lab3_RNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline \n",
    "%matplotlib nbagg\n",
    "import lasagne\n",
    "import theano\n",
    "import theano.tensor as T\n",
    "import matplotlib\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from IPython import display\n",
    "from data_generator import get_batch, print_valid_characters\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Recurrent Neural Networks\n",
    "\n",
    "Recurrent neural networks are the natural choice for sequential data i.e. time series analysis, translation, speech recognition, biological sequence analysis etc. Recurrent neural networks works by recursively applying the same operation at each time step of the data sequence and having layers that pass information from previous time steps to the current. It can therefore naturally handle input of varying length. Recurrent networks are used for several prediction tasks including: sequence-to-class, sequence tagging, and sequence-to-sequence predictions.\n",
    "\n",
    "In this exercise we'll implement an Encoder-Decoder RNN build with the GRU unit. This type of models have shown impressive performance in Neural Machine Translation and Image Caption generation. \n",
    "\n",
    "For more in depth background material on RNNs please cf. [Supervised Sequence Labelling with Recurrent\n",
    "Neural Networks](https://www.cs.toronto.edu/~graves/preprint.pdf) by Alex Graves.\n",
    "\n",
    "We know that LSTMs and GRUs are difficult to understand. A very good non-mathematical introduction is [Chris Olahs blog](http://colah.github.io/posts/2015-08-Understanding-LSTMs/) (All the posts are nice and cover various topics within machine-learning)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Encoder-Decoder\n",
    "In the encoder-decoder architecture, one RNN (blue) encodes the input and a second RNN (red) calculates the target values. One essential step is to let the encoder and decoder communicate. In the simplest approach you use the last hidden state of the encoder to initialize the decoder. Other approaches lets the decoder attend to different parts of the encoded input at different timesteps in the decoding process. \n",
    "\n",
    "<img src=\"files/enc-dec.png\", width=400>\n",
    "\n",
    "In our implementation we use a RNN with gated recurrent units (GRU) as encoder. We then use the last hidden state of the encoder ($h^{enc}_T$) as input to the decoder which is also a RNN with GRU units. \n",
    "\n",
    "### RNNs in Lasagne\n",
    "Lasagne has implementations of LSTM and GRU units. Both layers assume that the input from the layer below have the shape **(Batch_size, seq_len, num_features)**. In this excercise we will use the GRU unit since it only stores a single hidden value per neuron (LSTMs stores two) and is approximately twice as fast.\n",
    "\n",
    "As stated above we will implement an Encoder-Decoder model. The simplest way to do this is to encode the input sequence using the Encoder model. We will then use the last hidden state of the Encoder $h^{enc}_T$ as input to the decoder model, which then uses this information (simply a fixed length vector of numbers) to produce the targets. There is (at least) two ways to input $h^{enc}_T$ into the decoder\n",
    "\n",
    "1. Repeatly use $h^{enc}_T$ as input to the Decoder at each decode time step\n",
    "2. Intialize the decoder using $h^{enc}_T$ and run the decoder without any inputs\n",
    "\n",
    "In this exercise we follow the first approach because it's easier to implement. To do this need to create a lasagne layer that takes $h^{enc}_T$ and repeat it *N_decode_step* times. Below is an implementation of the RepeatLayer. You don't need to know the exact way it works, however make sure that you understand that it takes an input of size *(Batch_size x num_units)* and produces an output of size *(Batch_size x n_decode_steps x num_units)*.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class RepeatLayer(lasagne.layers.Layer):\n",
    "    def __init__(self, incoming, n, **kwargs):\n",
    "        '''\n",
    "        The input is expected to be a 2D tensor of shape \n",
    "        (num_batch, num_features). The input is repeated\n",
    "        n times such that the output will be \n",
    "        (num_batch, n, num_features)\n",
    "        '''\n",
    "        super(RepeatLayer, self).__init__(incoming, **kwargs)\n",
    "        self.n = n\n",
    "\n",
    "    def get_output_shape_for(self, input_shape):\n",
    "        return tuple([input_shape[0], self.n] + list(input_shape[1:]))\n",
    "\n",
    "    def get_output_for(self, input, **kwargs):\n",
    "        #repeat the input n times\n",
    "        tensors = [input]*self.n\n",
    "        stacked = theano.tensor.stack(*tensors)\n",
    "        dim = [1, 0] + range(2, input.ndim + 1)\n",
    "        return stacked.dimshuffle(dim)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The Data\n",
    "Since RNN models can be very slow to train on large datasets we will generate simpler training data for this exercise. The task for the RNN is simply to translate a string of letters spelling the numbers between 0-9 into the corresponding numbers i.e\n",
    "\n",
    "\"one two five\" --> \"125#\" (we use # as a special end-of-sequence character)\n",
    "\n",
    "To input the strings into the RNN model we translate the characters into a vector of integers using a simple translation table (i.e. 'h'->16, 'o'-> 17 etc). The code below prints a few input/output pairs using the *get_batch* function which randomly produces the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "input types: int32 float32 int32 float32\n",
      "Number of valid characters: 27\n",
      "'0'=0,\t'1'=1,\t'2'=2,\t'3'=3,\t'4'=4,\t'5'=5,\t'6'=6,\t'7'=7,\t'8'=8,\t'9'=9,\t'#'=10,\t' '=11,\t'e'=12,\t'g'=13,\t'f'=14,\t'i'=15,\t'h'=16,\t'o'=17,\t'n'=18,\t's'=19,\t'r'=20,\t'u'=21,\t't'=22,\t'w'=23,\t'v'=24,\t'x'=25,\t'z'=26,\t\n",
      "None\n",
      "Stop character = #\n",
      "\n",
      "SAMPLE 0\n",
      "TEXT INPUTS:\t\teight five\n",
      "TEXT TARGETS:\t\t85#\n",
      "ENCODED INPUTS:\t\t[12 15 13 16 22 11 14 15 24 12  0]\n",
      "MASK INPUTS:\t\t[ 1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  0.]\n",
      "ENCODED TARGETS:\t[ 8  5 10]\n",
      "MASK TARGETS:\t\t[ 1.  1.  1.]\n",
      "\n",
      "SAMPLE 1\n",
      "TEXT INPUTS:\t\tthree seven\n",
      "TEXT TARGETS:\t\t37#\n",
      "ENCODED INPUTS:\t\t[22 16 20 12 12 11 19 12 24 12 18]\n",
      "MASK INPUTS:\t\t[ 1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.]\n",
      "ENCODED TARGETS:\t[ 3  7 10]\n",
      "MASK TARGETS:\t\t[ 1.  1.  1.]\n",
      "\n",
      "SAMPLE 2\n",
      "TEXT INPUTS:\t\tthree\n",
      "TEXT TARGETS:\t\t3#\n",
      "ENCODED INPUTS:\t\t[22 16 20 12 12  0  0  0  0  0  0]\n",
      "MASK INPUTS:\t\t[ 1.  1.  1.  1.  1.  0.  0.  0.  0.  0.  0.]\n",
      "ENCODED TARGETS:\t[ 3 10  0]\n",
      "MASK TARGETS:\t\t[ 1.  1.  0.]\n"
     ]
    }
   ],
   "source": [
    "batch_size = 3\n",
    "inputs, input_masks, targets, target_masks, text_inputs, text_targets = \\\n",
    "    get_batch(batch_size=batch_size,max_digits=2,min_digits=1)\n",
    "\n",
    "print \"input types:\", inputs.dtype,  input_masks.dtype, targets.dtype, target_masks.dtype\n",
    "print print_valid_characters()\n",
    "print \"Stop character = #\"\n",
    "\n",
    "for i in range(batch_size):\n",
    "    print \"\\nSAMPLE\",i\n",
    "    print \"TEXT INPUTS:\\t\\t\", text_inputs[i]\n",
    "    print \"TEXT TARGETS:\\t\\t\", text_targets[i]\n",
    "    print \"ENCODED INPUTS:\\t\\t\", inputs[i]\n",
    "    print \"MASK INPUTS:\\t\\t\", input_masks[i]\n",
    "    print \"ENCODED TARGETS:\\t\", targets[i]\n",
    "    print \"MASK TARGETS:\\t\\t\", target_masks[i]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Encoder Decoder model setup\n",
    "Below is the Lasagne model implementation. We use an embedding layer to map from integer to multinomial vector representation.\n",
    "\n",
    "Note that the layer has a lot of print statements which we used for debugging during setup."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 20, 27)\n",
      "(100, 20, 10)\n",
      "(100, 10)\n",
      "(100, 21, 10)\n",
      "(100, 21, 10)\n",
      "(2100, 10)\n",
      "(2100, 11)\n",
      "(100, 21, 11)\n"
     ]
    }
   ],
   "source": [
    "BATCH_SIZE = 100\n",
    "NUM_UNITS_ENC = 10\n",
    "NUM_UNITS_DEC = 10\n",
    "MAX_DIGITS = 20 \n",
    "MIN_DIGITS = MAX_DIGITS #currently only support for same length outputs - we'll leave it for an exercise to add support for varying length targets\n",
    "NUM_INPUTS = 27\n",
    "NUM_OUTPUTS = 11 #(0-9 + '#')\n",
    "\n",
    "\n",
    "#symbolic theano variables. Note that we are using imatrix for X since it goes into the embedding layer\n",
    "x_sym = T.imatrix()\n",
    "y_sym = T.imatrix()\n",
    "xmask_sym = T.matrix()\n",
    "\n",
    "#dummy data to test implementation - We advise to check the output-dimensions of all layers.\n",
    "#One way to do this in lasagne/theano is to forward pass some data through the model and \n",
    "#check the output dimensions of these.\n",
    "#Create some random testdata\n",
    "X = np.random.randint(0,10,size=(BATCH_SIZE,MIN_DIGITS)).astype('int32')\n",
    "Xmask = np.ones((BATCH_SIZE,MIN_DIGITS)).astype('float32')\n",
    "\n",
    "##### ENCODER START #####\n",
    "l_in = lasagne.layers.InputLayer((None, None))\n",
    "l_emb = lasagne.layers.EmbeddingLayer(l_in, NUM_INPUTS, NUM_INPUTS, \n",
    "                                      W=np.eye(NUM_INPUTS,dtype='float32'),\n",
    "                                      name='Embedding')\n",
    "# Embedding layer transform integer input to words. Outputs size of words\n",
    "\n",
    "#Here we'll remove the trainable parameters from the embeding layer to constrain \n",
    "#it to a simple \"one-hot-encoding\". You can experiment with removing this line\n",
    "l_emb.params[l_emb.W].remove('trainable') \n",
    "#forward pass some data throug the inputlayer-embedding layer and print the output shape\n",
    "print lasagne.layers.get_output(l_emb, inputs={l_in: x_sym}).eval({x_sym: X}).shape\n",
    "\n",
    "l_mask_enc = lasagne.layers.InputLayer((None, None))\n",
    "l_enc = lasagne.layers.GRULayer(l_emb, num_units=NUM_UNITS_ENC, name='GRUEncoder', mask_input=l_mask_enc)\n",
    "print lasagne.layers.get_output(l_enc, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "\n",
    "# slice last index of dimension 1\n",
    "l_last_hid = lasagne.layers.SliceLayer(l_enc, indices=-1, axis=1)\n",
    "print lasagne.layers.get_output(l_last_hid, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "##### END OF ENCODER######\n",
    "\n",
    "\n",
    "##### START OF DECODER######\n",
    "l_in_rep = RepeatLayer(l_last_hid, n=MAX_DIGITS+1) #we add one to allow space for the end of sequence character\n",
    "print lasagne.layers.get_output(l_in_rep, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "\n",
    "l_dec = lasagne.layers.GRULayer(l_in_rep, num_units=NUM_UNITS_DEC, name='GRUDecoder')\n",
    "print lasagne.layers.get_output(l_dec, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "\n",
    "\n",
    "# We need to do some reshape voodo to connect a softmax layer to the decoder.\n",
    "# See http://lasagne.readthedocs.org/en/latest/modules/layers/recurrent.html#examples \n",
    "# In short this line changes the shape from \n",
    "# (batch_size, decode_len, num_dec_units) -> (batch_size*decodelen,num_dec_units). \n",
    "# We need to do this since the softmax is applied to the last dimension and we want to \n",
    "# softmax the output at each position individually\n",
    "l_reshape = lasagne.layers.ReshapeLayer(l_dec, (-1, [2]))\n",
    "print lasagne.layers.get_output(l_reshape, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "\n",
    "l_softmax = lasagne.layers.DenseLayer(l_reshape, num_units=NUM_OUTPUTS, \n",
    "                                      nonlinearity=lasagne.nonlinearities.softmax,\n",
    "                                      name='SoftmaxOutput')\n",
    "print lasagne.layers.get_output(l_softmax, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "\n",
    "# reshape back to 3d format (batch_size, decode_len, num_dec_units). Here we tied the batch size to the shape of the symbolic variable for X allowing \n",
    "#us to use different batch sizes in the model.\n",
    "l_out = lasagne.layers.ReshapeLayer(l_softmax, (x_sym.shape[0], -1, NUM_OUTPUTS))\n",
    "print lasagne.layers.get_output(l_out, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "###END OF DECODER######"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 20, 10)\n"
     ]
    }
   ],
   "source": [
    "#emb=lasagne.layers.get_output(l_emb, inputs={l_in: x_sym}).eval({x_sym: X})\n",
    "#print emb\n",
    "\n",
    "enc=lasagne.layers.get_output(l_enc, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "print enc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Defining the cost function and theano functions\n",
    "The targets are categorical, hence we use cross entropy error. We use the Adam optimizer but you\n",
    "can experiment with the different optimizers implemented in [Lasagne](http://lasagne.readthedocs.org/en/latest/modules/updates.html). "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Trainable Model Parameters\n",
      "----------------------------------------\n",
      "GRUEncoder.W_in_to_updategate (27, 10)\n",
      "GRUEncoder.W_hid_to_updategate (10, 10)\n",
      "GRUEncoder.b_updategate (10,)\n",
      "GRUEncoder.W_in_to_resetgate (27, 10)\n",
      "GRUEncoder.W_hid_to_resetgate (10, 10)\n",
      "GRUEncoder.b_resetgate (10,)\n",
      "GRUEncoder.W_in_to_hidden_update (27, 10)\n",
      "GRUEncoder.W_hid_to_hidden_update (10, 10)\n",
      "GRUEncoder.b_hidden_update (10,)\n",
      "GRUDecoder.W_in_to_updategate (10, 10)\n",
      "GRUDecoder.W_hid_to_updategate (10, 10)\n",
      "GRUDecoder.b_updategate (10,)\n",
      "GRUDecoder.W_in_to_resetgate (10, 10)\n",
      "GRUDecoder.W_hid_to_resetgate (10, 10)\n",
      "GRUDecoder.b_resetgate (10,)\n",
      "GRUDecoder.W_in_to_hidden_update (10, 10)\n",
      "GRUDecoder.W_hid_to_hidden_update (10, 10)\n",
      "GRUDecoder.b_hidden_update (10,)\n",
      "SoftmaxOutput.W (10, 11)\n",
      "SoftmaxOutput.b (11,)\n",
      "----------------------------------------\n"
     ]
    }
   ],
   "source": [
    "output_decoder_train = lasagne.layers.get_output(l_out, inputs={l_in: x_sym, l_mask_enc: xmask_sym}, \n",
    "                                                deterministic=False)\n",
    "\n",
    "#cost function\n",
    "total_cost = T.nnet.categorical_crossentropy(\n",
    "    T.reshape(output_decoder_train, (-1, NUM_OUTPUTS)), y_sym.flatten())\n",
    "mean_cost = T.mean(total_cost)\n",
    "#accuracy function\n",
    "argmax = T.argmax(output_decoder_train,axis=-1)\n",
    "eq = T.eq(argmax,y_sym)\n",
    "acc = T.mean(eq)  # gives float64 because eq is uint8, T.cast(eq, 'float32') will fix that...\n",
    "\n",
    "#Get parameters of both encoder and decoder\n",
    "all_parameters = lasagne.layers.get_all_params([l_out], trainable=True)\n",
    "\n",
    "print \"Trainable Model Parameters\"\n",
    "print \"-\"*40\n",
    "for param in all_parameters:\n",
    "    print param, param.get_value().shape\n",
    "print \"-\"*40\n",
    "\n",
    "#add grad clipping to avoid exploding gradients\n",
    "all_grads = [T.clip(g,-3,3) for g in T.grad(mean_cost, all_parameters)]\n",
    "all_grads = lasagne.updates.total_norm_constraint(all_grads,3)\n",
    "\n",
    "#Compile Theano functions.\n",
    "updates = lasagne.updates.adam(all_grads, all_parameters, learning_rate=0.005)\n",
    "train_func = theano.function([x_sym, y_sym, xmask_sym], [mean_cost, acc, output_decoder_train], updates=updates)\n",
    "#since we don't have any stochasticity in the network we will just use the training graph without any updates given\n",
    "test_func = theano.function([x_sym, y_sym, xmask_sym], [acc, output_decoder_train])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Xval (5000, 112)\n",
      "Yval (5000, 21)\n"
     ]
    }
   ],
   "source": [
    "#Generate validation data\n",
    "Xval, Xmask_val, Yval, Ymask_val, text_inputs_val, text_targets_val = \\\n",
    "    get_batch(batch_size=5000, max_digits=MAX_DIGITS,min_digits=MIN_DIGITS)\n",
    "print \"Xval\", Xval.shape\n",
    "print \"Yval\", Yval.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAyAAAAJYCAYAAACadoJwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZmbWNa6AmJmZmZlZ1/w/EJaq6mxL4pMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "val_interval = 5000\n",
    "samples_to_process = 3e5\n",
    "samples_processed = 0\n",
    "\n",
    "val_samples = []\n",
    "costs, accs = [], []\n",
    "plt.figure()\n",
    "try:\n",
    "    while samples_processed < samples_to_process:\n",
    "        x_, x_masks_, ys_, y_masks_, _, _ = \\\n",
    "            get_batch(batch_size=BATCH_SIZE,max_digits=MAX_DIGITS,min_digits=MIN_DIGITS)\n",
    "        batch_cost, batch_acc, batch_output = train_func(x_, ys_, x_masks_)\n",
    "        costs += [batch_cost]\n",
    "        samples_processed += BATCH_SIZE\n",
    "        #validation data\n",
    "        if samples_processed % val_interval == 0:\n",
    "            #print \"validating\"\n",
    "            val_acc, val_output = test_func(Xval, Yval, Xmask_val)\n",
    "            val_samples += [samples_processed]\n",
    "            accs += [val_acc]\n",
    "            plt.plot(val_samples,accs)\n",
    "            plt.ylabel('Validation Accuracy', fontsize=15)\n",
    "            plt.xlabel('Processed samples', fontsize=15)\n",
    "            plt.title('', fontsize=20)\n",
    "            plt.grid('on')\n",
    "            plt.savefig(\"out.png\")\n",
    "            display.display(display.Image(filename=\"out.png\"))\n",
    "            display.clear_output(wait=True)\n",
    "except KeyboardInterrupt:\n",
    "    pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support.' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        this.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width);\n",
       "        canvas.attr('height', height);\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option)\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'];\n",
       "    var y0 = fig.canvas.height - msg['y0'];\n",
       "    var x1 = msg['x1'];\n",
       "    var y1 = fig.canvas.height - msg['y1'];\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width, fig.canvas.height);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x;\n",
       "    var y = canvas_pos.y;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to  previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overriden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot of validation accuracy for each target position\n",
    "plt.figure(figsize=(7,7))\n",
    "plt.plot(np.mean(np.argmax(val_output,axis=2)==Yval,axis=0))\n",
    "plt.ylabel('Accuracy', fontsize=15)\n",
    "plt.xlabel('Target position', fontsize=15)\n",
    "#plt.title('', fontsize=20)\n",
    "plt.grid('on')\n",
    "plt.show()\n",
    "#why do the plots look like this? See exercise 4 below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# plot training cost\n",
    "#plt.figure(figsize=(7,7))\n",
    "#plt.plot(costs)\n",
    "#plt.ylabel('Cost', fontsize=15)\n",
    "#plt.xlabel('Number of updates', fontsize=15)\n",
    "#plt.title('Training', fontsize=20)\n",
    "#plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "from IPython.display import Image\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X [[9 0 4 1 4 2 9 2 6 3 3 5 9 8 1 4 3 7 2 1]]\n",
      "dim (1, 20)\n",
      "Xmask [[ 1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.\n",
      "   1.  1.]]\n",
      "dimension outer layer (1, 21, 11)\n",
      "out num [[ 8  7  2  2  2  2  2  7  7  8  8  8  8  8  8  8  4 10 10 10 10]]\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n",
    "X = np.random.randint(0,10,size=(1,MIN_DIGITS)).astype('int32')\n",
    "print \"X\", X\n",
    "print \"dim\",X.shape\n",
    "Xmask = np.ones((1,MIN_DIGITS)).astype('float32')\n",
    "print \"Xmask\",Xmask\n",
    "\n",
    "out = lasagne.layers.get_output(l_out, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval({x_sym: X, xmask_sym: Xmask})\n",
    "#print \"out\", out\n",
    "out_num = np.argmax(out,axis=2)\n",
    "print 'dimension outer layer', out.shape\n",
    "print \"out num\",out_num"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 21, 11)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "out.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exercises:\n",
    "1. Perform an explorative test of the network by inputting a defined input and evaulate the outputs.\n",
    "2. The model has two GRU networks. The ```GRUEncoder``` and the ```GRUDecoder```. A GRU is parameterized by the  ```{W_in_to_updategate, W_hid_to_updategate, b_updategate, W_in_to_resetgate, W_hid_to_resetgate, b_resetgate,\n",
    "W_in_to_hidden_update, W_hid_to_hidden_update,b_hidden_update}```. Try to explain the shape of ```W_in_to_updategate``` and ```hid_to_updategate```. Why are they different? You can find the equations for the gru at: [GRU](http://lasagne.readthedocs.io/en/latest/modules/layers/recurrent.html#lasagne.layers.GRULayer). \n",
    "3. The GRU unit is able to ignore the input and just copy the previous hidden state. In the beginning of training this might be desireable behavior, since it helps the model learn long range dependencies. You can make the model ignore the input by modifying initial bias values. What bias would you modify and how would you modify it? Again you'll need to refer to the GRU equations:  [GRU](http://lasagne.readthedocs.io/en/latest/modules/layers/recurrent.html#lasagne.layers.GRULayer).\n",
    "\n",
    "4. What is the final validation performance? Why do you think it is not better? Comment on the accuracy for each position in of the output symbols?\n",
    "\n",
    "5. Why do you think the validation performance looks more \"jig-saw\"-like compared to FFNN and CNN models?\n",
    "\n",
    "6. In the example we stack a softmax layer on top of a Recurrent layer. In the code snippet below explain how we can do that?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Answers\n",
    "\n",
    "##### Perform an explorative test of the network by inputting a defined input and evaulate the outputs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X [[6 5 9 1 7 5 8 8 1 0 9 7 2 3 4 4 4 4 5 4]]\n",
      "dim (1, 20)\n",
      "Xmask [[ 1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.  1.\n",
      "   1.  1.]]\n",
      "(1, 21, 11)\n",
      "out num [[ 3  7  7  7  7  7  7  7  7  7  7  7  5  8  8  8  8  8 10 10 10]]\n"
     ]
    }
   ],
   "source": [
    "X = np.random.randint(0,10,size=(1,MIN_DIGITS)).astype('int32')\n",
    "print \"X\", X\n",
    "print \"dim\",X.shape\n",
    "Xmask = np.ones((1,MIN_DIGITS)).astype('float32')\n",
    "print \"Xmask\",Xmask\n",
    "\n",
    "out = lasagne.layers.get_output(l_out, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval({x_sym: X, xmask_sym: Xmask})\n",
    "#print \"out\", out\n",
    "out_num = np.argmax(out,axis=2)\n",
    "print out.shape\n",
    "print \"out num\",out_num"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The network was trained on random integers. In this task we investigate the output of a vector X with length 20 and with integers between zero and nine. This means that X has 1 sequenze with length 20.\n",
    "The Xmask indicates that all numbers should be included for the sequenze with the indicating value 1 in all elements.\n",
    "Next we use *lasagne.layers.get_output* to get the output softmax probability for each class. In order to evaluate the prediction of each class we use *argmax*.\n",
    "\n",
    "The output of the network is listed in the vector \"out num\". The network does not predecit the pattern quite well, however, the network was trained on random data and therefore it is not generalized to a specific task.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Answers\n",
    "##### The model has two GRU networks. The GRUEncoder and the GRUDecoder. A GRU is parameterized by the  {W_in_to_updategate, W_hid_to_updategate, b_updategate, W_in_to_resetgate, W_hid_to_resetgate, b_resetgate,\n",
    "##### W_in_to_hidden_update, W_hid_to_hidden_update,b_hidden_update}. Try to explain the shape of W_in_to_updategate and hid_to_updategate. Why are they different? You can find the equations for the gru at: GRU.\n",
    "\n",
    "The size of *W_in_to_updategate* is 27 times 10. The first dimension equal 27 is the numbers of input numbers indexing letters, eg. \"e\"=12.\n",
    "The second dimension equal 10 is the number of encoding hidden units in the model. This means the variable *W_in_to_updategate* contains 27 input variables and 10 weights. Looking at the following equation we have the update gate:\n",
    "\n",
    "$$ u_t= \\sigma_u (x_t \\cdot W_{xu} +h_{t-1} \\cdot W_{hu}+b_u)$$\n",
    "\n",
    "the input vector x_t dimension (1 X 27}, in which $W_{xu} $ is equal to \n",
    "*W_in_to_updategate*. \n",
    "\n",
    "\n",
    "The variable $h_{t-1} $ is dimension (1 X 10). \n",
    "The size of *W_hid_in_to_updategate* ($W_{hr}$) is 10 times 10 indicates the weights for mapping from one time instance to the next.\n",
    "\n",
    "\n",
    "##### 3. The GRU unit is able to ignore the input and just copy the previous hidden state. In the beginning of training this might be desireable behavior, since it helps the model learn long range dependencies. You can make the model ignore the input by modifying initial bias values. What bias would you modify and how would you modify it? Again you'll need to refer to the GRU equations:  [GRU](http://lasagne.readthedocs.io/en/latest/modules/layers/recurrent.html#lasagne.layers.GRULayer).\n",
    "\n",
    "To understand how we can ignore a input in the GRU-unit model we can kook at the the hidden layer state:\n",
    "\n",
    "$$ h_t = (1-u_t)\\odot h_{t-1} + u_t \\odot c_t $$\n",
    "\n",
    "we can see if would like ignore a desired input $c_t$ we have to set $$u_t$ to zero.\n",
    "\n",
    "We can do that by changing the argument in $$ u_t= \\sigma_u (x_t \\cdot W_{xu} +h_{t-1} \\cdot W_{hu}+b_u) $$ to returning zero. This can be done by changing the bias $b_u$.\n",
    "If we have a ReLu nonlinear activation function we need a $b_u$ with high negative values or close to zero.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### 4. What is the final validation performance? Why do you think it is not better? Comment on the accuracy for each position in of the output symbols?\n",
    "\n",
    "\n",
    "The final validation performance for the 10 digit classifyer is approximately 0.42 validation accuracy.\n",
    "![title](validation_accuracy_10.jpeg)\n",
    "\n",
    "The final validation performance for the 20 digit classifyer is approximately 0.25 validation accuracy.\n",
    "\n",
    "![title](validation_acc_20.jpeg)\n",
    "\n",
    "We also notice the last digit in the sequence always have 1 validation accuracy due to the fact that the last digit always is denoted \"#\"."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### 6. In the example we stack a softmax layer on top of a Recurrent layer. In the code snippet below explain how we can do that?\n",
    "\n",
    "The layer is reshaped to a coloumn vector. This approach acts as an efficient way to apply softmax to each elenebt and reshape it back to the desired discrete output layer.\n",
    "The outer layer resturns the probability from the softmax function as printed below.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "out [  1.11453187e-01   9.53880162e-02   8.39029900e-02   1.29550124e-01\n",
      "   8.88013983e-02   1.04281965e-01   5.75887462e-02   1.20998152e-01\n",
      "   1.18651008e-01   8.93136729e-02   7.07402434e-05]\n"
     ]
    }
   ],
   "source": [
    "print \"out\",out[0][0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "l_input_ (16, 140, 40)\n",
      "l_gru_ (16, 140, 10)\n",
      "l_reshape_ (2240, 10)\n",
      "l_softmax_ (2240, 11)\n",
      "l_softmax_seq_ (16, 140, 11)\n"
     ]
    }
   ],
   "source": [
    "bs_, seqlen_, numinputs_ = 16, 140, 40\n",
    "l_input_ = lasagne.layers.InputLayer((bs_, seqlen_, numinputs_))\n",
    "l_gru_ = lasagne.layers.GRULayer(l_input_, num_units=10)\n",
    "l_reshape_ = lasagne.layers.ReshapeLayer(l_gru_, (-1, [2])) # the [2] tells lasagne to use the dim of the 2. \n",
    "                                                            # dim of the input here. -1 is a wildcard\n",
    "l_softmax_ = lasagne.layers.DenseLayer(l_reshape_, num_units=11, \n",
    "                                      nonlinearity=lasagne.nonlinearities.softmax)\n",
    "l_softmax_seq_ = lasagne.layers.ReshapeLayer(l_softmax_,(bs_, seqlen_, -1))\n",
    "print \"l_input_\", l_input_.output_shape\n",
    "print \"l_gru_\", l_gru_.output_shape\n",
    "print \"l_reshape_\", l_reshape_.output_shape\n",
    "print \"l_softmax_\", l_softmax_.output_shape\n",
    "print \"l_softmax_seq_\", l_softmax_seq_.output_shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "6. Optional: If you are interested in doing sentiment analysis on tweets, i.e classification as positive or negative. Use RNN to read over the twitter seqeuence and use the last hidden state to do the classification. How can you modify the small network above to only output binary classification? Hints: Look at the SliceLayer or the GRU docs...\n",
    "7. Optional: Bidirectional Encoder, In Lasagne, bidirectional RNNs are implementated by running a forward model and a backward model separately and then concatenating them before parsing them to the next layer. You can experiment with using a different merging layer than concat e.g. sum or multiplication see [lasagne merge layers [lasagne merge layers](http://lasagne.readthedocs.org/en/latest/modules/layers/merge.html).\n",
    "```\n",
    "l_rec_fwd = lasagne.layers.GRULayer(...,backwards=False)\n",
    "l_rec_bwd = lasagne.layers.GRULayer(...,backwards=True)\n",
    "l_rec = lasagne.layers.ConcatLayer([l_rec_fwd, l_rec_bwd], axis=2))\n",
    "```\n",
    "8. Optional: Add support for different lengths of targets (hint: add the target_mask to the cost function and only calculate the cost for the non-masked targets)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Attention Decoder (LSTM)\n",
    "Selective attention for recurrent neural networks have recently attracted a lot of interest. These methods let the Decoder model selectively focus on which sub-sequence of the encoder sequence it will use for each decoded output symbol. This relieves the encoder from having to compress the input sequence into a fixed size vector representation. Secondly, we can investigate which part of the input that the decoder network emphazises. Below we'll implement an LSTM-decoder with selective attention and show that it significantly improves the performance of the toy translation task.\n",
    "\n",
    "The attention paper is found here: https://arxiv.org/pdf/1409.0473v7.pdf.\n",
    "\n",
    "The principles of the attention model are simple. \n",
    "\n",
    "1. Use the encoder to get the hidden represention $\\{h^1_e, ...h^n_e\\}$ for each position in the input sequence. \n",
    "2. For timestep $t$ in the decoder $m = 1...n$ : $a_m = f(h^m_e, h^d_t)$. Where f is a function returning a scalar value.\n",
    "3. You can then normalize the sequence of scalars $\\{a_1, ... a_n\\}$ to get probablities $\\{p_1, ... p_n\\}$.\n",
    "4. Weigh each $h^e_t$ by its probablity $p_t$ and sum to get $h_{in}$.\n",
    "5. Use $h_{in}$ as an additional input to the decoder. $h_{in}$ is recalculated each time the decoder is updated."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 20, 27)\n",
      "(1, 20, 10)\n",
      "(1, 21, 10)\n",
      "(1, 21, 11)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from decoder_attention import LSTMAttentionDecodeFeedbackLayer\n",
    "\n",
    "# you can acces the attetion weights alpha by adding l_dec.alpha \n",
    "# to the output variables in the theano function\n",
    "\n",
    "BATCH_SIZE = 100\n",
    "NUM_UNITS_ENC = 10\n",
    "NUM_UNITS_DEC = 10\n",
    "MAX_DIGITS = 20 \n",
    "MIN_DIGITS = MAX_DIGITS #currently only support for same length outputs - we'll leave it for an exercise to add support for varying length targets\n",
    "NUM_INPUTS = 27\n",
    "NUM_OUTPUTS = 11 #(0-9 + '#')\n",
    "\n",
    "\n",
    "x_sym = T.imatrix()\n",
    "y_sym = T.imatrix()\n",
    "xmask_sym = T.matrix()\n",
    "    \n",
    "\n",
    "#dummy data to test implementation\n",
    "#X = np.random.randint(0,10,size=(BATCH_SIZE,15)).astype('int32')\n",
    "#Xmask = np.ones((BATCH_SIZE,NUM_INPUTS)).astype('float32')\n",
    "\n",
    "l_in = lasagne.layers.InputLayer((None, None))\n",
    "l_emb = lasagne.layers.EmbeddingLayer(l_in, NUM_INPUTS, NUM_INPUTS, \n",
    "                                      W=np.eye(NUM_INPUTS,dtype='float32'),\n",
    "                                      name='Embedding')\n",
    "##### ENCODER START #####\n",
    "l_in = lasagne.layers.InputLayer((None, None))\n",
    "l_emb = lasagne.layers.EmbeddingLayer(l_in, NUM_INPUTS, NUM_INPUTS, \n",
    "                                      W=np.eye(NUM_INPUTS,dtype='float32'),\n",
    "                                      name='Embedding')\n",
    "#Here we'll remove the trainable parameters from the embeding layer to constrain \n",
    "#it to a simple \"one-hot-encoding\". You can experiment with removing this line\n",
    "l_emb.params[l_emb.W].remove('trainable') \n",
    "print lasagne.layers.get_output(l_emb, inputs={l_in: x_sym}).eval(\n",
    "    {x_sym: X}).shape\n",
    "T.grad(lasagne.layers.get_output(l_emb, inputs={l_in: x_sym}).sum(), \n",
    "       lasagne.layers.get_all_params(l_emb, trainable=True))\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "l_mask_enc = lasagne.layers.InputLayer((None, None))\n",
    "l_enc = lasagne.layers.GRULayer(l_emb, num_units=NUM_UNITS_ENC, name='GRUEncoder', mask_input=l_mask_enc)\n",
    "print lasagne.layers.get_output(l_enc, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "T.grad(lasagne.layers.get_output(l_enc, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).sum(), \n",
    "       lasagne.layers.get_all_params(l_enc, trainable=True))\n",
    "####END OF ENCODER######\n",
    "\n",
    "\n",
    "####START OF DECODER######\n",
    "#note that the decoder have its own input layer, we'll use that to plug in the output \n",
    "#from the encoder later\n",
    "l_dec = LSTMAttentionDecodeFeedbackLayer(l_enc,\n",
    "                                        num_units=NUM_UNITS_DEC, \n",
    "                                        aln_num_units=20,\n",
    "                                        n_decodesteps=MAX_DIGITS+1,\n",
    "                                        name='LSTMDecoder')\n",
    "print lasagne.layers.get_output(l_dec, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "T.grad(lasagne.layers.get_output(l_dec, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).sum(), \n",
    "       lasagne.layers.get_all_params(l_dec, trainable=True))\n",
    "\n",
    "# We need to do some reshape voodo to connect a softmax layer to the decoder.\n",
    "# See http://lasagne.readthedocs.org/en/latest/modules/layers/recurrent.html#examples \n",
    "l_reshape = lasagne.layers.ReshapeLayer(l_dec, (-1, [2]))\n",
    "l_softmax = lasagne.layers.DenseLayer(l_reshape, num_units=NUM_OUTPUTS, \n",
    "                                      nonlinearity=lasagne.nonlinearities.softmax,\n",
    "                                      name='SoftmaxOutput')\n",
    "# print lasagne.layers.get_output(l_softmax, x_sym).eval({x_sym: X}).shape\n",
    "# reshape back to 3d format (here we tied the batch size to the shape of the symbolic variable for X allowing \n",
    "#us to use different batch sizes in the model)\n",
    "l_out = lasagne.layers.ReshapeLayer(l_softmax, (x_sym.shape[0], -1, NUM_OUTPUTS))\n",
    "print lasagne.layers.get_output(l_out, inputs={l_in: x_sym, l_mask_enc: xmask_sym}, deterministic=False).eval(\n",
    "    {x_sym: X, xmask_sym: Xmask}).shape\n",
    "T.grad(lasagne.layers.get_output(l_out, inputs={l_in: x_sym, l_mask_enc: xmask_sym}).sum(), \n",
    "       lasagne.layers.get_all_params(l_dec, trainable=True))\n",
    "\n",
    "print \"\"\n",
    "###END OF DECODER######\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Generate some validation data\n",
    "Xval, Xmask_val, Yval, Ymask_val, text_inputs_val, text_targets_val = \\\n",
    "    get_batch(batch_size=5000, max_digits=MAX_DIGITS,min_digits=MIN_DIGITS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Trainable Model Parameters\n",
      "----------------------------------------\n",
      "GRUEncoder.W_in_to_updategate (27, 10)\n",
      "GRUEncoder.W_hid_to_updategate (10, 10)\n",
      "GRUEncoder.b_updategate (10,)\n",
      "GRUEncoder.W_in_to_resetgate (27, 10)\n",
      "GRUEncoder.W_hid_to_resetgate (10, 10)\n",
      "GRUEncoder.b_resetgate (10,)\n",
      "GRUEncoder.W_in_to_hidden_update (27, 10)\n",
      "GRUEncoder.W_hid_to_hidden_update (10, 10)\n",
      "GRUEncoder.b_hidden_update (10,)\n",
      "LSTMDecoder.W_hid_to_ingate (10, 10)\n",
      "LSTMDecoder.b_ingate (10,)\n",
      "LSTMDecoder.W_hid_to_forgetgate (10, 10)\n",
      "LSTMDecoder.b_forgetgate (10,)\n",
      "LSTMDecoder.W_hid_to_cell (10, 10)\n",
      "LSTMDecoder.b_cell (10,)\n",
      "LSTMDecoder.W_hid_to_outgate (10, 10)\n",
      "LSTMDecoder.b_outgate (10,)\n",
      "LSTMDecoder.W_weightedhid_to_ingate (10, 10)\n",
      "LSTMDecoder.W_weightedhid_to_forgetgate (10, 10)\n",
      "LSTMDecoder.W_weightedhid_to_cell (10, 10)\n",
      "LSTMDecoder.W_weightedhid_to_outgate (10, 10)\n",
      "LSTMDecoder.W_cell_to_ingate (10,)\n",
      "LSTMDecoder.W_cell_to_forgetgate (10,)\n",
      "LSTMDecoder.W_cell_to_outgate (10,)\n",
      "LSTMDecoder.AlignSeqOutputLayer: (aln) W_a (10, 20)\n",
      "LSTMDecoder.AlignSeqOutputLayer: (aln) U_a (10, 20)\n",
      "LSTMDecoder.AlignSeqOutputLayer: v_a (20, 1)\n",
      "SoftmaxOutput.W (10, 11)\n",
      "SoftmaxOutput.b (11,)\n",
      "----------------------------------------\n"
     ]
    }
   ],
   "source": [
    "#get output of encoder using X and Xmask as input\n",
    "output_decoder_train = lasagne.layers.get_output(l_out, inputs={l_in: x_sym, l_mask_enc: xmask_sym}, \n",
    "                                                 deterministic=False)\n",
    "\n",
    "#cost function\n",
    "total_cost = T.nnet.categorical_crossentropy(\n",
    "    T.reshape(output_decoder_train, (-1, NUM_OUTPUTS)), y_sym.flatten())\n",
    "mean_cost = T.mean(total_cost)\n",
    "#accuracy function\n",
    "acc = T.mean(T.eq(T.argmax(output_decoder_train,axis=-1),y_sym))\n",
    "\n",
    "#Get parameters of both encoder and decoder\n",
    "all_parameters = lasagne.layers.get_all_params(l_out, trainable=True)\n",
    "\n",
    "print \"Trainable Model Parameters\"\n",
    "print \"-\"*40\n",
    "for param in all_parameters:\n",
    "    print param, param.get_value().shape\n",
    "print \"-\"*40\n",
    "\n",
    "#add grad clipping to avoid exploding gradients\n",
    "all_grads = [T.clip(g,-3,3) for g in T.grad(mean_cost, all_parameters)]\n",
    "all_grads = lasagne.updates.total_norm_constraint(all_grads,3)\n",
    "\n",
    "#Compile Theano functions\n",
    "updates = lasagne.updates.adam(all_grads, all_parameters, learning_rate=0.005)\n",
    "train_func = theano.function([x_sym, y_sym, xmask_sym], [mean_cost, acc, output_decoder_train], updates=updates)\n",
    "#since we don't have any stochasticity in the network we will just use the training graph without any updates given\n",
    "test_func = theano.function([x_sym, y_sym, xmask_sym], [acc, output_decoder_train, l_dec.alpha])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAyAAAAJYCAYAAACadoJwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "val_interval = 5000\n",
    "samples_to_process = 1.5e5\n",
    "samples_processed = 0\n",
    "val_samples = []\n",
    "costs, accs = [], []\n",
    "plt.figure()\n",
    "try:\n",
    "    while samples_processed < samples_to_process:\n",
    "        inputs, input_masks, targets, target_masks, _, _ = \\\n",
    "            get_batch(batch_size=BATCH_SIZE,max_digits=MAX_DIGITS,min_digits=MIN_DIGITS)\n",
    "        batch_cost, batch_acc, batch_output = train_func(inputs, targets, input_masks)\n",
    "        costs += [batch_cost]\n",
    "        samples_processed += BATCH_SIZE\n",
    "        if samples_processed % val_interval == 0:\n",
    "            #print \"validating\"\n",
    "            val_acc, val_output, alpha = test_func(Xval, Yval, Xmask_val)\n",
    "            val_samples += [samples_processed]\n",
    "            accs += [val_acc]\n",
    "            plt.plot(val_samples,accs)\n",
    "            plt.ylabel('', fontsize=15)\n",
    "            plt.xlabel('Processed samples', fontsize=15)\n",
    "            plt.title('Validation Accuracy', fontsize=20)\n",
    "            plt.grid('on')\n",
    "            plt.savefig(\"out.png\")\n",
    "            display.display(display.Image(filename=\"out.png\"))\n",
    "            display.clear_output(wait=True)\n",
    "except KeyboardInterrupt:\n",
    "    pass\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support.' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        this.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width);\n",
       "        canvas.attr('height', height);\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option)\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'];\n",
       "    var y0 = fig.canvas.height - msg['y0'];\n",
       "    var x1 = msg['x1'];\n",
       "    var y1 = fig.canvas.height - msg['y1'];\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width, fig.canvas.height);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x;\n",
       "    var y = canvas_pos.y;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to  previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overriden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot of validation accuracy for each target position\n",
    "plt.figure(figsize=(7,7))\n",
    "plt.plot(np.mean(np.argmax(val_output,axis=2)==Yval,axis=0))\n",
    "plt.ylabel('Accuracy', fontsize=15)\n",
    "plt.xlabel('Target position', fontsize=15)\n",
    "#plt.title('', fontsize=20)\n",
    "plt.grid('on')\n",
    "plt.show()\n",
    "#why do the plot look like this?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support.' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        this.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width);\n",
       "        canvas.attr('height', height);\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option)\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'];\n",
       "    var y0 = fig.canvas.height - msg['y0'];\n",
       "    var x1 = msg['x1'];\n",
       "    var y1 = fig.canvas.height - msg['y1'];\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width, fig.canvas.height);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x;\n",
       "    var y = canvas_pos.y;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to  previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overriden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot of average attention weight as a function of the sequence position for each of \n",
    "#the 21 targets in the output sequence i.e. each line is the mean postion of the \n",
    "#attention for each target position.\n",
    "\n",
    "np.mean(alpha,axis=0).shape\n",
    "plt.figure()\n",
    "plt.plot(np.mean(alpha,axis=0).T)\n",
    "plt.ylabel('alpha', fontsize=15)\n",
    "plt.xlabel('Input Sequence position', fontsize=15)\n",
    "plt.title('Alpha weights', fontsize=20)\n",
    "plt.legend(map(str,range(1,22)), bbox_to_anchor=(1.125,1.0), fontsize=10)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Assignments for the attention decoder\n",
    "1. Explain what the alpha weights show. \n",
    "2. Why are the alpha curve for the first digit narrow and peaked while later digits have alpha curves that are wider and less peaked?\n",
    "3. Why is attention a good idea for this problem? Can you think of other problems where attention is a good choice?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Answers\n",
    "\n",
    "##### 1. Explain what the alpha weights show. \n",
    "The idea of alpha weights is to give attention to certain parts of a sequence, e.g the first alpha wieght informs where to keep attention in the the sequence for classifying the first digit. For our model we have 21 alpha weights since we have 20 digits to classify + \"#\". \n",
    "\n",
    "##### 2. Why are the alpha curve for the first digit narrow and peaked while later digits have alpha curves that are wider and less peaked?\n",
    "For a probablity distribution the narrower the distribution is the more probability mass is concentrated within the curve. The alpha curve for the first digit is narrow because all the sequnce lengths are 3-5. \n",
    "In other words the attension points were to look in the sequence to gain best prediction\n",
    "\n",
    "##### 3. Why is attention a good idea for this problem? Can you think of other problems where attention is a good choice?\n",
    "\n",
    "Using attension it is possible to get information about the previous hidden state. This gives the model the oppotunity to find characters which are specific interesting in the sequency for classifying the digit. In any system where the current state is dependent on the previous state this approach is convenient. For example if we would like to predict the next word in a sentences it might be beneficial to look at both the previous state and special words placed somewhere else in the sentence.\n",
    "For example: The *clouds* are in the *sky*"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}