disentangled_directions.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "disentangled_directions.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/nick-monto/disentangled_directions/blob/master/disentangled_directions.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ccsSfR3KFC01",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%matplotlib inline"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "46AVv-HTUx8x",
        "colab_type": "code",
        "outputId": "f09eebea-55fe-4b1e-83ee-9fa00d67b5cd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 151
        }
      },
      "source": [
        "# for input generation\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from scipy.stats import norm\n",
        "from numpy.random import randint, choice, dirichlet\n",
        "from pandas import get_dummies\n",
        "from PIL import Image, ImageDraw\n",
        "\n",
        "# for input analysis\n",
        "import pandas as pd\n",
        "\n",
        "# for model\n",
        "import tensorflow.compat.v1 as tf\n",
        "tf.disable_v2_behavior()\n",
        "from tensorflow.compat.v1 import set_random_seed\n",
        "from tensorflow.compat.v1.keras.utils import to_categorical\n",
        "from tensorflow.compat.v1.keras.models import Model, Sequential\n",
        "from tensorflow.compat.v1.keras.layers import BatchNormalization, Conv3D, Dense, \\\n",
        "                                              Dropout, Flatten, Input, ReLU, Softmax"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<p style=\"color: red;\">\n",
              "The default version of TensorFlow in Colab will soon switch to TensorFlow 2.x.<br>\n",
              "We recommend you <a href=\"https://www.tensorflow.org/guide/migrate\" target=\"_blank\">upgrade</a> now \n",
              "or ensure your notebook will continue to use TensorFlow 1.x via the <code>%tensorflow_version 1.x</code> magic:\n",
              "<a href=\"https://colab.research.google.com/notebooks/tensorflow_version.ipynb\" target=\"_blank\">more info</a>.</p>\n"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow_core/python/compat/v2_compat.py:68: disable_resource_variables (from tensorflow.python.ops.variable_scope) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "non-resource variables are not supported in the long term\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RUSAmRcVgii8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def findMiddleIndex(input_list):\n",
        "    middle = float(len(input_list))/2\n",
        "    if middle % 2 != 0:\n",
        "        return [int(middle - .5)]\n",
        "    else:\n",
        "        return (int(middle), int(middle-1))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3cGL4wUliXzA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def centralSort(flat_array):\n",
        "  descSort = -np.sort(-flat_array)[0]\n",
        "  cen_sorted = []\n",
        "  for i in reversed(range(len(descSort))):\n",
        "    while i >= len(descSort) - 2:\n",
        "      cen_sorted.append(descSort[i])\n",
        "      i -= 1\n",
        "    while i == len(descSort) - 3:\n",
        "      cen_sorted.insert(1,descSort[i])\n",
        "      i -= 1\n",
        "    helpme = findMiddleIndex(cen_sorted)\n",
        "    if len(helpme) >= 2:\n",
        "      cen_sorted.insert(helpme[1], descSort[i])\n",
        "    else:\n",
        "      cen_sorted.insert(helpme[0], descSort[i])\n",
        "  uniques = [ii for n,ii in enumerate(cen_sorted) if ii not in cen_sorted[:n]]\n",
        "  return uniques    "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qxNdejx8cBXG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# big ugly function to generate stimuli\n",
        "def central_image(dims, text, movement, show = False, to_matrix = True):\n",
        "  x = randint(30,91)\n",
        "  y = randint(30,91)\n",
        "  base_coords = np.asarray((x, y))\n",
        "  img = Image.new('L', dims)\n",
        "  d = ImageDraw.Draw(img)\n",
        "  d.text((x, y), text, fill=(255))\n",
        "  if movement == 'R':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = centralSort(prob_list)\n",
        "    # random choice from uniform distribution of range 5-24\n",
        "    x += choice(np.arange(5,25), 1, p=probs)\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      y += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      y -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  elif movement == 'L':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = centralSort(prob_list)\n",
        "    x -= choice(np.arange(5,25), 1, p=probs)\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      y += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      y -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  elif movement == 'U':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = centralSort(prob_list)\n",
        "    y -= choice(np.arange(5,25), 1, p=probs)\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      x += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      x -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  elif movement == 'D':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = centralSort(prob_list)\n",
        "    y += choice(np.arange(5,25), 1, p=probs)\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      x += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      x -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  else:\n",
        "    print(\"Please use U, D, L, or R to indicate movement direction.\")\n",
        "  if show == True:\n",
        "    fig = plt.figure()\n",
        "    a=fig.add_subplot(1,2,1)\n",
        "    a.set_title('Before')\n",
        "    plt.imshow(np.asarray(img))\n",
        "    a=fig.add_subplot(1,2,2)\n",
        "    a.set_title('After '+ movement + ' shift')\n",
        "    plt.imshow(np.asarray(img1))\n",
        "  if to_matrix == True:\n",
        "    stimulus, label = np.array([np.asarray(img), np.asarray(img1)]), movement\n",
        "    shift_coords = np.asarray((x[0],y[0]))\n",
        "    change = shift_coords - base_coords\n",
        "    return stimulus, label, change"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7QmrbSI6lz-D",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# set up the probability space to be \"trivial\"\n",
        "# big ugly function to generate stimuli\n",
        "def trivial_image(dims, text, movement, show = False, to_matrix = True):\n",
        "  x = randint(30,91)\n",
        "  y = randint(30,91)\n",
        "  base_coords = np.asarray((x, y))\n",
        "  img = Image.new('L', dims)\n",
        "  d = ImageDraw.Draw(img)\n",
        "  d.text((x, y), text, fill=(255))\n",
        "  if movement == 'R':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # ascending sort\n",
        "    probs = np.sort(prob_list)[0]\n",
        "    # random choice from uniform distribution of range 5-24\n",
        "    x += choice(np.arange(5,25), 1, p=probs)\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      y += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      y -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  elif movement == 'L':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = np.sort(prob_list)[0]\n",
        "    x -= choice(np.arange(5,25), 1, p=probs)\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      y += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      y -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  elif movement == 'U':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = np.sort(prob_list)[0]\n",
        "    y -= choice(np.arange(5,25), 1, p=probs)\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      x += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      x -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  elif movement == 'D':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = np.sort(prob_list)[0]\n",
        "    y += choice(np.arange(5,25), 1, p=probs)\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      x += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      x -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  else:\n",
        "    print(\"Please use U, D, L, or R to indicate movement direction.\")\n",
        "  if show == True:\n",
        "    fig = plt.figure()\n",
        "    a=fig.add_subplot(1,2,1)\n",
        "    a.set_title('Before')\n",
        "    plt.imshow(np.asarray(img))\n",
        "    a=fig.add_subplot(1,2,2)\n",
        "    a.set_title('After '+ movement + ' shift')\n",
        "    plt.imshow(np.asarray(img1))\n",
        "  if to_matrix == True:\n",
        "    stimulus, label = np.array([np.asarray(img), np.asarray(img1)]), movement\n",
        "    shift_coords = np.asarray((x[0],y[0]))\n",
        "    change = shift_coords - base_coords\n",
        "    return stimulus, label, change"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tr1HfpJQ7kiu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# set up the probability space to overlap completely\n",
        "# big ugly function to generate stimuli\n",
        "def overlap_image(dims, text, movement, show = False, to_matrix = True):\n",
        "  x = randint(30,91)\n",
        "  y = randint(30,91)\n",
        "  base_coords = np.asarray((x, y))\n",
        "  img = Image.new('L', dims)\n",
        "  d = ImageDraw.Draw(img)\n",
        "  d.text((x, y), text, fill=(255))\n",
        "  if movement == 'R':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    # random choice from uniform distribution of range 5-24\n",
        "    x += choice(np.arange(5,25), 1, p=probs)\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      y += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      y -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  elif movement == 'L':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    x -= choice(np.arange(5,25), 1, p=probs)\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      y += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      y -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  elif movement == 'U':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    y -= choice(np.arange(5,25), 1, p=probs)\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      x += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      x -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  elif movement == 'D':\n",
        "    # generate probability list for weighted movement\n",
        "    prob_list = dirichlet(range(1,21), 1)\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    y += choice(np.arange(5,25), 1, p=probs)\n",
        "    prob_list = dirichlet(range(1,21), 1) # first arguement must be >0\n",
        "    # descending sort\n",
        "    probs = -np.sort(-prob_list)[0]\n",
        "    flip = randint(0,2)\n",
        "    if flip == 0:\n",
        "      x += choice(np.arange(0,20), 1, p=probs)\n",
        "    else:\n",
        "      x -= choice(np.arange(0,20), 1, p=probs)\n",
        "    img1 = Image.new('L', dims)\n",
        "    d = ImageDraw.Draw(img1)\n",
        "    d.text((x,y), text, fill=(255))\n",
        "  else:\n",
        "    print(\"Please use U, D, L, or R to indicate movement direction.\")\n",
        "  if show == True:\n",
        "    fig = plt.figure()\n",
        "    a=fig.add_subplot(1,2,1)\n",
        "    a.set_title('Before')\n",
        "    plt.imshow(np.asarray(img))\n",
        "    a=fig.add_subplot(1,2,2)\n",
        "    a.set_title('After '+ movement + ' shift')\n",
        "    plt.imshow(np.asarray(img1))\n",
        "  if to_matrix == True:\n",
        "    stimulus, label = np.array([np.asarray(img), np.asarray(img1)]), movement\n",
        "    shift_coords = np.asarray((x[0],y[0]))\n",
        "    change = shift_coords - base_coords\n",
        "    return stimulus, label, change"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "COY_lfaegIVl",
        "colab_type": "code",
        "outputId": "6c9c7fda-6e67-4079-a73f-fff19ee32bcd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 216
        }
      },
      "source": [
        "# in order to show the images, all you have to do is indicate show=True\n",
        "\n",
        "# base function\n",
        "test = central_image((128,128), 'Plz graduate', 'R', show=True)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAADHCAYAAADifRM/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pkZ74K5_piMW",
        "colab_type": "code",
        "outputId": "ef2d6823-5d9e-4ad7-b327-b2db9294febd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        }
      },
      "source": [
        "test"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([[[0, 0, 0, ..., 0, 0, 0],\n",
              "         [0, 0, 0, ..., 0, 0, 0],\n",
              "         [0, 0, 0, ..., 0, 0, 0],\n",
              "         ...,\n",
              "         [0, 0, 0, ..., 0, 0, 0],\n",
              "         [0, 0, 0, ..., 0, 0, 0],\n",
              "         [0, 0, 0, ..., 0, 0, 0]],\n",
              " \n",
              "        [[0, 0, 0, ..., 0, 0, 0],\n",
              "         [0, 0, 0, ..., 0, 0, 0],\n",
              "         [0, 0, 0, ..., 0, 0, 0],\n",
              "         ...,\n",
              "         [0, 0, 0, ..., 0, 0, 0],\n",
              "         [0, 0, 0, ..., 0, 0, 0],\n",
              "         [0, 0, 0, ..., 0, 0, 0]]], dtype=uint8), 'R', array([ 15, -10]))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gEZS-1gpOUFD",
        "colab_type": "code",
        "outputId": "8d514566-8d17-4356-9e66-a02a4b9cc7d5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 170
        }
      },
      "source": [
        "# can look at shape of input stimulus as so\n",
        "print(test[0].shape)\n",
        "\n",
        "# each layer like so\n",
        "print(test[0][0])\n",
        "test[0][1]\n",
        "\n",
        "# and its associated label like this\n",
        "print(test[1])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(2, 128, 128)\n",
            "[[0 0 0 ... 0 0 0]\n",
            " [0 0 0 ... 0 0 0]\n",
            " [0 0 0 ... 0 0 0]\n",
            " ...\n",
            " [0 0 0 ... 0 0 0]\n",
            " [0 0 0 ... 0 0 0]\n",
            " [0 0 0 ... 0 0 0]]\n",
            "R\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xlz3okg3XTRe",
        "colab_type": "code",
        "outputId": "d9f2cc6d-e09d-421a-ea36-2d610f824e96",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 233
        }
      },
      "source": [
        "# example of another direction\n",
        "test = central_image((128,128), 'Test', 'R', show=True)\n",
        "test[2]"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([23, 14])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAADHCAYAAADifRM/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_xwJYtn_XdUi",
        "colab_type": "code",
        "outputId": "223024ab-8314-471f-f0f0-9228d47301b7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 216
        }
      },
      "source": [
        "# set to_matrix as False and show to True if you're just playing around\n",
        "central_image((128,128), 'Apples', 'U', show=True, to_matrix=False)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAADHCAYAAADifRM/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JZj36ITY79BM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# generate 300 stimulus pairs for the central shifts\n",
        "directions = ['U','D','L','R']\n",
        "central_list = []\n",
        "\n",
        "for i in range(300):\n",
        "  central_list.append(central_image((128,128), 'Orange', choice(directions)))\n",
        "  \n",
        "# lets add some more variety to the written cases\n",
        "# add more variation if you want\n",
        "text_values = ['king','queen','list','hedge','dog','cat','truck','horse',\\\n",
        "               'prince', 'cow', 'smile','laugh','frown','boat','car',\\\n",
        "              'vehicle','red','silver','glitch','matrix','keanu','fin']\n",
        "\n",
        "# then generate an addition 300 cases for each item in the text list\n",
        "for x in text_values:\n",
        "  for i in range(300):\n",
        "    central_list.append(central_image((128,128), x, choice(directions)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M8ILo3Q5PbO_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# generate 300 stimulus pairs for the trivial shifts\n",
        "directions = ['U','D','L','R']\n",
        "trivial_list = []\n",
        "\n",
        "for i in range(300):\n",
        "  trivial_list.append(trivial_image((128,128), 'Orange', choice(directions)))\n",
        "  \n",
        "# lets add some more variety to the written cases\n",
        "# add more variation if you want\n",
        "text_values = ['king','queen','list','hedge','dog','cat','truck','horse',\\\n",
        "               'prince', 'cow', 'smile','laugh','frown','boat','car',\\\n",
        "              'vehicle','red','silver','glitch','matrix','keanu','fin']\n",
        "\n",
        "# then generate an addition 300 cases for each item in the text list\n",
        "for x in text_values:\n",
        "  for i in range(300):\n",
        "    trivial_list.append(trivial_image((128,128), x, choice(directions)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XFxU6U_RUDnf",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# generate 300 stimulus pairs for the overlap pairs\n",
        "overlap_list = []\n",
        "\n",
        "for i in range(300):\n",
        "  overlap_list.append(overlap_image((128,128), 'Orange', choice(directions)))\n",
        "  \n",
        "# lets add some more variety to the written cases\n",
        "# add more variation if you want\n",
        "text_values = ['king','queen','list','hedge','dog','cat','truck','horse',\\\n",
        "               'prince', 'cow', 'smile','laugh','frown','boat','car',\\\n",
        "              'vehicle','red','silver','glitch','matrix','keanu','fin']\n",
        "\n",
        "# then generate an addition 300 cases for each item in the text list\n",
        "for x in text_values:\n",
        "  for i in range(300):\n",
        "    overlap_list.append(overlap_image((128,128), x, choice(directions)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ojBwG-zpa6Db",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def plot_dataHist(stimuli_array, plot_title, plot=True): \n",
        "  df = []\n",
        "  for i in range(len(stimuli_array)):\n",
        "    df.append((stimuli_array[i][2], stimuli_array[i][1]))\n",
        "  df = pd.DataFrame(df, columns=[\"XY_change\", \"Label\"])\n",
        "  df['Difference'] = 0\n",
        "  for i in range(len(df)):\n",
        "    difference = abs(df[\"XY_change\"][i][1]) - abs(df[\"XY_change\"][i][0])\n",
        "    df.at[i,'Difference'] = difference\n",
        "  if plot:\n",
        "    plt.style.use('seaborn-deep')\n",
        "\n",
        "    R = df[df['Label'] == \"R\"]\n",
        "    L = df[df['Label'] == \"L\"]\n",
        "    U = df[df['Label'] == \"U\"]\n",
        "    D = df[df['Label'] == \"D\"]\n",
        "    bins = np.linspace(-25, 25, 30)\n",
        "\n",
        "    plt.hist([R['Difference'], L['Difference'], U['Difference'], D['Difference']], bins, label=['R','L','U','D'])\n",
        "    plt.title(plot_title)\n",
        "    plt.legend(loc='upper right')\n",
        "    plt.show()\n",
        "  return df"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LAzeoEMIojXN",
        "colab_type": "code",
        "outputId": "fc42590d-beab-4f37-cf1d-2cb878dfab11",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        }
      },
      "source": [
        "trivial_hist = plot_dataHist(trivial_list, \"Trivial\")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3Ny8EZmw8T8_",
        "colab_type": "code",
        "outputId": "7d46ad11-b413-4b77-e801-3336f225557f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        }
      },
      "source": [
        "central_hist = plot_dataHist(central_list, \"Central\")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhrskFchwl6QCGe6SVKD/BwDhg/Hv3MuOAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kgj2MAqKrqRd",
        "colab_type": "code",
        "outputId": "c4f37eb5-5b36-408a-ff07-21b44cdb94dc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        }
      },
      "source": [
        "overlap_hist = plot_dataHist(overlap_list, \"Overlap\")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wNOBbAC0HZxf",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def stim_massage(stimulus_list):\n",
        "  data = []\n",
        "  labels = []\n",
        "  data, labels, change = zip(*stimulus_list)\n",
        "\n",
        "  # reshape image input to include an additional channel \n",
        "  # the -1 is a placeholder that will take the left over number \n",
        "  # number of images in this case\n",
        "  data = np.reshape(np.asarray(data), (-1,2,128,128,1))\n",
        "  print(\"The shape of the data is: {}.\".format(data.shape))\n",
        "  # one hot encode variables\n",
        "  # because we are using categorical crossentropy for loss\n",
        "  labels = get_dummies(labels).values\n",
        "  print(\"The shape of the labels is: {}.\".format(labels.shape))\n",
        "  return data, labels"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "49NwX0-Dy3FT",
        "colab_type": "code",
        "outputId": "3096898b-0b1e-48f6-9389-75a72c931773",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dat, lab = stim_massage(central_list)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The shape of the data is: (6900, 2, 128, 128, 1).\n",
            "The shape of the labels is: (6900, 4).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FzRafX1DpMJd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Old method of building models with Keras\n",
        "\n",
        "# keras.io has documentation for all the available layers etc...\n",
        "\n",
        "model.add(Input(batch_shape=(None,2,128,128,1), name=\"Input\"))\n",
        "model.add(Conv3D(8, (1,3,3), (1,2,2), padding='same',\n",
        "                 activation='relu', name=\"Conv1\"))\n",
        "model.add(BatchNormalization(axis=-1))\n",
        "\n",
        "model.add(Conv3D(16, (1,3,3), (1,2,2), padding='same',\n",
        "                 activation='relu', name=\"Conv2\"))\n",
        "model.add(BatchNormalization(axis=-1))\n",
        "\n",
        "model.add(Conv3D(32, (1,3,3), (1,2,2), padding='same',\n",
        "                 activation='relu', name=\"Conv3\"))\n",
        "model.add(BatchNormalization(axis=-1))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dense(256, activation='relu', name='fc1'))\n",
        "model.add(Dropout(0.50, name='do1'))\n",
        "model.add(Dense(128, activation='relu', name='fc2'))\n",
        "model.add(Dense(len(lab[0]), activation='softmax', name=\"output\"))\n",
        "\n",
        "# compile the network with categorical cross entropy loss function\n",
        "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zxN_qTznjns_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# New method of building models with Keras\n",
        "model = Sequential([\n",
        "    Input(batch_shape=(None,2,128,128,1), name=\"Input\"),\n",
        "\n",
        "    Conv3D(8, (1,3,3), (1,2,2), padding='same', activation='relu', name=\"Conv1\"),\n",
        "    BatchNormalization(axis=-1),\n",
        "\n",
        "    Conv3D(16, (1,3,3), (1,2,2), padding='same', activation='relu', name=\"Conv2\"),\n",
        "    BatchNormalization(axis=-1),\n",
        "\n",
        "    Conv3D(32, (1,3,3), (1,2,2), padding='same', activation='relu', name=\"Conv3\"),\n",
        "    BatchNormalization(axis=-1),\n",
        "\n",
        "    Flatten(),\n",
        "    Dense(256, activation='relu', name='fc1'),\n",
        "    Dropout(0.50, name='do1'),\n",
        "    Dense(128, activation='relu', name='fc2'),\n",
        "    Dense(len(lab[0]), activation='softmax', name=\"output\")\n",
        "])\n",
        "\n",
        "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CA-CtS1dk8uW",
        "colab_type": "code",
        "outputId": "df64423b-2873-4b78-aaaf-ce80b9857855",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 527
        }
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "Conv1 (Conv3D)               (None, 2, 64, 64, 8)      80        \n",
            "_________________________________________________________________\n",
            "batch_normalization_3 (Batch (None, 2, 64, 64, 8)      32        \n",
            "_________________________________________________________________\n",
            "Conv2 (Conv3D)               (None, 2, 32, 32, 16)     1168      \n",
            "_________________________________________________________________\n",
            "batch_normalization_4 (Batch (None, 2, 32, 32, 16)     64        \n",
            "_________________________________________________________________\n",
            "Conv3 (Conv3D)               (None, 2, 16, 16, 32)     4640      \n",
            "_________________________________________________________________\n",
            "batch_normalization_5 (Batch (None, 2, 16, 16, 32)     128       \n",
            "_________________________________________________________________\n",
            "flatten_1 (Flatten)          (None, 16384)             0         \n",
            "_________________________________________________________________\n",
            "fc1 (Dense)                  (None, 256)               4194560   \n",
            "_________________________________________________________________\n",
            "do1 (Dropout)                (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "fc2 (Dense)                  (None, 128)               32896     \n",
            "_________________________________________________________________\n",
            "output (Dense)               (None, 4)                 516       \n",
            "=================================================================\n",
            "Total params: 4,234,084\n",
            "Trainable params: 4,233,972\n",
            "Non-trainable params: 112\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zUpGeZPTxVTp",
        "colab_type": "code",
        "outputId": "7e6eb6d5-4221-470d-af2b-3e0f158a320b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 241
        }
      },
      "source": [
        "# train the model on an 80/20 split for a user decided number of epochs\n",
        "# defaults to a 32 batch size\n",
        "model.fit(dat.astype('float16'), lab,\n",
        "          batch_size=32,\n",
        "          validation_split=0.2, epochs=5,\n",
        "          verbose=1)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 5520 samples, validate on 1380 samples\n",
            "Epoch 1/5\n",
            "5520/5520 [==============================] - 49s 9ms/sample - loss: 1.0598 - acc: 0.6270 - val_loss: 0.5097 - val_acc: 0.8283\n",
            "Epoch 2/5\n",
            "5520/5520 [==============================] - 48s 9ms/sample - loss: 0.5236 - acc: 0.8178 - val_loss: 0.4587 - val_acc: 0.8399\n",
            "Epoch 3/5\n",
            "5520/5520 [==============================] - 48s 9ms/sample - loss: 0.3428 - acc: 0.8793 - val_loss: 0.4387 - val_acc: 0.8580\n",
            "Epoch 4/5\n",
            "5520/5520 [==============================] - 50s 9ms/sample - loss: 0.2234 - acc: 0.9201 - val_loss: 0.5242 - val_acc: 0.8507\n",
            "Epoch 5/5\n",
            "5520/5520 [==============================] - 48s 9ms/sample - loss: 0.1712 - acc: 0.9424 - val_loss: 0.4894 - val_acc: 0.8601\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7fd23ccdce80>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "outputId": "711d6d2b-4cf4-42c4-adfa-c14a85e08375",
        "id": "hVK9qP99ldLz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 102
        }
      },
      "source": [
        "# test the model with the overlapped case\n",
        "test_data, test_lab = stim_massage(overlap_list)\n",
        "test_loss, test_acc = model.evaluate(test_data,  test_lab, verbose=1)\n",
        "\n",
        "print('\\nTest accuracy:', test_acc)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The shape of the data is: (6900, 2, 128, 128, 1).\n",
            "The shape of the labels is: (6900, 4).\n",
            "6900/6900 [==============================] - 22s 3ms/sample - loss: 1.1087 - acc: 0.7196\n",
            "\n",
            "Test accuracy: 0.7195652\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0hw_SRc9lsGH",
        "colab_type": "code",
        "outputId": "a5e2a958-8d97-4d7b-c580-92667f39e890",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 250
        }
      },
      "source": [
        "# set up a test case\n",
        "test_image = trivial_image((128,128), 'd(^-^)b','L',show=True)\n",
        "print(\"The pixel movement of the test case is {} to the {}.\".format(test_image[2], \n",
        "                                                                    test_image[1]))\n",
        "test = np.reshape(test_image[0], (-1,2,128,128,1))\n",
        "\n",
        "# outputs probability of class in the order below\n",
        "# down, left, right, up\n",
        "model.predict(test)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The pixel movement of the test case is [-11   0] to the L.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.21077012, 0.00340331, 0.7383659 , 0.04746066]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAADHCAYAAADifRM/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GVcRVR5fIoDR",
        "colab_type": "code",
        "outputId": "3e352e73-53d9-471e-b3d3-1518c25c001c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 221
        }
      },
      "source": [
        "model_vis = Model(inputs=model.inputs, outputs=model.layers[0].output)\n",
        "model_vis.summary()\n",
        "feature_maps = model_vis.predict(test)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "Input (InputLayer)           [(None, 2, 128, 128, 1)]  0         \n",
            "_________________________________________________________________\n",
            "Conv1 (Conv3D)               (None, 2, 64, 64, 8)      80        \n",
            "=================================================================\n",
            "Total params: 80\n",
            "Trainable params: 80\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aGAwZuDrvdTQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "plt.rcParams['figure.figsize'] = 20, 10"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ftxlZGGgJu0p",
        "colab_type": "code",
        "outputId": "31585207-e6f3-472e-e8a8-6b50757bf052",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 173
        }
      },
      "source": [
        "# TODO make bigger plots\n",
        "# plot all 8 maps in an 8x8 squares\n",
        "square = 8\n",
        "ix = 1\n",
        "for _ in range(square):\n",
        "  # specify subplot and turn of axis\n",
        "  ax = plt.subplot(1, square, ix)\n",
        "  ax.set_xticks([])\n",
        "  ax.set_yticks([])\n",
        "  # plot filter channel in grayscale\n",
        "  plt.imshow(feature_maps[0, 0, :, :, ix-1], cmap='gray', aspect=\"equal\")\n",
        "  ix += 1\n",
        "# show the figure\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABG0AAACICAYAAABdqBblAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAL2klEQVR4nO3daYxd4x8H8O/800QtpbboC8KLBolK\nRGoJJYRKS1qKhMY2iISUWkJEbLEFIRpENBEhFbEnligxxO6FJSFEbCOpJYqK9QVC7v/F/J97Zm5n\n2s7ttH2m/8/nzcTc3z3nTO7jOed+z+857Wm1WgEAAACgLv/Z0AcAAAAAwMqENgAAAAAVEtoAAAAA\nVEhoAwAAAFAhoQ0AAABAhYQ2AAAAABWaMJrinp4e/z74RqrVavWsq20bNxu1Fa1Wa/t1tXFjZ+Nl\nzqFL5hy6Ys6hS+YcumLOoUvDzjk6bYC1sWxDHwDwf8WcA6xP5hxgfRp2zhHaAAAAAFRIaAMAAABQ\nIaENAAAAQIWENgAAAAAVEtoAAAAAVEhoAwAAAFAhoQ0AAABAhYQ2AAAAABUS2gAAAABUSGgDAAAA\nUCGhDQAAAECFhDYAAAAAFRLaAAAAAFRIaAMAAABQIaENAAAAQIWENgAAAAAVEtoAAAAAVEhoAwAA\nAFAhoQ0AAABAhYQ2AAAAABUS2gAAAABUSGgDAAAAUCGhDQAAAECFhDYAAAAAFRLaAAAAAFRIaAMA\nAABQIaENAAAAQIWENgAAAAAVEtoAAAAAVEhoAwAAAFAhoQ0AAABAhYQ2AAAAABUS2gAAAABUSGgD\nAAAAUCGhDQAAAECFhDYAAAAAFRLaAAAAAFRIaAMAAABQIaENAAAAQIWENgAAAAAVEtoAAAAAVEho\nAwAAAFAhoQ0AAABAhYQ2AAAAABUS2gAAAABUSGgDAAAAUCGhDQAAAECFhDYAAAAAFRLaAAAAAFRI\naAMAAABQIaENAAAAQIUmbOgD+H/ywgsvJEm23HLLJMn+++/ffm3mzJlJkr6+viG1RxxxxBpvt7O2\nc5utVitbbbVVkuS3337r7o9gg9l6662TNOPmueeeG7F29uzZq60ZTe3UqVOTJF988cWaHSxVmDRp\nUpKkp6cnSXLssccmSe6///4Ra4877rgRa7qpLftutVprfuBsUOU88euvvyZpzlnDnTdKbfl8V3Vu\nGU0t49Pdd9+dJPnuu++SJJtuummS5LLLLhux9pdffhmxpptaxp/zzjsvSXLnnXdusFo2DhMmDHy1\n/eeff3L44YcnSV588cURa0azvSTDbnM019zUYbPNNkuSnHDCCUmS559/PsnAuevggw9Okrz22mtD\nass1TDm/rWq7nbWd25w9e3a75uGHH16jY9ZpAwAAAFChntHc/ezp6XGrdB3r7JopP8vvu60tTjnl\nlCTJU089laS509lqtXrW+uBHYNysncmTJydp7hgddNBBSQY+75LOnnjiiUmSu+66K0nywQcfJEm+\n/vrrJMOn/6urXbRoUbvu6quvTpJcc801SYZ03LzXarWmr+3fOBJjZ+3stttuSZJPPvkkSbL99tsn\nSVasWNGuKV0zjz76aJLmjlFvb2+SoV00q6t9+eWXkyTLli1LkixevDhnn312kpU7bsw59bv22muT\nJFdddVWSZOLEifnzzz+TNHeRdtpppyTJnnvumSR59tlnkwztolld7a677pokeffdd9vvKV1cTzzx\nROdhmXMqNn36wEdTPsvtttsuycAY6O/vT9J0zZxzzjlD3nvjjTcmGdpFs7raTz/9NEkzT9177705\n88wzhz02c874sckmmyRJ/vrrr3aXzGeffZZkbDpsynmonJd++OGH7LHHHkmSH3/8sbPcnDOOlNUF\nM2fOzIcffpikOecUO++8c5Lk22+/TbLqjpuRap988skkyTHHHJNbb701SfLSSy8laa6jzTn1uvnm\nm5Mk8+bNS9J8t1q4cGHeeOONJMmrr76aJPnyyy+TJBdeeGGS5JVXXkkytOOmdNiMVHvAAQckSXbf\nffckyZQpU/LTTz8laa7RB3XcDDvn6LQBAAAAqJBn2lSm87k0q+qaGU1t8cADDyRJpk2bliT56KOP\nRnuIrGfff/99kuTAAw9M0qTB06ZNyzbbbJMkmTVrVpJkwYIFSZIZM2YkSd56660kyR133NHe3i67\n7JIkmTt37pDat99+O0my1157JWnS4C222CInn3xykqZT65lnnkmSzJkzZ4z+SsbSDjvskKTpjOl8\n5syCBQvyyCOPJEnOOuusJMn111+fpOmmKnerBvv999+HrX366aeTJBdddFGS5Pzzz08ycJe0dNqU\n9d+ei1SvznX55Zk2xeTJk7N8+fIhr+27775JkoceeihJ08H38ccft99XftdZe9JJJyVJHnzwwSRD\nnxVQOmzKmCtjmTq9/vrrSZrzU6f+/v521035vDtdd911K/2us8OmKN04pWOizHFnnHHGSp0206dP\nHzIeqcepp56aJFmyZMmINRdffHGSgfNJ0nTolc/08ssvT5L8/fff7feUDpvO2nJ3/cgjjxyyj9NO\nO63dYbPtttsmSfsuOHWbP39+kua8Mlg5XxXl3HbJJZckaToBB9eVjqvSPVNqjzrqqCTNSoV99tmn\n/Z7SRVq2f+6557avsahL+e5cunm/+eabIa+ffvrp7e9M5ZplypQpSZrvPP/++2+SppsmabqLO2tL\nN07pQv3555+TJH/88Ud7+0X5PlY6fTpZHrUelZNIadvcEMqX7hLeFFr46nPYYYclaSaAHXfcMUnT\nqtnX19e+oF2TB1avqRLaPP744+3/7ny4ZPm5aNEibcMVKl9+rrzyyiTNUoVy0njnnXfaX9AHL5Va\nW7fffnuSpkX4q6++yvvvv5+kWTp16KGHJjHn1KyEJ2XJZFmCMn/+/GEvjMfKDTfckGTgS1i5oCoX\nLyW8jqUKVSrh7OLFi4d9fdKkSe0lL2M553Quc+nt7R3xwejmnHqVJZJlCVSxZMmS9o2FEtqMhXL+\nK9dX9913X3v5cAlvymvLly8351Ssc1nl4Dlh8FKpsVbmmd7e3nbAUwLGpUuXZuHChfn888/NOZW5\n4oorkjT/n5dzVnkg8bx583LBBRckGXLdsdbKOC3L6m677bb2+CzB4Oabb17KLY8CAAAAGC902lRi\n6dKlSVZu2Vxf3IGqV2mXmzhxYpJmmcmsWbPaSfF77703ZvsrdwoGtwd27nsQd6AqVpYv7b333kmS\nm266KcnAPz1Y7nbfc889Y7a/xx57bMh+5s6du9K+33zzzSTmnPGk3KXs6+tbqbNhLEydOjVJst9+\n+yUZunxm8L7/x5wzDs2YMWPElu/1xZxTr1tuuSVJ87D88rMsSVlXyvKYSy+9dFX7NudUrDzu4eij\nj07SXDOXbqp1peyvv79/xH2bc+pVli+V7r7SUXzIIYe0Hx48lo4//vgkzSqGOXPmrLTvQXTaAAAA\nAIwXOm1IIg2ma+5A0RVzDl0y59AVcw5dMufQFXMOXdJpAwAAADBeCG0AAAAAKiS0AQAAAKiQ0AYA\nAACgQkIbAAAAgAoJbQAAAAAqJLQBAAAAqJDQBgAAAKBCQhsAAACACgltAAAAACoktAEAAACokNAG\nAAAAoEJCGwAAAIAKCW0AAAAAKiS0AQAAAKiQ0AYAAACgQkIbAAAAgAoJbQAAAAAqJLQBAAAAqJDQ\nBgAAAKBCQhsAAACACgltAAAAACoktAEAAACokNAGAAAAoEJCGwAAAIAKCW0AAAAAKiS0AQAAAKiQ\n0AYAAACgQkIbAAAAgAoJbQAAAAAqJLQBAAAAqJDQBgAAAKBCQhsAAACACgltAAAAACoktAEAAACo\nkNAGAAAAoEJCGwAAAIAKCW0AAAAAKiS0AQAAAKiQ0AYAAACgQkIbAAAAgAoJbQAAAAAqJLQBAAAA\nqJDQBgAAAKBCQhsAAACACgltAAAAACoktAEAAACokNAGAAAAoEJCGwAAAIAKCW0AAAAAKiS0AQAA\nAKiQ0AYAAACgQkIbAAAAgAoJbQAAAAAqJLQBAAAAqNCEUdavSLJsXRwIG9TO63j7xs3Gy9ihG8YN\n3TJ26IZxQ7eMHbph3NCtYcdOT6vVWt8HAgAAAMBqWB4FAAAAUCGhDQAAAECFhDYAAAAAFRLaAAAA\nAFRIaAMAAABQIaENAAAAQIWENgAAAAAVEtoAAAAAVEhoAwAAAFCh/wJ+R9HnESFbWAAAAABJRU5E\nrkJggg==\n",
            "text/plain": [
              "<Figure size 1440x720 with 8 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}