print "training cost"
            cost_ij = train_model(minibatch_index)
            #if minibatch_index == 0:
            #print cost_ij

            if (iter + 1) % test_frequency == 0:

                    # test it on the test set
                    test_losses = [test_model(i) for i in xrange(n_test_batches)]
                    #print "test"
                    print test_losses
                    test_score = numpy.mean(test_losses)
                    print(('     epoch %i, minibatch %i/%i, test error of best '
                           'model %f %%') %
                          (epoch, minibatch_index + 1, n_train_batches,
                           test_score * 100.))

    end_time = time.clock()
    print('Optimization complete.')
    print >> sys.stderr, ('The code for file ' +
                          os.path.split(__file__)[1] +
                          ' ran for %.2fm' % ((end_time - start_time) / 60.))

if __name__ == '__main__':
    evaluate_lenet5()


def experiment(state, channel):
    evaluate_lenet5(state.learning_rate, dataset=state.dataset)