Proof_of_Concept_GenNet_Non_Linear.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.9"
    },
    "toc": {
      "base_numbering": 1,
      "nav_menu": {},
      "number_sections": true,
      "sideBar": true,
      "skip_h1_title": false,
      "title_cell": "Table of Contents",
      "title_sidebar": "Contents",
      "toc_cell": false,
      "toc_position": {},
      "toc_section_display": true,
      "toc_window_display": false
    },
    "colab": {
      "name": "Proof of Concept GenNet: Non Linear",
      "provenance": [],
      "collapsed_sections": [
        "VEOeUcbzpJll",
        "1MpsGk6LxDaY"
      ],
      "include_colab_link": true
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ArnovanHilten/GenNet/blob/dev/Proof_of_Concept_GenNet_Non_Linear.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cojwEwHAjJ4I"
      },
      "source": [
        "# Proof of Concept GenNet\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GLq3XanB-SzG"
      },
      "source": [
        "1. Copy to playground\n",
        "2. Press Runtime > Run all or ctrl + F9"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VEOeUcbzpJll"
      },
      "source": [
        "## Initialization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XtVzIkV7sTZM",
        "outputId": "2514a1f1-34b1-4b2b-ab4d-450d1d705207"
      },
      "source": [
        "pip install numpy==1.19.3"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: numpy==1.19.3 in /usr/local/lib/python3.7/dist-packages (1.19.3)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MN6IJvRfsyDV",
        "outputId": "2cc3f1b3-d459-46bf-d6a0-9707016b3d5f"
      },
      "source": [
        "pip install tensorflow==2.4.0"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: tensorflow==2.4.0 in /usr/local/lib/python3.7/dist-packages (2.4.0)\n",
            "Requirement already satisfied: keras-preprocessing~=1.1.2 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (1.1.2)\n",
            "Requirement already satisfied: tensorflow-estimator<2.5.0,>=2.4.0rc0 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (2.4.0)\n",
            "Requirement already satisfied: astunparse~=1.6.3 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (1.6.3)\n",
            "Requirement already satisfied: h5py~=2.10.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (2.10.0)\n",
            "Requirement already satisfied: six~=1.15.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (1.15.0)\n",
            "Requirement already satisfied: absl-py~=0.10 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (0.12.0)\n",
            "Requirement already satisfied: numpy~=1.19.2 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (1.19.3)\n",
            "Requirement already satisfied: protobuf>=3.9.2 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (3.17.3)\n",
            "Requirement already satisfied: flatbuffers~=1.12.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (1.12)\n",
            "Requirement already satisfied: google-pasta~=0.2 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (0.2.0)\n",
            "Requirement already satisfied: wheel~=0.35 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (0.37.0)\n",
            "Requirement already satisfied: wrapt~=1.12.1 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (1.12.1)\n",
            "Requirement already satisfied: tensorboard~=2.4 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (2.6.0)\n",
            "Requirement already satisfied: gast==0.3.3 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (0.3.3)\n",
            "Requirement already satisfied: opt-einsum~=3.3.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (3.3.0)\n",
            "Requirement already satisfied: grpcio~=1.32.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (1.32.0)\n",
            "Requirement already satisfied: termcolor~=1.1.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (1.1.0)\n",
            "Requirement already satisfied: typing-extensions~=3.7.4 in /usr/local/lib/python3.7/dist-packages (from tensorflow==2.4.0) (3.7.4.3)\n",
            "Requirement already satisfied: setuptools>=41.0.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard~=2.4->tensorflow==2.4.0) (57.4.0)\n",
            "Requirement already satisfied: tensorboard-data-server<0.7.0,>=0.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard~=2.4->tensorflow==2.4.0) (0.6.1)\n",
            "Requirement already satisfied: requests<3,>=2.21.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard~=2.4->tensorflow==2.4.0) (2.23.0)\n",
            "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard~=2.4->tensorflow==2.4.0) (1.8.0)\n",
            "Requirement already satisfied: google-auth<2,>=1.6.3 in /usr/local/lib/python3.7/dist-packages (from tensorboard~=2.4->tensorflow==2.4.0) (1.34.0)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.7/dist-packages (from tensorboard~=2.4->tensorflow==2.4.0) (3.3.4)\n",
            "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.7/dist-packages (from tensorboard~=2.4->tensorflow==2.4.0) (1.0.1)\n",
            "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /usr/local/lib/python3.7/dist-packages (from tensorboard~=2.4->tensorflow==2.4.0) (0.4.5)\n",
            "Requirement already satisfied: cachetools<5.0,>=2.0.0 in /usr/local/lib/python3.7/dist-packages (from google-auth<2,>=1.6.3->tensorboard~=2.4->tensorflow==2.4.0) (4.2.2)\n",
            "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.7/dist-packages (from google-auth<2,>=1.6.3->tensorboard~=2.4->tensorflow==2.4.0) (0.2.8)\n",
            "Requirement already satisfied: rsa<5,>=3.1.4 in /usr/local/lib/python3.7/dist-packages (from google-auth<2,>=1.6.3->tensorboard~=2.4->tensorflow==2.4.0) (4.7.2)\n",
            "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.7/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard~=2.4->tensorflow==2.4.0) (1.3.0)\n",
            "Requirement already satisfied: importlib-metadata in /usr/local/lib/python3.7/dist-packages (from markdown>=2.6.8->tensorboard~=2.4->tensorflow==2.4.0) (4.6.4)\n",
            "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in /usr/local/lib/python3.7/dist-packages (from pyasn1-modules>=0.2.1->google-auth<2,>=1.6.3->tensorboard~=2.4->tensorflow==2.4.0) (0.4.8)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests<3,>=2.21.0->tensorboard~=2.4->tensorflow==2.4.0) (1.24.3)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests<3,>=2.21.0->tensorboard~=2.4->tensorflow==2.4.0) (2.10)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests<3,>=2.21.0->tensorboard~=2.4->tensorflow==2.4.0) (3.0.4)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests<3,>=2.21.0->tensorboard~=2.4->tensorflow==2.4.0) (2021.5.30)\n",
            "Requirement already satisfied: oauthlib>=3.0.0 in /usr/local/lib/python3.7/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard~=2.4->tensorflow==2.4.0) (3.1.1)\n",
            "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata->markdown>=2.6.8->tensorboard~=2.4->tensorflow==2.4.0) (3.5.0)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yvR9Nzea8lPj"
      },
      "source": [
        "#@title Importing python modules { display-mode: \"form\" }\n",
        "\n",
        "import tensorflow as tf\n",
        "import os\n",
        "from scipy.sparse import coo_matrix\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "import tensorflow.keras as K\n",
        "import sklearn.metrics as skm\n",
        "import scipy\n",
        "import sys\n",
        "\n",
        "from __future__ import absolute_import\n",
        "from __future__ import division\n",
        "from __future__ import print_function\n",
        "\n",
        "from tensorflow.python.keras import activations\n",
        "from tensorflow.python.keras import backend as Kb\n",
        "from tensorflow.python.keras import constraints\n",
        "from tensorflow.python.keras import initializers\n",
        "from tensorflow.python.keras import regularizers\n",
        "from tensorflow.python.keras.engine.base_layer import InputSpec\n",
        "from tensorflow.python.keras.engine.base_layer import Layer\n",
        "from tensorflow.python.keras.utils import conv_utils\n",
        "from tensorflow.python.keras.utils import tf_utils\n",
        "from tensorflow.python.util.tf_export import tf_export\n"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NasN2T_r8plA"
      },
      "source": [
        "#@title Locally-Directed1D layer\n",
        "\"\"\"Locally-Directed1D layer.\n",
        "\"\"\"\n",
        "\n",
        "\n",
        "\n",
        "@tf_export('keras.layers.LocallyDirected1D')\n",
        "class LocallyDirected1D(Layer):\n",
        "    \"\"\"Locally-Directed1D layer for 1D inputs.\n",
        "\n",
        "  The `LocallyDirected1D` layer works similarly to\n",
        "  the `Conv1D` layer, except that weights are unshared,\n",
        "  that is, a different set of filters is applied at each different patch\n",
        "  of the input.\n",
        "\n",
        "  Example:\n",
        "  ```python\n",
        "      # apply a unshared weight convolution 1d of length 3 to a sequence with\n",
        "      # 10 timesteps, with 64 output filters\n",
        "      model = Sequential()\n",
        "      model.add(LocallyDirected1D(64, 3, input_shape=(10, 32)))\n",
        "      # now model.output_shape == (None, 8, 64)\n",
        "      # add a new conv1d on top\n",
        "      model.add(LocallyDirected1D(32, 3))\n",
        "      # now model.output_shape == (None, 6, 32)\n",
        "  ```\n",
        "\n",
        "  Arguments:\n",
        "      mask: sparse matrix with shape (input, output) connectivity matrix,\n",
        "            True defines connection between (in_i, out_j), should be sparse (False,0) >> True\n",
        "            should be scipy sparese matrix in COO Format!\n",
        "      filters: Integer, the dimensionality of the output space\n",
        "          (i.e. the number of output filters in the convolution).\n",
        "      kernel_size: An integer or tuple/list of a single integer,\n",
        "          specifying the length of the 1D convolution window.\n",
        "      strides: An integer or tuple/list of a single integer,\n",
        "          specifying the stride length of the convolution.\n",
        "          Specifying any stride value != 1 is incompatible with specifying\n",
        "          any `dilation_rate` value != 1.\n",
        "      padding: Currently only supports `\"valid\"` (case-insensitive).\n",
        "          `\"same\"` may be supported in the future.\n",
        "      data_format: A string,\n",
        "          one of `channels_last` (default) or `channels_first`.\n",
        "          The ordering of the dimensions in the inputs.\n",
        "          `channels_last` corresponds to inputs with shape\n",
        "          `(batch, length, channels)` while `channels_first`\n",
        "          corresponds to inputs with shape\n",
        "          `(batch, channels, length)`.\n",
        "          It defaults to the `image_data_format` value found in your\n",
        "          Keras config file at `~/.keras/keras.json`.\n",
        "          If you never set it, then it will be \"channels_last\".\n",
        "      activation: Activation function to use.\n",
        "          If you don't specify anything, no activation is applied\n",
        "          (ie. \"linear\" activation: `a(x) = x`).\n",
        "      use_bias: Boolean, whether the layer uses a bias vector.\n",
        "      kernel_initializer: Initializer for the `kernel` weights matrix.\n",
        "      bias_initializer: Initializer for the bias vector.\n",
        "      kernel_regularizer: Regularizer function applied to\n",
        "          the `kernel` weights matrix.\n",
        "      bias_regularizer: Regularizer function applied to the bias vector.\n",
        "      activity_regularizer: Regularizer function applied to\n",
        "          the output of the layer (its \"activation\")..\n",
        "      kernel_constraint: Constraint function applied to the kernel matrix.\n",
        "      bias_constraint: Constraint function applied to the bias vector.\n",
        "\n",
        "  Input shape:\n",
        "      3D tensor with shape: `(batch_size, steps, input_dim)`\n",
        "\n",
        "  Output shape:\n",
        "      3D tensor with shape: `(batch_size, new_steps, filters)`\n",
        "      `steps` value might have changed due to padding or strides.\n",
        "  \"\"\"\n",
        "\n",
        "    def __init__(self,\n",
        "                 mask,\n",
        "                 filters,\n",
        "                 padding='valid',\n",
        "                 data_format=None,\n",
        "                 activation=None,\n",
        "                 use_bias=True,\n",
        "                 kernel_initializer='glorot_uniform',\n",
        "                 bias_initializer='zeros',\n",
        "                 kernel_regularizer=None,\n",
        "                 bias_regularizer=None,\n",
        "                 activity_regularizer=None,\n",
        "                 kernel_constraint=None,\n",
        "                 bias_constraint=None,\n",
        "                 **kwargs):\n",
        "        super(LocallyDirected1D, self).__init__(**kwargs)\n",
        "        self.filters = filters\n",
        "        self.padding = conv_utils.normalize_padding(padding)\n",
        "        self.data_format = conv_utils.normalize_data_format(data_format)\n",
        "        self.activation = activations.get(activation)\n",
        "        self.use_bias = use_bias\n",
        "        self.kernel_initializer = initializers.get(kernel_initializer)\n",
        "        self.bias_initializer = initializers.get(bias_initializer)\n",
        "        self.kernel_regularizer = regularizers.get(kernel_regularizer)\n",
        "        self.bias_regularizer = regularizers.get(bias_regularizer)\n",
        "        self.activity_regularizer = regularizers.get(activity_regularizer)\n",
        "        self.kernel_constraint = constraints.get(kernel_constraint)\n",
        "        self.bias_constraint = constraints.get(bias_constraint)\n",
        "        self.input_spec = InputSpec(ndim=3)\n",
        "        self.mask = mask\n",
        "\n",
        "    @tf_utils.shape_type_conversion\n",
        "    def build(self, input_shape):\n",
        "        if self.data_format == 'channels_first':\n",
        "            input_dim, input_length = input_shape[1], input_shape[2]\n",
        "        else:\n",
        "            input_dim, input_length = input_shape[2], input_shape[1]\n",
        "\n",
        "        if input_dim is None:\n",
        "            raise ValueError('Axis 2 of input should be fully-defined. '\n",
        "                             'Found shape:', input_shape)\n",
        "        self.output_length = self.mask.shape[1]\n",
        "        if self.data_format == 'channels_first':\n",
        "            self.kernel_shape = (input_dim, input_length,\n",
        "                                 self.filters, self.output_length)\n",
        "        else:\n",
        "            self.kernel_shape = (input_length, input_dim,\n",
        "                                 self.output_length, self.filters)\n",
        "\n",
        "\n",
        "        self.kernel = self.add_weight(shape=(len(self.mask.data),),    #sum of all nonzero values in mask sum(sum(mask))\n",
        "                                      initializer=self.kernel_initializer,\n",
        "                                      name='kernel',\n",
        "                                      regularizer=self.kernel_regularizer,\n",
        "                                      constraint=self.kernel_constraint)\n",
        "        self.kernel_idx = sorted(get_idx(self.mask))\n",
        "\n",
        "\n",
        "        if self.use_bias:\n",
        "            self.bias = self.add_weight(\n",
        "                shape=(self.output_length, self.filters),\n",
        "                initializer=self.bias_initializer,\n",
        "                name='bias',\n",
        "                regularizer=self.bias_regularizer,\n",
        "                constraint=self.bias_constraint)\n",
        "        else:\n",
        "            self.bias = None\n",
        "\n",
        "        if self.data_format == 'channels_first':\n",
        "            self.input_spec = InputSpec(ndim=3, axes={1: input_dim})\n",
        "        else:\n",
        "            self.input_spec = InputSpec(ndim=3, axes={-1: input_dim})\n",
        "        self.built = True\n",
        "\n",
        "\n",
        "    def call(self, inputs):\n",
        "\n",
        "        # output = local_conv_matmul(inputs, self.kernel_mask,\n",
        "        #                            self.output_length)\n",
        "\n",
        "        output = local_conv_matmul_sparse(inputs, self.mask, self.kernel, self.kernel_idx, self.output_length, self.filters)\n",
        "\n",
        "\n",
        "\n",
        "        if self.use_bias:\n",
        "            output = Kb.bias_add(output, self.bias, data_format=self.data_format)\n",
        "\n",
        "        output = self.activation(output)\n",
        "        return output\n",
        "\n",
        "    # def get_config(self):  # delete this?\n",
        "    #     config = {\n",
        "    #         'filters':\n",
        "    #             self.filters,\n",
        "    #         'padding':\n",
        "    #             self.padding,\n",
        "    #         'data_format':\n",
        "    #             self.data_format,\n",
        "    #         'activation':\n",
        "    #             activations.serialize(self.activation),\n",
        "    #         'use_bias':\n",
        "    #             self.use_bias,\n",
        "    #         'kernel_initializer':\n",
        "    #             initializers.serialize(self.kernel_initializer),\n",
        "    #         'bias_initializer':\n",
        "    #             initializers.serialize(self.bias_initializer),\n",
        "    #         'kernel_regularizer':\n",
        "    #             regularizers.serialize(self.kernel_regularizer),\n",
        "    #         'bias_regularizer':\n",
        "    #             regularizers.serialize(self.bias_regularizer),\n",
        "    #         'activity_regularizer':\n",
        "    #             regularizers.serialize(self.activity_regularizer),\n",
        "    #         'kernel_constraint':\n",
        "    #             constraints.serialize(self.kernel_constraint),\n",
        "    #         'bias_constraint':\n",
        "    #             constraints.serialize(self.bias_constraint),\n",
        "    #     }\n",
        "    #     base_config = super(LocallyDirected1D, self).get_config()\n",
        "    #     return dict(list(base_config.items()) + list(config.items()))\n",
        "\n",
        "\n",
        "def local_conv_matmul_sparse(inputs, mask, kernel, kernel_idx, output_length, filters):\n",
        "    \"\"\"Apply N-D convolution with un-shared weights using a single matmul call.\n",
        "\n",
        "  Arguments:\n",
        "      inputs: (N+2)-D tensor with shape\n",
        "          `(batch_size, channels_in, d_in1, ..., d_inN)`\n",
        "          or\n",
        "          `(batch_size, d_in1, ..., d_inN, channels_in)`.\n",
        "      mask: sparse matrix COO format connectivity matrix, shape: (input layer, output layer)\n",
        "      kernel_idx =  the transposed mask matrix indices\n",
        "      kernel: the unshared weights for N-D convolution,\n",
        "          an (N+2)-D tensor of shape:\n",
        "          `(d_in1, ..., d_inN, channels_in, d_out2, ..., d_outN, channels_out)`\n",
        "          or\n",
        "          `(channels_in, d_in1, ..., d_inN, channels_out, d_out2, ..., d_outN)`,\n",
        "          with the ordering of channels and spatial dimensions matching\n",
        "          that of the input.\n",
        "          Each entry is the weight between a particular input and\n",
        "          output location, similarly to a fully-connected weight matrix.\n",
        "      output_shape: (mask.shape[1], mask.shape[0]) is used instead.\n",
        "      output_length = length of the output\n",
        "      filters =  standard 1\n",
        "\n",
        "  Returns:\n",
        "      Output (N+2)-D tensor with shape `output_shape`.\n",
        "  \"\"\"\n",
        "    output_shape = (mask.shape[1], mask.shape[0])\n",
        "    inputs_flat = Kb.reshape(inputs, (Kb.shape(inputs)[0], -1)) \n",
        "\n",
        "    print(\"kernel_idx\", len(kernel_idx))\n",
        "    print(\"inputs\", Kb.shape(inputs_flat))\n",
        "    print(\"kernel\", Kb.shape(kernel))\n",
        "\n",
        "    output_flat = Kb.sparse_ops.sparse_tensor_dense_mat_mul(\n",
        "        kernel_idx, kernel, (mask.shape[1], mask.shape[0]), inputs_flat, adjoint_b=True)\n",
        "    # tf.print(kernel)\n",
        "    \n",
        "    output_flat_transpose= Kb.transpose(output_flat)\n",
        "    output_reshaped = Kb.reshape(output_flat_transpose, [-1, output_length, filters])\n",
        "    # output_reshaped = Kb.reshape(output_flat_transpose,[Kb.shape(output_flat_transpose)[0],] + output_shape.as_list()[1:])\n",
        "    return output_reshaped\n",
        "\n",
        "def get_idx(mask):\n",
        "  \"\"\"\"returns the transposed coordinates in tuple form:\n",
        "   [(mask.col[0], mask,row[0])...[mask.col[n], mask.row[n])]\"\"\"\n",
        "  coor_list = []\n",
        "  for i, j in zip(mask.col, mask.row):\n",
        "    coor_list.append((i,j))\n",
        "\n",
        "  return coor_list\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NeWVBhf1VxlH"
      },
      "source": [
        "#@title Utility functions { display-mode: \"form\" }\n",
        "#\n",
        "\n",
        "tf.keras.backend.set_epsilon(0.00001)\n",
        "global weight_negative_class\n",
        "global weight_possitive_class\n",
        "global data_path\n",
        "\n",
        "def weighted_binary_crossentropy(y_true, y_pred):\n",
        "    y_true = K.backend.clip(y_true, 0.0001, 1)\n",
        "    y_pred = K.backend.clip(y_pred, 0.0001, 1)\n",
        "\n",
        "    return K.backend.mean(\n",
        "        -y_true * K.backend.log(y_pred + 0.0001) * weight_possitive_class - (1 - y_true) * K.backend.log(\n",
        "            1 - y_pred + 0.0001) * weight_negative_class)\n",
        "\n",
        "\n",
        "\n",
        "def progressBar(value, end_value, bar_length=50):\n",
        "    percent = float(value + 1) / end_value\n",
        "    arrow = '-' * int(round(percent * bar_length) - 1) + '>'\n",
        "    spaces = ' ' * (bar_length - len(arrow))\n",
        "\n",
        "    sys.stdout.write(\"\\r [{0}] {1}%\".format(arrow + spaces, int(round(percent * 100))))\n",
        "\n",
        "    if value >= end_value:\n",
        "        sys.stdout.write(\"\\n\")\n",
        "\n",
        "    sys.stdout.flush()\n",
        "\n",
        "\n",
        "def sensitivity(y_true, y_pred):\n",
        "    y_true = tf.cast(y_true, tf.float32)\n",
        "    y_pred = tf.cast(y_pred, tf.float32)\n",
        "    y_pred = tf.keras.backend.clip(y_pred, 0, 1)\n",
        "    y_true = tf.keras.backend.clip(y_true, 0, 1)\n",
        "\n",
        "    y_pred = tf.keras.backend.round(y_pred)\n",
        "\n",
        "    true_p = K.backend.sum(K.backend.round(y_pred) * y_true)\n",
        "    pos = tf.keras.backend.sum(y_true)\n",
        "    sensitivity1 = tf.keras.backend.clip((true_p / (pos + 0.00001)), 0, 1)\n",
        "    return sensitivity1\n",
        "\n",
        "def specificity(y_true, y_pred):\n",
        "    y_true = tf.cast(y_true, tf.float32)\n",
        "    y_pred = tf.cast(y_pred, tf.float32)\n",
        "    y_pred = tf.keras.backend.clip(y_pred, 0, 1)\n",
        "    y_true = tf.keras.backend.clip(y_true, 0, 1)\n",
        "\n",
        "    neg_y_true = 1 - y_true\n",
        "    neg_y_pred = 1 - K.backend.round(y_pred)\n",
        "    fp = K.backend.sum(neg_y_true * K.backend.round(y_pred))\n",
        "    tn = K.backend.sum(neg_y_true * neg_y_pred)\n",
        "    specificity1 = tn / (tn + fp + 0.00001)\n",
        "    return tf.keras.backend.clip(specificity1, 0, 1)\n",
        "\n",
        "\n",
        "def evaluate_performance(y,p):\n",
        "    print(\"\\n\")\n",
        "    print(\"Confusion matrix\")\n",
        "    confusion_matrix = skm.confusion_matrix(y, p.round())\n",
        "    print(confusion_matrix)\n",
        "\n",
        "    fpr, tpr, thresholds = skm.roc_curve(y, p)\n",
        "    roc_auc = skm.auc(fpr, tpr)\n",
        "    print(\"\\n\")\n",
        "    print(\"Area under the Curve (AUC) = \", roc_auc)\n",
        "\n",
        "    specificity = confusion_matrix[0, 0] / (confusion_matrix[0, 0] + confusion_matrix[0, 1])\n",
        "    print('Specificity = ', specificity)\n",
        "\n",
        "    sensitivity = confusion_matrix[1, 1] / (confusion_matrix[1, 0] + confusion_matrix[1, 1])\n",
        "    print('Sensitivity = ', sensitivity)\n",
        "    print(\"F_1 score = \" + str(skm.f1_score(y, p.round())))\n",
        "    plt.figure()\n",
        "    plt.plot(fpr, tpr, color='darkorange',\n",
        "            lw=2, label='ROC curve (area = %0.2f)' % roc_auc)\n",
        "    plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--')\n",
        "    plt.plot(1 - specificity, sensitivity, color='b', marker='o')\n",
        "    plt.xlim([0.0, 1.0])\n",
        "    plt.ylim([0.0, 1.05])\n",
        "    plt.xlabel('False Positive Rate')\n",
        "    plt.ylabel('True Positive Rate')\n",
        "    plt.title('Receiver operating characteristic example')\n",
        "    plt.legend(loc=\"lower right\")\n",
        "    plt.show()\n",
        "\n"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wd8fZ2KzpY8S"
      },
      "source": [
        "## Parameters\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2020-04-26T15:18:54.995418Z",
          "start_time": "2020-04-26T15:18:53.807863Z"
        },
        "scrolled": true,
        "id": "2zxl6koE3n3Z"
      },
      "source": [
        "# Hyperparameters of the simulation and the network training.\n",
        "\n",
        "jobid = 1\n",
        "modeltype = \"GenNet_proof_of_concept_network\"\n",
        "weight_possitive_class = 1 \n",
        "weight_negative_class = 1\n",
        "batch_size = 32\n",
        "startindex = 0\n",
        "stopindex = -1\n",
        "inputsize = 100\n",
        "max_id = 10\n",
        "train_size = 10000\n",
        "val_size = 2000\n",
        "test_size = 2000\n",
        "epochs = 25\n",
        "optimizer = tf.keras.optimizers.Adam(learning_rate = 0.001)"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Pny2nqV-pl10"
      },
      "source": [
        "## Create data, mask and network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2020-04-26T15:18:55.011984Z",
          "start_time": "2020-04-26T15:18:54.997483Z"
        },
        "scrolled": true,
        "id": "M6-buUse3n3d",
        "cellView": "form"
      },
      "source": [
        "#@title Functions for creating the network, simulations and mask\n",
        "\n",
        "def GenNet_proof_of_concept_network(inputsize, mask):\n",
        "    ''' The function that creates the neural network. \n",
        "    The locallyDirected1D layer let's you define all the connections yourself \n",
        "    with the help of a mask, the rest is standard Keras\n",
        "    The created network is plotted at the end of the notebook'''\n",
        "\n",
        "    inputs_ = K.Input((inputsize,), name='inputs_')\n",
        "\n",
        "    Input_layer = K.layers.Reshape(input_shape=(inputsize,), target_shape=(inputsize, 1))(inputs_)\n",
        "    \n",
        "    # next line we use the new Gennet layer (LocallyDirected1D) to define all the connections ourself\n",
        "    Gene_layer = LocallyDirected1D(mask=mask, filters=1, input_shape=(inputsize, 1), name=\"gene_layer\",)(Input_layer)\n",
        "    Gene_layer = K.layers.Flatten()(Gene_layer)\n",
        "    Gene_layer = K.layers.Activation(\"tanh\")(Gene_layer) #gene layer\n",
        "\n",
        "    BatchNorm = K.layers.BatchNormalization(center=False, scale=False, name=\"inter_out\")(Gene_layer)\n",
        "    \n",
        "    Prediction_layer = K.layers.Dense(units=1,    kernel_regularizer=K.regularizers.l1(0.01) )(BatchNorm)\n",
        "    Prediction_layer = K.layers.Activation(\"sigmoid\")(Prediction_layer)\n",
        "    model = K.Model(inputs=inputs_, outputs=Prediction_layer)\n",
        "    return model\n",
        "\n",
        "def make_mask_gene_layer(inputsize):\n",
        "    '''We create a simple mask for this network, the first 5 are connected to the \n",
        "    first neuron. SNPs 5 to 10 are connected to the second neuron etc. \n",
        "    We save it as a sparse matrix.\n",
        "    The created network is plotted at the end of the notebook'''\n",
        "    mask_d  = np.zeros((inputsize,10), np.bool)\n",
        "    mask_d[0:5,0]= True\n",
        "    mask_d[5:10,1]= True\n",
        "    mask_d[10:20,2]=True\n",
        "    mask_d[20:30,3]=True\n",
        "    mask_d[30:50,4]=True\n",
        "    mask_d[50:70,5]= True\n",
        "    mask_d[70:80,6]= True\n",
        "    mask_d[80:90,7]=True\n",
        "    mask_d[90:95,8]=True\n",
        "    mask_d[95:100,9]=True\n",
        "    mask =  coo_matrix(mask_d)\n",
        "    gene_end =[0,5,10,20,30,50,70,80,90,95,100]  \n",
        "    return mask, gene_end\n",
        "\n",
        "\n",
        "def get_simdata(num_patients=100, num_features=100,\n",
        "                    ind_linked=[   [[0, 0], [4, 2]],   [[6, 2], [9, 0]], \n",
        "                                   [[12, 2], [19, 2]],   [[21, 2], [29, 0]], [[35, 0], [45, 2]] ] # causal SNPs\n",
        "                    , n=2, p=0.3,\n",
        "                    random_seed=42):\n",
        "    '''A function to create some simulated non-linear data.  \n",
        "    [[0, 0], [4, 2]] means that whenever the first value/SNP is 0 and the 4th has \n",
        "    value 2 then there is an effect'''\n",
        "    \n",
        "    np.random.seed(random_seed)\n",
        "    basis = np.zeros([num_patients, num_features])\n",
        "    effectsize = 1\n",
        "\n",
        "    for k in range(num_features):\n",
        "        basis[:, k] = np.random.binomial(n, p, num_patients)\n",
        "\n",
        "    status = np.zeros(num_patients)\n",
        "    for patient in range(num_patients):\n",
        "        for linked in ind_linked:\n",
        "            temp = np.zeros([len(linked)])\n",
        "            i = 0\n",
        "            for element in linked:\n",
        "                if basis[patient, element[0]] == element[1]:\n",
        "                    temp[i] = 1\n",
        "                i += 1\n",
        "            if np.min(temp) > 0:\n",
        "                status[patient] = 1\n",
        "\n",
        "\n",
        "\n",
        "    num_diseased = np.sum(status)\n",
        "    causal_snps = [[x[0][0] for x in ind_linked],[x[1][0] for x in ind_linked]]\n",
        "    print((\"Created dataset[\", num_patients, \" x \", num_features, \"] with\", num_diseased, \"diseased\"))\n",
        "    return basis, status,causal_snps\n",
        "\n",
        "\n"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "l0sUfPo8p5E1"
      },
      "source": [
        "## Train and evaluate network\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2020-04-26T15:18:56.936114Z",
          "start_time": "2020-04-26T15:18:55.013160Z"
        },
        "scrolled": true,
        "id": "IOnoELnN3n3f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "5fa644c2-06e9-41fa-8f4a-bf222e915e75"
      },
      "source": [
        "# all functions used here can be found under create data. You can easily change the causal SNPs there and run the notebook to see the effect of such a change.\n",
        "\n",
        "genemask, gene_end = make_mask_gene_layer(inputsize)  # get the mask to know which SNP connect to which gene\n",
        "\n",
        "model = GenNet_proof_of_concept_network(inputsize=int(inputsize), mask=genemask)  # create the neural network\n",
        "\n",
        "model.compile(loss=weighted_binary_crossentropy, optimizer=optimizer, metrics=[\"accuracy\", sensitivity, specificity]) # compile the network\n",
        "print(model.summary())\n",
        "\n",
        "xtrain, ytrain, causal_snps =get_simdata(num_features=inputsize,num_patients=train_size, random_seed=1+ jobid) # get the training data, we use a different seed than for the validation and test to get unique data\n",
        "xval, yval, _ =get_simdata(num_features=inputsize,num_patients=val_size, random_seed=21 + jobid) # get validation data, note that it has a different random seed\n",
        "\n",
        "earlystop = K.callbacks.EarlyStopping(monitor='val_loss', min_delta=0.0001, patience=50, verbose=1, mode='auto', restore_best_weights=True) # we stop training after not improving after 50 epochs on the validation data\n",
        "model.fit(x=xtrain, y=ytrain, batch_size=batch_size, epochs=epochs, verbose=1, callbacks=[earlystop], validation_data=(xval,yval), shuffle=True) # train the neural network\n",
        "\n",
        "\n",
        "pval = model.predict(xval) # Get the validation predictions\n",
        "\n",
        "evaluate_performance(yval, pval) # evaluate the validation predictions\n",
        "\n"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "kernel_idx 100\n",
            "inputs Tensor(\"gene_layer/Shape_1:0\", shape=(2,), dtype=int32)\n",
            "kernel Tensor(\"gene_layer/Shape_2:0\", shape=(1,), dtype=int32)\n",
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "inputs_ (InputLayer)         [(None, 100)]             0         \n",
            "_________________________________________________________________\n",
            "reshape (Reshape)            (None, 100, 1)            0         \n",
            "_________________________________________________________________\n",
            "gene_layer (LocallyDirected1 (None, 10, 1)             110       \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 10)                0         \n",
            "_________________________________________________________________\n",
            "activation (Activation)      (None, 10)                0         \n",
            "_________________________________________________________________\n",
            "inter_out (BatchNormalizatio (None, 10)                20        \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 1)                 11        \n",
            "_________________________________________________________________\n",
            "activation_1 (Activation)    (None, 1)                 0         \n",
            "=================================================================\n",
            "Total params: 141\n",
            "Trainable params: 121\n",
            "Non-trainable params: 20\n",
            "_________________________________________________________________\n",
            "None\n",
            "('Created dataset[', 10000, ' x ', 100, '] with', 1744.0, 'diseased')\n",
            "('Created dataset[', 2000, ' x ', 100, '] with', 350.0, 'diseased')\n",
            "Epoch 1/25\n",
            "kernel_idx 100\n",
            "inputs Tensor(\"model/gene_layer/Shape_1:0\", shape=(2,), dtype=int32)\n",
            "kernel Tensor(\"model/gene_layer/Shape_2:0\", shape=(1,), dtype=int32)\n",
            "kernel_idx 100\n",
            "inputs Tensor(\"model/gene_layer/Shape_1:0\", shape=(2,), dtype=int32)\n",
            "kernel Tensor(\"model/gene_layer/Shape_2:0\", shape=(1,), dtype=int32)\n",
            "308/313 [============================>.] - ETA: 0s - loss: 0.7556 - accuracy: 0.5444 - sensitivity: 0.5330 - specificity: 0.5483kernel_idx 100\n",
            "inputs Tensor(\"model/gene_layer/Shape_1:0\", shape=(2,), dtype=int32)\n",
            "kernel Tensor(\"model/gene_layer/Shape_2:0\", shape=(1,), dtype=int32)\n",
            "313/313 [==============================] - 2s 4ms/step - loss: 0.7546 - accuracy: 0.5453 - sensitivity: 0.5331 - specificity: 0.5493 - val_loss: 0.5704 - val_accuracy: 0.8070 - val_sensitivity: 0.3976 - val_specificity: 0.8977\n",
            "Epoch 2/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.5822 - accuracy: 0.7459 - sensitivity: 0.5333 - specificity: 0.7928 - val_loss: 0.4485 - val_accuracy: 0.8535 - val_sensitivity: 0.2276 - val_specificity: 0.9878\n",
            "Epoch 3/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.5067 - accuracy: 0.8340 - sensitivity: 0.4113 - specificity: 0.9268 - val_loss: 0.4079 - val_accuracy: 0.8530 - val_sensitivity: 0.1955 - val_specificity: 0.9938\n",
            "Epoch 4/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.4542 - accuracy: 0.8612 - sensitivity: 0.4007 - specificity: 0.9623 - val_loss: 0.3676 - val_accuracy: 0.8510 - val_sensitivity: 0.1513 - val_specificity: 0.9981\n",
            "Epoch 5/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.4072 - accuracy: 0.8714 - sensitivity: 0.4004 - specificity: 0.9752 - val_loss: 0.3335 - val_accuracy: 0.8590 - val_sensitivity: 0.2057 - val_specificity: 0.9969\n",
            "Epoch 6/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.3651 - accuracy: 0.8854 - sensitivity: 0.4851 - specificity: 0.9779 - val_loss: 0.2996 - val_accuracy: 0.8875 - val_sensitivity: 0.3872 - val_specificity: 0.9962\n",
            "Epoch 7/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.3311 - accuracy: 0.8963 - sensitivity: 0.5574 - specificity: 0.9782 - val_loss: 0.2670 - val_accuracy: 0.9160 - val_sensitivity: 0.5512 - val_specificity: 0.9956\n",
            "Epoch 8/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.3026 - accuracy: 0.9123 - sensitivity: 0.6124 - specificity: 0.9839 - val_loss: 0.2444 - val_accuracy: 0.9335 - val_sensitivity: 0.6356 - val_specificity: 0.9988\n",
            "Epoch 9/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.2765 - accuracy: 0.9228 - sensitivity: 0.6629 - specificity: 0.9877 - val_loss: 0.2156 - val_accuracy: 0.9540 - val_sensitivity: 0.7503 - val_specificity: 0.9988\n",
            "Epoch 10/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.2573 - accuracy: 0.9327 - sensitivity: 0.7142 - specificity: 0.9905 - val_loss: 0.1925 - val_accuracy: 0.9760 - val_sensitivity: 0.8728 - val_specificity: 0.9993\n",
            "Epoch 11/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.2361 - accuracy: 0.9439 - sensitivity: 0.7507 - specificity: 0.9924 - val_loss: 0.1765 - val_accuracy: 0.9860 - val_sensitivity: 0.9235 - val_specificity: 1.0000\n",
            "Epoch 12/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.2205 - accuracy: 0.9534 - sensitivity: 0.8025 - specificity: 0.9947 - val_loss: 0.1656 - val_accuracy: 0.9895 - val_sensitivity: 0.9430 - val_specificity: 1.0000\n",
            "Epoch 13/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.2065 - accuracy: 0.9599 - sensitivity: 0.8230 - specificity: 0.9963 - val_loss: 0.1566 - val_accuracy: 0.9885 - val_sensitivity: 0.9340 - val_specificity: 1.0000\n",
            "Epoch 14/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1972 - accuracy: 0.9655 - sensitivity: 0.8474 - specificity: 0.9955 - val_loss: 0.1416 - val_accuracy: 0.9935 - val_sensitivity: 0.9625 - val_specificity: 1.0000\n",
            "Epoch 15/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1896 - accuracy: 0.9672 - sensitivity: 0.8514 - specificity: 0.9970 - val_loss: 0.1347 - val_accuracy: 0.9935 - val_sensitivity: 0.9625 - val_specificity: 1.0000\n",
            "Epoch 16/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1769 - accuracy: 0.9725 - sensitivity: 0.8836 - specificity: 0.9979 - val_loss: 0.1334 - val_accuracy: 0.9935 - val_sensitivity: 0.9625 - val_specificity: 1.0000\n",
            "Epoch 17/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1723 - accuracy: 0.9765 - sensitivity: 0.9062 - specificity: 0.9966 - val_loss: 0.1244 - val_accuracy: 0.9935 - val_sensitivity: 0.9625 - val_specificity: 1.0000\n",
            "Epoch 18/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1705 - accuracy: 0.9753 - sensitivity: 0.8883 - specificity: 0.9967 - val_loss: 0.1225 - val_accuracy: 0.9935 - val_sensitivity: 0.9625 - val_specificity: 1.0000\n",
            "Epoch 19/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1660 - accuracy: 0.9770 - sensitivity: 0.9029 - specificity: 0.9991 - val_loss: 0.1172 - val_accuracy: 0.9935 - val_sensitivity: 0.9625 - val_specificity: 1.0000\n",
            "Epoch 20/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1580 - accuracy: 0.9819 - sensitivity: 0.9190 - specificity: 0.9990 - val_loss: 0.1148 - val_accuracy: 0.9935 - val_sensitivity: 0.9625 - val_specificity: 1.0000\n",
            "Epoch 21/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1512 - accuracy: 0.9856 - sensitivity: 0.9262 - specificity: 0.9991 - val_loss: 0.1118 - val_accuracy: 0.9935 - val_sensitivity: 0.9625 - val_specificity: 1.0000\n",
            "Epoch 22/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1512 - accuracy: 0.9842 - sensitivity: 0.9327 - specificity: 0.9980 - val_loss: 0.1104 - val_accuracy: 0.9935 - val_sensitivity: 0.9625 - val_specificity: 1.0000\n",
            "Epoch 23/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1471 - accuracy: 0.9859 - sensitivity: 0.9371 - specificity: 0.9988 - val_loss: 0.1081 - val_accuracy: 0.9940 - val_sensitivity: 0.9678 - val_specificity: 1.0000\n",
            "Epoch 24/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1486 - accuracy: 0.9851 - sensitivity: 0.9267 - specificity: 0.9998 - val_loss: 0.1085 - val_accuracy: 0.9950 - val_sensitivity: 0.9718 - val_specificity: 1.0000\n",
            "Epoch 25/25\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1462 - accuracy: 0.9860 - sensitivity: 0.9336 - specificity: 0.9997 - val_loss: 0.1063 - val_accuracy: 0.9950 - val_sensitivity: 0.9718 - val_specificity: 1.0000\n",
            "kernel_idx 100\n",
            "inputs Tensor(\"model/gene_layer/Shape_1:0\", shape=(2,), dtype=int32)\n",
            "kernel Tensor(\"model/gene_layer/Shape_2:0\", shape=(1,), dtype=int32)\n",
            "\n",
            "\n",
            "Confusion matrix\n",
            "[[1650    0]\n",
            " [  10  340]]\n",
            "\n",
            "\n",
            "Area under the Curve (AUC) =  0.9999705627705628\n",
            "Specificity =  1.0\n",
            "Sensitivity =  0.9714285714285714\n",
            "F_1 score = 0.9855072463768115\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2020-04-26T15:18:57.301175Z",
          "start_time": "2020-04-26T15:18:56.938441Z"
        },
        "scrolled": true,
        "id": "qDlABZcV3n3i",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 516
        },
        "outputId": "36e31aa0-0609-4d6c-cdcf-dd353d889262"
      },
      "source": [
        "print(\"evaluate over test patients\")\n",
        "xtest, ytest, _ = get_simdata(num_features=inputsize, num_patients=test_size, random_seed=12+ jobid) # get the test set\n",
        "ptest = model.predict(xtest) # predict with the trained neural network, optimized with the validation set\n",
        "evaluate_performance(ytest, ptest) # get the unbiased predictions"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "evaluate over test patients\n",
            "('Created dataset[', 2000, ' x ', 100, '] with', 344.0, 'diseased')\n",
            "\n",
            "\n",
            "Confusion matrix\n",
            "[[1656    0]\n",
            " [  13  331]]\n",
            "\n",
            "\n",
            "Area under the Curve (AUC) =  0.9998525446579036\n",
            "Specificity =  1.0\n",
            "Sensitivity =  0.9622093023255814\n",
            "F_1 score = 0.9807407407407407\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gkyAgncIj8D8"
      },
      "source": [
        "## Compare with logistic regression "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2020-04-26T15:18:57.608024Z",
          "start_time": "2020-04-26T15:18:57.303251Z"
        },
        "scrolled": true,
        "id": "x4Rw4U493n3k",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 482
        },
        "outputId": "c707ac06-7fb4-41af-fe5c-08c3cdb21ad8"
      },
      "source": [
        "import sklearn\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "\n",
        "clf = LogisticRegression(random_state=0, solver='lbfgs').fit(xtrain, ytrain)\n",
        "ptest_lr = clf.predict(xtest)\n",
        "\n",
        "evaluate_performance(ytest, ptest_lr)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "\n",
            "Confusion matrix\n",
            "[[1582   74]\n",
            " [ 220  124]]\n",
            "\n",
            "\n",
            "Area under the Curve (AUC) =  0.6578895629704528\n",
            "Specificity =  0.9553140096618358\n",
            "Sensitivity =  0.36046511627906974\n",
            "F_1 score = 0.4575645756457565\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1MpsGk6LxDaY"
      },
      "source": [
        "### Inspect the betas of the logistic regression to see which SNPs are important for prediction"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2020-04-26T15:18:57.619345Z",
          "start_time": "2020-04-26T15:18:57.609432Z"
        },
        "scrolled": true,
        "id": "RAoo6Bxp3n3n",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 455
        },
        "outputId": "a4004415-c7bc-45f8-8da1-5e574377a6e3"
      },
      "source": [
        "#@title Show the important weights for logistic regression\n",
        "\n",
        "betas_lr = pd.DataFrame(clf.coef_[:].T, columns=[\"Betas logistic regression\"])\n",
        "betas_lr_s = betas_lr.sort_values(\"Betas logistic regression\",ascending=False)\n",
        "print(\"Causal SNPs are:\", causal_snps)\n",
        "betas_lr_s.head(12)\n",
        "\n",
        "\n"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Causal SNPs are: [[ 0  6 12 21 35]\n",
            " [ 4  9 19 29 45]]\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Betas logistic regression</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>45</th>\n",
              "      <td>1.215506</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>1.110963</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>1.103864</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1.082567</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>0.293083</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>0.284680</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>33</th>\n",
              "      <td>0.099352</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>80</th>\n",
              "      <td>0.096045</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>76</th>\n",
              "      <td>0.094810</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>48</th>\n",
              "      <td>0.093763</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>70</th>\n",
              "      <td>0.086982</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>0.080717</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    Betas logistic regression\n",
              "45                   1.215506\n",
              "21                   1.110963\n",
              "6                    1.103864\n",
              "4                    1.082567\n",
              "12                   0.293083\n",
              "19                   0.284680\n",
              "33                   0.099352\n",
              "80                   0.096045\n",
              "76                   0.094810\n",
              "48                   0.093763\n",
              "70                   0.086982\n",
              "25                   0.080717"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cDdddKoUq1kl"
      },
      "source": [
        "On this toy example the network performs thus substantially better than logistic regression. We will inspect interpretability next by plotting the weights\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aJYhQeshqOTc"
      },
      "source": [
        "## Plot network"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "81ZwaoC8qXWR"
      },
      "source": [
        "We plot the network below. The line-thickness between the nodes represent the strenght of the connections (magnitude of the weight)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2020-04-26T15:22:35.773662Z",
          "start_time": "2020-04-26T15:22:33.432553Z"
        },
        "id": "rgrAmkDL3n33",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "cellView": "form",
        "outputId": "3209570b-ea7c-4e5d-875d-a88c17c9f20d"
      },
      "source": [
        "#@title Plot network \n",
        "from matplotlib import cm\n",
        "import matplotlib.patches as mpatches\n",
        "\n",
        "weight_locally = model.layers[2].get_weights()[0]\n",
        "weights_dense = model.layers[5].get_weights()[0]\n",
        "\n",
        "causal_snps = np.array(causal_snps)\n",
        "colors = plt.cm.tab10(np.linspace(0,1,10))\n",
        "fig = plt.figure(figsize=(18, 18))\n",
        "\n",
        "n_layers = 3\n",
        "gene_mask = genemask.todense()\n",
        "layer_sizes = [genemask.shape[0], genemask.shape[1], 1]\n",
        "\n",
        "\n",
        "ax = fig.gca()\n",
        "top=0.9\n",
        "bottom=0.1\n",
        "left = 0.1\n",
        "right = 0.9\n",
        "                 \n",
        "v_spacing = (top - bottom)/float(max(layer_sizes))\n",
        "h_spacing = (right - left)/float(len(layer_sizes) - 1)\n",
        "\n",
        "\n",
        "causal_red = \"firebrick\" #\"#7dcfe2\"\n",
        "causal_circle_red = \"tomato\" #\"#4b78b5\"\n",
        "gray1 = \"darkgrey\"\n",
        "gray2 = \"dimgray\"\n",
        "\n",
        "linewith_genes = (weights_dense * 5 + 0.5)\n",
        "linewith_snps = (weight_locally* 5+0.5)\n",
        "colormap_snps = np.zeros(inputsize, dtype=object)\n",
        "colormap_snps[:] =gray2\n",
        "colormap_snps[causal_snps] = causal_red\n",
        "colormap_genes = np.zeros((gene_mask.shape[1]), dtype=object)\n",
        "colormap_genes[:int(len(colormap_genes)/2)] = causal_red\n",
        "colormap_genes[int(len(colormap_genes)/2):] = gray2\n",
        "colormap_circle_genes = np.zeros((gene_mask.shape[1]), dtype=object)\n",
        "colormap_circle_genes[:int(len(colormap_circle_genes)/2)] = causal_circle_red\n",
        "colormap_circle_genes[int(len(colormap_circle_genes)/2):] = gray1\n",
        "\n",
        "\n",
        "\n",
        "#SNPs\n",
        "n=0    \n",
        "layer_size = layer_sizes[0]\n",
        "layer_top = v_spacing*(layer_size - 1)/2. + (top + bottom)/2\n",
        "\n",
        "y_snps=[]\n",
        "layer_size_a = gene_mask.shape[0]\n",
        "layer_size_b = gene_mask.shape[1]\n",
        "layer_top_a = v_spacing*(layer_size_a - 1)/2. + (top + bottom)/2.\n",
        "layer_top_b = v_spacing*(layer_size_b - 1)/2. + (top + bottom)/2.\n",
        "for m in range(layer_size):\n",
        "    y_snps.append(layer_top_a - m*v_spacing) \n",
        "        \n",
        "line_vert_1 =plt.Line2D(xdata=[n*h_spacing + left, n*h_spacing + left],\n",
        "                                  ydata=[top, ((top-bottom)/2)+bottom ],\n",
        "                                  c=causal_circle_red, linewidth = \"18\")\n",
        "\n",
        "\n",
        "line_vert_2 =plt.Line2D(xdata=[n*h_spacing + left, n*h_spacing + left],\n",
        "                                  ydata=[((top-bottom)/2)+bottom, bottom],\n",
        "                                  c=gray1 , linewidth = \"18\")\n",
        "\n",
        "\n",
        "\n",
        "## Gene layer\n",
        "layer_sizes2 = [genemask.shape[1], 1]\n",
        "v_spacing = (top - bottom)/float(max(layer_sizes2))\n",
        "\n",
        "\n",
        "n=1\n",
        "layer_size = layer_sizes[1]\n",
        "layer_top = v_spacing*(layer_size - 1)/2. + (top + bottom)/2\n",
        "\n",
        "layer2xy = []\n",
        "for m in range(layer_size):\n",
        "    layer2xy.append( (n*h_spacing + left, layer_top - m*v_spacing))\n",
        "    circle = plt.Circle(xy=layer2xy[m], radius=v_spacing/5,\n",
        "                        color=colormap_circle_genes[m], ec='k', zorder=4)\n",
        "    ax.add_artist(circle)\n",
        "\n",
        "    \n",
        "## Dense layer\n",
        "\n",
        "n=2\n",
        "layer_size = layer_sizes[2]\n",
        "layer_top = v_spacing*(layer_size - 1)/2. + (top + bottom)/2\n",
        "\n",
        "\n",
        "for m in range(layer_size):\n",
        "    circle = plt.Circle(xy=(n*h_spacing + left, layer_top - m*v_spacing), radius=v_spacing/5,\n",
        "                         color = causal_circle_red, ec='k', zorder=4)\n",
        "    ax.add_artist(circle)   \n",
        "\n",
        "\n",
        "\n",
        "# draw lines    \n",
        "teller_snp_lines = 0\n",
        "for column in range(gene_mask.shape[1]):\n",
        "    rows = np.where(gene_mask[:,column]!=0)\n",
        "    for row in rows[0]:\n",
        "            line = plt.Line2D(xdata=[left,        layer2xy[column][0]],\n",
        "                              ydata=[y_snps[row], layer2xy[column][1]],\n",
        "                              c = colormap_snps[teller_snp_lines], linewidth = linewith_snps[teller_snp_lines])\n",
        "            ax.add_artist(line)\n",
        "            teller_snp_lines +=1\n",
        "            \n",
        "for column in range(gene_mask.shape[1]):\n",
        "        line = plt.Line2D(xdata=[layer2xy[column][0],right],\n",
        "                          ydata=[layer2xy[column][1], layer_top - m*v_spacing],\n",
        "                          c = colormap_genes[column], linewidth = linewith_genes[column])\n",
        "        ax.add_artist(line)\n",
        "\n",
        "\n",
        "\n",
        "plt.annotate(\"Control region\" ,xy=(left/2,0.24), size = 24, rotation = \"90\")\n",
        "plt.annotate(\"Causal region\", xy=(left/2,0.64), size=24, rotation = \"90\")\n",
        "plt.annotate(\"Input SNPs\", xy=(left-0.05,0.95), size=24, rotation = \"0\")\n",
        "plt.annotate(\"Gene layer\", xy=(layer2xy[0][0]-0.05,0.92), size=24)\n",
        "plt.annotate(\"Output\", xy=(right-0.05,0.57), size=24)\n",
        "        \n",
        "ax.add_artist(line_vert_1)\n",
        "ax.add_artist(line_vert_2)\n",
        "ax.set_yticklabels([])\n",
        "ax.set_xticklabels([])\n",
        "\n",
        "plt.rcParams['axes.spines.right'] = False\n",
        "plt.rcParams['axes.spines.top'] = False\n",
        "plt.rcParams['axes.spines.bottom'] = False\n",
        "plt.rcParams['axes.spines.left'] = False\n",
        "\n",
        "plt.savefig(\"simulation_non_linear4.pdf\")\n",
        "plt.show()\n"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1296x1296 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8F1M2jwswxmL"
      },
      "source": [
        "The network assigns high weights to causal SNP pairs ([0, 4], [6, 9],[12, 19], [21, 29],[35, 45]) The network is thus fully interpretable because each node and connection represents a contribution of the SNP to the gene or the contribution of the gene on the output. The weights gives an indication on how strong this contribition is.\n",
        "\n",
        "The network is interpretable because it is not fully connected, instead prior knowledge is used to create an interpretable neural network.\n"
      ]
    }
  ]
}