DTI_nr.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "DTI_nr.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "TlLJCYZ_Pycd",
        "lxmM2NsoPk21",
        "lMMMeflsPKkz",
        "BcTKFGbuPVQJ",
        "ZTqYrNPzFbs-"
      ]
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "F7m-HY9V3vfk"
      },
      "source": [
        "from google.colab import drive\n",
        "\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-rnaW0aq4avv"
      },
      "source": [
        "%cd /content/drive/My\\ Drive/DTI/Input"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "us_v2tZniW6D"
      },
      "source": [
        "!pip install git+https://github.com/palash1992/GEM.git\n",
        "!pip install tensorflow==1.13.1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_RF7F_ts5Bl9"
      },
      "source": [
        "from data_loading import *\n",
        "from utils import *\n",
        "import networkx as nx\n",
        "import matplotlib.pyplot as plt\n",
        "from copy import deepcopy\n",
        "from gem.embedding.sdne import SDNE\n",
        "from gem.embedding.hope import HOPE\n",
        "from gem.embedding.lap import LaplacianEigenmaps\n",
        "from gem.embedding.gf import GraphFactorization\n",
        "from gem.embedding.lle import LocallyLinearEmbedding\n",
        "from gem.evaluation import visualize_embedding as viz\n",
        "from gem.evaluation import evaluate_graph_reconstruction as gr\n",
        "from gem.utils import graph_util, plot_util\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import csv\n",
        "from sklearn.model_selection import StratifiedKFold\n",
        "from sklearn.preprocessing import MaxAbsScaler\n",
        "from sklearn.ensemble import AdaBoostClassifier, RandomForestClassifier\n",
        "from sklearn.neural_network import MLPClassifier\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score, precision_recall_curve, auc, confusion_matrix, classification_report\n",
        "from xgboost import XGBClassifier\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense\n",
        "from time import time\n",
        "%matplotlib inline"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CbR5EDVQ5NKm",
        "outputId": "17f1c65b-8836-4c7e-d6e6-6c7ae52f1705"
      },
      "source": [
        "all_drugs, all_targets, drug_similarity_matrices, target_similarity_matrices, DTIs, DT_labeled, X, Y = load_datasets('nr')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of drugs: 54\n",
            "Number of targets: 26\n",
            "Loading only selected similarity matrices\n",
            "Dimensions of all pairs: (1404, 2)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eBIePo3C8lsL"
      },
      "source": [
        "all_negative_pairs = []\n",
        "all_positive_pairs = []\n",
        "for target_id in range(len(all_targets)):\n",
        "  for drug_id in range(len(all_drugs)):\n",
        "    if DTIs[target_id][drug_id] == 1:\n",
        "      all_positive_pairs.append((drug_id, target_id))\n",
        "    else:\n",
        "      all_negative_pairs.append((drug_id, target_id))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XbEqHUORwea1"
      },
      "source": [
        "models = []\n",
        "filenames = ['hope', 'lap', 'lle']#, 'sdne']\n",
        "# HOPE takes embedding dimension (d) and decay factor (beta) as inputs\n",
        "models.append(HOPE(d=12, beta=0.01))\n",
        "# LE takes embedding dimension (d) as input\n",
        "models.append(LaplacianEigenmaps(d=12))\n",
        "# LLE takes embedding dimension (d) as input\n",
        "models.append(LocallyLinearEmbedding(d=12))\n",
        "# SDNE takes embedding dimension (d), seen edge reconstruction weight (beta), first order proximity weight (alpha), lasso regularization coefficient (nu1), ridge regreesion coefficient (nu2), number of hidden layers (K), size of each layer (n_units), number of iterations (n_ite), learning rate (xeta), size of batch (n_batch), location of modelfile and weightfile save (modelfile and weightfile) as inputs\n",
        "# models.append(SDNE(d=12, beta=5, alpha=1e-5, nu1=1e-6, nu2=1e-6, K=3, n_units=[16, 8,], n_iter=10, xeta=0.01, n_batch=50,\n",
        "#                 modelfile=['enc_model.json', 'dec_model.json'], weightfile=['enc_weights.hdf5', 'dec_weights.hdf5']))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S3Bc1JsxXFSd"
      },
      "source": [
        "ada_boost = AdaBoostClassifier(DecisionTreeClassifier(criterion='gini', splitter='best', max_depth=6, min_samples_split=2,\n",
        "                                                      min_samples_leaf=1, max_features=None, random_state=1,\n",
        "                                                      max_leaf_nodes=None, min_impurity_split=None), algorithm=\"SAMME\",\n",
        "                                                      n_estimators=100, random_state=0)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rlGCZlwUee0R"
      },
      "source": [
        "xgb = XGBClassifier(max_depth=6, min_child_weight=1, n_estimators=100, gamma=0.1, random_state=0)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9iFBDB_vmLyG"
      },
      "source": [
        "ann = MLPClassifier(activation='relu', solver='adam', alpha=1e-5, hidden_layer_sizes=(8, 4), learning_rate='adaptive', random_state=0)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "73Gut1Hc8cLP"
      },
      "source": [
        "fused_DD_similarities_SNF = SNF(drug_similarity_matrices, K=5, t=3, alpha=1.0)\n",
        "fused_TT_similarities_SNF = SNF(target_similarity_matrices, K=5, t=3, alpha=1.0)\n",
        "\n",
        "fused_DD_similarities_avg = avg_matrix(drug_similarity_matrices)\n",
        "fused_TT_similarities_avg = avg_matrix(target_similarity_matrices)\n",
        "\n",
        "fused_DD_similarities_max = max_matrix(drug_similarity_matrices)\n",
        "fused_TT_similarities_max = max_matrix(target_similarity_matrices)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HEwD9IO09TIi"
      },
      "source": [
        "chosen_negative_pairs_SNF = choose_negative_pairs(all_negative_pairs, DTIs, fused_DD_similarities_SNF, fused_TT_similarities_SNF, len(all_positive_pairs))\n",
        "\n",
        "chosen_negative_pairs_avg = choose_negative_pairs(all_negative_pairs, DTIs, fused_DD_similarities_avg, fused_TT_similarities_avg, len(all_positive_pairs))\n",
        "\n",
        "chosen_negative_pairs_max = choose_negative_pairs(all_negative_pairs, DTIs, fused_DD_similarities_max, fused_TT_similarities_max, len(all_positive_pairs))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sAqDRICCBFpV",
        "outputId": "a682b5f6-1cb4-4d66-f955-36ffb7d88bb8"
      },
      "source": [
        "print(f'Number of positive pairs: {len(all_positive_pairs)}')\n",
        "\n",
        "print(f'[SNF] Number of negative pairs: {len(chosen_negative_pairs_SNF)}')\n",
        "print(f'[AVG] Number of negative pairs: {len(chosen_negative_pairs_avg)}')\n",
        "print(f'[MAX] Number of negative pairs: {len(chosen_negative_pairs_max)}')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of positive pairs: 90\n",
            "[SNF] Number of negative pairs: 90\n",
            "[AVG] Number of negative pairs: 90\n",
            "[MAX] Number of negative pairs: 90\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_Ib5RzTXpXX_"
      },
      "source": [
        "Since both the drugs and the targets have ids starting from 0, we need to add the length of all_drugs to the target ids so they will have a unique identifier. This is important when creating a graph with NetworkX, otherwise we'll have a drug and a target represented with the same node."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iq_QbvUaqRSf"
      },
      "source": [
        "all_positive_pairs_edges = [(drug_id, target_id + len(all_drugs)) for (drug_id, target_id) in all_positive_pairs]\n",
        "\n",
        "chosen_negative_pairs_edges_SNF = [(drug_id, target_id + len(all_drugs)) for (drug_id, target_id) in chosen_negative_pairs_SNF]\n",
        "chosen_negative_pairs_edges_avg = [(drug_id, target_id + len(all_drugs)) for (drug_id, target_id) in chosen_negative_pairs_avg]\n",
        "chosen_negative_pairs_edges_max = [(drug_id, target_id + len(all_drugs)) for (drug_id, target_id) in chosen_negative_pairs_max]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TlLJCYZ_Pycd"
      },
      "source": [
        "---\n",
        "### Graph initialization and primary visualizations"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hqHF9NpHmpAn",
        "outputId": "7c3c8e2b-2ed3-4305-be74-f9e0ce366373"
      },
      "source": [
        "# all graphs are the same as of right now, that's why we are showing only graph_SNF info and visualizations\n",
        "graph_SNF = nx.Graph()\n",
        "graph_avg = nx.Graph()\n",
        "graph_max = nx.Graph()\n",
        "for u, v in all_positive_pairs_edges:\n",
        "    graph_SNF.add_edge(u, v, weight=1)\n",
        "    graph_avg.add_edge(u, v, weight=1)\n",
        "    graph_max.add_edge(u, v, weight=1)\n",
        "\n",
        "color_map = []\n",
        "for node in graph_SNF:\n",
        "  # drug\n",
        "  if node < len(all_drugs):\n",
        "    color_map.append(\"blue\")\n",
        "  # target\n",
        "  else:\n",
        "    color_map.append(\"red\")\n",
        "\n",
        "node_sizes = [30] * graph_SNF.number_of_nodes()\n",
        "\n",
        "calculate_statistics(graph_SNF)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Nodes: 80\n",
            "Edges: 90\n",
            "Number of connected components: 10\n",
            "Size of each component (number of nodes): [42, 13, 6, 4, 3, 3, 3, 2, 2, 2]\n",
            "Nodes in largest connected component: 42\n",
            "Edges in largest connected component: 50\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": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tQy80a7nnT1O",
        "outputId": "63cc5561-d404-457d-c2f8-51a6360533b6"
      },
      "source": [
        "nx.draw(graph_SNF, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fAXw0_SXOQcG",
        "outputId": "12e0afb1-00ff-48f9-a9a0-86b661e9e19a"
      },
      "source": [
        "nx.draw_spring(graph_SNF, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NRaaWimyORON",
        "outputId": "fd89af25-fa1d-45a5-cca9-3171ee7b68bc"
      },
      "source": [
        "nx.draw_kamada_kawai(graph_SNF, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lxmM2NsoPk21"
      },
      "source": [
        "---\n",
        "### Add strongest similarities edges to graph"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nouqRJYCcbVn"
      },
      "source": [
        "strongest_drug_similarities_edges_SNF = get_strongest_all_p_similarities(fused_DD_similarities_SNF, 5)\n",
        "strongest_target_similarities_edges_SNF = get_strongest_all_p_similarities(fused_TT_similarities_SNF, 5)\n",
        "\n",
        "strongest_drug_similarities_edges_avg = get_strongest_all_p_similarities(fused_DD_similarities_avg, 5)\n",
        "strongest_target_similarities_edges_avg = get_strongest_all_p_similarities(fused_TT_similarities_avg, 5)\n",
        "\n",
        "strongest_drug_similarities_edges_max = get_strongest_all_p_similarities(fused_DD_similarities_max, 5)\n",
        "strongest_target_similarities_edges_max = get_strongest_all_p_similarities(fused_TT_similarities_max, 5)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "isgeT6nBlnFS"
      },
      "source": [
        "for drug_id1, drug_id2 in strongest_drug_similarities_edges_SNF:\n",
        "  graph_SNF.add_edge(drug_id1, drug_id2, weight=fused_DD_similarities_SNF[drug_id1][drug_id2])\n",
        "for target_id1, target_id2 in strongest_target_similarities_edges_SNF:\n",
        "  graph_SNF.add_edge(target_id1 + len(all_drugs), target_id2 + len(all_drugs), weight=fused_TT_similarities_SNF[target_id1][target_id2])\n",
        "\n",
        "for drug_id1, drug_id2 in strongest_drug_similarities_edges_avg:\n",
        "  graph_avg.add_edge(drug_id1, drug_id2, weight=fused_DD_similarities_avg[drug_id1][drug_id2])\n",
        "for target_id1, target_id2 in strongest_target_similarities_edges_avg:\n",
        "  graph_avg.add_edge(target_id1 + len(all_drugs), target_id2 + len(all_drugs), weight=fused_TT_similarities_avg[target_id1][target_id2])\n",
        "\n",
        "for drug_id1, drug_id2 in strongest_drug_similarities_edges_max:\n",
        "  graph_max.add_edge(drug_id1, drug_id2, weight=fused_DD_similarities_max[drug_id1][drug_id2])\n",
        "for target_id1, target_id2 in strongest_target_similarities_edges_max:\n",
        "  graph_max.add_edge(target_id1 + len(all_drugs), target_id2 + len(all_drugs), weight=fused_TT_similarities_max[target_id1][target_id2])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lMMMeflsPKkz"
      },
      "source": [
        "---\n",
        "### Graph visualizations (after adding strongest similarities edges)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TGw0j7scoQmu",
        "outputId": "dc98e931-c945-47ad-cd85-4462f0d9b06d"
      },
      "source": [
        "print('SNF graph')\n",
        "calculate_statistics(graph_SNF)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "SNF graph\n",
            "Nodes: 80\n",
            "Edges: 177\n",
            "Number of connected components: 1\n",
            "Size of each component (number of nodes): [80]\n",
            "Nodes in largest connected component: 80\n",
            "Edges in largest connected component: 177\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": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "X1sCJkxROgv7",
        "outputId": "0ba15701-1869-4e34-c071-fbbfa0853dd6"
      },
      "source": [
        "print('SNF graph')\n",
        "nx.draw(graph_SNF, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "SNF graph\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2H-9tYYfOj-b",
        "outputId": "eb395970-748e-4caf-885c-383181f5c70f"
      },
      "source": [
        "print('SNF graph')\n",
        "nx.draw_spring(graph_SNF, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "SNF graph\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VdX4_8ZvOlu9",
        "outputId": "f9ec6729-5b92-44fe-da07-75190644a534"
      },
      "source": [
        "print('SNF graph')\n",
        "nx.draw_kamada_kawai(graph_SNF, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "SNF graph\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAb4AAAEuCAYAAADx63eqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd1iT5/oH8DubhBVICHsPAREREcWFOHCg1olVUXHWqjiqLW7FVXfdIs566jxqcVWLnlr33qPWUTcgDqqCsvJ+f3/wIzUFBGSE8Xyui+sc37x53zsp5M6z7ocHAMQwDMMwVQRf1wEwDMMwTFliiY9hGIapUljiYxiGYaoUlvgYhmGYKoUlPoZhGKZKYYmPYRiGqVJY4mMYhmGqFJb4GIZhmCqFJT6GYRimSmGJj2EYhqlSWOJjGIZhqhSW+BiGYZgqhSU+hmEYpkphiY9hGIapUljiYxiGYaoUlvgYpoq4cYNo+3aiP//UdSQMo1tCXQfAMEzpyswk6tKF6PBhIoGAKCuLqFMnoo0bifjsqy9TBbFfe4ap5KKjs5Pe+/dE794RffhAFBtLtHmzriNjGN1giY9hKrmNG7OT3sdSU7OPM0xVxBIfw1RyhoZ5HzcwKNs4GKa8YImPYSq5ESOI9PW1j8lkRBERuomHYXSNJT6GqeS++IIoKopIJEojkSiTjIyIFi4kCgrSdWQMoxs8ANB1EAzDlL66dRvT6NFzqWPHeiQS6ToahtEdlvgYpgrIzMwkuVxOCQkJZGRkpOtwGEanWFcnw1QBt27dIjs7O5b0GIZY4iuc48eJatTIXv1rbc3mgTMVzoULF8jPz0/XYTBMucAqtxTk3j2iVq3+WQgVH0/09ddEcjlR+/a6jY1hCoklPob5B2vxFWTlSqKMDO1j798TzZihm3gY5jNcuHCB6tSpo+swGKZcYImvIM+eZRc3/LekpLKPhWE+Q3p6Ot28eZN8fHx0HQrDlAss8RXkiy9yrf6FSMS6OZkK48aNG+Ti4kIymUzXoTBMucASX0G6diVq3Dg7+QmFlC4WU4JIRJgyRdeRMUyhsPE9htHGEl9BhEKi/fuJ9u0jmjGDBBs3Usdq1WhdbKyuI2OYQmGJj2G0sQXsn+H69evUtGlTunjxItnZ2ek6HIb5pFq1atGqVavI399f16EwTLnAEt9nmjlzJh09epR+/fVX4vF4ug6HYfL04cMHUigU9Pr1a9LT09N1OAxTLrCuzs8UGRlJr1+/pjVr1ug6FIbJ17Vr18jd3Z0lPYb5CEt8n0koFNKPP/5I48ePp0ePHuk6HIbJExvfY5jcWOIrhurVq9Po0aOpf//+xHqMmfKIJT6GyY0lvmIaM2YMvXv3jlatWqXrUBgml/Pnz7PExzD/wia3lIBbt25RYGAgnTt3jhwdHXUdDsMQEVFKSgqpVCr6+++/SSwW6zochik3WIuvBHh6etK3335L/fv3J47jdB0OwxAR0ZUrV8jLy4slPYb5F5b4Ssjo0aPpw4cPFB0dretQGIaI2Pgew+SHJb4SIhAIaP369TR58mT666+/dB0Ow7AdGRgmHyzxlSB3d3caN24c9evXj3V5MjrHWnwMkzeW+ErYyJEjKTMzk5YvX67rUJgq7O3bt/T06VPy8PDQdSgMU+6wxFfCcro8o6Ki6N69e7oOh6miLl26RDVr1iShUKjrUBim3GGJrxS4ubnRhAkTWJcnozOsm5Nh8scSXykZPnw4AaClS5fqOhSmCmKJj2Hyxxawl6J79+5RvXr16PTp0+Tq6qrrcJgqxMXFhfbu3cvG+BgmDyzxlbIlS5bQ9u3b6ejRoyQQCHQdDlMFJCcnk729PSUnJ7PfOYbJA+vqLGXDhg0jgUBAixcv1nUoTBVx8eJFqlWrFkt6DJMPlvhKGZ/Pp3Xr1tGsWbPozz//1HU4TBXAxvcY5tNY4isDzs7ONHXqVOrbty+p1Wpdh8NUcizxMcynscRXRoYMGUISiYR++OEHXYfCVHJsKyKG+TQ2uaUMPXjwgOrUqUPHjx9ns+2YUpGUlERubm70+vVr4vPZ91qGyYtu/zKePCGKiCDy9ycaOpTo8WOdhlPaHB0dafr06RQeHk5ZWVm6DqdSePeOaPVqokmTiOLiiKp6vYCLFy+Sn58fS3oM8wm6a/E9eUJUs2b2J1dWFpFIRKSvT3T1KpGdnU5CKgscx1GLFi0oODiYIiMjdR1OhfboUfZ3ptTU7B8DA6LAQKI9e4gq6+d+aipRSgqRSkXE4+V+fPr06ZSamkqzZ88u++AYpoLQ3cfDvHnZf8E5LZ/MzOy/6jlzdBZSWeDz+bR27VqaP38+3bp1S9fhVGijRxO9fJn9a0OU/et09CjR/v26jas0pKcT9e1LpFAQ2dsTubgQnT2b+zw2sYVhCqa7xHf+fHay+1hmZvbxSs7BwYFmzJhBffr0YV2exfDbb7m7NlNSiA4e1E08pQUg6tePaOvW7ASYnk70119ELVoQ/f239rks8TFMwco28anVRNu2EXXtSvT+PXH/WmCbQUR7ExMpLi6OKvucm0GDBpFcLqd58+bpOpQKy9w89zE9vewWUWVx6VJ2z//mzURpadqPcRzRzp3//Ds+Pp7S09PJvjK9AQxTCsou8QFEXboQ9e9PtGMH4eZN4qnVpM4ZjBGJSCSXU9qIETRmzBiqUaMGrV27ltL+/ddeSfB4PFq7di0tXLiQbty4oetwKqQpU4hkso+PcCQWE4WH6yigzwQQHTlCNG0a0fr12a1WouyWXfPmRE+f5v28rKx/ziX6Z2ILL6/BP4ZhNMou8Z07R3TokGZAhqdWE4+IBAYG2TMUBg0i3tWr1HX0aLp69SotWrSIdu3aRfb29jR16lR6/vx5mYVaVuzs7GjWrFkUHh5Omf/u9mUK9OWXRNHRRI6ORBIJR0RH6ccf75FKpevICo/jsr8PtmtHNHVq9iRnOzuigAAiCwuit2/zfy6PB2rX7p9/s25Ohimcsk18eVQtQUoK0bFjRMuWaWZz8ng8at68Oe3fv59+//13SkhIIHd3d+rfv3+lax0NGDCAFAoFzankk3pKS69e2eNdaWl8srTsSTExw3UdUpH8+mv2T2pqdssvNZUoOZnozJns8Tu1Oq8ufxCfn0m1am0hJ6d/jrLExzCFU3aJz9WVKI/doF8DFDlpEt2+fTvPp3l4eNCqVavo7t275OjoSC1atKCWLVvSwYMHK8U4II/HozVr1tDixYvp2rVrug6nQouIiKDDhw9XqN6BuLh/ZqXm7d/dlmpydCS6cyedXr6cSlu2bCEiIgAs8TFMIZVd4mvRIqdP6p9jMhllTJ5MxONRUFAQ1a9fn9asWUNv8+jfUSqVNHHiRHr48CH16NGDIiMjycvLi1avXk0fPnwos5dRGmxtbWnOnDmsy7OYBg0aRDwejxYsWKDrUArNxiZ7Qk5BeDwiAwOQWJxAvXptIGdnA9q+fTsNHz6c7t69S0+fPiUej0fW1talHzTDVHQoS3//DXz7LeDsDPj7Azt3ah7KyMjAnj170LFjRxgbG6N37944cuQI1Gp1npfiOA7/+9//EBISApVKhcmTJyMxMbGsXkmJ4zgOrVu3RlRUlK5DqdBatGgBAwMDpKam6jqUQklKAoyMgOyOzvx/BAJg3z7g3r37UKlUOHHiBABg2bJlqFWrFrZt24aQkBAdvxqGqRjKNvEV0vPnz7Fw4UJUr14dTk5OmD59Oh4/fpzv+X/88QcGDx4MuVyOvn374tq1a2UYbcl58uQJlEolLl++rOtQKqzY2FiYmppi+fLlug6l0K5dAxo1Ani8DBgbf8gz8Rkacprvifv27YONjQ0SExPBcRw6deoEPz8/TJkyRaevg2EqinKZ+HJwHIdz587h66+/hqmpKYKDg7FlyxZ8+PAhz/NfvHiBGTNmwNLSEs2bN8cvv/ySb4uxvFq/fj1q1qyJ9PR0XYdSIaWnp0Mul8POzg5ZWVm6DqdIOnXqhI0bd0EiyZ34RKI0nD//z7mTJk1CkyZNkJmZieTkZOjp6WH8+PG6C55hKpByXdGQx+NRnTp1aMWKFfT06VMKDw+ntWvXko2NDQ0bNowuXryoNcFFqVTShAkT6MGDB9SrVy8aP348eXl5UUxMTIUZB+zTpw/Z2NjQzJkzdR1KhSQWi6lXr16kVqtp9+7dug6nSGQyGQHvqH//LOLx/vl9FYtBPN5NevZsj+bYlClTSCwW04QJE8jY2JjEYjHFxMTQo0ePdBE6w1Qo5TrxfUwqlVL37t3p0KFDdPHiRVKpVNSlSxfy8fGhRYsW0YsXLzTnSiQS6t27N126dImWL19Oe/fuJXt7e5o0aRIlJibq8FUUjMfjUUxMDK1cuZIuXbqk63AqpPDwcMrIyKhwVXH09fUpNTWViL6hGjV2kp0dSKUi+uorHu3fn0kDBw6g+/fvExGRQCCgTZs20datWyk6OpqMjIzou+++o+7du7MJUgxTEF03OYtDrVbjt99+Q1hYGIyNjdG5c2fs27cPmZmZuc69ffs2vv76a8jlcvTp0wdXrlzRQcSFt3HjRtSoUYN1eX4GjuPg5eUFCwsLnDx5UtfhFNqoUaPQt29f2NvbIzk5OdfjS5YsQc2aNfH+/XvNsXPnzsHIyAjNmzeHWq1G69atERkZWZZhM0yFU6ET38f+/vtvrFq1CnXr1oWVlRUiIyNx+/btXOe9fPkSs2bNgpWVFZo1a4b9+/eXy3FAjuPQrl07TJw4UdehVEjz5s1DvXr10LFjR12HUmgjRoyAgYGBZsbmv3Ech+7duyM8PBwcx2mOt2jRAubm5khJSUFSUhJsbGxw4MCBsgqbYSqcSpP4Pnbz5k2MGTMG5ubmaNCgAdasWYO3b99qnZOeno7//Oc/qFWrFqpVq4bo6OhyNwU+Pj4eKpUK5z+e1cAUSnx8PORyORQKBe7cuaPrcAqkVqvh7OyMhg0bfvK8d+/ewdPTEzExMZpjTZo0QfPmzdGzZ09wHIfff/8dFhYWePbsWWmHzTAVUqVMfDkyMjKwe/dufPHFF5ouzqNHj2p9W+Y4DkeOHEH79u1hZmaGCRMmID4+XodRa/vpp59QvXp1pKWl6TqUCqd169Zo3749vv76a12HUqDZs2fDyckJQ4cOLfDcP/74A0qlEufPn4darYaRkREePXoEb29vLFu2DAAQFRWFoKCgCjezlWHKQqVOfB9LTEzE/Pnz4enpCRcXF8yYMQNPnjzROufPP//EkCFDIJfL0bt373Kxno7jOHTo0AHjxo3TdSgVztatW9G4cWPI5XIkJSXpOpx8nT17FiqVCrNnz0b//v0L9Zz//ve/cHBwwJkzZ+Dg4AAAuHfvHszMzHD69GlkZWUhKCiIFURgmDxUmcSXg+M4nD17Fl999RVMTEzQsmVLbNu2TatF9erVK3z//fewsrJC06ZNsXfvXp2OAyYkJEClUuHcuXM6i6Ei+vDhA0xNTdGtW7dymwDevHkDJycn7NixA5s2bcKXX35Z6Od+88038Pb2RufOnTXH9uzZA1tbWzx//hzPnj2DhYUFjhw5UgqRM0zFVeUS38dSU1Px008/oVmzZlAoFBg2bBguXbqkeTw9PR0//fQTfH194ebmhhUrVuhsHHDLli3w8PDId/E+k7dBgwZh5MiRUKlUWrMhy4uwsDAMHDgQAPDzzz+jffv2hX5uRkYGrK2t0bx5c63j48ePR9OmTZGVlYWDBw/C2tq6XLd4GaasVZh1fKVBJpNRz5496fDhw3ThwgVSKpXUsWNH8vHxoSVLltC7d++oZ8+edOHCBVq9ejX9+uuvZG9vTxMmTKD4+PgyjbVbt27k4eFBU6dOLdP7VnR9+vShAwcOUJ06deg///mPrsPR8tNPP9HFixdp0aJFRJS9ju/9+/eFfr5IJCIbGxu6fPkyHTx4UHN82rRpxOPxaNKkSdSyZUsKCwujPn36EMdxJf4aGKZC0nXmLW/UajUOHz6Mnj17wtjYGF26dMEvv/yimSRw9+5dDBs2DHK5HL169dJqIZa258+fw9zcHKdPny6ze1Z0HMfB1dUVK1asgJubW7lZunLv3j0olUqt9aQnTpxAQEBAoa+RlZUFAwMD7Nu3DyqVCg8fPtQ8lpSUBFtbW8TGxiIjIwMBAQGYO3duib4GhqmoWOL7hOTkZERHR8Pf3x9WVlYYN26cZmr869evMXv2bFhbW6NJkybYs2dPmXyobtu2De7u7uWy2668mj59OgYPHgw/Pz/s3r1b1+EgIyMD/v7+WLx4sdbxS5cuoWbNmoW+zs2bN+Hi4gIAmD9/Pvz8/LTGqs+cOQMzMzPcvXsXDx8+hEqlYl+amPLnxAnA1xeQSAA3N2DPnlK/JUt8hXTjxg2MHj0aKpUKDRs2xLp16/Du3TtkZGRg06ZNqF27NlxdXbF8+XKkpKSUaixdu3bFmDFjSvUelcnDhw+hUCiwceNGNGrUSNfhYOzYsQgJCdFaVgNkzyp2dXUt9HV+/PFHzWSYnF0avvrqK61zli9fDm9vb6SmpuLnn3+Gvb09Xr9+XfwXwTAl4fZtQCbTrsgukwG//Vaqt2WJr4gyMjIQGxuL9u3ba7ZBOnbsGNRqNY4dO4YOHTpAqVRi7NixePr0aanEkJSUVOHKcela06ZNsWXLFtjb2+PMmTM6i+Pw4cOwsrLKc7LJkydPYGVlVehrRUREYMGCBZp/v3nzBm5ubtiwYYPmGMdxCAsLQ69evcBxHIYPH46OHTvmSroMoxNff5292eS/tyP514StksYSXzEkJCRg3rx58PDwgKurK2bNmoWnT5/i3r17iIiIgImJCXr27ImLFy+W+L137NgBNzc31uVZSD/++CNCQkKwaNEidO3aVScxJCUlwdraGnFxcXk+/urVK8jl8kJfLyAgAEePHtU6dv369VxjhykpKfDy8sLKlSuRlpYGX19fzUJ3htGFlJQUDBo0CAfzSnpEgLt7qd6fJb4SwHEczpw5g0GDBsHExAStW7fG9u3bkZCQgLlz58LGxgaBgYHYvXt3iY4DduvWDd98802JXa8ye/fuHeRyOe7duweFQoH79++X6f1zaq9+++23+Z6TlpYGkUhUqOtlZGRAX18/Vyk+ILvaj7Ozs1ah6zt37sDMzAxnz57F3bt3oVQqy3RiFlO1paamYtq0abCwsACPxwMRgYgwiAipPJ520hOLgdGjSzUelvhKWGpqKv7zn/8gKCgISqUSw4cPx7lz57Blyxb4+fnBxcUFy5YtK5FxwBcvXsDCwgLHjx8vgcgrvz59+mDBggUYO3YsIiIiyvTeS5cuhZ+f3yd32+A4Dnw+HxkZGQVe78qVK/Dw8Mj38aFDh6J9+/ZaX7R+/vln2NnZ4cWLF9i8eTNcXV3zTJwMU1wpKSlYsmQJXFxcwOfzNYmOiCCXyzFy5Mjs3qq0NKBePcDAIDvpGRgAzs7Aq1elGh9LfKXor7/+wuTJk2FnZ4datWph8eLF2LdvHzp16gSFQoHIyMhcZdOKateuXXBxcSl3BbbLo99++w3e3t549uwZTExM8KqU/7hyXL16FUqlslDFsg0MDPDmzZsCz1uzZg169eqV7+Pp6emoW7cuvv/+e63jkZGRaN68ObKysjBgwAD06NGDjfcxxfb27VvExMTAx8cHQqFQK9FJJBIEBwfnvxVcVhawezcweTKwaRNQBkU6WOIrA2q1GocOHUL37t1hbGyM0NBQrF27FsOGDYOJiQl69OhRrB0YevTogREjRpRgxJWTWq2Gvb09Ll++jPDwcMycObPU75mamgpPT0/8+OOPhTrf3Ny8UEXSBw8enGs5xL89fvwY5ubm+N///qc5lpmZiaCgIEycOBGpqamoXr061q5dW6jYGCbHmzdvsHbtWgQEBEAikWglOj6fD3d3d6xevbrcfiFnia+MvX79GitWrECdOnVgY2ODb775BpGRkbC1tUXjxo3x888/F7mi/suXL2FpaZlrogOT28SJEzFy5Ehcv34dFhYWpb7rxeDBg4vUqnJ0dMS9e/cKPM/Pz69Qs3oPHToECwsLrRnGiYmJsLGxwd69e3Hz5k0olUrcvHmzUPExVVNycjLWrl2LJk2aQCqVaiU6IoKpqSkGDhyImzdvVogeBJb4dOjatWsYNWoUVCoVGjVqhMGDB6N27dpwdnbGkiVL8O7du0JfKzY2Fs7OzoUaO8zKAl6+zP7fqubOnTtQqVTIyMhAq1atSrW1s3PnTjg5ORWq6zJH9erVce3atU+ek5aWBqlUWuhv0zNnzkRAQIDW+OLJkydhZmaG+/fvY+3atahevXq5/XbOlL2XL19i3bp1aN68OfT19XMlOpFIhAYNGmDTpk1F+v0uL1jiKwfS09Oxa9cutGvXDnK5HG3btkVgYCBMTU3x3Xff4fHjx4W6TlhYWIGTNtasAUxMsoskmJhk/7uqqV+/Pvbs2YPDhw/Dw8OjVCruPH78+LMqpfj7+xf4nAsXLqBGjRqFvqZarUa7du0wfPhwreNLliyBj48PUlNT0aNHD02xbKbqSUpKwrp16xAcHAxDQ8NciY7H48Ha2hpjxozBlStXKkSr7lNY4itn4uPjMWfOHLi7u8PR0RENGjSAsbExunfvXuC2RK9evYKVlZXWNjQcB8yfD5ibAyIRwOfnLpLw0RBQlbBq1Sp07twZHMfBx8cH+/fvL9HrZ2VlITAw8LPGEJs0aaI1JpeX6Oho9O3bt0jXff36NZycnLBlyxbNMY7j0L17d4SHh+PNmzdwdXXF5s2bixwzU/EkJCRgzZo1aNmyJYyMjHIlOoFAALFYjODgYGzatKnMJoKVFZb4yimO43Dq1CkMHDgQcrkcHh4eUCqVqF+/Pnbt2pXvOODevXvh6OiId+/eIT0dmDABkErzXiOa89O2bRm/OB1LTk6GsbExXr16hZ9++glBQUElev3p06ejSZMmn7X7eUhICPbu3fvJcwYMGIDly5cX+dqXL1/ONZ6XkpKC6tWrIyYmBpcuXYJSqcTdu3eLfG2mfHv69CnWrFmDVq1awdjYWKsll5PoBAIBHB0dMXbsWJw9e/azfn8rCpb4KoCUlBRs3LgRgYGBMDIygoWFBaytrbF48eI812H17t0HdevGapbGFPTj7Pyqyu3z161bNyxfvhwZGRmwtbXFhQsXSuS6p06dgkql+uxydV27dsXWrVs/eY6Pjw/Onj37Wddft24dqlWrpvV7c/v2bZiZmeH8+fNYunQpfH19S33SD1O6Hj16hNWrV2sSXU6Cy1lTx+PxoKenBz09PbRp0wYbNmxAYmKirsMuMyzxVTD379/HpEmTYG5uDrlcDn19fQwdOhSPHj3SnLN6dQp4vJRCJT0iDvr6t2FiYoJ+/frh8OHDlfqbXo5ffvkF/v7+ALJ3NujevXuxr/n333/DwcEBsbGxn32NPn36YN26dfk+/v79e0il0mJ9URk4cCC6du2qNU6zY8cO2Nvb48WLF+jYsWOu8UCm/OI4Dn/99RdiYmI0XZc5CU4oFILH44HH40Emk0EikcDZ2Rnfffcdjh07hszMTF2HrxMs8VVQWVlZiIuLQ7t27SCRSCAWixEYGIhTp07Bz69wLb2cH0PDTDx58gTz58+Hr68vrKys8M033+DChQsVfhA7P5mZmbC0tMQff/yBN2/ewNTUVGs/u6LiOA7dunXDkCFDihXXkCFDPllH88yZM6hVq1ax7vHhwwf4+vpi4cKFWsfHjBmDli1b4sWLF7C3t8fPP/9crPswpYPjONy9exfR0dGaySgikQg8Hg9isRh8Ph98Ph8SiQRyuVzTqlu1alWxC2ZUFizxVQKvX7/G/PnzYWNjA4FAAKHwbpESH1F2i69v3764fPkybt68iUmTJsHJyQnVqlVDVFRUpRz3GTNmDMaOHav5/6NGjfrsa61btw5eXl7FLho+ZswYzJkzJ9/Hly1bhkGDBhXrHkB2VSGVSoVjx45pjmVmZiIwMBBTpkzB6dOnc21uy+gGx3G4ffs2li9fjhYtWsDQ0BBisVjTXSkSiTRjdEqlEgYGBnB0dMSoUaNw6NAh1m2dB5b4KpmLFy9CpToJIq5QSU8i4RAVdR99+vSBXC6HUCiEiYkJwsPDcfHiRZw+fRoRERFQqVSazVMTEhJ0/TJLxPXr12FtbY2srCw8fvwYJiYmWoWdC+vPP/+EUqnE9evXix3T5MmTMWXKlHwfDw8PR0xMTLHvA2R391pbW2v990xISIC1tTV++eUXzJ07FwEBAYWqHcqUHI7jcOPGDSxduhQtWrSAgYEBJBIJ+Hw+pFIppFKpplVnYmICS0tL6OnpITg4GEuXLi1UAYSqjiW+SiYmpuBZnB+P73Xq9M9zOY7D5cuX0a9fP5iYmEAgEEAul6NXr144c+YMDhw4gN69e0MulyM4OBgbNmyokItXP+br66vZJigsLOyTra285Gzz8zmzLPMyZ86cT+7g4OXlVaK7KkyePBmBgYFaYz3Hjx+HSqXCvXv30Lp1a02rmCkdarUaV69exaJFizQLxqVSKYRCIfT19WFsbAw+nw+RSAR9fX04ODhAoVDA1tYWQ4YMwb59+1jxgSJiia+SsbEpShdnCr78clmem6JyHIdr165h4MCBUCgUEAgEMDY2Rs+ePXH8+HFs3boV7du3h5GREbp27YrY2NgK2aWyePFi9OzZE0D2dH9ra+tP7qDwb6NHj8YXX3xRYmOhS5cuzXecMCUlBTKZrEjxFSQrKwvBwcG5ku0PP/wAX19fPH78GNbW1jh48GCJ3bOqy8rKwqVLl7Bw4UI0a9YMMpkMMpkMYrEYBgYGUCqVEIlEkEgkEIlEsLe3h5ubG2QyGZo0aYL58+fj1q1blXb8vSywxFfJSCR5J7nshetqEKX9/79T0bjxC3z99RCYmJhg7NixePnyZb7XvXHjBgYPHgwzMzMIBAIYGRmhe/fuOHToEKKjozWVZgYMGIAjR46USjWU0pCUlARjY2NNy7V58+aFLih98OBB2NjY4MWLFyUWz9q1axEeHp7nY8ePH9fMRC1JOZNZdu7cqTmWM1mnf//+OHLkCCwsLApVPJvJLTMzE+fPn8e8efMQFBQEqbxhhWEAACAASURBVFQKQ0NDSKVSGBgYwNraGvr6+hCLxRAKhbCwsICPjw+sra1hYWGB/v37Y+fOnRW+d6U8YYmvkgkKAv69ryOfDwwcCFy9mgVX162wtj4Ac/OvQMRHgwYNcPjwYQwaNAimpqaYMGFCgVUabt++jaFDh8Lc3BwCgQCGhobo1q0bdu/ejTlz5qBmzZqa8kaXL18u999MO3TogDX/X7vt4MGDqFGjRoExP3/+HJaWlvjtt99KNJatW7ciNDQ0z8d++OGHYs8azc+5c+dgZmaGP//8U3Ps3bt38PDwwJo1azB16lQEBQVV6KUuqanAyZNAac/TysjIwJkzZzB79mw0adIEenp6MDY2hr6+PgwMDODs7AyFQgGhUAg9PT0YGRmhVq1a8PX1hYGBARo0aICZM2dWiL+dioolvkrm1i1ALgdkMu7/1+hxCA3NLl0GZH+Y+fj4YO7cuTh06JBmo8h69erhwIED6N+/PxQKBaZMmVKoiR53797F8OHDYWFhoUmCXbt2xaZNmzBu3Dg4ODjAw8MDM2bMKPNdzwvr559/RqNGjQBkt3S8vLzw66+/5nu+Wq1G69atMX78+BKPZc+ePQgJCcnzsZ49e35yjV9xrVy5El5eXlqFzm/dugWlUolz586hSZMmiIqKKrX7l7Tk5OyNvJ2dAReX7N4QI6PsMfCgIKCgPXjPnAEiIoBRo4ArV4Do6OzSf3w+4OsL5Owklp6ejhMnTmDmzJlo3Lgx9PT0YGJiAkNDQ+jr68Pd3R22trYQCoUwMDCAWCxG9erVERQUBBcXFygUCoSFhWHLli2VrjRYecUSXyX0+jWwdCmgVK7GqlX38e8vjU+ePIG1tbVmnVZcXBycnZ3B5/NRp04d7N69G3369IFCocC0adMK3cXy119/YdSoUbCysgKfz4eBgQE6deqE6OhoDBkyBGZmZggICMDSpUvx/Pnzkn7Zny09PV2zUwEAbNiwAS1atMj3/B9++AF169YtldmOhw8fzreEWrVq1QrcuaE4OI5D7969ERYWptXS2L59OxwdHXHjxg1YWFjg999/L7UYSkpmJuDhkX/Xv0QC9O+f//PnzMmuY8vjZf8IBNm1bj++hlicBn//jpBIJFAqlTAxMYFMJoOXlxfc3d2hp6enSXQODg5o2bIlAgMDYWxsDD8/P0yePBlnzpyp0K3oioolvkqsTZs22L17d56PnT9/HkqlEhcvXtQc+/XXX+Hi4gKBQIBatWph27ZtCAsLg1KpxKxZs/Isj5afhw8fYsyYMbCxsQGfz4e+vj6++OILzJs3Dz179oSxsTFatWqF//znP0W6bmmJiIjQLCNIT0+HlZVVnjtG59SzLK3W6+nTp/Mcx3vz5g309fVLvdJGamoqvL29c81SHTVqFNq0aYP9+/fDxsYmzwlR5cnWrYBY/OnJXWIxcn0pBLK37NLTK8zksA8QCqeiZs2a8PPzg6mpKaRSKfT19WFiYoJmzZqhY8eO8Pb2hlwuR9euXatcabDyiiW+SmzIkCFYsmRJvo/v2LEDNjY2uepKHjhwAE5OThAKhfD29sbGjRvx5ZdfQqVSYc6cOYXa8+9jT548QWRkJOzs7DRJMCQkBJMmTULbtm1hbGyML7/8Env27CnRGYtFceHCBTg6Omom5cyePRu9evXSOiclJQXVqlXDpk2bSi2Oq1evwsvLK9fxI0eOoH79+qV234/dvXsXZmZmOHPmjOZYRkYGGjZsiKioKHz33Xdo3bp1uZ3A9P49oFSiUGtZ27QBIiOzuzVzbN+eu3WX3w+fvxJyuRwSiQT16tVDWFgY2rZtC4VCgRo1aiAyMhJHjx5layHLGZb4KrG5c+fim2+++eQ5s2bNgq+vb65kxnEc9u3bB0dHR4hEInh6emL16tXo2rUrzM3NsWDBgs9aOxQfH4/x48fD3t4efD4fMpkMwcHBGDVqFBo2bAiFQoGvvvoKx44dK9MPVo7jUL16dc0u9snJyTAxMdEq8dS/f3/07t27VOO4e/cunJycch2fN29emdbPjI2Nha2trVbLLj4+HlZWVti3bx/q1auHefPmlVk8hZWSkoLw8KMg+lDoZT08Xna3Zr9+wMqVgKlpYZ+XgqCg+ejbty/q1q0LQ0NDtG/fHqtWrSr0HpqMbrDEV4lt374dnT5eoZ4HjuPQp08fdOzYMc9Ew3Ecdu/eDQcHB0gkElSrVg1Lly5Fx44dYWlpiUWLFn12weTExERMnjwZjo6OmiQYFBSEAQMGoEaNGrC1tcV3332Hq1evftb1i2ru3Lno16+f5t8jR47UrG/bvn07XFxcSr1bNj4+Hubm5rmOd+vWDRs3bizVe/9bZGQkmjdvrjUGdfToUZibm+PEiRNQqVRarUJdunHjBoYOHQpDQ0PIZBvyLdhQUDIrbEuPiINEsgeurtUwcuRIxMXFVch1rFUVS3yV2Llz5+Dr61vgeWlpaWjUqBEiIyPzPYfjOOzatQv29vbQ09ODk5MT5s6di7Zt28La2hrLli0r1h9+UlISpk6dqplkI5VK0bBhQ3z55Zews7ODl5cXZs2ahQcPHnz2PQoSHx8PuVyuack+ePAApqamuH79OszMzArcCLgk/P333zA0NMx13NnZGbdu3Sr1+38sMzMTQUFBmDhxotbx+fPnw8/PD9u2bYODgwNev35dpnHlSEtLw6ZNm1CvXj0YGhpCLpfD3NwcJibf5rM7SVYRijt8+kcgyMSJE3/p5HUzxccSXyWWlJQEU1PTQp374sULODk5FThdXq1WY/v27bCzs4O+vj7s7OwQFRWF1q1bw9bWFitXriz2ON3Lly8xffp0uLq6aqrM+/v7IyQkBEqlEg0aNMCKFStKdOF4jlatWuGnn37S/Ds0NBQODg5FLmX2uTIyMiAQCLRmVb5+/RqGhoY6GVNLTEyEjY0N9u3bpznGcRy6dOmCQYMGISIiAp06dSrT9WZ3797FmDFjNIlOT08Pzs7O0NfXR6tWrdCxY08IBA813Z08nhpSqRqtW3PQ08v6/0IOhatlm9+PVJo9lshUTCzxVWIcx0EmkxV6OcIff/wBMzOzQk1Xz8rKwubNm2FjYwNDQ0NYWVlhwoQJCA4Ohr29PVavXl0iA/rJycmYNWsWqlWrpkmCPj4+aNKkCYyNjRESEoJNmzYVecJNfrZs2aK1lGHgwIGQSCRl2o0lEom07nfo0CEEBgaW2f3/7eTJk1CpVFozWd++fYtq1aohJiYGtWrVKrFapfnJzMzErl270KRJE+jr68PIyAhmZmZQqVRwcHBAixYtYGNjA5FIBD09PYSEhKFfv2fw8+MQHPwKvXsvgI2NLezt+6J+/dPw83tfqJq2IlH2uj2B4J9jMln2+kCm4mKJr5Lz9PQs0tqvQ4cOwdzcHHfu3CnU+ZmZmfjxxx9hZWWl+QY+evRoBAUFwdHREevXry+xKfhv3rzB3Llz4e7uDj6fr1kIXLt2bRgbG6NHjx7Yv39/sRLu+/fvNZNajh07BgsLC9SvX1+rFVjajI2NtboPv//+e4zW8Sft4sWLUatWLa1tl27evAmlUonY2FgolUpcvny5xO/7+PFjTJo0CQqFAgqFAhKJBHZ2dpBKpfD19YWNjY1mP8qGDRti165dyMjIwO3btzFlyhS4ubnB0dER48eP19o9IysLmDVLDR7vGaRSdZ5Jr2ZNYMQI4PJl4KuvsmeK2tkB8+YB5XRCK1NILPFVcm3atMGePXuK9JyVK1fCzc2tSGM3GRkZWLt2LSwtLaFUKqFUKjF06FA0bNgQLi4u2LhxY4ku1H337h0WLFgAT09PTRJ0c3ODu7s7zMzMMGTIEJw4ceKzugcHDhyIyZMnw87ODvv27cP+/fvh4+NTZt15VlZWWktMOnfujC1btpTJvfOTU7vz48k/QHYL2cnJCTExMXBzcyuRyT9qtRoHDhxAy5YtNfUsTU1NNT0LlpaWkEqlkMlkcHNzw6JFi/Dy5Us8evQIc+bMgY+PDywtLTFixAicOXMm3/9uN27cgIuLC8LDc5f5IwK8vYGPlrkylQhLfJVcQWv58jNy5Eg0bdq0yK2n9PR0REdHw9zcHBYWFjAxMUH//v0REBCAatWqYfPmzSVeqSIlJQWLFi2Cl5eXZvsWBwcH2Nrawt7eHuPGjSvSXnnHjx+HoaEhIiIiAGR/EHt6euLw4cMlGnd+XFxctGpm2tvbl4uNgD+u3fmx4cOHo23btujXrx969uz52V8Qnj9/jlmzZsHCwgJyuRxisRhmZmaQSCRQKBSQyWQwMzODQqHAmDFjcPPmTSQmJmLZsmVo0KABFAoFBg4ciN9++63A37HLl9PRoMEJWFpugr7+m3y7OmUyYPXqz3o5TDnGEl8lV5i1fHnJyspCSEgIBg4c+FkfZGlpaVi6dClUKhVsbGxgZGSEnj17ws/PDx4eHti2bVupTNZ4//49li5dCm9vb/B4PAiFQtjY2EChUMDb2xuzZ8/Go0ePPnmNVatWQSwWa+1OvnbtWrRq1arE482Lt7e3ptswKSkJcrm83BQr/uOPP3JV/ElPT0f9+vUxefJkeHp6FqmeKMdx+P3339GhQwfo6elBJpPB0NBQs/ecTCaDo6Mj9PX1ERoaigMHDuDFixdYt24dWrRoodkqa+/evXlOqsrKysKdO3ewa9cuREVFoWvXrrC0HA2iVBBlFGqii0gEPHtWIm8fU06wxFfJbdu2rcC1fPl5+/YtvL29sWDBgs++//v37/HDDz/AzMwMDg4Omv37fHx84OXlhZ07d5babMW0tDSsXLkSPj4+4PP5EAgEsLCwgKGhIRo1aoTo6OhcRYFzijJHRERo7YSQlpYGCwuLEtllvSABAQE4efIkgOxd0ps1a1bq9yyKbdu2wdHRUeu9e/r0KSwtLRETEwOlUombN29+8hqvX7/GDz/8AFtbWxgaGkIoFEIqlUIgEEAqlcLd3R1yuRwBAQGIiYnBs2fPsHXrVnzxxRcwMjJChw4dsG3bNs3SE47jEB8fj19//RULFixAeHg4ateuDZlMBjs7O4SEhCAyMhLr12+Gvn7RlzXY2QE6KirElAKW+Cq5s2fPonbt2p/9/EePHsHKyirfmp+FlZKSgrlz50KpVMLFxQWGhoZo164datSoAR8fH8TGxpZqqyY9PR0xMTHw9fXVJEGFQgGpVIq2bdti69atePXqFWrWrIlVq1bhwYMHUCgUWrMrZ86cme9eeSWpWbNmml3hp02b9sn1lboycuRItGnTRutLS86+fXPnzoWXl5fWRBggOzmdPXsWoaGh0NPTg1gshkQiAY/Hg1gshoeHB5ycnGBra4uJEyfixo0b2LNnD3r06AFjY2MEBwdj/fr1ePjwIU6ePIno6GgMHTpUsxekQqFAkyZNMGzYMKxatQonT57E33//DSC7BunVq1exYMGvkEjSipz49PWzS5kxlQNLfJXc8+fPoVAoinWNM2fOlNisvbdv32LmzJkwNTWFh4cHDA0N0bJlS3h4eKB27drYt29fqXfr5UzE8fPzA5/PB5/Ph7GxMYRCIWxtbfHLL78gMzMTTZo0wY4dOzTPe/XqFUxMTPCslPu92rVrh9jYWABA+/bt8d///rdU7/c5MjIy0KBBA0yfPl3r+Jw5c1CnTh2EhoZi0KBBALLHBleuXAlnZ2fo6elp3nM+nw83NzfUrl0bRkZG6N27N+Li4nDo0CEMGDAAJiYm8PHxQe/evTFs2DC0adMGdnZ2kMlk8PPzQ9++fbFgwQLExcUhISEBaWlpuH37Nvbs2YMFCxbgq6++QtOmTTUzP7M3dq2OopQzy/nh84EZM3TxTjOlgQcAxFRaAEhfX5+eP39OhoaGn32d7du305gxY+js2bNkaWlZ7LjevHlDixYtoiVLlpCtrS09ePCAfH196dmzZ2RiYkJRUVHUsmVL4vF4xb7Xp2RlZdGmTZto1qxZdOfOHeLxeCSTyUggEJCfnx9lZGTQsWPHNHEMHz6c9PX16fvvvy+1mLp3707t27en7t27k7W1NZ08eZIcHBxK7X6fKz4+nurUqUPr16+n4OBgIsr+fevcuTOZmprSoUOHyN7ens6dO0dZWVkEgDiOIxsbG3JwcKBbt25RzZo1KSwsjEQiEW3bto2OHj1KQqGQxGIxvXnzhpycnMjLy4tq1KhBNWrUIA8PDxIKhfTXX3/RnTt36O7du3Tnzh26c+cOPXnyhGxsbMjNzY2cnJxIIBDQkydP6OLFi5SQkEBisZj4fD4ZG++ixMTGpFZLCv1aDQyIduwgatmytN5NpiyxxFcFeHp60vbt28nLy6tY15kxYwbt3r2bjh49SjKZrERiS05OpgULFtCKFSvI2dmZ7t27R56enpSYmEjm5uYUFRVFzZs3L9UEmJCQQL6+vrRlyxZ69uwZLVu2jM6fP6/5oFYoFDRo0CDq1asXSSQS8vf3pwcPHhTri8Sn9O/fnwICAqhNmzbk7e1NL168KPUvAJ/r6NGj1K1bNzp37hzZ2dlRWloarV+/nkaMGEGZmZma84yMjKhGjRqUkJBAHz58IG9vb+I4jq5fv07Pnz8ngUBAjo6O1KRJE2rUqBFZWVkRx3H0+PFjTWK7e/cu3b9/nxQKBbm6upKbmxu5ubmRq6srubq60vv37+nw4cO0c+dOunr1KolEIsrMzCQfHx/q1KkTtWjRgtRqNf3++wmaNi2dUlJ6EpE1EfE/+RqlUiJ/f6LffiPif/pUpoJgia8KaNOmDQ0ZMoTatm1brOsAoF69elF6ejpt27aN+CX4KfDy5UuaN28excTEkKenJ/3555/k4OBAL168IDs7O5o2bRoFBQWV2P1ycBxHLVu2pPr161NUVJTW8e3bt1NERAS9evWKiIiEQiFZWFiQsbExdenShaZMmVLi8RARRUREkKurKzk6OtLy5cvp4MGDpXKfkjJv3jz66aefyMXFhfbu3auV8AwMDCgtLY0AkFAoJJFIREqlkt6+fUscx5G/vz/VqVOHeDwe3bt3T5PgxGKxVmLL+f8uLi6kr69PRNlfWOLi4mj79u10/PhxysrKoqysLHJ0dKT27dtT69atSSwW06lTp+jw4cN06tQpkkgkJBQKKSkpiYiI6tRZQefP9yEiCREJiAhElPMlQ01KpYDmzCEKCyMSi8vyXWVKE0t8VcCQIUPI09OThg0bVuxrpaWlUbNmzSgoKIhmzJhRAtFpe/78Oc2ZM4c2bNhA3t7edPv2bTI3N6dXr16Rq6srRUVFUePGjUvsfvPmzaPdu3fT77//TkKhMNfjv/32G40aNYomTpxIixcvprNnz1JWVhYREbm6utLo0aMpNDSUTExMSiymsWPHklwupw8fPpBarS6V97kkZGZm0g8//ECzZ8+m5ORkzXGBQEDInj9APB6PTExMSCqVUlZWFiUnJ5OJiYnmtf07seW03kxNTXPd7/3793T06FHasWMHHTx4UNMSNjU1peDgYGrfvj0ZGBjQxYsXKS4ujs6dO0dSqVTzXH9/f2revDn5+flR27ZtSSAQ0JYtW6h37xWkVg+gNm260+HDh+n9ez8CiLp0Sad168zp//MsU5mU/bAiU9bmzJlToiWvkpKS4OjoiB9//LHErvlvz549Q0REBExMTDR1GN3d3WFpaYlmzZpppvsXx/nz52FmZoaHDx/me45arYadnZ1mN3aO47Bz504YGhqCz+eDiMDn81GzZk1s3rw510zGzxEVFYWJEyeiTZs2mkkuuqZWq3Hv3j3ExsYiPDwcpqamoOzmEYgIPB5PU0HH0NAQ9vb2MDIy0jwuEAjQoEEDREdH4+jRo4iPjy9wEpNarcaFCxcwfvx4eHp6QigUQiKRwNDQEK1bt8aqVavw3//+F1OnTkW9evUgkUhgamoKIyMjzTnz58/HuXPntAoxHDt2TDOjdOrUqTAyMkJoaCjevXsHPp+PmTNnlmmlHqbsscRXBWzbtg2dO3cu0WveuHEDZmZmWou8S8Pjx48xePBgmJqaom3btnBycoKTkxPMzMzQsmXLz94P7u3bt3BxccG2bdsKPHfChAkYNWqU1rHdu3fD19cXsbGxaNiwIYRCoeYD3t/fXzMz9HPMmzcPo0aNgkql0ipdVhY4jkNCQgIOHTqEhQsXol+/fqhTpw6kUin09PTA4/G0Ep5QKNQkPXNzcwiFQjg5OUEmkyEkJARbtmyBv78/Bg8eDJVKVWDxgIcPH2LZsmVo3Lgx9PT0NHU4/f39MX36dKxduxYTJ05E7dq1IRaLYWJiAplMBoVCgS5dumDlypW4cePGJ9eGzps3TxNzixYtQEQ4evQoZsyYAZlMhtDQUCxbtqyk31qmHGGJrwoo7lq+/Bw8eBDm5ua4d+9eiV/73x48eID+/ftDoVCgc+fOqFatGmxtbaFQKNC6dWucP3++SNfr06dPrrqT+fnzzz9hbm6u1WpQq9Vwc3PT7GTBcRz279+PBg0aaJKgUChE48aNceLEiSK1HlasWIGePXvCwsKiVFsdb9++xalTpxATE4OIiAgEBQVpamLm1D1VKpWalm1Oy04oFMLCwgJSqRSBgYEYNWoUOnbsCCMjI9SsWRMWFhZaO9c/efIEFhYWGDhwIOrXr6/1Pr558wY7d+5E586doVAoIBKJIBQK4ebmhoiICCxcuBBjxoxBjRo1IBaLIZfLNUsTevfujY0bN+LBgwdFep9atWoFgUAAPT096OnpQSQSAQAcHR3RunVrGBsbIzk5ueTeaKbcYYmvCkhMTCz2Wr78LF++HO7u7mX2QXHv3j307t0bSqUSPXr0gLe3t6YmaLt27XDp0qUCr7Fp0ya4ubkVaSujgIAA7N27V+vYqlWr0LZt21znchyHAwcOICAgQJMExWIxWrRoUaidMjZs2IDAwMA8r/050tPTce3aNWzatAljx45Fy5YtYWFhAbFYDEtLSzg4OMDa2hp6enowNjaGtbU19PX1IRAINAmPz+fDyckJKpUK3t7eGDRoEEJDQyGXyxEUFISYmBhNJZchQ4agQ4cOWsno8OHDMDc3R+PGjdGrVy8MHToUjo6OEAgEmkQaGhqKKVOmYOjQoXB3d4dIJIKRkRFEIhFcXFwwZMgQ7NixA4mJicV6PxQKBfh8PlQqFQwNDeHn54fnz5+Dx+Phm2++Qa9evYp1fab8Y4mvCuA4DlKpFO/evSuV60dERKBFixYlsv9eYd2+fRvdu3eHmZkZ+vXrBz8/PyiVShgbG6NDhw75Jpj79+9DqVQWKkF+LDo6Gl26dNE69v79e6hUqgJ3Rv/1119Rt25dTSLR09ND27Zttfa301Cr8fvEiYi0sMDi4cOLFKNarcb9+/cRGxuLyZMno2XLlrC1tYVQKISRkRFMTU01XZbVq1dHaGgoJk2ahFmzZqFfv35wcnKCkZGRppqKQCCAsbEx7O3toVQqERoaiq5du8LMzAx169bFokWLEB8fnyuOtLQ0+Pv7Y+7cueA4Dnfu3MGsWbNgZWWlSaRSqRRNmzbF8OHD0adPHzg7O0MkEmnqdHp5eeHbb7/FgQMHCr2fZGG8evUKAoEASqUS5ubmEIlEWLBgAUaMGAG5XA5PT08cPXq0xO7HlE8s8VUR7u7uuHHjRqlcOzMzE61bt8bgwYPLfELAjRs30KVLF5ibm2PIkCFo1KgRTExMYGRkhE6dOmm95oyMDNStWxcLFy4s8n2Sk5NhbGycq7ZnVFQUBgwYUOjrHDp0CHXq1NEkQZlMhs6dOyMhIQFISgLc3JApleIdEbLEYiCfcmWJiYk4cOAAxo4di+bNm2sSnEQigUQigUAggLW1NYKCgvDNN99gzZo1OHbsGBISEqBWq3Ht2jVMnDgR1apVg7m5OWxsbDRdmgKBAFZWVjAwMEDTpk3RoUMHWFtbo0aNGpg5c2beCfsjL1++xIIFCyAWiyGVSiEUCiEWi1GzZk3Y2NjAyspK09LT19fXjOFNnToVx44dw4cPHwr9fhZVXFwcxGIxateuDX19fRAR3r17BwsLC4SEhMDV1ZVNaqkCWOKrIlq1aoV9+/aV2vXfvHmD6tWrY9GiRaV2j0+5cuUKOnToACsrK4waNQrNmzeHoaEhDAwM0LVrV/zxxx8YP348WrVq9dlFsUNDQ7FixQqtYy9evIBcLs9OXEV06NAh+Pr6apLgZoEAmf/aGI6TyfDb999j5MiRCAwMhI2NDcRiMfh8Png8HoyNjVG9enV06dIF8+bNQ1xcHB4+fJjrNXIch6tXr2qSna2tLRo0aKCZecnn82FkZAQjIyN4enqidevWmolEEyZM+OSXprS0NMTFxSEsLAyWlpYQCASaguASiQSNGzeGlZUVRCKRpjWpUCjg4uKCs2fPlthGxYUxadIk8Pl8tGvXDnw+H5aWlrhz5w54PB66du2KOXPmlFksjO6wxFdFDB48uNRnqj148ACWlpalmmALcuHCBYSEhMDGxgbjxo1DSEgI9PX1NRMZTp069dnX3r9/P+rWrZvr+Ndff42JEycWJ2z88ssveMvn5yoSmUWEOTweLCwsUK9ePQwcOBAbNmzA7du3C+xa5jgOV65cwYQJE+Dm5gZ7e3v07NkTderU0SROgUAAExMTKBQKBAYGolq1apovD2fPns2z9cNxHK5du4Zvv/0W7u7umkRnZmaGevXqoX79+jAzM4NIJNI8FhwcjJUrV+KPP/7AlStXoFAoUKdOHUybNq1Y71tR1a9fHzKZDM2aNYNAIEB4eDjCwsJgYWEBY2PjYo8fMhUDS3xVxOzZszFmzJhSv8+pU6egVCpx9erVUr/Xp5w+fRrBwcGwt7dHZGQkpFIpJBIJpFIpunfvXmB3XV4yMzNhYWGB27dvax2/c+cOlEplkSbL5MnKKlfi48Ri4PvvC32JnGQ3fvx4uLq6wt7eHiNGjEBERATMzMzA4/HA4/E0XwZq1aoF+h/VXwAAIABJREFUDw8PmJqaYtCgQThy5Eiem7gmJCRg0aJFqF+/PvT09CAQCKCvrw8PDw94e3tDLpdrZmSam5ujc+fOmhmXISEhGDlypNb1Nm7cCEdHR6hUqjIdUzM0NNRse0REuHTpkmbnh8/dvoupeFjiqyK2bt2aa3JGadm8eTPs7Ow+q/uvpB07dkwz6WXOnDn48ssvNa2/Hj164MGDB0W63ujRozFu3Lhcxzt27Fj8FvXs2dlbfv97P5yPlgbkheM4XL58WZPsHBwc8O2332LLli1o2bKlpitVIBBAIpHAwcEBHh4eMDIyQlhYGPbv359rE9fU1FTs2LED7du3h6mpKXg8HkQiEWxsbODk5KSZhCIQCGBnZ4devXph586dePHiRa74Xr9+DUdHx1xrJgcPHoz69evDxsYmz+eVtPj4eIhEIkilUojFYgiFQpw+fRo8Hg++vr745ZdfSj0Gpnxgia+KOHPmDPz8/MrsflOmTEHdunVLpJJJcaxYsQK1atXCwYMH0bBhQ1SrVg2LFy9Gnz59NAkwLCyswIXVOa5duwYbG5tcraKTJ0/Cyckpz9ZSoanVwHffQS2VIp0IcHAAjhzJ81SO43Dp0iWMGzcOLi4ucHR0xHfffYdTp05h8eLFMDc318ygzKmm4urqCgMDA3Ts2BHbt2/XbOKafWs1Tp8+jQEDBsDe3l6zwFuhUMDKykozSUUgEMDV1RWDBw/GwYMH8fbt20K9tIsXL0KpVGrNgM2Z/RkYGIiQkJBS25A4R2xsLIRCIby8vMDn8+Hq6oqQkBDY2dnl+d+UqbxY4qsiEhMToVQqy+x+HMehe/fuCA0NLfUPtPxcv34dCoVC0zXJcRzi4uJQt25deHp6YtWqVfjqq680MyHDwsIKVSmlVq1aOHToUK7jAQEBWvv3fa7k589hb2gI/Gt8jeM4XLx4EWPHjoWzs7Mm2V24cAG3b99GSEiIpnWXs9munZ0d9PX1ERwcjA0bNmg2ZgWyx2QnT56MmjVratYb6uvrw9TUVNMiyllaMHr0aBw/flxrY96iWrt2LTw8PLSW1Tx69AgqlQqenp6YP3/+Z1+7MIYMGQI9PT00bNgQRIThw4dDT08PDRo0wOTJk0v13kz5whJfFVHaa/ny8uHDB9SrV08nHyrv379H9erVsW7dulyP5VRZqV27Nry9vbF+/XpERERodgXv1atXnuvTcixatAhhYWG5ju/cuRN169Yt9nT49PR0CIVCTawXL15EZGQknJ2d4eTkhMjISFy4cAEZGRmIjo6GpaWl1kJzU1NTyGQyBAQEYPny5Xj+/DmA7Jm3MTExCAoKgoGBgaY1aGBgoGnN6enpwc/PD5MnT8alS5dKvBXUv39/dOvWTes9iouLg5mZGRQKxWeXoCuMGjVqQKVSwdnZGUSkKV2mUPwfe+cdX+Pdv/HrrGwyzjmZMkjIQBJBYkaQGEGNqj1qtrbUqFlKUKsoajxGpWZVjZoPHhQVIyJCkyCEBiESI7JzzvX7I3V+jRghJzkh9/v1yovc93dcd5JzrvP5js9X/sZ8rQIfH4LxlSNKci/f60hKSqKTkxM3btxYqv0OHTq00Bvsy6jVau7atYteXl708fHh1q1b+dVXXxUwwFet8nv48CFNTU0LDfPl5eXR2dmZp06dKpZ2lUpFsVjMcePGsUqVKnR2duaECRMYERFBtVrNhIQEtmvXThPdiUQi6unp0djYmNWrV+f8+fN5584d5ubmcv/+/ezSpQstLS01xvjvU9BNTEzYqFEjzp07l3FxcSW+hy0jI4O1atUqtO0lNDSUrq6urFy5colkAVKr1dTX16eNjQ2lUinFYjHr1q1LJycntmzZUuv9CZRtBOMrR7Rs2VInWw2io6OpUCiKbQhFZdeuXXRyciryG6hKpeKvv/7K6tWr08/Pj7/99hsnTpxIQ0NDymQy9u7dmw8fPixQ55NPPnllNLl8+XJ26NDhnTWr1WqeP3+e48ePZ5UqVSgSiRgSEsKLFy9SrVZTpVJxzZo1tLa2LpAk2tDQkA4ODpw2bRpjY2MZHR3N4cOH08XFRZNQWiaTaVZzmpubMzAwkMuXLy/1BNgvuHnzJi0tLQv8PahUKrZt25ZeXl789NNPtW7A8fHxlMlkmiHcF9ss3NzcuH37dq32JVD2EYyvHPHFF19w+fLlOul73759tLa25s2bN0u0n8TERFpZWb3XsUV5eXncsmULXV1d2ahRI+7fv5/Tpk2jsbGxxgAfPXpEMn9Ys0mTJoXaSE9Pp1KpZFxc3Fv7e2F248aNY+XKleni4sKJEyfy4sWLVCqVTEpK4r1799i2bdsCeTOlUinlcjlDQkJ49OhRzpo1i7Vr16ZMJtNEdS8iQUtLS7Zr144bNmzQaC8L7N27l3Z2dgUi6tTUVDo5OdHR0bFQooDiEhYWpolwX/wMpVIplUploRWtAh8/gvGVI+bMmcNx48bprP8lS5bQw8OjwAILbZKXl8emTZsWe1N0bm4uw8LC6OzszKZNm2rM5cVcWK9evXj//n0qFIpXGvnUqVP55ZdfvrJttVrNc+fOcdy4cXRycmLVqlU5adIkRkZGaqIctVpNuVxOuVxe4FSEFwtwpkyZwsDAQE3KrReRnUgkoq2tLbt06cIdO3YUf19hCTNlyhQGBAQUyNwSGRlJc3NzmpubMzIyUmt9devWjRYWFgVOZHBwcNDqOZUCHw6C8ZUjtmzZws8++0xn/avVag4dOpQtW7YskTRVs2fPpr+/v9YWZOTk5HDt2rV0cnJiUFAQT548yXnz5rFixYqao3MmTJhQqF5SUhLNzMw0w6NqtZpnz57l2LFjC5jdpUuXCgzpPXz4kK1atSp0DFD9+vXZpEkTmpubFzr81d7enn379uWhQ4c+uMglLy+PQUFB/PqlfKTr16+njY0NXVxctLYYq0qVKprh31q1ahEALSws3ppgXODjRDC+csSZM2dYt25dnWrIzc1lixYtOGzYMK22e+bMGVpaWvLOnTtabZfMX2W5cuVKVqpUicHBwQwPD+eiRYvYxsCAZwE+1dNjTosWZEyMps7AgQM5cOBAjhkzho6OjqxWrRonT55cyOzI/GOIXj7RXCKRaIbl/m10Dg4O/PLLL3n69GmdbRPRJsnJyXRwcODOnTsLXB88eDCdnJzYq1evYs/35eXlUSKR0MnJiQDo7+9PkUjEhg0bFqtdgQ8XwfjKEffv36dSqdS1DD5+/Jju7u5cunSpVtp78uQJK1euzB07dmilvdeRlZXFpUuX0tbWllMaNGCegYEmw4oKYLpUyj/+WRlqY2NDS7GYW9u0YcL8+VS/NOz46NEjNm7cuMCJ5i/yZ/7b6BwdHTl69GhGRUV9tKcGhIeHU6lU8tq1a5prmZmZ9Pb2ppWVFdevX1+s9q9cuUKJREJDQ0OKRCIqlUrq6+vzp59+KqZygQ8VwfjKEWq1mgYGBmVi7ic+Pp7W1tY8cOBAsdpRq9Xs0aMHv/jiCy0pezsZGRlMrFy5UF7NTIDTAHp4eDBm5kxmicXM1tcnK1QgTU3J8+e5dOlSGhgYFIjk/m189vb2HDduHBs2bMhDhw6V2jPpmh9//JE1a9YskE0mISGBcrmcpqamxRqSXLhwoWbhT6VKlTQrXcvC60BANwjGV85wdXXl1atXdS2DJHny5EkqlUpGR0e/dxsbNmygh4dHgTfMkkCtVvPu3bs8cOAA586dy0cv59T85+t/Vla0NTHh81fcu/kKo7Ozs+P48eML5TVt164dd+3aVaLPVJZQq9Xs1asXe/fuXSCyPXjwIE1NTenm5vbe6e9atmypMbxatWpRKpWydu3a2pIu8AEihkC5wsnJCQkJCbqWAQBo1KgRvv/+e7Rr1w4PHz7Mv6hSFbn+jRs3MGbMGGzZsgVGRkZa05WZmYkLFy5g3bp1GD16NJo1awalUgkvLy8sWLAASUlJeOLlBYoLvnzy9PWR3qABBlarhlc9hRUAz4oVMWrUKKSmpkKlUiExMRFz586FtbV1gbKGhobIzMzU2jOVdUQiEVauXInIyEisWrVKc71ly5YICQlBSkoKRo4c+V5tR0VFQSqVAgD+/vtviEQiNG3aVCu6BT5MpLoWIFC6lCXjA4BevXohLi4OKxo2xDfPn0OUlARUqQIsXw60avXaejk5OejevTu++eYbeHp6vlffJHHnzh1cvny5wFdCQgKqVasGT09PeHp6onXr1vDy8oKVlRVEIhEAIPv6dah8fYGMDEhzcpAOIEEsRq8DB9C8QgVIJZJCJm6kr4+o+HhAoXirNiMjI2RkZLzXc32oGBsb47fffkPDhg3h4+MDX19fAMDUqVNx6tQp/PrrrwgMDETXrl2L3GZOTg4ePnwIQ0NDSKVSPHr0CBUrVoS3t3dJPYbAB4BgfOWMsmZ8APBt/frImTMHohdGcfMm8OmnwJkzwGtMberUqbC2tsbw4cOL1Mfz589x5coVjblFRUUhOjoaRkZG8PLygqenJz755BNMmTIFrq6u0NPTAwA8efIEsbGxOHDgAGJiYhATE4PY2Fj8/fff8HJwwBf6+qguFmPNtWto89NP+Lt1a+hJpXiqUEAvI0PzAqO+PkTNmxfJ9IDyF/G9oGrVqli1ahU+++wzREREQKFQQCwWY9u2bahRowYGDx6MOnXqwNnZuUjtXbx4ESSRnp4Oc3NzZGVlwcbGBooi/h4EPk4E4ytnODo64uLFi7qWUQDxvHkweHmIMysL+OEHYM2a/O/Dw4ElS4DUVFzx9MSWTZsQERWlicBeoFarcevWLURFRRWI4u7duwcPDw9NFNepUyfUrFkTSqUSJHHv3j3ExMTg+PHjWLFiBWJjYxETE4O0tDS4ubnB3d0d7u7u6N+/P9zc3ODs7KwxRwDY9tVXuHD5Mpq2bIlWwcFw9ffHj0+fgufPIzcvD0/8/GC5ZUuRfybl1fgAoGPHjjhz5gx69uyJ/fv3QyKRwMLCAnv37oW/vz/at2+PiIgI6Ovrv7Wt3bt3w8jICOnp6ZoIWk9PTzC+co5gfOUMJycn3L59W9cyCpKcXPiaWo07585BLykJ1sePAwMGAJmZAInK//0vLtSvD5lMhpMnTxYwuCtXrsDCwkJjcF27dsXs2bPh4uICAIiPj0dMTAzOnj2LDRs2aCI4Q0NDjbm5ubmhQ4cOcHd3R6VKlQqZ66vo378/goKCsGfPHvj7+2PJkiUQSyRAejpqenjA18EBP1esWOQfSXkc6vw3s2fPRmBgIGbMmIFvv/0WAODj44PFixcjJCQEY8aMwbJly97azvHjx6Gnp4f09HRkZ2eja9eu+PPPPyGXy0v6EQTKMILxlTPK4lAnOnUC4uOB7GzNpVx9few1MMAUNzdcT0+HPC9Pc88YgCg8HC1sbJDn5aUZquzVqxdq1qwJmUyG2NhYxMbGIjo6Gr/88gtiYmJw8+ZN2NnZaczN398fX3zxBdzc3GBhYVGsR5DJZEhNTUWLFi2wdOnS/zdLY2M0aNYMR44ceaf2DA0NkZaWVixNHzJSqRRbt25FnTp14Ofnh+DgYADAwIEDceLECaxfvx5BQUFo3779G9uJjY1Feno6RCIRZDIZhg4dij179ggRXzlHRJK6FiFQeqjVahgbGyMlJUWrKyGLRVoa0LgxEB8P5uRAJRbjrqMjQuvVQ9zFizgcHY2XB7XUhoZICw1FVJ06mrm3F9FbcnIyqlatWmCI0t3dHVWrVoWhoaHW5YeHh6NDhw5o2bIlMjIysH379gL3T5w4gWbNmiEtLa3IP/PFixfj1q1bWLJkidb1fkicPn0anTp1Qnh4OCpXrgwgf9Wtl5cX7t+/j6tXr8LBweGVddPT02FiYqL5XqFQICEhAXK5HJmZmUWK5AU+ToSIr5whFovh4OCA27dvw93dXWc6cnJyEBcX9/9DlNbWMP/7b9jn5iKtalWwUSPU8fJCvwEDIGnTJt8c/0VGZia6TJuGtJo1NcYWGBgId3d3ODk5QSKRlMpz7N27F/369cNPP/2ERo0awdHREY8ePSoQUTRq1AgikQi7d+9G9+7di9RueZ7j+zcNGzbExIkT0blzZ5w+fRoGBgYwNDTEoUOHULNmTbRt2xYXL17UbFf4N8eOHSvwfUhICFJTUyGXywXTK+cIxlcOeTHcWRrGRxIPHjwotGUgLi4Ojo6OmmHK/oMGwSQkBE+ePEFcXBxiYmKwYsUKXLt2DSP09TFdLIahWg0AyAAQa2yMXQ8ewFCHUeu6deswadIk7N27F35+fgCAdu3aYePGjRg9erSmnEQigYuLCzZu3CgY33swatQonDlzBiNHjsTq1asBAJUrV8bWrVvRuXNnjBkz5pWR8e7duyGVSpGXlwexWIwBAwbg3r17wjCngGB85ZGSmufLyspCTExMoRWVKpVKY3C+vr4ICAhATk4Obt68iZiYGKxfvx6JiYmoXLmyZv4tODgYY8aMgaurKypUqADs3w8sWAA8fozE2rURuHEj6rZvj3379hVYXVkakMSsWbOwdu1anDhxAq6urpp7AwYMwMiRIzFq1KgCUUWrVq0QFhZW5D7K++KWfyMSibBmzRr4+vpi/fr16NevHwCgbdu2GD58OJYuXYrWrVuj1Uv7Ps+cOQPVP6uFPT09YWVlhejoaGFhi4BgfOWR4hofSSQmJhaK4m7evAkXFxd4enrC0dERzZo1Q0BAAO7evYvY2Fhs27YNz58/h5ubm2b+7cX2ABcXF8hkstd3Ghyc/wWgGoDh1tZYtWoVunfvjm3btr1yqKskUKlUGDFiBP7880/8+eefsLGxKXDf398f6enpiIiIQJ06dTTXu3fvjmXLluHOnTuvnZP6N0LEV5AKFSpgx44daNKkCby9vVGrVi0AwNy5c3HixAl07twZN27cKJAB5/r163ixhCE0NBQACg1DC5RPBOMrhzg5OWHXrl1FKpueno6rV68WMjk9PT14enrC3t4eVlZWaN68Oby9vXH9+nX8/vvvMDIyKrC4pGPHju+0PeBtfPvtt5qtDH379kVYWFiJz+tlZWWhZ8+eePLkCU6cOAFTU9NCZcRiMfr164e1a9cWML7atWtDJBJhx44dCAkJeWtfgvEVxsPDA8uWLUPnzp1x4cIFmJubQyKR4ODBg3B2dkaLFi0QGRkJiUSClJQU5OTkAMj/nbyIBlNSUoSIT0AwvvLIqyI+tVqN27dva7KavDC4xMREVKtWDfb29qhYsSLs7OxgbGyMW7du4Y8//tBsD3B3d0ezZs0wbNgwrWwPeBsSiQRbt26Fj48P/vrrLwwePBj/+c9/IBaXTPrZx48fo3379rC1tcX+/fvfuHn6888/h6enJxYuXKhZxSmVSuHh4YFt27YVyfiEoc5X07VrV5w5cwZ9+vTB7t27IRaLIZfLcfDgQfj7+2PMmDFYvHhxgQ92Hh4emg9FQsQnAAjGVy6Ry+W4fv06VqxYoTG46OhoVKhQAQ4ODjA1NYVIJIJSqURmZiauXbsGkpr5t5LeHlBUbGxsEBYWhr59+yI6OhojRozAsmXLtL5iLzExEa1atUJgYCC+//77t5prpUqV4Ofnh507d6Jnz56a6+3atcOCBQuQl5f31qFZIeJ7PfPmzUPTpk3x3XffYdKkSQCAevXqYebMmZgyZQqCg4Oxb98+TflBgwZp/p+SkoIqVaqUumaBsoWwj+8jRqVSIT4+vsAQZVRUFJKSkpCdnQ0fHx/IZDKkpaXh7t27EIlEhczNzc2tVLcHvA8vkhinpaUhICAA8+fP15r5/fXXX2jdujWGDx+OsWPHFrnd7du3Y8WKFfjf//6nuXb27FkEBATgyJEjaNiw4RvrX79+Ha1bt8aNGzeKpf9j5e7du6hbty7CwsIQGBgIIH/uuUWLFvjzzz+hp6eHJ0+eQCwW4/Lly6hevToAoEePHggODkavXr10KV9AxwjG95GQmpqK6OhojcFFRkbi6tWrqFixIszNzSGVSvH8+XMkJSVBoVDg8ePH6NSpExo2bKgxOktLyw9yf1NeXh4CAwPRoEED7N+/H+3atcPMmTOL3e6LzdMLFixA796936ludnY2KlWqhLNnz2oijNzcXJiYmGDkyJGYP3/+G+snJibCz88Pd+/efW/9HzvHjh1Djx49cO7cOdjb2wPI39zu6OiI5H/S4EmlUmRkZGgWTrVs2RKjR49G69atdaZboAyggzMABYpBbm4ur1y5ws2bN3PChAkMCgqipaUlDQwMaGtrS0dHRyoUCspkMlatWpUdO3bkpEmTGBYWxvPnz/PZs2ckyaCgoGKffl6WuHv3Lq2trblz5056eHgwNDS0WO3t2rWLSqWSBw8efO82Ro0axalTpxa4VqdOHVatWvWtdR89ekRzc/P37ru88N1339HPz4/Z2dmaawcOHNAc9luzZs0C5X18fHju3LnSlilQxhDm+MowDx8+1AxPhoeH49KlS7h9+zaMjIwgk8mQmZkJlUqFqlWrokWLFgXSczk7O79xe0CZzNlZDGxtbbFhwwb0798f+/fvR+fOnWFoaIivvvrqndtavXo1pk+fjv379xdYmfmu9OvXD23btsW0adM0Q8WffPIJQkND37q6UFjcUjTGjx+PM2fO4KuvvtIkrT5+/Ljm/ssLWYRVnQKAsLilTJCdnY3Y2FhcvHgRp06dQkREBOLj45GTkwN9fX1kZWXBwMAALi4u6NOnDzw9PTXzb++7PeBjMz4AaNGiBfr27YuxY8fi8OHDCAgIgIGBAYYOHVqk+iQxY8YMhIWF4Y8//tCc6PC+vDi89siRI2jZsiUAIDAwEAsWLMDhw4fRrVu319Y1MDBATk4O1Gp1ia1U/RgQiUTYsGED6tSpg80//4weNjZI//FHVAGQbWeHkydPFljJKazqFAAgDHUWh6dPycGDSQsL0sqKnDqVzMl5fXm1Ws3ExETu2LGDw4cPZ/369alUKikWi6mnp0exWExzc3PWqVOHgwcP5po1a3j69GmmpKRoXfumTZvYtWtXrbera3Jzc9m4cWOGhoby5s2btLe359q1a4tUb/DgwfTx8WFSUpLW9Pz444/s0qWL5vucnBzq6+uze/fub61rYGDA9PR0rWn5mLly/Divi8XMMzLiM4AZALc4OFChUNDV1ZVqtZqZmZmUyWRUq9W6liugYwTjKyIqFRkbS96/T+7bR7ZoQZqYkBIJCeR/GRqS/frll09PT+ehQ4c4fvx4Nm3alJUqVaJMJqNYLKZYLKZCoaCvry+/+OILhoWFMSoqipmZmaX2PKdPn6afn1+p9VeaJCYm0srKisePH2dcXBxtbW25adOm15bPyMhg+/btGRQUpJkD1RaPHz+mqakpHz16pLnWsGFDWlhYvPUN2MLCgsnJyVrV89HSuzfz/v1iBJgO8OratZTJZPy6Vy8+mTyZi4yNyfBwXasV0DGC8b2C+/fJAQNIBweyQQNywQLS2po0Ns43OrG4wOvrpa8sGhhYEwAlEgmVSiXr1q3LgQMH8ueff+aNGzeYl5en60fk3bt3aWVlpWsZJcaBAwdoZ2fHBw8e8MqVK7S2tuavv/5aqFxKSgobNGjAnj17FlggoU169OjBJUuWaL4PDQ2lqakpL1++/MZ6dnZ2vHPnTolo+ugwNS30YlQBVI8fz/99+y3TAOZKpcwDSCMjcsIEXSsW0CGC8b3E8+ekrS0plb7J3F7/JZVmcs6cX3nnzp0yPaSiUqmor6/PjIwMXUspMSZOnMgWLVpQpVIxMjKSlpaW/P333zX3b9++TXd3d44dO5YqlarEdBw9epSenp6av4dTp05RqVRy/vz5b6xXtWpVxsbGlpiujwm1rW2hF2MmwMXW1nz0ClOkgQF544auZQvoCGHW/CW2bgWePgX+deD3O1GxogHGjv0U9vb2ZXpP3L/P5ftYmTFjBjIyMjB37lx4e3vj999/R//+/XH48GFcuXIFDRs2xMCBAzF//vwSXUASEBCAZ8+eITIyEgBQt25dpKWlYe/evW+sJ2RvKRqXL1/GYhJZLyVZkIrFaDp1KkyfPy9cSSIB/vijlBQKlDUE43uJa9eA9PR3rUVIpYCREbBuHVBKBwUUG0dHx49uZee/kUql2LJlC5YsWYKTJ0/C19cXv/32Gz777DM0btwY8+bNe6/tDu/KvxNXA4Cenh78/Pxw7tw5pL/hj00wvjejUqkwd+5cNG/eHOahodAfNgwQiaDJyCESocY334Cv+gAqFgOVKpWmXIEyhGB8L1G/PmBi8rZS/052kwmJ5BqqV/8VM2bshJ9fUgmq0y5OTk4fdcQH5OfNXL9+PXr06IHk5GQ8ePBAc8/JyanUdHz++efYunWrxsiaN28OuVyOEydOvLaOsJfv9dy4cQP+/v44dOgQLly4gM/794fI1BSQyfDC5qQqFfJSUpBYsSLU/84pK5UC1tZAs2Y60S6gewTje4m2bQFvb8DYOP97PT1AXx8wMAAqVgSMjdWwsMhApUqxMDI6D2AYatbsCUfHn7Fnz2K4u7vDy8sL48aNw+HDh8v0J/aPcS/fq2jdujV69uyJgIAAjBgxAseOHcOWLVvQvn17RERElIoGBwcH1KlTR3NqQEBAAEji0KFDr60jRHyFIYkVK1agXr166NKlC44cOQJHR8f8mydOAP8cRfQCPQCVK1WCeNo0wMYm/0XctStw6lT+cKdAuUQwvpeQSoGjR4Hly4EuXYCQECAuDkhOBrZvB44dE+PRI2P8/bcb0tPrYto0B5iYGKJq1arIyMhATk4OxGIxzp07h5CQEFhaWqJly5ZYuHAhoqOjNQdjlgXKi/GRhEgkws2bN9G7d2/UqlULrVq1wurVq9GmTRtcvny5VHT0799fM9zp6+uL1NRUHDhw4LXlBeMryItTMtavX49Tp05h1KhRBeZmj6ekIOflSmIxUKMG8PXXwL17+RPhuiu2AAAgAElEQVT4GzcClpalql2gjKHbtTUfPmlpabS0tOTVq1dJkk+fPuWBAwc4YcIENmjQgEZGRnR1dWWtWrVoY2NDKysr9unThz///LNWN0q/D6dOnWK9evV0qqGkyc3NZf/+/Vm3bl1GRkbSysqKJ0+e1NzfunUrbWxsGBMTU+JasrKyKJfLeevWLZJkQEAAzczMNN+/TO/evfnTTz+VuK6yjlqt5saNG6lUKjljxgzm5uYWuL9r1y7q6+vTCeAzkYjqF/uNRKL8PUj/vDYFBF4gGJ8WmDNnDrt16/bKe+np6Tx69Ci/+eYbBgQE0NDQkA4ODnRxcaGRkRFr1KjBcePG8fDhw6W6gZ3M3+htbW1dqn2WJs+fP2ebNm3YunVrpqWlkST37t1Le3v7AhvDN2zYwEqVKvFGKSxvHzFiBKdNm0aSnD59Oj08PLhy5cpXlh08eDBXrFhR4prKMsnJyezcuTM9PDx44cKFAveuX79OR0dHzZ7ZLVu2kDExZOfOpJMT+cknZFSUjpQLlGUE49MCz549o1Kp5F9//fXWsllZWTx58iRDQ0MZGBhIIyMjWlpa0tramgYGBmzatCkXLlzI6OjoEt8HqFKpaKCnx4znz0u0H12QnJxMPz8/9u3blzkv5ZEbN24cg4ODC+zdW7VqFR0dHZmQkFCiuiIjI+ng4MC8vDweP36cVapUYceOHV9ZdvTo0Vy4cGGJ6inL7NmzhzY2NhwzZkyBD4Xp6els06aN5gSGoKAgZmVl6VCpwIeGYHxaYvbs2ezRo8c718vJyeHZs2c5b948tmzZkkZGRjQ1NWWFChVoZmbGzp07c+PGjdofFk1NJTt3Zg5AtVRKdulCPnmi3T50xK1bt1itWjVOnDjxlR8ecnJyWL9+fc6bN6/A9SVLlrBKlSpMTEwsUX21atXif//7X2ZmZmp+3y+bM5m/Ab+4xyt9iDx9+pQDBgxg5cqVeeLECc11tVrNqVOnUiKREABNTEx4/PhxHSoV+FARjE9LvIj6ijtXlJeXx8jISC5evJgtWrSgsbExjYyMKJPJ6ODgwMGDB2tnWLRePVJP7/8zWejpkU2aFK/NMsClS5doZ2fHH3744Y3lbt++TUtLS54+fbrA9e+++46urq4lOv+6bNkyzdC4v78/nZ2dC8w7vmDGjBmcPHlyiekoixw7doxOTk4cNGhQgbypv/32G01NTQmAIpGI/fv3F6I8gfdGMD4tEhoayp49e2q1TbVazatXr3Lp0qVs3rw5TUxMKJPJKJVKWb16dX799ddvHhZ9/pwMCyPnzCHPnCHVajIuLj9f4ctpnAwNyfh4reovTf73v/9RqVTyl19+KVL5PXv20MHBoUACaZKcNm0aa9asWei6tkhJSaGpqSlTUlL4zTffsF69epwyZUqhcgsWLGBISEiJaChrZGRkMCQkhLa2tty7d6/m+tWrV+nm5qYxPGtra+EgWYFiIxifFnn69CkVCgXj4uJKrA+1Ws3r169zyZIlbNKkCU1MTCgWi2lgYEBfX1/OmTOH9+/fzy986xZpaZl/jIRUmm92/frx+X//y2wDg0LGl21gwMw//igx7SXJtm3bqFQqeezYsXeqN2bMGLZp06bAfJ9arebXX39NHx8fPn78WMtK8+nWrRuXLl3Ko0eP0sPDg3Xr1i1UZvny5fzyyy9LpP+yxPnz5+nu7s4uXbpoPmw8efKEbdq0oUgkolQqpVQqZUhISIklEhcoXwjGp2VmzJjB3r17l2qft27d4vz581m/fn0aGxtTJBLR1NSUp+VyqkSigol7pVJ+YmTE9Fdk4U6TSmllbs4hQ4bw/PnzZTrJ9r9ZsmQJ7ezsGPUeK/hycnLo5+dXKGG0Wq3myJEjWa9ePa0fVUSShw8fZq1atZienk4jIyNWrFix0BFE69atY9++fbXed1khJyeH06ZNo6WlZf6KTOYP9Y8fP55SqZT6+vqUyWR0cnJiZGSkjtUKfEwIxqdlnjx5QoVCwWvXrulMQ0JCAqdPn86MV5yfpAaYNmoUuXt3/tCmkdH/f+3bxzt37nDGjBl0cnKip6cnlyxZUmJDfsXlRWTm6uparNWYCQkJtLS05J9//lmo/cGDB9Pf31/rB8KqVCo6ODjw4sWLbNSoEevVq8fNmzcXKLN161Z+9tlnWu23rHD16lXWrl2brVq14t27d0mSGzdupKmpKaVSKfX09GhgYMBJkya9cuGPgEBxEIyvBPj222/Lxif1SpVePY+3dGn+/aQkcuVKctUq8sGDAlVVKhWPHj3KHj160NTUlF26dOGhQ4fKxFmCZH600KdPH9arV08rxrxr1y46ODgUOu1epVKxT58+DAwM1Po+y2nTpnH48OGcMmUKW7RoUehvZvfu3Wzbtq1W+9Q1KpWK33//PRUKBVeuXEm1Ws2LFy+yatWqFIvFtLCwYIUKFejm5vZeEbyAQFEQjK8EePz4MeVyOa9fv65bIcuWFV7EYmpKvvTm/jZSU1O5fPly+vj40N7ent988w1v3rxZQqLfTlpaGlu2bMm2bdtqNRILCQlhu3btCg3x5ubmsmvXrmzTpo1W55gSEhIol8u5b98++vj40MbGpkDfhw8fZrNmzbTWn665desWmzRpwkaNGvHGjRtMTk5mYGCgZtGKoaEhjY2NOWPGDCHKEyhRhFydJYCZmRmGDx+OWbNm6VbI0KHArFmAUpmfkLdBg/zkvBYW79SMubk5hg4dioiICOzZswePHz9G3bp1ERgYiC1btiArK6uEHqAwDx8+RNOmTVGpUiXs3LkTRkZGWmv7u+++w4MHD7Bo0aIC16VSKX7++WfIZDJ0794dee97WONLODo6olatWkhOTkZcXBwMDQ0RHR2tuf+x5OokibVr16Ju3bpo06YNDh8+jMWLF8PGxgYRERFQKpUgCRcXF4SHh2Pq1KmQyWS6li3wMaNr5/1YeRH1lUYaLF2QmZnJrVu3MigoiHK5nMOGDePFixdLtM/4+Hi6uLhw6tSpJbbw5ubNm1QqlQwPDy90Lysri61atWL37t21NuS7efNmBgUFsUGDBvzkk084d+5czb2IiAh6e3trpR9dcf/+fbZt25be3t68fPkyV69eTRMTExoYGLBWrVq0sLBgxYoVOXv27EI5OAUESgrB+EqQb775hv3799e1jBLnxWIaR0dHent7c+nSpUxNTdVqHxEREbSxseGPP/6o1XZfxc6dO+no6PjKZ8jIyGCzZs3Yr1+/Alsg3pfMzEzNB4cuXboUGNqMiYlhtWrVit2Hrti+fTutrKw4ZcoUHjt2jI6OjhSLxaxXrx7Nzc3p6OhIHx8fXrlyRddSBcoZgvGVIKmpqbSwsGD8B7wp/F1QqVQ8fPgwu3XrRlNTU3br1o2HDx8utkH897//pVKp5I4dO7Sk9O2MGjWK7du3f2Vk+fz5czZs2JBDhgzRSuQ5bNgw9u7dm/Xr16eJiQmf/5M7NSEhgfb29sVuv7RJTU1ljx49WK1aNe7Zs4eNGjWiWCymp6cnfXx8WLlyZZqZmXHu3LlClCegEwTjK2GmTJnCAQMG6FpGqZOSksKlS5fS29ubjo6OnD59Om/fvv3O7WzatImWlpb8o5Q31mdnZ7NOnTpctGjRK+8/ffqUvr6+DAkJKbb5RURE0NHRkUZGRmzcuLEmc8mDBw+oUCiK1XZpc/DgQVaqVIlDhgzh559/TolEQltbW3bv3p1yuZzVqlWjr69vkRK6CwiUFILxlTApKSm0sLB47Zlr5YGIiAgOGzaMFhYWDAoK4tatW4uUZ3HhwoW0t7dndHR0KagsTHx8PJVKJc+ePfvK+6mpqfT29uakSZOK1Y9araaXlxfd3d05cOBAjhgxgmT+6lUjI6NitV1aPH/+nEOGDKGDgwOHDBlCQ0NDGhkZ8csvv6STkxNr165NCwsLzp8/v8xsiREovwjGVwpMnjyZgwYN0rUMnZORkcFNmzaxefPmlMvlHDFiBC9dulSonEql4pgxY+ju7s47d+7oQOn/s2PHDjo5Ob02ddnDhw/p4eHBmTNnFqufH374gR4eHhw0aJBmXi8vL48ikajMZ9A5deoUnZ2dGRQURGtra0okEvbo0YOdOnWivb09a9asyQYNGjA2NlbXUgUESArGVyo8evSo3Ed9L3Pz5k1OnTqV9vb2rF27NpcvX87Hjx8zOzubPXr0YMOGDQttJtcVI0aMYMeOHV9rQPfv32fVqlW5YMGC9+7j0aNHNDY2Zv369WlpaanZJ6mnp1fqBxQXlaysLH799ddUKpV0cXGhWCxmo0aNGBoaSoVCwaCgIFpYWPD7778XojyBMoVgfKXExIkTOXjwYF3LKHPk5eXx4MGD7NKlC01NTWltbc0GDRpoFniUBbKysli7dm0uWbLktWX+/vtvVq5cmcuWLXvvfjp16kR9fX12795dc/K6qamp1lfIaoNLly6xevXqrFSpEsViMe3t7blu3Tr6+fmxdu3a9PHxYaNGjXSauk9A4HUIxldKJCcn08LC4r0WeJQHkpKS6Onpyfr167NGjRqsUqUKZ86cqfOhzhe8mO87f/78a8vcunWLDg4OXLNmzXv1cejQIRoZGXHSpEns0KEDSdLa2lqTy7IskJuby9DQUBoZGVEqldLIyIiLFi3imDFjqFQq2aVLF8rlci5ZskQr2z0EBEoCwfhKkQkTJpSLY2belWvXrrFKlSr89ttvqVarqVaref78eX755Zc0Nzdnq1at+Msvv+j84NHt27ezcuXKbzyq6Nq1a7Szs+PPP//8zu3n5eWxQoUK7Nu3r+ZU9ipVqug+9d0/XLt2ja6urpTJZJRIJBwyZAh//fVXOjo68pNPPmGdOnXo7+9fZvQKCLwOwfhKkRdRX1mJYsoC586do7W1NVevXv3K++np6fz5558ZEBBApVLJ0aNH8/Lly6Ws8v8ZNmwYO3Xq9MYFJ1evXqW1tTW3b9/+zu1369aNlSpVYq1atfjHH3+wevXqOn1eMn+x0aRJkyiRSCgSiRgQEMDz58/z008/pYuLCwcPHky5XM6lS5cKUZ7AB4FgfKXM+PHjOWTIEF3LKBMcOHCACoWCu3fvLlL5GzducPLkybSzs2PdunW5cuVKPnnypIRVFiQzM5M+Pj5c+uKEi9cQGRlJS0tL7tmz553aj4qKIgCGhIRw8uTJrFOnzmu3U5QGUVFRtLS0JADa2dnxjz/+4A8//ECFQsGhQ4fS19eXAQEB5SZJg8DHgWB8pczDhw9pbm7Ov//+W9dSdEpYWBgtLS15+vTpd66bl5fH/fv389NPP6WpqSl79+7N48ePl9qy/+vXr1OhUPDChQtvLHfu3DkqlUoeOnTondqvUKEC+/Xrx9q1a9Pf3/+dT5XXBllZWWzXrh0BUE9PjytWrGBERATr1KnDxo0bc8yYMVQoFFy+fLkQ5Ql8cAjGpwPGjh3LYcOG6VqGTlCr1Zw7dy4dHBy0kr3j4cOH/P7771m9enU6Oztz1qxZTExM1ILSN7Nt2zZWqVLlrRHnqVOnqFAo3sm8goOD6ezszIoVK7Jp06bcv39/MdUWHbVazRUrVlAmkxEAu3XrxocPHzIkJISWlpYMDQ1l3bp12axZM50eTSUgUBwE49MBDx48oLm5eam8QZclVCoVR40axRo1amj92dVqNc+ePcvBgwfT3NycwcHB3LFjh1bPz3uZIUOGsHPnzm+NNP/3v/9RoVAUObr99ddfKZVKGRQUxDp16vDXX3/Vhty3cv78edra2hIAHR0defPmTc0Bvb179+bkyZMLHCArIPChIhifjhgzZgyHDx+uaxmlRlZWFrt06UJ/f/83rorUBs+fP+eGDRvo7+9PS0tLfvXVV7x69arW+8nMzKS3tzeXL1/+1rIHDhx463aIFzx+/JhSqZTBwcGsXLnye60QfRfu3r3L5s2bUyQSUSqVcsWKFbxz5w47dOjAatWqce3ataxTpw4DAwOZkJBQoloEBEoDwfh0RFJSEs3NzcvUHq2S4smTJ2zatCk//fTTUs9Ccu3aNU6cOJG2trasV68eV69ezadPn2q1fYVCwYiIiLeW3bVrF62srBgVFfXWsq6urrS0tKS+vj5XrlypDamFeP78OUePHk2pVEqRSMQmTZowNTWVixYtolwu5zfffMPp06dToVDwP//5jxDlCXw0CManQ0JCQjhy5EhdyyhR7t27Ry8vLw4bNkynaatyc3O5d+9eduzYkaampuzbty//+OMPrbyZb9myhc7OzkUy1F9++YU2NjZvnd8cPXo0raysaGxszHHjxhVb479RqVRcuXIlTUxMKBaLaWJiwp07d/L8+fP08fFh06ZNuXv3bvr4+LBly5ZC0gWBjw7B+HTI/fv3aW5uznv37ulaSokQGxtLJycnzpo1q0xFC0lJSZw/fz7d3NxYtWpVzpkzp9i/gy+++IJdunQp0nOGhYXRzs7ujRu9d+/eTVdXV80Gfm1x/PhxOjk5USKR0NjYmO3bt+etW7c4cuRIWllZcd26dfz222+pUCi4du3aMvV7ExDQFoLx6ZjRo0dz1KhRupahdcLDwzVvpGUVtVrNP//8kwMHDqSZmRnbtm3LnTt3Micn553bysjIoJeXlybH5ttYvXo1HRwcXpu4PDU1lcbGxpTJZFo5jDY+Pp5NmjShVCqlhYUFzc3NuXXrVu7YsYOVKlVi//79efz4cXp7e7N169blfruNwMeNYHw65t69ex9d1Ld3714qFArNgaofAmlpaVy3bh0bNWpEKysrjhs3jjExMe/URlxcHBUKBSMjI4tU/ocffmCVKlVeu8LV29ublStXplgsZlpa2jtpecHjx485aNAg6unp0cTEhI6OjgwODua5c+fYrl07uru788iRI5w+fTqVSiXXr18vRHkCHz2C8ZUBRo4cyZCQEF3L0Apr166llZUVw8PDdS3lvYmNjeX48eM1J0WsWbOGz549K1LdzZs3s2rVqkVeQDNv3jy6uroyKSmp0L1Ro0bRz8+PMpmMv//++zs9Q25uLhcuXEgjIyPq6ekxICCACoWCK1as4Pz58ymXyxkaGsqzZ8/Sy8uLbdq0KXfbawTKL4LxlQHu3r1Lc3Nz3r9/X9dS3hu1Ws3Q0FA6OTl9NAeO5uTkcPfu3Wzfvj3NzMzYv39/njp16q0R0aBBg9itW7ciR07ffvsta9SoweTk5ALXd+7cSQ8PD8pkMnbr1q3Iuvfu3UsbGxvq6emxWbNm9PX1ZePGjblz5056e3szMDCQV69e5dSpU6lUKhkWFiZEeQLlCsH4yggjRozgV199pWsZ70VeXh6HDh1KLy+vj2rI9t/cv3+fc+fOZbVq1ejm5sZ58+a99oNKRkYGa9asyVWrVhWpbbVazQkTJrBWrVoFzt67d+8RZbLPaGKymWJxFD08yBkzyIwM8vlz8rvvyPr1ya5dyQsXyOjoaPr6+lJfX5+urq6cMGECFQoFZ86cySFDhtDa2pqbNm3i+fPnWbNmTX7yyScf7e9LQOBNCMZXRkhMTKS5ufkrh7zKMpmZmezUqRObNWtW6gmjdYFareapU6fYr18/mpmZsX379tyzZw9zc3MLlIuNjaVCoeClS5eK3O6Loc1nz57x779JOzsSyCSgJkACpJ4eaWubf8/AIP+aSKSmRJJFsXgJJZJI2ts/YPXqP9DZeRg7d/4vlcpADh48mPfv3+ekSZNoaWnJTZs2CVGeQLlFRJIQKBMMHz4choaGmD9/vq6lFIknT56gffv2sLGxwYYNG6Cvr69rSaVKWloafvnlF6xduxYJCQno06cP+vfvj2rVqgEANm3ahG+//RYRERGoUKHCW9sjiSFDhuCvv/5ChQr/w4EDUrzbq1MNQPyv/6sgEuVBT08PzZs/RkJCAKpVq4oVK1bA2tr6HZ9WQODjQTC+MkRiYiI8PT0RGxsLS0tLXct5I4mJiWjdujWaN2+O77//HmKx+O2VPmJiYmKwbt06/Pzzz6hatSoGDBiAzp07Y/To0cjIyMCmTZtAiiASASLR69tRq9Xo128QwsJWA5BoTZ9IlI6xYy9g7lx/iN4kQECgHCAYXxlj2LBhMDExwdy5c3Ut5bX89ddfaN26NYYPH46xY8cKb6T/Ijc3F/v27cPatWtx6tQpdOjQAUeP3oa+/ibcvGkDQ0OgaVPg0SMgNxfo0wdQKIBbt4C6dQGVCpgxQ43wcBEA7f5c+/UD1q3TapMCAh8kgvGVMf7++294eXkhLi4OSqVS13IKcfr0aXTq1AkLFixA7969dS2nTHPv3j2sXbsZ06f3g1pthldFcCIRIJEAajUgFuf/q1ZrX4tMBoweDcybp/22BQQ+NMr3+FQZxN7eHl27dsXChQt1LaUQu3fvRseOHREWFiaYXhGwtbWFm9tYGBtb4HXDliSQl5dvdi/+LQlkMmDQoJJpW0DgQ0OI+Mogd+7cQa1atRAXFweFQqFrOQCA1atXY/r06dizZw/q1KmjazkfDD/+CIwdC2RmarPVf79k84dDjYyA4GBoFsOYmQG//gqkpwP29sDq1UBgoDY1CAh8uAjGV0b54osvIJfLMXv2bJ3qIIkZM2YgLCwMhw4dgouLi071fGjcuAHUrAlkZb1/GxJJ/pCosXF+O59++ghnzw5BcvICZGQ4QCoVYdAg4PvvAan0/+up1fmGa2T05gU1AgLlDcH4yii3b9+Gj48Prl27BrlcXiJ9ZGQAu3YBKSlAixaAq2vB+yqVCkOHDsWFCxewf/9+WFlZlYiOj53QUGDWrBc78YCcnH/fJd60iOWTT4DNm4GkJODatXwTrVQJyMrKwoQJE7B9+wH89NMqBAUFlPBTCAh8PAjGV4YZPHgwlEolZs2apfW24+KAhg2B7Oz8uSUA6NQJaNIEaNQIcHLKRIcOo5GZmY79+1cUaR+awOuJiwMOHgQsLPKHIX/6CcjKUuPs2c0QiTri0SMjFDZAYsoUEWbOfH27Bw8eRP/+/dG7d2/MnDkTenp6JfgUAgIfB4LxlWESEhJQu3ZtXL9+HRYWFlptu0EDIDwchTZIGxgAajWhUqUB0IdMpgdXVxH27AEcHLQqodyzfPlybNy4Ec7Ozti0KRSAU4H7IlEG+vS5iFWr6r4xOUBycjIGDBiAu3fvYvPmzXB9OXQXEBAogGT69OnTdS1C4NWYmZnh+vXruHLlCpo1a6a1dlUq4MsvC5sekB/9qVQikHogpcjLE+HRI+DQIWDIEGGuSBukpqZi3rx5mDBhAhITE3Hnzh1kZd0H0ALAi4hNBVNTEYyMRmPatAlQq9WoUaPGKw3Q2NgY3bp1g1qtRq9evWBmZgYfHx9hf6WAwOso3QxpAu9KfHw8LSwsmJKSorU2b9++Q5ksi/8/6/T2LyMjMi5OaxLKHVlZWdyxYwebNm1KmUxGkUhEOzs7Hj58mBYWFpRIJAS609z8AYEUAr/x4sX8hNURERHs2rUr5XI5J06c+MZ8rn/99Re9vb3Zvn37Qqc9CAgI5CPs4yvjVKlSBR06dMDixYuL3VZkZCR69eoFb28veHmdhKFh0TeNiUQvL8oQeBskcerUKQwaNAgKhQL9+vXDpUuX0KtXL1haWuLq1atYs2YNnj17BpVKBReX89i79wYAOfT1u0OligcA+Pj4YOvWrTh79iweP34MNzc3DBkyBPHx8YX6dHd3R3h4OFxcXODt7Y3Dhw+X8lMLCHwA6Np5Bd5OfHw85XJ5gSNriopKpeK+ffvYrFkz2tnZce7cuXz8+DHz8sgJE0hjY1IkIsXiN0d89vakkMy/aFy7do1Tp06lvb095XI5TUxM2KBBA+7YsYPZ2dmsW7cuf/rpJ+7fv59SqZQAKJFIeOHCBZ4/f54AaGBgwO3bt7+y/aSkJE6aNIlyuZxdunRhRETEK8sdPnyYdnZ2/Oqrr5iVlVWSjywg8EEhGN8Hwueff85p06YVuXxmZibXrFlDDw8Penl5MSwsjNnZ2YXKqdVkbi45c+brzU8mI6OitPgwHyHJyclcunQpfX19aW5uThcXF1asWJFDhw7l1atXNeXWrFnD+vXrMyUlhRUrVqSRkREBsHPnziTzhzUBUF9fn/Pnz39jn8+ePeOCBQtoa2vLoKAgHjlypNBRQ48ePWLHjh3p5eVVQIeAQHlGML4PhOvXr1Mul/Px48dvLPfo0SPOnDmT1tbWbNWq1SvfDF9Hnz6kRPKy8WXT2bnweXMC+QfObtu2je3atWPFihVZp04dOjg40N3dncuWLePTp08LlE9NTaWVlRUjIiIYHBxMmUxGADQyMmJCQgJJ8uLFiwRAqVTKYcOGFUlHVlYW165dS1dXV9auXZu//PIL8/LyNPfVajX/85//UC6Xc/ny5cI5fALlHsH4PiD69u3L6dOnv/LetWvXOGTIEJqZmbFfv36Mjo5+5/avXycrVCBFItU/B5zm/rPQwpZSqZTfffddcR/hg0elUvH48eMcMGAAzc3NWb9+fTZv3pxmZmbs1KkTjx49+lpjGTFiBL/88ktu2rSJ+vr6FIlENDY25rhx4zRlIiMjKRKJKBKJ2KZNm3fWtnPnTtarV48uLi5ctWoVMzMzNfdjY2NZu3ZttmnThg8ePHi/H4CAwEeAYHwfENeuXaNcLtecdK5Wq3ny5El26NCBCoWCkydP5v3794vVx40bpFy+i8bG1/jFF2o2atSDAGhqakoANDMz4759+7TxOB8Uf/31FydOnEgHBwfWqFGDffr0YaNGjWhlZcXJkyfzzp07b6wfFRVFS0tLRkdH09jYmIaGhgRACwuLAlH8pUuXKBaLCYDu7u7vpVWtVvPEiRMMDg6mtbU158yZo/mbyc7O5oQJE2hjY8P9+/e/V/sCAh86gvF9YPTu3ZvTp0/nL7/8Qj8/Pzo7O3PZsmV8/vy5Vtp//Pgx9fT0GBoaSpLMy8ujn58fRSIR69evr8W8qh4AACAASURBVJmTcnV1ZUxMjFb6LKskJSVx0aJF9PHxoa2tLYcMGcIRI0bQ3t6e9evX56ZNm4q0aEStVrNx48b88ccf2bBhQ812Bmtray5atKhA2aioKIrFYopEIhoaGhZ7WDIqKoo9e/akhYUFx40bx7t375Ikjx07Rnt7e44YMaJAVCggUB4QtjN8QDx//hwODg6YMWMGFixYgPHjxyMuLg7Dhg2DsbGxVvrYunUrLCws4OnpCQCQSCQ4duwY3NzcEB4eDj8/PwwbNgw3btyAu7s7WrdujcePH2ul77JARkYGNm/ejODgYLi5ueHixYvo27cvAgICsHnzZqSnp2PXrl34888/0aNHjzdmVHnBli1bkJ6ejuzsbFy6dAl5eXmQSqXQ19fH0KFDC5QViUSaL6lUikePHhXreTw9PbFx40ZEREQgKysL1atXx6BBg2Bra4uoqCgkJSWhbt26iI6OLlY/AgIfFLp2XoG3k5iYyK+//ppyuZyffvopW7ZsyZkzZ5ZIX76+vrSzsyu0AjA5OZn29vaUSCSsW7cur1y5Qi8vL0qlUkokEg4bNqzAgooPiby8PB45coR9+/almZkZW7VqxXXr1nHFihX08fFhlSpVuGDBgvdKIvDs2TPa2dlx06ZNNDY2pkwmo56eHm1sbLht27ZC5aOjo//ZzA7a2dnx3Llz2nhEDcnJyZw2bRoVCgU7derE8PBwrl+/ngqFgosXLxYWvgiUCwTjK8NERUWxT58+NDMz44gRIxgfH08yf5GCQqEotGqwuERHR9POzo56enrMyMgodP/GjRtUKpXU09Oju7s7nz17xo0bN9LExIRSqZRGRkZctmyZVjWVJJcvX+a4ceNoZ2fHWrVq8fvvv+fZs2c5btw4KhQKtm7dmvv27aNKpXrvPsaNG8fevXuzRo0alEqllMlkNDExYe3atV9pMleuXNEMhTo5Ob3SHLXB8+fPuXjxYjo4ODAgIIBr166lr68vW7VqVex5YgGBso4w1FnGIIlDhw6hRYsWaNWqFdzc3BAfH48ffvgBVapUAQC4urqiRYsWWLZsmVb7Xr9+PTp27AiFQgFDQ8NC952dnbFv3z4YGxsjISEBNWrUQKtWrXD79m307dsXADBy5EjY2dnhyJEjWtWmLe7du4cFCxbAy8sLbdq0gUQiwYEDBzBr1iwcPXoUwcHBUKvVCA8Px/79+xEcHAyx+P1eJrGxsVi/fj0sLCxw584dAIBCoYCenh4WLVr0ylyaL/qSSCTIy8tDQkLCez/rmzA2NsaoUaNw48YN9O/fH4sWLUJWVhYMDAxQq1Yt/P777yXSr4BAmUDXziuQT1ZWFtevX8+aNWuyRo0aXL9+/RsXTvz1119UKpV89uyZVvrPzs6mpaUlN27cyEaNGr2x7O+//06lUskKFSrQ0tKSiYmJJMlTp07R1dWVcrmcIpGI3t7evHbtmlb0FYe0tDRu2LCBQUFBNDc3Z//+/Xns2DE+evSICxcupIuLC729vblmzRqmp6drpU+1Ws3AwECOGjWKFSpUoEgkooWFBRUKBdu1a/faejExMZrhUDMzMw4ZMkQreoqid+/evWzcuDFtbGxoYWHBQYMGae3nISBQlhCMT8ekpKRw9uzZtLGxYVBQEA8ePFjkeZZu3bpx9uzZWtHx22+/0d/fn2vXrmXfvn3fWn7lypV0cnKiXC6nqakp4/7JYJ2dnc05c+bQ1NSUlpaWFIlEbNu2rdaHZd9Gbm4uDx48yJ49e9LU1JRt27bltm3bmJGRwcjISA4cOJBmZmbs0aMHT58+rfW5rR07dtDd3Z2Ojo6USCQ0MDCgs7MzK1SowNjY2NfWi42NpUwmo6GhISUSCVu2bKlVXUXh9OnTbN26NQ0MDKhUKnn8+PFS1yAgUJIIxqcj4uPjOWLECJqbm7NPnz68dOnSO7dx9epVKpVKpqWlFVtP27ZtuX79ek6aNIkzZswoUp1JkybRx8eHNjY2NDY2LpAz8ubNm2zVqhXt7e1pbm5OqVTK0aNHF2u+7G2o1WpevHiRISEhtLa2pq+vL5cuXcqHDx8yOzubW7ZsYcOGDWlnZ8eZM2e+8ZSD4pCenk5HR0e2adOG5ubmNDQ0ZJs2bWhlZcUvv/zyjXXj4uKop6dHIyMjWlpa0tnZuUQ0FoWrV6+ycePGFIlE9Pf312SXERD40BGMr5T5888/+emnn1Iul3PChAmaYcL3pWvXrsXOqHLv3j2amZkxLS2NXbt25caNG4tUT61Ws3fv3mzVqhUrV65MQ0NDHjt2rMD9bdu20dbWlr6+vjQwMKCJiQlXrVpVLL0vc+fOHc6ZM4ceHh50cnLilClTNFFVYmIip06dSmvr/2vvvqOiOrc+AO8zhWGGOkxhULoELCBEELgGC4q5CvbIZzexB40magJLvbHHFuuNvcXYozF4LUFjS4jGEitKFAOixhJEgwhImzm/7w/CiQMYqdLeZy0Xy1P3UZjNect+dQgODsY333xT5eXXPv30UwQFBcHGxkYYpOLo6AhLS8tXJtubN28Kia9FixaQyWTVPtLy5MmTsLOzg0QiQXh4OKv5ydR6LPG9Bnq9Hnv27EGrVq3g7OyMZcuWVcpbGlAwClCr1VboevPnz8ewYcMAAC1btsTp06dLfW5ubi46dOiAoUOHonHjxjA1NUV0dLTRMU+fPsWYMWOg1WrRtm1biMVi2NvbGyXJskpPT8eGDRsQHBwMGxsbjBw5Ej/99BMMBgN4nseJEyfQu3dvWFtbY/To0bh27Vq571UWiYmJUCqVUKlUEIlEMDMzQ2RkJGxtbUs1BeW3336DiYmJ8JZoZmZWZW+mZZGfn4+oqCiYmZnBysoK3bp1w6lTp6o7LIYpF5b4qlBmZiaWL1+ORo0awd/fH7t27aqSt43w8HAsWLCgXOfyPA8PDw/hQ8zGxqbMdRyfPn2K5s2bY8aMGXjzzTchl8uxcePGYsedPXsW3t7eCAoKQmBgIDiOg6+vL27dulWq++Tl5eHAgQPo06cPrKys0KNHD+zZs0eoPJKRkYGVK1eiWbNmaNy4Mb744ovX3rfYpUsXNGnSBNbW1jA3N8fEiROhVquh0WhKNVAkMTERJiYmkMlkePfdd6HVanHmzJnXEHnpnDp1Cs7Oznjrrbfg6OiIoKAg7N+/v0qbsBmmsrHEVwUePnyIKVOmQK1Wo0ePHvjpp5+qtLkqLi4OWq22XGXLTp06BQ8PD/A8j7S0NJibm5cr1nv37sHR0REbN25EYGAgFApFick4Pz8fixYtgkqlQkREBNzd3cFxHHr27FniWyvP8zh37hzGjh0LrVaLVq1aYdWqVXj8+LFwzPXr14X+0p49e/5joeiqdPDgQWi1Wmi1WohEInh4eGDGjBlQq9XYsGFDqa6RlJQEExMTSKVSfPjhh9DpdNi5c2cVR1426enpGDx4MNzd3TF79mx4e3ujWbNm+Oqrr5CXl1fd4THMK7HEV4muXbuGIUOGCM1rr3Mof+/evV+5fltJhg0bJvQRXrhwAc2bNy93DFevXoVWq0VMTAyCg4Nhbm6OqKioEpPQnTt30K1bN3h4eGDmzJlQq9WQSqWIjIyEwWBAcnIyZs2aBQ8PD7i5uWH69OlITEwUzs/Pz0d0dDRCQkKg1WpLVSi6KuXk5MDJyQnm5ubgOA7m5uY4duwYLCws4O7uXuqqNsnJycLKDVOmTIG1tXWNXRVjx44d0Gg0mDt3Lg4ePIjg4GA4Ojpi6dKllVY7lmGqAkt8FcTzPI4cOYJOnToJ/TipqamvPY64uDjY2tqWad5VZmYmrK2t8eDBAwDArl270LNnzwrFcfz4cWg0Gpw/fx6dO3eGlZUVhg0b9tIP/ujoaNjb2+Pdd9/FhAkThBJo5ubmGD16NE6fPm2UOB89eoQ5c+bAwcEBgYGB2Lp1a41YXXzWrFlQKpWwsLCApaUlpk2bhgkTJsDKygoxMTGlvs7t27dhYmICkUiE6dOnQywWY8SIEVUYecXcvn0brVu3Rrt27XD37l2cPXsWvXr1glqtxtSpU6vlZ4FhXoUlvnLKzc3F5s2b4e3tjSZNmmD9+vXVXuW+V69eWLRoUamP37RpE7p06SL8fd68efj4448rHMe2bdvg4OCAW7duoVevXkITZEkJqnCagZubGziOQ4sWLRAcHAyRSAQnJyehmfjMmTMYNGgQrKysMHToUKOpE9Xt7t27kMvl0Gg0EIvFaNasGRITE2FmZobWrVuXqdn1zp07Qh/fxIkTodFo0Lp16yqMvuL0ej0+++wzaDQa7N69G0DBfMThw4dDqVRi7NixbCoEU6OwxFdGaWlpmD9/Pho2bIj27dtXuJZjZbp8+TJ0Ol2p3/ratGmDb7/9Vvj7iBEjsHLlykqJZf78+fDy8sKTJ08wYMAAqNVqBAcHIyMjAzzP4+eff0ZERATUajXatGmDtWvX4ocffoCfnx/atm2Ln376CW+99ZawQrm9vT0+//xzo769miIkJAQymUwYxXnlyhUMHDgQCoUCly5dKtO17t69C6lUCktLSwwbNgwBAQGwt7evosgr19mzZ+Hm5oYhQ4YIFYXu37+PyMhI2NjYYMCAAbhy5Uo1R8kwLPGVWnJyMj766CMolUoMGDAAFy9erO6QStSzZ08sXrz4lcfdvHkTWq0Wubm5wrb27dvj8OHDlRIHz/MYM2YM2rdvj+zsbAwfPhwqlQp2dnZwdnaGh4cHZs+ejeTkZKPz9Ho9pk2bBrlcDrlcjsDAQNjZ2UEkEiE8PLzGldD67rvvIJFIIJPJYGlpiVmzZiE+Ph4KhQJ9+/Yt8/V+//13SKVSaLVavPPOOxgyZAikUmm1z+UrrYyMDAwdOhSNGjUyGo2alpaGuXPnQqfToXPnzvjxxx9rzTMxdQ9LfK9w7tw59OnTBzY2Nvj444+rdQBFaVy6dAl2dnYlrq7wosmTJ2P8+PFG25ydnfHbb79VWix6vR6dO3eGv78/AgICIJfLoVAo4OjoWOzf0WAw4NChQ+jSpQtUKhVGjBiBf//733Bzc8ORI0ewc+dOKJVKmJiYYMqUKTXiLTsvLw82NjawsrKCVCpF8+bNkZ+fj06dOkEul5fre+XevXuQSqVwdHREx44dsWDBAsjl8lq3YsLu3buh1Woxe/Zso/7d7OxsrFmzBm5ubggMDER0dHSN+L9k6heW+EpgMBiwd+9etG7dGo6Ojli8ePFrnw9WEd27d8fSpUtful+v16Nhw4a4evWqsC03NxcmJiaVMhw9Ozsb33zzDbp37w5LS0uoVCr83//9H3Jzc/HJJ5/Azs4O9vb2SEhIQFpaGpYsWSIUil63bp3RW92BAwfg5OSE/v374+HDh5g+fbpQwHnLli0VjrUiRo0aJbztmZmZIT4+HqdPn4ZCocDEiRPLdc379+9DKpWicePGCAwMxLfffgtra2v8/PPPlRx91fv9998RHByM1q1bF+vj0+v12L17N/z8/NC4cWNs2LDBqPWBYaoSS3wvyMrKwqpVq+Du7g5fX1/s2LGjystbVYULFy7841tfTEwMWrZsabTtt99+g4uLS7nvaTAYEBsbixEjRsDGxgbt27fHl19+ifT0dDx69Ahubm5YtWoVeJ7Hp59+CrVaLZQw69evH06ePPnSpq/MzExERkZCo9FgzZo1yMrKQv/+/SESieDi4lItE7x//fVXcBwHExMTWFhYYP78+eB5Hn5+frCwsMDTp0/Ldd0HDx5AIpGgRYsW8PLyQlxcHCwsLLB9+/ZKfoLXQ6/XY/78+dBoNNixY0ex/TzP49ixY+jYsSMaNmyIhQsXVtqKIwzzMizxAUhJScHUqVOh0WjQtWtX/PDDD7W+/6Fr165YtmxZifvCw8OxatUqo22HDh1Chw4dynyfGzduYMqUKXB2doanpyfmz5+P33//vdhxiYmJ0Ol0+PjjjxEUFARLS0uYm5vD2tq61KXLrly5gsDAQLRq1QpxcXG4d+8eWrVqBY7j0KpVqwrXPS0tnufh5OQEmUwGmUwGHx8f6PV6xMTEwMzMDEuWLCn3tf/44w9IJBK0bdsWrq6uyMzMhFgsxuzZsyvxCV6/8+fPw93dHYMGDXpp68mFCxfQp08fqFQqTJ48uUaUamPqpnqd+K5fv44RI0bA2toaI0eOxPXr16s7pEpz/vx5NGjQoNgUi9TUVFhZWSEtLc1o+4oVKzBy5MhSXTslJQX//e9/0bJlS+h0OkyYMAGXLl166S8L9+/fx9SpU6FSqSCVSjF37lzk5eVh8eLFwtpve/fuLdW9DQYDVq1aBbVajaioKGRlZeHMmTNwcXGBSCRC//79q3xaydSpU8FxHKytrSGXy5GQkACDwQBXV1fodLoKNdmlpKRALBYjLCwMdnZ2AAArKyv079+/ssKvNpmZmRg1ahRcXFz+sc5nYmIiIiIioFQq8f777xsVLmCYylDvEl9hAeOwsDBotVpMnz4djx49qu6wqkSXLl3wxRdfGG1btmxZiR+iEydO/McKIc+fP8fOnTsRFhYGKysrDBw4EIcPH37pxHSe5/Hjjz8iPDwcSqUSERERuHbtGvbt2wedTicMolmxYgVsbW2hVqvx5ZdflvrZHj58iH79+sHZ2RkHDx4EAGzZsgXW1tYwMTHBjBkzquStPTk5GSKRCKampjAzM8PChQuFeysUCmEeW3k9evQIYrEY/fr1g1KpBAB4enrCz8+vwrHXFNHR0bC1tcX06dP/sSvhjz/+wOTJk6FSqdCnT58aO5KaqX3qTeLLy8vDtm3b0KJFC7i7u2PNmjWvHPlY2/3yyy9o2LCh0RuQt7c3jh49+vdBN24AgwYhycoKN0NDgb+quAAFb1fHjx/HkCFDoFQq8fbbb2PLli3/uBJERkYGVq1aBU9Pz5cWil69ejXc3NyEXzg2bNgArVYLOzu7Mk3AB4DDhw/D1dUV4eHhuH//PgwGA6ZMmQKpVAobG5tKrXNpMBjQqFEjiEQiyGQytGjRAgaDAbm5udBoNGjWrFmFk21qairEYjEiIiJgZmYGoKAwgVarrYxHqDHu37+Pjh074l//+heSkpL+8dhnz55h4cKFaNiwITp27IijR4/W+q4IpnrV+cSXnp6OhQsXwsHBAW3btsW+ffvq1fDp0NBQLF++HABw8eJFODk5/f38164B5uaAWAwQwSCRABoNrsfGIioqCvb29vDx8cGiRYuEsmYvc+PGDYwbN06o0vKqD6fJkycjMDBQGMG5detWaDQaODs7Y9KkSWX6YHv+/LlQFHz58uXQ6/XIyspCeHg4RCIR3NzccP78+VJf72VmzpwJjuNgZ2cHuVwuNMEtXrwYpqamlbJMz+PHjyEWizFp0iRIpVIAwIwZMyAWi+vc963BYMCiRYugVquxefPmV/6f5+TkYMOGDfDw8ICfnx92795d6hqoDPOiOpv47t69i4kTJ8LGxgZ9+/bFL7/8Ut0hVYtzsbGYbWkJg7c3btjbY8uAAX/v7NUL4DiASPiTzXH43NISUVFRRtMdSqLX67F371507NgRWq0WkydPxp07d0oVV+Eitt27dxc+vHbv3g2NRoMmTZpgxIgRZf5QK1wxvGXLlkLFlNu3byMgIEBYRby88+GuXr0KiUQCCwsLyOVyYbpIRkYGLCwsyjUwqCRPnjyBSCTC559/Do7jwPM8du3aBZlMhvv371fKPWqaS5cuoUmTJujXr1+xvueSGAwGREdHIzAwEG+88QbWrFlT7eUCmdqlziW+CxcuoH///lAqlRg/fnz9rhHI80BQELJFIiGxGeRyYN06ZGZmIkunM0p6wjFhYf942UePHmHu3LlwcnJCQEAAtmzZUq5C0YWL2I4ePVr4bX/fvn1Qq9Xw9fVF7969y3xdg8EgNJ1OmDBBaJaNjY2Fk5MTRCIR3n333TINQMnJyYGjo6NQPs3X11d4+4qKioKJiUmlrcTx559/QiQSYd26deA4Djk5Obh48SLkcnmdXvg1KysLY8aMgZOTE2JjY0t1TmE/cmhoKOzs7DB37txyTyNh6pc6kfgMBgMOHDiAdu3awd7eHgsWLCjVb4513k8/AWZmxRLbU4kE5goFjqrVMBR544OpKfDZZyVe7uzZsxg8eDCsra0xZMiQSmk+fPr0Kby8vDB//nxh26FDh4QaniEhIeVaXf7Ro0cYPHgwHB0djUaMbty4EZaWlpDJZJgzZ06pmlTHjx8PkUgEBwcHmJqaCr9MpaamwtTUFIMGDSpzfC+TlpYGkUiEHTt2QCwW48mTJ0hPT4dYLK72Cfuvw/79+6HT6fCf//ynTMUUrly5ggEDBsDGxgaRkZGvbJpn6rdanfiys7Oxdu1aNG7cGD4+Pti6dStbCPNFa9YACkWxxKcXi5Fx/37BwBYLi7/7+IgAmQzYs0e4RHZ2NjZt2gQ/Pz+4uLhgwYIFlV4o+vfff4eDgwO2bdsmbDt+/DjUarVQ8qy89zx+/Djc3d3Ro0cPoYSYwWDAJ598AqlUCrVabVSou6jY2FhIpVKYm5tDKpUajZJ97733IJPJKnXpnfT0dIhEIuzfvx9isViIWaFQIDIystLuU5M9fPgQnTp1gr+/f5lL6CUnJwuLEg8fPhwJCQlVFCVTm9XKxJeamooZM2bA1tYWoaGh1bbido137lyJb3wpIhE8mzXDzJkzkXT0KODuDv7FNz+FAo9nzEBUVBQ0Gg06deqE/fv3V+lAgqtXr0Kj0eD48ePCtpMnT0Kj0aB3795o0qRJiRPjSyMnJwfTp0+HSqXC4sWLhSH0GRkZ6NmzJziOg4eHR7GVFNLT06FWq4VVF/z8/ITvs8J18yZPnlzOJy7Zs2fPwHEcTpw4AbFYLMwtdXV1RdgrmqDrEp7nsWzZMqjVamzcuLHMP9+pqamYNm0a1Go1evXqhbNnz1ZRpExtVKsSX0JCAt5//31YW1tj2LBhiI+Pr+6QajaeB7p0+futj+MAhQKGnTtx8uRJjBs3Dm9pNMgu2txJhGdE+Hjs2Ne6ivzx48eh1WqNBtWcPXsWWq0WgwYNgpOTE27cuFHu6yckJKB9+/Z48803ce7cOWH7rVu34OvrC47j0L59e2GaRXh4OMRiMRo0aABTU1OjxBsaGgpLS8tKnxKTkZEBjuNw/vx54SsAdOjQAU2bNq3Ue9UGcXFx8PT0RHh4OJ48eVLm8zMzM7F06VI4OjoiODgYhw4dYr8kMzU/8fE8j9jYWHTr1g0ajQb/+c9/al2l+mqVnw9s3AiEhAB9+wKnTxvtNuzahfwSmkN5hQIosmTQ67Bt2zY4OjoalR+7dOkSdDodRo4cCZ1OV6G+RZ7nsXnzZtja2uKDDz4wmmN4/Phx2NvbQywWo3379kIBaolEghUrVgjHXb16FVKptFjZt8qQmZkJjuNw7do1EJEw57Jwqkh9lJ2djQ8//BAODg6lLm9XVF5eHjZv3gxPT0/4+PjU2jq8TOWosYkvPz8fX3/9NVq2bAk3NzesXLmyxq3FVif8+isglxcf3WlmBlTTEPF58+ahefPmRiP0rl27hgYNGuCDDz6ARqMp9wdgoSdPnmD48OFo2LAhdu/ebfQWsGDBAhARiAgSiQS+vr5G+/38/NCgQYMqmVeXlZUFjuPw66+/QiqVCoWdN23aBJFIVK/nrcXExMDOzg5RUVHlLgvH8zwOHDiA1q1bw8XFBStWrCh4a8/KAkaNKvi+NzUFBgwA2AC5OqvGJb5nz55hyZIlcHZ2RlBQEKKjo+v1D/tr0auX8SAYhQJYsKDawuF5HqNHj0aHDh2MPuASEhLg4OCAjz76CBqNBtHR0RW+V2xsLJo2bYrQ0FAkJyeD53kEBgZCKpXCxMQERAS1Wo39+/cDAE6cOAGJRCKUSatsz58/B8dxuH79OhQKhVB84Ny5c5BIJOXu56wrUlJS0KVLF/j6+lao2RsATp06hW7dusHW1hY3PDzAy2R//wyYmAABAZUUNVPT1JjEd+/ePURGRkKlUiE8PLxalpqpt/LzgRUrAH9/oH174H//q+6IoNfr0b17dwwaNMjobSspKQnOzs6YMGECdDodNm7cWOF75ebmYs6cOVCpVOjSpQskEgksLS0hEomwaNEidOnSBRzHoWnTpmjYsCE8PT0rfM+Xyc7OBhEhISEBSqUSM2bMAPD3/L7SznGry3iex8qVK6FWq7F27doK99kl/PAD8l6Y62r0C+C1a5UUNVOTvJ7E9+wZcPAgcPIkUKR56PLlyxg0aBCUSiXGjRuHW7duvZaQmJovKysLAQEBmDJlitH2O3fuoFGjRhg/fjycnJyEQtEVdeTIEXAcB5lMBqlUajSK8+bNm3BycgIRITAwsFwDLUojJycHRISbN29Cp9Pho48+EvaZmJgY9TXWd/Hx8fD29kaPHj0qNqUkLq5gWk/RxGdpCbxY15apM0RU1fbsIdLpiPr1I+rcmcjNjXD3Lh06dIg6duxIoaGh1LRpU0pKSqJly5aRi4tLlYfE1A4KhYL2799PX3/9Na1Zs0bY7ujoSLGxsRQTE0M9e/ak9evX06RJkwhAue+l1+tp1KhRpFKpyGAwUH5+PjVt2pSePn1KRESurq70/Plz8vHxoTt37pBWq6WIiAjS6/UVfs4XcRwnfJXL5cL9iYhUKhVdvny5Uu9XmzVt2pTOnj1LjRo1Ih8fHzp69Gj5LtS4MZFEUnx7fj6Rv3/FgmRqpipNq6mpxQZOGEQinFco4OXlha+++qpCa5cx9UPhIraF/WyFUlJS4OXlhbFjx8LPzw/Dhw8vd3/wBx98AKlUiubNm4PjOCxcuBARERGws7PD9u3bsXjxYqM+ti+++AJmZmZQKBRC3c7KkJ+fDyJCYmIiPD090atXL2Gfv78/2rRpU2n3qkuOHDmChg0bYuLEicg9dgzw9gYkEsDFBdi169UXOHCg4LPK1LSgf08uB776quoDro/xggAACjFJREFUZ6pF1Sa+zZsLqv8XrQUpFoMvRxkqpv46e/Ys1Gp1sYnIjx8/RosWLfD++++jQ4cOeOedd8pc3/PkyZOQSCQYOHAgJBIJ3nzzTWHf6dOn4eXlBZFIhB49ehidl5+fjzFjxkAsFkOn0yEmJqb8D/gXvV4PIkJSUhICAgIQEhIi7BswYACcnZ0rfI+6KjU1FaNDQvC86LxUhQI4fPjVF7h/H1i2DFi0CGBdLnVa1TZ1ymREfzXdvEgkFhNXUtMCw7yEv78/bdiwgbp3705JSUnCdpVKRceOHaPLly8LzeRhYWGUkZFRqus+f/6cunbtSt7e3rRv3z4iIjp8+LCwPzAwkDp27Egcx1FsbCx99tlnlJeXR0REEomEli9fTqmpqeTj40OhoaHk7e1NN27cKPdzvtjUaWFhQZmZmcI+b29vevz4cbmvXdep1Wpa7uFBsqKfOc+fE3322asv0KAB0bhxRBMmELEulzqtahNfWBiRqMgtTEyIunYlMjWt0lszdU+3bt1o2rRp1LlzZ6MEYG1tTd9//z0lJCSQmZkZubi4UIcOHUqVJPr27Uu5ubnk7e1Nz549o2XLlpFGoxH2P3v2jJYvX05RUVF08eJFOnPmDPn4+FBsbKxwjFKppJiYGIqPjyee56lp06YUFhZm1D9XWi8mPktLS6PE5+/vT1lZWWQwGMp83fqCu3ePRDxffMeDB68/GKbmqvJ3ynPnAEfHguYGmQzo3BkosiI3w5TFpEmTjBaxLZSZmYmQkBD07dsXkZGRaNy4sVDkuSTbt2+HSCTC2rVrIRKJ0KxZs2LHDBgwABYWFkLxc57nsWfPHtjb22Po0KElFs8+ePAgbG1tIZFIMHbs2DJXCCEiJCcnY8iQIWjUqJGw/dGjR+A47h+fqd7buLF4fVoTE2DChOqOjKlBXs90Bp4HkpKAv2ogMkxF8DyPgQMHokePHsUGs2RnZyM0NBQ9e/bEvHnzXlrf89GjR5DJZOjfvz/c3NwgEomEGp2FHjx4ALFYjNWrVxc7Pz09HePGjYOtrS02bdpUbC4Zz/NYvHgx5HI5zMzMyjQNoTDxjR8/Hg0aNDC6plgsLjbIh3lBXh7Qpk3B2AKOK/jaqBFQRdNPmNqpxkxgZ5iyKFzEdsyYMcWSTk5ODnr06IGwsDCsXbu2WH1PnufRvHlz6HQ6rF69GkSExYsXF7tHhw4dYGdn948TpM+fP48WLVqgXbt2JSbYvLw8jBgxQih2feTIkVc+W2HimzlzJlQqldE+pVKJadOmvfIa9ZrBAMTEALNmAV9/DbCR40wRLPExtVbhIrYLSiivlpeXhz59+iAkJAQ7d+40WvJo1qxZEItEuHjxIiQSCVxcXIqdHx8fD47j8N13370yjvz8fCxduhQqlQrTpk1Ddgk1TlNTUxESEgKO4+Dj44PExMSXXq8w8a1cuRLm5uZG+5o0aYLevXu/MiaGYV6OJT6mVitcxHb79u3F9un1egwePBht2rTBwYMHodFosH/YMCT/1ffz0NQU3Ynwxx9/FDvX29u7xD6/V8XSq1cvvPHGGzh27FiJx8TFxaFJkybgOA7dunUzWh2iUGHi2717N2QymdG+0NBQo+kWDMOUHUt8TK0XFxf30hUbDAYDRo4cicDAQPzy3/8is8ic0lyJBDh1yuic48ePg+M4XL58uVzx7Nu3D46Ojhg4cCBSUlJKPGbv3r3QaDSQSCSYMGGCUV8lEeHWrVv44YcfIJFIjM6bOHEidDpdueJiGKZA1ZcsY5gq5uXlRTt37qQ+ffpQfHy80T6RSESrV68mf39/ejhxIsmLnGtiMBAtXGi0bciQIdSmTRvy9vYuVzxdu3al+Ph40ul05OnpSevXrye+yBD77t27U0pKCs2ePZtWrlxJSqWS1q9fb3SMVqMhC72e6IWyaL6+vpSWllauuBiG+Ut1Z16GqSwlLWJb6NChQzhVtAhx4Z/AQOG4rVu3QiQS4cGDB5US0+XLlxEQEICgoCBce0ml/9zcXLz33nsQiURwcHAAEeHhunUw6HTIIQJvaQn8VYj79u3bICK2iCrDVABLfEydUriIbXp6ekEJqgsXkPXkCczMzLDa3R3Pi640L5cLSYXneSiVSvTp06dSY9Lr9cIyOpMmTXrpgsopKSlo164dmhGVXHZr+3bwPA+O43D16tVKjZFh6hPW1MnUKZGRkdTuX/+iy25uBFdXouBgIo2G+hgMNPvJE3rm7Ey5JiaUR0TPxWJ6oFYTjR5NRETz5s2jzMxM2rBhQ6XGJBaLKSIiguLi4ujWrVvk5eVFhw4dKnacVqulEydO0PtEJC260sTz50Sffy6s2nDmzJlKjZFh6hOW+Jg6heM4WmJhQQFPnhCXm0v07BkpeJ6W5+TQwVWryDYpiWT799Ppbt1opLU1zUtNpfymTQkqFTl/+ilNGTaMzMzMqiQ2Ozs72rlzJ61YsYLGjBlDffv2pYcPHxY7TklEJVWyRXo6UXQ0nc7Pp17jxxONGUP05EmVxMowdRlLfEydI/rqK5IVGUwiE4vJ/cKFgtqxb79Nbf/3P3o/OJjm5eSQ9PZt4v78k3obDDT1u++I/ipCXVU6depEV69eJVdXV/L29qZVq1YVDH7R64m+/ZZkRMSbmBidk01E39+9S/p+/ah5fj7ZZGYSrVtHFBBQ5fEyTF3DARVYvZNhaiK1utibkJ6IFkmltMDCglxdXcnV1ZW+3LePFDk5xudaWBB9+SXRO++8llDj4+Np1KhRJMrLo5jnz8nszh2izEzScxyJAeIUCiKRiHh3d3oaH082ubnVGi/D1AXsjY+pewYOLLb6h0Qup8gLF+jGjRu0fPly6tGtG8mLJj0iotxcort3X1OgRM2aNaPY2Fia4+5OXHw80V+rMUgAgkRCfzZpQvtGjaJP2rYl66JJj4goO5vohWWaGIZ5NZb4mLpnzhyi1q2J5HIiK6uCr0uWEOflRRqNhgICAqjfgAHEeXkVP1ciIQoKeq3hikQiCkpLI0XR7Xo9ZcTH0//S0shWp6MMR8fiJ8tkBc2dDMOUGlsNlql7FAqi778nSkwkun+f6M03iSwtix+3di1Rx45E+fkFb3pmZkQ9exK1bPn6Y3Z2JhKLiYqstecUFPT3KNNWrYg6dSro08vPL4i3XTuiNm1ee7gMU5uxPj6mfrt3j2jTJqI//ihYIPntt4mKruD9Oty8SdSiBVFW1t/b5HKimBiitm3/3paURLR6dcHCqt27F/TticWvP16GqcVY4mOYmuLMGaIPPyS6epXIxYVowQKisLDqjoph6hyW+BiGYZh6hQ1uYRiGYeoVlvgYhmGYeoUlPoZhGKZeYYmPYRiGqVdY4mMYhmHqFZb4GIZhmHqFJT6GYRimXmGJj2EYhqlXWOJjGIZh6hWW+BiGYZh6hSU+hmEYpl5hiY9hGIapV1jiYxiGYeoVlvgYhmGYeoUlPoZhGKZeYYmPYRiGqVf+H1AxiWSR8xh1AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JGgCXrK3OxN5",
        "outputId": "5ea9a6e6-554f-48f8-f606-58b6d4b62626"
      },
      "source": [
        "print('AVG graph')\n",
        "calculate_statistics(graph_avg)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "AVG graph\n",
            "Nodes: 80\n",
            "Edges: 177\n",
            "Number of connected components: 1\n",
            "Size of each component (number of nodes): [80]\n",
            "Nodes in largest connected component: 80\n",
            "Edges in largest connected component: 177\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": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F7sHgAPNO1Vf",
        "outputId": "5b623e8d-f7f7-4dd2-9cec-abc3118dbd50"
      },
      "source": [
        "print('AVG graph')\n",
        "nx.draw(graph_avg, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "AVG graph\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zLq3Rem0O2r9",
        "outputId": "500da598-b22d-4cff-ce35-9782b15c7418"
      },
      "source": [
        "print('AVG graph')\n",
        "nx.draw_spring(graph_avg, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "AVG graph\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "r1x9xV8yO4TN",
        "outputId": "b734e846-9f4f-4a3f-d7cf-a4470b4cddd5"
      },
      "source": [
        "print('AVG graph')\n",
        "nx.draw_kamada_kawai(graph_avg, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "AVG graph\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "P3bwA1lUO7YH",
        "outputId": "8ff8aed2-c464-4bdc-8073-48904aa5018c"
      },
      "source": [
        "print('MAX graph')\n",
        "calculate_statistics(graph_max)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MAX graph\n",
            "Nodes: 80\n",
            "Edges: 177\n",
            "Number of connected components: 2\n",
            "Size of each component (number of nodes): [78, 2]\n",
            "Nodes in largest connected component: 78\n",
            "Edges in largest connected component: 176\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": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SzWWkmUcO79j",
        "outputId": "cea6913c-e9aa-48e0-81d0-95701bbb264b"
      },
      "source": [
        "print('MAX graph')\n",
        "nx.draw(graph_max, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MAX graph\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sRz4lsZoO8Wk",
        "outputId": "e09efe67-cc36-4d5c-950b-84d796166eea"
      },
      "source": [
        "print('MAX graph')\n",
        "nx.draw_spring(graph_max, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MAX graph\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pe2vGK1RO8sG",
        "outputId": "a81a5177-f3b9-411f-9afb-259f18b43247"
      },
      "source": [
        "print('MAX graph')\n",
        "nx.draw_kamada_kawai(graph_max, node_color=color_map, node_size=node_sizes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MAX graph\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BcTKFGbuPVQJ"
      },
      "source": [
        "---\n",
        "### Initialize skf and generate test edges and target labels"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jBQkVqgDlcUK"
      },
      "source": [
        "skf = StratifiedKFold(n_splits=10, shuffle=True, random_state=0)\n",
        "\n",
        "test_edges_SNF = np.array(all_positive_pairs_edges + chosen_negative_pairs_edges_SNF)\n",
        "target_label_SNF = np.array([1] * len(all_positive_pairs_edges) + [0] * len(chosen_negative_pairs_edges_SNF))\n",
        "\n",
        "test_edges_avg = np.array(all_positive_pairs_edges + chosen_negative_pairs_edges_avg)\n",
        "target_label_avg = np.array([1] * len(all_positive_pairs_edges) + [0] * len(chosen_negative_pairs_edges_avg))\n",
        "\n",
        "test_edges_max = np.array(all_positive_pairs_edges + chosen_negative_pairs_edges_max)\n",
        "target_label_max = np.array([1] * len(all_positive_pairs_edges) + [0] * len(chosen_negative_pairs_edges_max))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZTqYrNPzFbs-"
      },
      "source": [
        "---\n",
        "### Generate embeddings and similarities, and validate generated files"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PtbiywwFxn06"
      },
      "source": [
        "for (embedding, filename) in zip(models, filenames):\n",
        "  count = 1\n",
        "  for train_index, test_index in skf.split(test_edges_SNF, target_label_SNF):\n",
        "    test_fold_edges, test_fold_labels = test_edges_SNF[test_index], target_label_SNF[test_index]\n",
        "    graph_copy = graph_SNF.copy()\n",
        "    with open(f\"embeddings/nr/{filename}/SNF/folds/fold{count}.csv\", \"w\") as file:\n",
        "      for edge, label in zip(test_fold_edges, test_fold_labels):\n",
        "        file.write(f\"{edge[0]},{edge[1]},{label}\\n\")\n",
        "        if graph_copy.has_edge(edge[0], edge[1]):\n",
        "          graph_copy.remove_edge(edge[0], edge[1])\n",
        "      file.close()\n",
        "    graph_copy = graph_copy.to_directed()\n",
        "    # Learn embedding - accepts a networkx graph or file with edge list\n",
        "    embs, t = embedding.learn_embedding(graph=graph_copy, edge_f=None, is_weighted=True, no_python=True)\n",
        "    save_embeddings(f\"embeddings/nr/{filename}/SNF/fold{count}.emb\", embs, list(graph_copy.nodes()))\n",
        "    count += 1\n",
        "    # Evaluate on graph reconstruction\n",
        "    MAP, prec_curv, err, err_baseline = gr.evaluateStaticGraphReconstruction(graph_copy, embedding, embs, None)\n",
        "    #---------------------------------------------------------------------------------\n",
        "    print((\"\\tMAP: {} \\t precision curve: {}\\n\\n\\n\\n\"+'-'*100).format(MAP, prec_curv[:5]))\n",
        "\n",
        "for (embedding, filename) in zip(models, filenames):\n",
        "  count = 1\n",
        "  for train_index, test_index in skf.split(test_edges_avg, target_label_avg):\n",
        "    test_fold_edges, test_fold_labels = test_edges_avg[test_index], target_label_avg[test_index]\n",
        "    graph_copy = graph_avg.copy()\n",
        "    with open(f\"embeddings/nr/{filename}/AVG/folds/fold{count}.csv\", \"w\") as file:\n",
        "      for edge, label in zip(test_fold_edges, test_fold_labels):\n",
        "        file.write(f\"{edge[0]},{edge[1]},{label}\\n\")\n",
        "        if graph_copy.has_edge(edge[0], edge[1]):\n",
        "          graph_copy.remove_edge(edge[0], edge[1])\n",
        "      file.close()\n",
        "    graph_copy = graph_copy.to_directed()\n",
        "    # Learn embedding - accepts a networkx graph or file with edge list\n",
        "    embs, t = embedding.learn_embedding(graph=graph_copy, edge_f=None, is_weighted=True, no_python=True)\n",
        "    save_embeddings(f\"embeddings/nr/{filename}/AVG/fold{count}.emb\", embs, list(graph_copy.nodes()))\n",
        "    count += 1\n",
        "    # Evaluate on graph reconstruction\n",
        "    MAP, prec_curv, err, err_baseline = gr.evaluateStaticGraphReconstruction(graph_copy, embedding, embs, None)\n",
        "    #---------------------------------------------------------------------------------\n",
        "    print((\"\\tMAP: {} \\t precision curve: {}\\n\\n\\n\\n\"+'-'*100).format(MAP, prec_curv[:5]))\n",
        "\n",
        "for (embedding, filename) in zip(models, filenames):\n",
        "  count = 1\n",
        "  for train_index, test_index in skf.split(test_edges_max, target_label_max):\n",
        "    test_fold_edges, test_fold_labels = test_edges_max[test_index], target_label_max[test_index]\n",
        "    graph_copy = graph_max.copy()\n",
        "    with open(f\"embeddings/nr/{filename}/MAX/folds/fold{count}.csv\", \"w\") as file:\n",
        "      for edge, label in zip(test_fold_edges, test_fold_labels):\n",
        "        file.write(f\"{edge[0]},{edge[1]},{label}\\n\")\n",
        "        if graph_copy.has_edge(edge[0], edge[1]):\n",
        "          graph_copy.remove_edge(edge[0], edge[1])\n",
        "      file.close()\n",
        "    graph_copy = graph_copy.to_directed()\n",
        "    # Learn embedding - accepts a networkx graph or file with edge list\n",
        "    embs, t = embedding.learn_embedding(graph=graph_copy, edge_f=None, is_weighted=True, no_python=True)\n",
        "    save_embeddings(f\"embeddings/nr/{filename}/MAX/fold{count}.emb\", embs, list(graph_copy.nodes()))\n",
        "    count += 1\n",
        "    # Evaluate on graph reconstruction\n",
        "    MAP, prec_curv, err, err_baseline = gr.evaluateStaticGraphReconstruction(graph_copy, embedding, embs, None)\n",
        "    #---------------------------------------------------------------------------------\n",
        "    print((\"\\tMAP: {} \\t precision curve: {}\\n\\n\\n\\n\"+'-'*100).format(MAP, prec_curv[:5]))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bHN63t_91CP3"
      },
      "source": [
        "# cosine similarity\n",
        "for filename in filenames:\n",
        "  for count in range(10):\n",
        "    per_fold_dd_scores = np.empty(shape=(len(all_drugs), len(all_drugs)))\n",
        "    per_fold_tt_scores = np.empty(shape=(len(all_targets), len(all_targets)))\n",
        "    per_fold_dt_scores = np.empty(shape=(len(all_drugs), len(all_targets)))\n",
        "    embeddings = read_embeddings(f\"embeddings/nr/{filename}/SNF/fold{count + 1}.emb\")\n",
        "    for i in range(graph_SNF.number_of_nodes() - 1):\n",
        "      for j in range(i, graph_SNF.number_of_nodes()):\n",
        "        score = cosine_similarity(embeddings[i], embeddings[j])\n",
        "        if i < len(all_drugs) and j < len(all_drugs):\n",
        "          per_fold_dd_scores[i, j] = score\n",
        "          per_fold_dd_scores[j, i] = score\n",
        "        elif i >= len(all_drugs) and j >= len(all_drugs):\n",
        "          per_fold_tt_scores[i - len(all_drugs), j - len(all_drugs)] = score\n",
        "          per_fold_tt_scores[j - len(all_drugs), i - len(all_drugs)] = score\n",
        "        else:\n",
        "          per_fold_dt_scores[i, j - len(all_drugs)] = score\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/SNF/cosine_sim/DDsim_fold{count + 1}.csv\", per_fold_dd_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/SNF/cosine_sim/TTsim_fold{count + 1}.csv\", per_fold_tt_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/SNF/cosine_sim/DTsim_fold{count + 1}.csv\", per_fold_dt_scores, delimiter=\",\")\n",
        "\n",
        "for filename in filenames:\n",
        "  for count in range(10):\n",
        "    per_fold_dd_scores = np.empty(shape=(len(all_drugs), len(all_drugs)))\n",
        "    per_fold_tt_scores = np.empty(shape=(len(all_targets), len(all_targets)))\n",
        "    per_fold_dt_scores = np.empty(shape=(len(all_drugs), len(all_targets)))\n",
        "    embeddings = read_embeddings(f\"embeddings/nr/{filename}/AVG/fold{count + 1}.emb\")\n",
        "    for i in range(graph_avg.number_of_nodes() - 1):\n",
        "      for j in range(i, graph_avg.number_of_nodes()):\n",
        "        score = cosine_similarity(embeddings[i], embeddings[j])\n",
        "        if i < len(all_drugs) and j < len(all_drugs):\n",
        "          per_fold_dd_scores[i, j] = score\n",
        "          per_fold_dd_scores[j, i] = score\n",
        "        elif i >= len(all_drugs) and j >= len(all_drugs):\n",
        "          per_fold_tt_scores[i - len(all_drugs), j - len(all_drugs)] = score\n",
        "          per_fold_tt_scores[j - len(all_drugs), i - len(all_drugs)] = score\n",
        "        else:\n",
        "          per_fold_dt_scores[i, j - len(all_drugs)] = score\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/AVG/cosine_sim/DDsim_fold{count + 1}.csv\", per_fold_dd_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/AVG/cosine_sim/TTsim_fold{count + 1}.csv\", per_fold_tt_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/AVG/cosine_sim/DTsim_fold{count + 1}.csv\", per_fold_dt_scores, delimiter=\",\")\n",
        "\n",
        "for filename in filenames:\n",
        "  for count in range(10):\n",
        "    per_fold_dd_scores = np.empty(shape=(len(all_drugs), len(all_drugs)))\n",
        "    per_fold_tt_scores = np.empty(shape=(len(all_targets), len(all_targets)))\n",
        "    per_fold_dt_scores = np.empty(shape=(len(all_drugs), len(all_targets)))\n",
        "    embeddings = read_embeddings(f\"embeddings/nr/{filename}/MAX/fold{count + 1}.emb\")\n",
        "    for i in range(graph_max.number_of_nodes() - 1):\n",
        "      for j in range(i, graph_max.number_of_nodes()):\n",
        "        score = cosine_similarity(embeddings[i], embeddings[j])\n",
        "        if i < len(all_drugs) and j < len(all_drugs):\n",
        "          per_fold_dd_scores[i, j] = score\n",
        "          per_fold_dd_scores[j, i] = score\n",
        "        elif i >= len(all_drugs) and j >= len(all_drugs):\n",
        "          per_fold_tt_scores[i - len(all_drugs), j - len(all_drugs)] = score\n",
        "          per_fold_tt_scores[j - len(all_drugs), i - len(all_drugs)] = score\n",
        "        else:\n",
        "          per_fold_dt_scores[i, j - len(all_drugs)] = score\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/MAX/cosine_sim/DDsim_fold{count + 1}.csv\", per_fold_dd_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/MAX/cosine_sim/TTsim_fold{count + 1}.csv\", per_fold_tt_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/MAX/cosine_sim/DTsim_fold{count + 1}.csv\", per_fold_dt_scores, delimiter=\",\") "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WOdJHHYls53M"
      },
      "source": [
        "# Pearson correlation similarity\n",
        "for filename in filenames:\n",
        "  for count in range(10):\n",
        "    per_fold_dd_scores = np.empty(shape=(len(all_drugs), len(all_drugs)))\n",
        "    per_fold_tt_scores = np.empty(shape=(len(all_targets), len(all_targets)))\n",
        "    per_fold_dt_scores = np.empty(shape=(len(all_drugs), len(all_targets)))\n",
        "    embeddings = read_embeddings(f\"embeddings/nr/{filename}/SNF/fold{count + 1}.emb\")\n",
        "    for i in range(graph_SNF.number_of_nodes() - 1):\n",
        "      for j in range(i, graph_SNF.number_of_nodes()):\n",
        "        score = Pearson_correlation_similarity(embeddings[i], embeddings[j])\n",
        "        if i < len(all_drugs) and j < len(all_drugs):\n",
        "          per_fold_dd_scores[i, j] = score\n",
        "          per_fold_dd_scores[j, i] = score\n",
        "        elif i >= len(all_drugs) and j >= len(all_drugs):\n",
        "          per_fold_tt_scores[i - len(all_drugs), j - len(all_drugs)] = score\n",
        "          per_fold_tt_scores[j - len(all_drugs), i - len(all_drugs)] = score\n",
        "        else:\n",
        "          per_fold_dt_scores[i, j - len(all_drugs)] = score\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/SNF/Pearson_corr_sim/DDsim_fold{count + 1}.csv\", per_fold_dd_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/SNF/Pearson_corr_sim/TTsim_fold{count + 1}.csv\", per_fold_tt_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/SNF/Pearson_corr_sim/DTsim_fold{count + 1}.csv\", per_fold_dt_scores, delimiter=\",\")\n",
        "\n",
        "for filename in filenames:\n",
        "  for count in range(10):\n",
        "    per_fold_dd_scores = np.empty(shape=(len(all_drugs), len(all_drugs)))\n",
        "    per_fold_tt_scores = np.empty(shape=(len(all_targets), len(all_targets)))\n",
        "    per_fold_dt_scores = np.empty(shape=(len(all_drugs), len(all_targets)))\n",
        "    embeddings = read_embeddings(f\"embeddings/nr/{filename}/AVG/fold{count + 1}.emb\")\n",
        "    for i in range(graph_avg.number_of_nodes() - 1):\n",
        "      for j in range(i, graph_avg.number_of_nodes()):\n",
        "        score = Pearson_correlation_similarity(embeddings[i], embeddings[j])\n",
        "        if i < len(all_drugs) and j < len(all_drugs):\n",
        "          per_fold_dd_scores[i, j] = score\n",
        "          per_fold_dd_scores[j, i] = score\n",
        "        elif i >= len(all_drugs) and j >= len(all_drugs):\n",
        "          per_fold_tt_scores[i - len(all_drugs), j - len(all_drugs)] = score\n",
        "          per_fold_tt_scores[j - len(all_drugs), i - len(all_drugs)] = score\n",
        "        else:\n",
        "          per_fold_dt_scores[i, j - len(all_drugs)] = score\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/AVG/Pearson_corr_sim/DDsim_fold{count + 1}.csv\", per_fold_dd_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/AVG/Pearson_corr_sim/TTsim_fold{count + 1}.csv\", per_fold_tt_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/AVG/Pearson_corr_sim/DTsim_fold{count + 1}.csv\", per_fold_dt_scores, delimiter=\",\")\n",
        "\n",
        "for filename in filenames:\n",
        "  for count in range(10):\n",
        "    per_fold_dd_scores = np.empty(shape=(len(all_drugs), len(all_drugs)))\n",
        "    per_fold_tt_scores = np.empty(shape=(len(all_targets), len(all_targets)))\n",
        "    per_fold_dt_scores = np.empty(shape=(len(all_drugs), len(all_targets)))\n",
        "    embeddings = read_embeddings(f\"embeddings/nr/{filename}/MAX/fold{count + 1}.emb\")\n",
        "    for i in range(graph_max.number_of_nodes() - 1):\n",
        "      for j in range(i, graph_max.number_of_nodes()):\n",
        "        score = Pearson_correlation_similarity(embeddings[i], embeddings[j])\n",
        "        if i < len(all_drugs) and j < len(all_drugs):\n",
        "          per_fold_dd_scores[i, j] = score\n",
        "          per_fold_dd_scores[j, i] = score\n",
        "        elif i >= len(all_drugs) and j >= len(all_drugs):\n",
        "          per_fold_tt_scores[i - len(all_drugs), j - len(all_drugs)] = score\n",
        "          per_fold_tt_scores[j - len(all_drugs), i - len(all_drugs)] = score\n",
        "        else:\n",
        "          per_fold_dt_scores[i, j - len(all_drugs)] = score\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/MAX/Pearson_corr_sim/DDsim_fold{count + 1}.csv\", per_fold_dd_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/MAX/Pearson_corr_sim/TTsim_fold{count + 1}.csv\", per_fold_tt_scores, delimiter=\",\")\n",
        "    np.savetxt(f\"embeddings/nr/{filename}/MAX/Pearson_corr_sim/DTsim_fold{count + 1}.csv\", per_fold_dt_scores, delimiter=\",\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8Pieb-ZX4aqh"
      },
      "source": [
        "fused_sims = ['SNF', 'AVG', 'MAX']\n",
        "files = []\n",
        "for filename in filenames:\n",
        "  for fused_sim in fused_sims:\n",
        "    for count in range(10):\n",
        "      files.append(f\"embeddings/nr/{filename}/{fused_sim}/cosine_sim/DDsim_fold{count + 1}.csv\")\n",
        "      files.append(f\"embeddings/nr/{filename}/{fused_sim}/cosine_sim/TTsim_fold{count + 1}.csv\")\n",
        "      files.append(f\"embeddings/nr/{filename}/{fused_sim}/cosine_sim/DTsim_fold{count + 1}.csv\")\n",
        "      files.append(f\"embeddings/nr/{filename}/{fused_sim}/Pearson_corr_sim/DDsim_fold{count + 1}.csv\")\n",
        "      files.append(f\"embeddings/nr/{filename}/{fused_sim}/Pearson_corr_sim/TTsim_fold{count + 1}.csv\")\n",
        "      files.append(f\"embeddings/nr/{filename}/{fused_sim}/Pearson_corr_sim/DTsim_fold{count + 1}.csv\")\n",
        "\n",
        "for file_path in files:\n",
        "  validate_file(file_path)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M2gu7QUqFtDM"
      },
      "source": [
        "---\n",
        "### Classification"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "prcLfEQNfUzt"
      },
      "source": [
        "AdaBoost Classifier - SNF"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zLg-cfNn6c7n"
      },
      "source": [
        "with open(f\"../Output/nr/summary.csv\", \"w\") as summary:\n",
        "  summary.write(\"model,accuracy,precision,recall,f1,auc,aupr\\n\")\n",
        "summary.close()\n",
        "\n",
        "fused_sims = ['cosine_sim', 'Pearson_corr_sim']\n",
        "for fused_sim in fused_sims:\n",
        "  for filename in filenames:\n",
        "    with open(f\"../Output/nr/summary.csv\", \"a\") as summary:\n",
        "      with open(f\"../Output/nr/SNF_AdaBoost_{fused_sim}_{filename}_results.txt\", \"w\") as output_file:\n",
        "        accuracy = []\n",
        "        precision = []\n",
        "        recall = []\n",
        "        f1 = []\n",
        "        roc_auc = []\n",
        "        aupr = []\n",
        "        TP = []\n",
        "        TN = []\n",
        "        FP = []\n",
        "        FN = []\n",
        "        count = 1\n",
        "        for train_index, test_index in skf.split(test_edges_SNF, target_label_SNF):\n",
        "          output_file.write(\"\\n---------------------------------------------\")\n",
        "          output_file.write(f\"\\n------------------- Fold {count} ------------------\")\n",
        "          train_fold_edges, train_fold_labels = test_edges_SNF[train_index], target_label_SNF[test_index]\n",
        "          test_fold_edges, test_fold_labels = test_edges_SNF[test_index], target_label_SNF[test_index]\n",
        "          \n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G1)\n",
        "          DTI_train = DTIs.transpose()\n",
        "          DTI_train = mask_test_edges(test_fold_edges, test_fold_labels, DTI_train, len(all_drugs))\n",
        "\n",
        "          sumDDD1, maxDDD1 = DDD_TTT_sim(fused_DD_similarities_SNF)\n",
        "          sumTTT1, maxTTT1 = DDD_TTT_sim(fused_TT_similarities_SNF)\n",
        "\n",
        "          sumDDT1, maxDDT1 = calculate_drug_similarities_and_DTIs_meta_path(fused_DD_similarities_SNF, DTI_train, 2)\n",
        "          sumDTT1, maxDTT1 = calculate_DTIS_and_target_similarities_meta_path(fused_TT_similarities_SNF, DTI_train, 2)\n",
        "\n",
        "          sumDDDT1, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD1, DTI_train, 3)\n",
        "          _, maxDDDT1 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD1, DTI_train, 3)\n",
        "\n",
        "          sumDTTT1, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT1, DTI_train, 3)\n",
        "          _, maxDTTT1 = calculate_DTIS_and_target_similarities_meta_path(maxTTT1, DTI_train, 3)\n",
        "\n",
        "          sumDTDT1, maxDTDT1 = metaPath_DTDT(DTI_train)\n",
        "          sumDDTT1, maxDDTT1 = metaPath_DDTT(DTI_train, fused_DD_similarities_SNF, fused_TT_similarities_SNF)\n",
        "\n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G2)\n",
        "          DTI_new = np.genfromtxt(f\"embeddings/nr/{filename}/SNF/{fused_sim}/DTsim_fold{count}.csv\", delimiter=',')\n",
        "          DD_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/SNF/{fused_sim}/DDsim_fold{count}.csv\", delimiter=',')\n",
        "          TT_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/SNF/{fused_sim}/TTsim_fold{count}.csv\", delimiter=',')\n",
        "\n",
        "          DTI_train_new = mask_test_edges(test_fold_edges, test_fold_labels, DTI_new, len(all_drugs))\n",
        "\n",
        "          sumDDD2, maxDDD2 = DDD_TTT_sim(DD_similarities_new)\n",
        "          sumTTT2, maxTTT2 = DDD_TTT_sim(TT_similarities_new)\n",
        "\n",
        "          sumDDT2, maxDDT2 = calculate_drug_similarities_and_DTIs_meta_path(DD_similarities_new, DTI_train_new, 2)\n",
        "          sumDTT2, maxDTT2 = calculate_DTIS_and_target_similarities_meta_path(TT_similarities_new, DTI_train_new, 2)\n",
        "\n",
        "          sumDDDT2, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD2, DTI_train_new, 3)\n",
        "          _, maxDDDT2 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTTT2, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT2, DTI_train_new, 3)\n",
        "          _, maxDTTT2 = calculate_DTIS_and_target_similarities_meta_path(maxTTT2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTDT2, maxDTDT2 = metaPath_DTDT(DTI_train_new)\n",
        "          sumDDTT2, maxDDTT2 = metaPath_DDTT(DTI_train_new, DD_similarities_new, TT_similarities_new)\n",
        "\n",
        "          features = []\n",
        "          class_labels = []\n",
        "          for pair, label in zip(test_edges_SNF, target_label_SNF):\n",
        "            features.append([ # path scores from G1\n",
        "                              sumDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT1[pair[0]][pair[1] - len(all_drugs)], sumDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT1[pair[0]][pair[1] - len(all_drugs)], maxDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              # path scores from G2\n",
        "                              sumDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT2[pair[0]][pair[1] - len(all_drugs)], sumDTDT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT2[pair[0]][pair[1] - len(all_drugs)], maxDTDT2[pair[0]][pair[1] - len(all_drugs)]])\n",
        "            class_labels.append(label)\n",
        "\n",
        "          X = np.asarray(features)\n",
        "          y = np.array(class_labels)\n",
        "          X_train, y_train = X[train_index], y[train_index]\n",
        "          X_test, y_test = X[test_index], y[test_index]\n",
        "          scaler = MaxAbsScaler()\n",
        "          scaler.fit(X_train)\n",
        "          X_train = scaler.transform(X_train)\n",
        "          X_test = scaler.transform(X_test)\n",
        "\n",
        "          ada_boost.fit(X_train, y_train)\n",
        "          predictions = ada_boost.predict(X_test)\n",
        "          predicted_probabilities = ada_boost.predict_proba(X_test)[:, 1]\n",
        "\n",
        "          cm = confusion_matrix(y_test, predictions)\n",
        "          TN.append(cm[0][0])\n",
        "          FP.append(cm[0][1])\n",
        "          FN.append(cm[1][0])\n",
        "          TP.append(cm[1][1])\n",
        "          output_file.write(\"\\nConfusion Matrix for this fold\")\n",
        "          output_file.write(f\"\\n{cm}\")\n",
        "          output_file.write(f\"\\nAccuracy score: {accuracy_score(y_test, predictions)}\")\n",
        "          accuracy.append(accuracy_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nPrecision score: {precision_score(y_test, predictions)}\")\n",
        "          precision.append(precision_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nRecall score: {recall_score(y_test, predictions)}\")\n",
        "          recall.append(recall_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nF1 score: {f1_score(y_test, predictions)}\")\n",
        "          f1.append(f1_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nArea ROC: {roc_auc_score(y_test, predicted_probabilities)}\")\n",
        "          roc_auc.append(roc_auc_score(y_test, predicted_probabilities))\n",
        "          p, r, _ = precision_recall_curve(y_test, predicted_probabilities, pos_label=1)\n",
        "          output_file.write(f\"\\nAUPR: {auc(r, p)}\")\n",
        "          aupr.append(auc(r, p))\n",
        "          count += 1\n",
        "        \n",
        "        conf_mat = np.zeros(shape=(2, 2))\n",
        "        conf_mat[0][0] = np.array(TN).sum()\n",
        "        conf_mat[0][1] = np.array(FP).sum()\n",
        "        conf_mat[1][0] = np.array(FN).sum()\n",
        "        conf_mat[1][1] = np.array(TP).sum()\n",
        "        output_file.write(\"\\n-------------- Evaluation Metrics for All Folds -----------------\")\n",
        "        output_file.write(f\"\\nAccuracy score: {np.array(accuracy).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nPrecision score: {np.array(precision).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRecall score: {np.array(recall).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nF1 score: {np.array(f1).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRoc_auc: {np.array(roc_auc).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nAUPR: {np.array(aupr).mean().round(decimals=3)}\")\n",
        "        output_file.write(\"\\nConfusion matrix for all folds:\")\n",
        "        output_file.write(f\"\\n{conf_mat}\")\n",
        "\n",
        "        summary.write(f\"SNF_AdaBoost_{fused_sim}_{filename},\")\n",
        "        summary.write(f\"{np.array(accuracy).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(precision).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(recall).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(f1).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(roc_auc).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(aupr).mean().round(decimals=3)}\\n\")\n",
        "      output_file.close()\n",
        "    summary.close()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QePHsLzZfzrG"
      },
      "source": [
        "XGBoost Classifier - SNF"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VEH-CbP3f5-5"
      },
      "source": [
        "fused_sims = ['cosine_sim', 'Pearson_corr_sim']\n",
        "for fused_sim in fused_sims:\n",
        "  for filename in filenames:\n",
        "    with open(f\"../Output/nr/summary.csv\", \"a\") as summary:\n",
        "      with open(f\"../Output/nr/SNF_{filename}_{fused_sim}_XGBoost_results.txt\", \"w\") as output_file:\n",
        "        accuracy = []\n",
        "        precision = []\n",
        "        recall = []\n",
        "        f1 = []\n",
        "        roc_auc = []\n",
        "        aupr = []\n",
        "        TP = []\n",
        "        TN = []\n",
        "        FP = []\n",
        "        FN = []\n",
        "        count = 1\n",
        "        for train_index, test_index in skf.split(test_edges_SNF, target_label_SNF):\n",
        "          output_file.write(\"\\n---------------------------------------------\")\n",
        "          output_file.write(f\"\\n------------------- Fold {count} ------------------\")\n",
        "          train_fold_edges, train_fold_labels = test_edges_SNF[train_index], target_label_SNF[test_index]\n",
        "          test_fold_edges, test_fold_labels = test_edges_SNF[test_index], target_label_SNF[test_index]\n",
        "          \n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G1)\n",
        "          DTI_train = DTIs.transpose()\n",
        "          DTI_train = mask_test_edges(test_fold_edges, test_fold_labels, DTI_train, len(all_drugs))\n",
        "\n",
        "          sumDDD1, maxDDD1 = DDD_TTT_sim(fused_DD_similarities_SNF)\n",
        "          sumTTT1, maxTTT1 = DDD_TTT_sim(fused_TT_similarities_SNF)\n",
        "\n",
        "          sumDDT1, maxDDT1 = calculate_drug_similarities_and_DTIs_meta_path(fused_DD_similarities_SNF, DTI_train, 2)\n",
        "          sumDTT1, maxDTT1 = calculate_DTIS_and_target_similarities_meta_path(fused_TT_similarities_SNF, DTI_train, 2)\n",
        "\n",
        "          sumDDDT1, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD1, DTI_train, 3)\n",
        "          _, maxDDDT1 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD1, DTI_train, 3)\n",
        "\n",
        "          sumDTTT1, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT1, DTI_train, 3)\n",
        "          _, maxDTTT1 = calculate_DTIS_and_target_similarities_meta_path(maxTTT1, DTI_train, 3)\n",
        "\n",
        "          sumDTDT1, maxDTDT1 = metaPath_DTDT(DTI_train)\n",
        "          sumDDTT1, maxDDTT1 = metaPath_DDTT(DTI_train, fused_DD_similarities_SNF, fused_TT_similarities_SNF)\n",
        "\n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G2)\n",
        "          DTI_new = np.genfromtxt(f\"embeddings/nr/{filename}/SNF/{fused_sim}/DTsim_fold{count}.csv\", delimiter=',')\n",
        "          DD_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/SNF/{fused_sim}/DDsim_fold{count}.csv\", delimiter=',')\n",
        "          TT_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/SNF/{fused_sim}/TTsim_fold{count}.csv\", delimiter=',')\n",
        "\n",
        "          DTI_train_new = mask_test_edges(test_fold_edges, test_fold_labels, DTI_new, len(all_drugs))\n",
        "\n",
        "          sumDDD2, maxDDD2 = DDD_TTT_sim(DD_similarities_new)\n",
        "          sumTTT2, maxTTT2 = DDD_TTT_sim(TT_similarities_new)\n",
        "\n",
        "          sumDDT2, maxDDT2 = calculate_drug_similarities_and_DTIs_meta_path(DD_similarities_new, DTI_train_new, 2)\n",
        "          sumDTT2, maxDTT2 = calculate_DTIS_and_target_similarities_meta_path(TT_similarities_new, DTI_train_new, 2)\n",
        "\n",
        "          sumDDDT2, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD2, DTI_train_new, 3)\n",
        "          _, maxDDDT2 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTTT2, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT2, DTI_train_new, 3)\n",
        "          _, maxDTTT2 = calculate_DTIS_and_target_similarities_meta_path(maxTTT2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTDT2, maxDTDT2 = metaPath_DTDT(DTI_train_new)\n",
        "          sumDDTT2, maxDDTT2 = metaPath_DDTT(DTI_train_new, DD_similarities_new, TT_similarities_new)\n",
        "\n",
        "          features = []\n",
        "          class_labels = []\n",
        "          for pair, label in zip(test_edges_SNF, target_label_SNF):\n",
        "            features.append([ # path scores from G1\n",
        "                              sumDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT1[pair[0]][pair[1] - len(all_drugs)], sumDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT1[pair[0]][pair[1] - len(all_drugs)], maxDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              # path scores from G2\n",
        "                              sumDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT2[pair[0]][pair[1] - len(all_drugs)], sumDTDT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT2[pair[0]][pair[1] - len(all_drugs)], maxDTDT2[pair[0]][pair[1] - len(all_drugs)]])\n",
        "            class_labels.append(label)\n",
        "\n",
        "          X = np.asarray(features)\n",
        "          y = np.array(class_labels)\n",
        "          X_train, y_train = X[train_index], y[train_index]\n",
        "          X_test, y_test = X[test_index], y[test_index]\n",
        "          scaler = MaxAbsScaler()\n",
        "          scaler.fit(X_train)\n",
        "          X_train = scaler.transform(X_train)\n",
        "          X_test = scaler.transform(X_test)\n",
        "\n",
        "          xgb.fit(X_train, y_train)\n",
        "          predictions = xgb.predict(X_test)\n",
        "          predicted_probabilities = xgb.predict_proba(X_test)[:, 1]\n",
        "\n",
        "          cm = confusion_matrix(y_test, predictions)\n",
        "          TN.append(cm[0][0])\n",
        "          FP.append(cm[0][1])\n",
        "          FN.append(cm[1][0])\n",
        "          TP.append(cm[1][1])\n",
        "          output_file.write(\"\\nConfusion Matrix for this fold\")\n",
        "          output_file.write(f\"\\n{cm}\")\n",
        "          output_file.write(f\"\\nAccuracy score: {accuracy_score(y_test, predictions)}\")\n",
        "          accuracy.append(accuracy_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nPrecision score: {precision_score(y_test, predictions)}\")\n",
        "          precision.append(precision_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nRecall score: {recall_score(y_test, predictions)}\")\n",
        "          recall.append(recall_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nF1 score: {f1_score(y_test, predictions)}\")\n",
        "          f1.append(f1_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nArea ROC: {roc_auc_score(y_test, predicted_probabilities)}\")\n",
        "          roc_auc.append(roc_auc_score(y_test, predicted_probabilities))\n",
        "          p, r, _ = precision_recall_curve(y_test, predicted_probabilities, pos_label=1)\n",
        "          output_file.write(f\"\\nAUPR: {auc(r, p)}\")\n",
        "          aupr.append(auc(r, p))\n",
        "          count += 1\n",
        "        \n",
        "        conf_mat = np.zeros(shape=(2, 2))\n",
        "        conf_mat[0][0] = np.array(TN).sum()\n",
        "        conf_mat[0][1] = np.array(FP).sum()\n",
        "        conf_mat[1][0] = np.array(FN).sum()\n",
        "        conf_mat[1][1] = np.array(TP).sum()\n",
        "        output_file.write(\"\\n-------------- Evaluation Metrics for All Folds -----------------\")\n",
        "        output_file.write(f\"\\nAccuracy score: {np.array(accuracy).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nPrecision score: {np.array(precision).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRecall score: {np.array(recall).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nF1 score: {np.array(f1).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRoc_auc: {np.array(roc_auc).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nAUPR: {np.array(aupr).mean().round(decimals=3)}\")\n",
        "        output_file.write(\"\\nConfusion matrix for all folds:\")\n",
        "        output_file.write(f\"\\n{conf_mat}\")\n",
        "\n",
        "        summary.write(f\"SNF_{filename}_{fused_sim}_XGBoost,\")\n",
        "        summary.write(f\"{np.array(accuracy).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(precision).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(recall).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(f1).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(roc_auc).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(aupr).mean().round(decimals=3)}\\n\")\n",
        "      output_file.close()\n",
        "    summary.close()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K-a2RY-pnDYN"
      },
      "source": [
        "Artificial Neural Network Classifier - SNF"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cXvoW11onHA2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2019571e-426e-4a69-da11-7d92c7241d0c"
      },
      "source": [
        "fused_sims = ['cosine_sim', 'Pearson_corr_sim']\n",
        "for fused_sim in fused_sims:\n",
        "  for filename in filenames:\n",
        "    with open(f\"../Output/nr/summary.csv\", \"a\") as summary:\n",
        "      with open(f\"../Output/nr/SNF_ANN_{fused_sim}_{filename}_results.txt\", \"w\") as output_file:\n",
        "        accuracy = []\n",
        "        precision = []\n",
        "        recall = []\n",
        "        f1 = []\n",
        "        roc_auc = []\n",
        "        aupr = []\n",
        "        TP = []\n",
        "        TN = []\n",
        "        FP = []\n",
        "        FN = []\n",
        "        count = 1\n",
        "        for train_index, test_index in skf.split(test_edges_SNF, target_label_SNF):\n",
        "          output_file.write(\"\\n---------------------------------------------\")\n",
        "          output_file.write(f\"\\n------------------- Fold {count} ------------------\")\n",
        "          train_fold_edges, train_fold_labels = test_edges_SNF[train_index], target_label_SNF[test_index]\n",
        "          test_fold_edges, test_fold_labels = test_edges_SNF[test_index], target_label_SNF[test_index]\n",
        "          \n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G1)\n",
        "          DTI_train = DTIs.transpose()\n",
        "          DTI_train = mask_test_edges(test_fold_edges, test_fold_labels, DTI_train, len(all_drugs))\n",
        "\n",
        "          sumDDD1, maxDDD1 = DDD_TTT_sim(fused_DD_similarities_SNF)\n",
        "          sumTTT1, maxTTT1 = DDD_TTT_sim(fused_TT_similarities_SNF)\n",
        "\n",
        "          sumDDT1, maxDDT1 = calculate_drug_similarities_and_DTIs_meta_path(fused_DD_similarities_SNF, DTI_train, 2)\n",
        "          sumDTT1, maxDTT1 = calculate_DTIS_and_target_similarities_meta_path(fused_TT_similarities_SNF, DTI_train, 2)\n",
        "\n",
        "          sumDDDT1, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD1, DTI_train, 3)\n",
        "          _, maxDDDT1 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD1, DTI_train, 3)\n",
        "\n",
        "          sumDTTT1, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT1, DTI_train, 3)\n",
        "          _, maxDTTT1 = calculate_DTIS_and_target_similarities_meta_path(maxTTT1, DTI_train, 3)\n",
        "\n",
        "          sumDTDT1, maxDTDT1 = metaPath_DTDT(DTI_train)\n",
        "          sumDDTT1, maxDDTT1 = metaPath_DDTT(DTI_train, fused_DD_similarities_SNF, fused_TT_similarities_SNF)\n",
        "\n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G2)\n",
        "          DTI_new = np.genfromtxt(f\"embeddings/nr/{filename}/SNF/{fused_sim}/DTsim_fold{count}.csv\", delimiter=',')\n",
        "          DD_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/SNF/{fused_sim}/DDsim_fold{count}.csv\", delimiter=',')\n",
        "          TT_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/SNF/{fused_sim}/TTsim_fold{count}.csv\", delimiter=',')\n",
        "\n",
        "          DTI_train_new = mask_test_edges(test_fold_edges, test_fold_labels, DTI_new, len(all_drugs))\n",
        "\n",
        "          sumDDD2, maxDDD2 = DDD_TTT_sim(DD_similarities_new)\n",
        "          sumTTT2, maxTTT2 = DDD_TTT_sim(TT_similarities_new)\n",
        "\n",
        "          sumDDT2, maxDDT2 = calculate_drug_similarities_and_DTIs_meta_path(DD_similarities_new, DTI_train_new, 2)\n",
        "          sumDTT2, maxDTT2 = calculate_DTIS_and_target_similarities_meta_path(TT_similarities_new, DTI_train_new, 2)\n",
        "\n",
        "          sumDDDT2, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD2, DTI_train_new, 3)\n",
        "          _, maxDDDT2 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTTT2, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT2, DTI_train_new, 3)\n",
        "          _, maxDTTT2 = calculate_DTIS_and_target_similarities_meta_path(maxTTT2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTDT2, maxDTDT2 = metaPath_DTDT(DTI_train_new)\n",
        "          sumDDTT2, maxDDTT2 = metaPath_DDTT(DTI_train_new, DD_similarities_new, TT_similarities_new)\n",
        "\n",
        "          features = []\n",
        "          class_labels = []\n",
        "          for pair, label in zip(test_edges_SNF, target_label_SNF):\n",
        "            features.append([ # path scores from G1\n",
        "                              sumDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT1[pair[0]][pair[1] - len(all_drugs)], sumDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT1[pair[0]][pair[1] - len(all_drugs)], maxDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              # path scores from G2\n",
        "                              sumDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT2[pair[0]][pair[1] - len(all_drugs)], sumDTDT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT2[pair[0]][pair[1] - len(all_drugs)], maxDTDT2[pair[0]][pair[1] - len(all_drugs)]])\n",
        "            class_labels.append(label)\n",
        "\n",
        "          X = np.asarray(features)\n",
        "          y = np.array(class_labels)\n",
        "          X_train, y_train = X[train_index], y[train_index]\n",
        "          X_test, y_test = X[test_index], y[test_index]\n",
        "          scaler = MaxAbsScaler()\n",
        "          scaler.fit(X_train)\n",
        "          X_train = scaler.transform(X_train)\n",
        "          X_test = scaler.transform(X_test)\n",
        "\n",
        "          ann.fit(X_train, y_train)\n",
        "          predictions = ann.predict(X_test)\n",
        "          predicted_probabilities = ann.predict_proba(X_test)[:, 1]\n",
        "\n",
        "          cm = confusion_matrix(y_test, predictions)\n",
        "          TN.append(cm[0][0])\n",
        "          FP.append(cm[0][1])\n",
        "          FN.append(cm[1][0])\n",
        "          TP.append(cm[1][1])\n",
        "          output_file.write(\"\\nConfusion Matrix for this fold\")\n",
        "          output_file.write(f\"\\n{cm}\")\n",
        "          output_file.write(f\"\\nAccuracy score: {accuracy_score(y_test, predictions)}\")\n",
        "          accuracy.append(accuracy_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nPrecision score: {precision_score(y_test, predictions)}\")\n",
        "          precision.append(precision_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nRecall score: {recall_score(y_test, predictions)}\")\n",
        "          recall.append(recall_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nF1 score: {f1_score(y_test, predictions)}\")\n",
        "          f1.append(f1_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nArea ROC: {roc_auc_score(y_test, predicted_probabilities)}\")\n",
        "          roc_auc.append(roc_auc_score(y_test, predicted_probabilities))\n",
        "          p, r, _ = precision_recall_curve(y_test, predicted_probabilities, pos_label=1)\n",
        "          output_file.write(f\"\\nAUPR: {auc(r, p)}\")\n",
        "          aupr.append(auc(r, p))\n",
        "          count += 1\n",
        "        \n",
        "        conf_mat = np.zeros(shape=(2, 2))\n",
        "        conf_mat[0][0] = np.array(TN).sum()\n",
        "        conf_mat[0][1] = np.array(FP).sum()\n",
        "        conf_mat[1][0] = np.array(FN).sum()\n",
        "        conf_mat[1][1] = np.array(TP).sum()\n",
        "        output_file.write(\"\\n-------------- Evaluation Metrics for All Folds -----------------\")\n",
        "        output_file.write(f\"\\nAccuracy score: {np.array(accuracy).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nPrecision score: {np.array(precision).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRecall score: {np.array(recall).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nF1 score: {np.array(f1).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRoc_auc: {np.array(roc_auc).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nAUPR: {np.array(aupr).mean().round(decimals=3)}\")\n",
        "        output_file.write(\"\\nConfusion matrix for all folds:\")\n",
        "        output_file.write(f\"\\n{conf_mat}\")\n",
        "\n",
        "        summary.write(f\"SNF_ANN_{fused_sim}_{filename},\")\n",
        "        summary.write(f\"{np.array(accuracy).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(precision).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(recall).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(f1).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(roc_auc).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(aupr).mean().round(decimals=3)}\\n\")\n",
        "      output_file.close()\n",
        "    summary.close()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iJ0eqtQxCMoL"
      },
      "source": [
        "AdaBoost Classifier - AVG"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ang4tfnSCG_4"
      },
      "source": [
        "fused_sims = ['cosine_sim', 'Pearson_corr_sim']\n",
        "for fused_sim in fused_sims:\n",
        "  for filename in filenames:\n",
        "    with open(f\"../Output/nr/summary.csv\", \"a\") as summary:\n",
        "      with open(f\"../Output/nr/AVG_AdaBoost_{fused_sim}_{filename}_results.txt\", \"w\") as output_file:\n",
        "        accuracy = []\n",
        "        precision = []\n",
        "        recall = []\n",
        "        f1 = []\n",
        "        roc_auc = []\n",
        "        aupr = []\n",
        "        TP = []\n",
        "        TN = []\n",
        "        FP = []\n",
        "        FN = []\n",
        "        count = 1\n",
        "        for train_index, test_index in skf.split(test_edges_avg, target_label_avg):\n",
        "          output_file.write(\"\\n---------------------------------------------\")\n",
        "          output_file.write(f\"\\n------------------- Fold {count} ------------------\")\n",
        "          train_fold_edges, train_fold_labels = test_edges_avg[train_index], target_label_avg[test_index]\n",
        "          test_fold_edges, test_fold_labels = test_edges_avg[test_index], target_label_avg[test_index]\n",
        "          \n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G1)\n",
        "          DTI_train = DTIs.transpose()\n",
        "          DTI_train = mask_test_edges(test_fold_edges, test_fold_labels, DTI_train, len(all_drugs))\n",
        "\n",
        "          sumDDD1, maxDDD1 = DDD_TTT_sim(fused_DD_similarities_avg)\n",
        "          sumTTT1, maxTTT1 = DDD_TTT_sim(fused_TT_similarities_avg)\n",
        "\n",
        "          sumDDT1, maxDDT1 = calculate_drug_similarities_and_DTIs_meta_path(fused_DD_similarities_avg, DTI_train, 2)\n",
        "          sumDTT1, maxDTT1 = calculate_DTIS_and_target_similarities_meta_path(fused_TT_similarities_avg, DTI_train, 2)\n",
        "\n",
        "          sumDDDT1, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD1, DTI_train, 3)\n",
        "          _, maxDDDT1 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD1, DTI_train, 3)\n",
        "\n",
        "          sumDTTT1, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT1, DTI_train, 3)\n",
        "          _, maxDTTT1 = calculate_DTIS_and_target_similarities_meta_path(maxTTT1, DTI_train, 3)\n",
        "\n",
        "          sumDTDT1, maxDTDT1 = metaPath_DTDT(DTI_train)\n",
        "          sumDDTT1, maxDDTT1 = metaPath_DDTT(DTI_train, fused_DD_similarities_avg, fused_TT_similarities_avg)\n",
        "\n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G2)\n",
        "          DTI_new = np.genfromtxt(f\"embeddings/nr/{filename}/AVG/{fused_sim}/DTsim_fold{count}.csv\", delimiter=',')\n",
        "          DD_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/AVG/{fused_sim}/DDsim_fold{count}.csv\", delimiter=',')\n",
        "          TT_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/AVG/{fused_sim}/TTsim_fold{count}.csv\", delimiter=',')\n",
        "\n",
        "          DTI_train_new = mask_test_edges(test_fold_edges, test_fold_labels, DTI_new, len(all_drugs))\n",
        "\n",
        "          sumDDD2, maxDDD2 = DDD_TTT_sim(DD_similarities_new)\n",
        "          sumTTT2, maxTTT2 = DDD_TTT_sim(TT_similarities_new)\n",
        "\n",
        "          sumDDT2, maxDDT2 = calculate_drug_similarities_and_DTIs_meta_path(DD_similarities_new, DTI_train_new, 2)\n",
        "          sumDTT2, maxDTT2 = calculate_DTIS_and_target_similarities_meta_path(TT_similarities_new, DTI_train_new, 2)\n",
        "\n",
        "          sumDDDT2, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD2, DTI_train_new, 3)\n",
        "          _, maxDDDT2 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTTT2, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT2, DTI_train_new, 3)\n",
        "          _, maxDTTT2 = calculate_DTIS_and_target_similarities_meta_path(maxTTT2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTDT2, maxDTDT2 = metaPath_DTDT(DTI_train_new)\n",
        "          sumDDTT2, maxDDTT2 = metaPath_DDTT(DTI_train_new, DD_similarities_new, TT_similarities_new)\n",
        "\n",
        "          features = []\n",
        "          class_labels = []\n",
        "          for pair, label in zip(test_edges_avg, target_label_avg):\n",
        "            features.append([ # path scores from G1\n",
        "                              sumDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT1[pair[0]][pair[1] - len(all_drugs)], sumDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT1[pair[0]][pair[1] - len(all_drugs)], maxDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              # path scores from G2\n",
        "                              sumDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT2[pair[0]][pair[1] - len(all_drugs)], sumDTDT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT2[pair[0]][pair[1] - len(all_drugs)], maxDTDT2[pair[0]][pair[1] - len(all_drugs)]])\n",
        "            class_labels.append(label)\n",
        "\n",
        "          X = np.asarray(features)\n",
        "          y = np.array(class_labels)\n",
        "          X_train, y_train = X[train_index], y[train_index]\n",
        "          X_test, y_test = X[test_index], y[test_index]\n",
        "          scaler = MaxAbsScaler()\n",
        "          scaler.fit(X_train)\n",
        "          X_train = scaler.transform(X_train)\n",
        "          X_test = scaler.transform(X_test)\n",
        "\n",
        "          ada_boost.fit(X_train, y_train)\n",
        "          predictions = ada_boost.predict(X_test)\n",
        "          predicted_probabilities = ada_boost.predict_proba(X_test)[:, 1]\n",
        "\n",
        "          cm = confusion_matrix(y_test, predictions)\n",
        "          TN.append(cm[0][0])\n",
        "          FP.append(cm[0][1])\n",
        "          FN.append(cm[1][0])\n",
        "          TP.append(cm[1][1])\n",
        "          output_file.write(\"\\nConfusion Matrix for this fold\")\n",
        "          output_file.write(f\"\\n{cm}\")\n",
        "          output_file.write(f\"\\nAccuracy score: {accuracy_score(y_test, predictions)}\")\n",
        "          accuracy.append(accuracy_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nPrecision score: {precision_score(y_test, predictions)}\")\n",
        "          precision.append(precision_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nRecall score: {recall_score(y_test, predictions)}\")\n",
        "          recall.append(recall_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nF1 score: {f1_score(y_test, predictions)}\")\n",
        "          f1.append(f1_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nArea ROC: {roc_auc_score(y_test, predicted_probabilities)}\")\n",
        "          roc_auc.append(roc_auc_score(y_test, predicted_probabilities))\n",
        "          p, r, _ = precision_recall_curve(y_test, predicted_probabilities, pos_label=1)\n",
        "          output_file.write(f\"\\nAUPR: {auc(r, p)}\")\n",
        "          aupr.append(auc(r, p))\n",
        "          count += 1\n",
        "        \n",
        "        conf_mat = np.zeros(shape=(2, 2))\n",
        "        conf_mat[0][0] = np.array(TN).sum()\n",
        "        conf_mat[0][1] = np.array(FP).sum()\n",
        "        conf_mat[1][0] = np.array(FN).sum()\n",
        "        conf_mat[1][1] = np.array(TP).sum()\n",
        "        output_file.write(\"\\n-------------- Evaluation Metrics for All Folds -----------------\")\n",
        "        output_file.write(f\"\\nAccuracy score: {np.array(accuracy).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nPrecision score: {np.array(precision).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRecall score: {np.array(recall).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nF1 score: {np.array(f1).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRoc_auc: {np.array(roc_auc).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nAUPR: {np.array(aupr).mean().round(decimals=3)}\")\n",
        "        output_file.write(\"\\nConfusion matrix for all folds:\")\n",
        "        output_file.write(f\"\\n{conf_mat}\")\n",
        "\n",
        "        summary.write(f\"AVG_AdaBoost_{fused_sim}_{filename},\")\n",
        "        summary.write(f\"{np.array(accuracy).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(precision).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(recall).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(f1).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(roc_auc).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(aupr).mean().round(decimals=3)}\\n\")\n",
        "      output_file.close()\n",
        "    summary.close()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G1RNA2eYgnPO"
      },
      "source": [
        "XGBoost Classifier - AVG"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6GCl22ltgqTD"
      },
      "source": [
        "fused_sims = ['cosine_sim', 'Pearson_corr_sim']\n",
        "for fused_sim in fused_sims:\n",
        "  for filename in filenames:\n",
        "    with open(f\"../Output/nr/summary.csv\", \"a\") as summary:\n",
        "      with open(f\"../Output/nr/AVG_XGBoost_{fused_sim}_{filename}_results.txt\", \"w\") as output_file:\n",
        "        accuracy = []\n",
        "        precision = []\n",
        "        recall = []\n",
        "        f1 = []\n",
        "        roc_auc = []\n",
        "        aupr = []\n",
        "        TP = []\n",
        "        TN = []\n",
        "        FP = []\n",
        "        FN = []\n",
        "        count = 1\n",
        "        for train_index, test_index in skf.split(test_edges_avg, target_label_avg):\n",
        "          output_file.write(\"\\n---------------------------------------------\")\n",
        "          output_file.write(f\"\\n------------------- Fold {count} ------------------\")\n",
        "          train_fold_edges, train_fold_labels = test_edges_avg[train_index], target_label_avg[test_index]\n",
        "          test_fold_edges, test_fold_labels = test_edges_avg[test_index], target_label_avg[test_index]\n",
        "          \n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G1)\n",
        "          DTI_train = DTIs.transpose()\n",
        "          DTI_train = mask_test_edges(test_fold_edges, test_fold_labels, DTI_train, len(all_drugs))\n",
        "\n",
        "          sumDDD1, maxDDD1 = DDD_TTT_sim(fused_DD_similarities_avg)\n",
        "          sumTTT1, maxTTT1 = DDD_TTT_sim(fused_TT_similarities_avg)\n",
        "\n",
        "          sumDDT1, maxDDT1 = calculate_drug_similarities_and_DTIs_meta_path(fused_DD_similarities_avg, DTI_train, 2)\n",
        "          sumDTT1, maxDTT1 = calculate_DTIS_and_target_similarities_meta_path(fused_TT_similarities_avg, DTI_train, 2)\n",
        "\n",
        "          sumDDDT1, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD1, DTI_train, 3)\n",
        "          _, maxDDDT1 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD1, DTI_train, 3)\n",
        "\n",
        "          sumDTTT1, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT1, DTI_train, 3)\n",
        "          _, maxDTTT1 = calculate_DTIS_and_target_similarities_meta_path(maxTTT1, DTI_train, 3)\n",
        "\n",
        "          sumDTDT1, maxDTDT1 = metaPath_DTDT(DTI_train)\n",
        "          sumDDTT1, maxDDTT1 = metaPath_DDTT(DTI_train, fused_DD_similarities_avg, fused_TT_similarities_avg)\n",
        "\n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G2)\n",
        "          DTI_new = np.genfromtxt(f\"embeddings/nr/{filename}/AVG/{fused_sim}/DTsim_fold{count}.csv\", delimiter=',')\n",
        "          DD_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/AVG/{fused_sim}/DDsim_fold{count}.csv\", delimiter=',')\n",
        "          TT_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/AVG/{fused_sim}/TTsim_fold{count}.csv\", delimiter=',')\n",
        "\n",
        "          DTI_train_new = mask_test_edges(test_fold_edges, test_fold_labels, DTI_new, len(all_drugs))\n",
        "\n",
        "          sumDDD2, maxDDD2 = DDD_TTT_sim(DD_similarities_new)\n",
        "          sumTTT2, maxTTT2 = DDD_TTT_sim(TT_similarities_new)\n",
        "\n",
        "          sumDDT2, maxDDT2 = calculate_drug_similarities_and_DTIs_meta_path(DD_similarities_new, DTI_train_new, 2)\n",
        "          sumDTT2, maxDTT2 = calculate_DTIS_and_target_similarities_meta_path(TT_similarities_new, DTI_train_new, 2)\n",
        "\n",
        "          sumDDDT2, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD2, DTI_train_new, 3)\n",
        "          _, maxDDDT2 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTTT2, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT2, DTI_train_new, 3)\n",
        "          _, maxDTTT2 = calculate_DTIS_and_target_similarities_meta_path(maxTTT2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTDT2, maxDTDT2 = metaPath_DTDT(DTI_train_new)\n",
        "          sumDDTT2, maxDDTT2 = metaPath_DDTT(DTI_train_new, DD_similarities_new, TT_similarities_new)\n",
        "\n",
        "          features = []\n",
        "          class_labels = []\n",
        "          for pair, label in zip(test_edges_avg, target_label_avg):\n",
        "            features.append([ # path scores from G1\n",
        "                              sumDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT1[pair[0]][pair[1] - len(all_drugs)], sumDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT1[pair[0]][pair[1] - len(all_drugs)], maxDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              # path scores from G2\n",
        "                              sumDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT2[pair[0]][pair[1] - len(all_drugs)], sumDTDT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT2[pair[0]][pair[1] - len(all_drugs)], maxDTDT2[pair[0]][pair[1] - len(all_drugs)]])\n",
        "            class_labels.append(label)\n",
        "\n",
        "          X = np.asarray(features)\n",
        "          y = np.array(class_labels)\n",
        "          X_train, y_train = X[train_index], y[train_index]\n",
        "          X_test, y_test = X[test_index], y[test_index]\n",
        "          scaler = MaxAbsScaler()\n",
        "          scaler.fit(X_train)\n",
        "          X_train = scaler.transform(X_train)\n",
        "          X_test = scaler.transform(X_test)\n",
        "\n",
        "          xgb.fit(X_train, y_train)\n",
        "          predictions = xgb.predict(X_test)\n",
        "          predicted_probabilities = xgb.predict_proba(X_test)[:, 1]\n",
        "\n",
        "          cm = confusion_matrix(y_test, predictions)\n",
        "          TN.append(cm[0][0])\n",
        "          FP.append(cm[0][1])\n",
        "          FN.append(cm[1][0])\n",
        "          TP.append(cm[1][1])\n",
        "          output_file.write(\"\\nConfusion Matrix for this fold\")\n",
        "          output_file.write(f\"\\n{cm}\")\n",
        "          output_file.write(f\"\\nAccuracy score: {accuracy_score(y_test, predictions)}\")\n",
        "          accuracy.append(accuracy_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nPrecision score: {precision_score(y_test, predictions)}\")\n",
        "          precision.append(precision_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nRecall score: {recall_score(y_test, predictions)}\")\n",
        "          recall.append(recall_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nF1 score: {f1_score(y_test, predictions)}\")\n",
        "          f1.append(f1_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nArea ROC: {roc_auc_score(y_test, predicted_probabilities)}\")\n",
        "          roc_auc.append(roc_auc_score(y_test, predicted_probabilities))\n",
        "          p, r, _ = precision_recall_curve(y_test, predicted_probabilities, pos_label=1)\n",
        "          output_file.write(f\"\\nAUPR: {auc(r, p)}\")\n",
        "          aupr.append(auc(r, p))\n",
        "          count += 1\n",
        "        \n",
        "        conf_mat = np.zeros(shape=(2, 2))\n",
        "        conf_mat[0][0] = np.array(TN).sum()\n",
        "        conf_mat[0][1] = np.array(FP).sum()\n",
        "        conf_mat[1][0] = np.array(FN).sum()\n",
        "        conf_mat[1][1] = np.array(TP).sum()\n",
        "        output_file.write(\"\\n-------------- Evaluation Metrics for All Folds -----------------\")\n",
        "        output_file.write(f\"\\nAccuracy score: {np.array(accuracy).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nPrecision score: {np.array(precision).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRecall score: {np.array(recall).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nF1 score: {np.array(f1).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRoc_auc: {np.array(roc_auc).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nAUPR: {np.array(aupr).mean().round(decimals=3)}\")\n",
        "        output_file.write(\"\\nConfusion matrix for all folds:\")\n",
        "        output_file.write(f\"\\n{conf_mat}\")\n",
        "\n",
        "        summary.write(f\"AVG_XGBoost_{fused_sim}_{filename},\")\n",
        "        summary.write(f\"{np.array(accuracy).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(precision).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(recall).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(f1).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(roc_auc).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(aupr).mean().round(decimals=3)}\\n\")\n",
        "      output_file.close()\n",
        "    summary.close()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HVtXqDY4nRhj"
      },
      "source": [
        "Artificial Neural Network Classifier - AVG"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4KFYs_MInVGq",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7f2d3268-ce7b-4e5c-aea1-231cda29bcf2"
      },
      "source": [
        "fused_sims = ['cosine_sim', 'Pearson_corr_sim']\n",
        "for fused_sim in fused_sims:\n",
        "  for filename in filenames:\n",
        "    with open(f\"../Output/nr/summary.csv\", \"a\") as summary:\n",
        "      with open(f\"../Output/nr/AVG_ANN_{fused_sim}_{filename}_results.txt\", \"w\") as output_file:\n",
        "        accuracy = []\n",
        "        precision = []\n",
        "        recall = []\n",
        "        f1 = []\n",
        "        roc_auc = []\n",
        "        aupr = []\n",
        "        TP = []\n",
        "        TN = []\n",
        "        FP = []\n",
        "        FN = []\n",
        "        count = 1\n",
        "        for train_index, test_index in skf.split(test_edges_avg, target_label_avg):\n",
        "          output_file.write(\"\\n---------------------------------------------\")\n",
        "          output_file.write(f\"\\n------------------- Fold {count} ------------------\")\n",
        "          train_fold_edges, train_fold_labels = test_edges_avg[train_index], target_label_avg[test_index]\n",
        "          test_fold_edges, test_fold_labels = test_edges_avg[test_index], target_label_avg[test_index]\n",
        "          \n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G1)\n",
        "          DTI_train = DTIs.transpose()\n",
        "          DTI_train = mask_test_edges(test_fold_edges, test_fold_labels, DTI_train, len(all_drugs))\n",
        "\n",
        "          sumDDD1, maxDDD1 = DDD_TTT_sim(fused_DD_similarities_avg)\n",
        "          sumTTT1, maxTTT1 = DDD_TTT_sim(fused_TT_similarities_avg)\n",
        "\n",
        "          sumDDT1, maxDDT1 = calculate_drug_similarities_and_DTIs_meta_path(fused_DD_similarities_avg, DTI_train, 2)\n",
        "          sumDTT1, maxDTT1 = calculate_DTIS_and_target_similarities_meta_path(fused_TT_similarities_avg, DTI_train, 2)\n",
        "\n",
        "          sumDDDT1, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD1, DTI_train, 3)\n",
        "          _, maxDDDT1 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD1, DTI_train, 3)\n",
        "\n",
        "          sumDTTT1, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT1, DTI_train, 3)\n",
        "          _, maxDTTT1 = calculate_DTIS_and_target_similarities_meta_path(maxTTT1, DTI_train, 3)\n",
        "\n",
        "          sumDTDT1, maxDTDT1 = metaPath_DTDT(DTI_train)\n",
        "          sumDDTT1, maxDDTT1 = metaPath_DDTT(DTI_train, fused_DD_similarities_avg, fused_TT_similarities_avg)\n",
        "\n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G2)\n",
        "          DTI_new = np.genfromtxt(f\"embeddings/nr/{filename}/AVG/{fused_sim}/DTsim_fold{count}.csv\", delimiter=',')\n",
        "          DD_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/AVG/{fused_sim}/DDsim_fold{count}.csv\", delimiter=',')\n",
        "          TT_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/AVG/{fused_sim}/TTsim_fold{count}.csv\", delimiter=',')\n",
        "\n",
        "          DTI_train_new = mask_test_edges(test_fold_edges, test_fold_labels, DTI_new, len(all_drugs))\n",
        "\n",
        "          sumDDD2, maxDDD2 = DDD_TTT_sim(DD_similarities_new)\n",
        "          sumTTT2, maxTTT2 = DDD_TTT_sim(TT_similarities_new)\n",
        "\n",
        "          sumDDT2, maxDDT2 = calculate_drug_similarities_and_DTIs_meta_path(DD_similarities_new, DTI_train_new, 2)\n",
        "          sumDTT2, maxDTT2 = calculate_DTIS_and_target_similarities_meta_path(TT_similarities_new, DTI_train_new, 2)\n",
        "\n",
        "          sumDDDT2, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD2, DTI_train_new, 3)\n",
        "          _, maxDDDT2 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTTT2, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT2, DTI_train_new, 3)\n",
        "          _, maxDTTT2 = calculate_DTIS_and_target_similarities_meta_path(maxTTT2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTDT2, maxDTDT2 = metaPath_DTDT(DTI_train_new)\n",
        "          sumDDTT2, maxDDTT2 = metaPath_DDTT(DTI_train_new, DD_similarities_new, TT_similarities_new)\n",
        "\n",
        "          features = []\n",
        "          class_labels = []\n",
        "          for pair, label in zip(test_edges_avg, target_label_avg):\n",
        "            features.append([ # path scores from G1\n",
        "                              sumDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT1[pair[0]][pair[1] - len(all_drugs)], sumDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT1[pair[0]][pair[1] - len(all_drugs)], maxDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              # path scores from G2\n",
        "                              sumDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT2[pair[0]][pair[1] - len(all_drugs)], sumDTDT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT2[pair[0]][pair[1] - len(all_drugs)], maxDTDT2[pair[0]][pair[1] - len(all_drugs)]])\n",
        "            class_labels.append(label)\n",
        "\n",
        "          X = np.asarray(features)\n",
        "          y = np.array(class_labels)\n",
        "          X_train, y_train = X[train_index], y[train_index]\n",
        "          X_test, y_test = X[test_index], y[test_index]\n",
        "          scaler = MaxAbsScaler()\n",
        "          scaler.fit(X_train)\n",
        "          X_train = scaler.transform(X_train)\n",
        "          X_test = scaler.transform(X_test)\n",
        "\n",
        "          ann.fit(X_train, y_train)\n",
        "          predictions = ann.predict(X_test)\n",
        "          predicted_probabilities = ann.predict_proba(X_test)[:, 1]\n",
        "\n",
        "          cm = confusion_matrix(y_test, predictions)\n",
        "          TN.append(cm[0][0])\n",
        "          FP.append(cm[0][1])\n",
        "          FN.append(cm[1][0])\n",
        "          TP.append(cm[1][1])\n",
        "          output_file.write(\"\\nConfusion Matrix for this fold\")\n",
        "          output_file.write(f\"\\n{cm}\")\n",
        "          output_file.write(f\"\\nAccuracy score: {accuracy_score(y_test, predictions)}\")\n",
        "          accuracy.append(accuracy_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nPrecision score: {precision_score(y_test, predictions)}\")\n",
        "          precision.append(precision_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nRecall score: {recall_score(y_test, predictions)}\")\n",
        "          recall.append(recall_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nF1 score: {f1_score(y_test, predictions)}\")\n",
        "          f1.append(f1_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nArea ROC: {roc_auc_score(y_test, predicted_probabilities)}\")\n",
        "          roc_auc.append(roc_auc_score(y_test, predicted_probabilities))\n",
        "          p, r, _ = precision_recall_curve(y_test, predicted_probabilities, pos_label=1)\n",
        "          output_file.write(f\"\\nAUPR: {auc(r, p)}\")\n",
        "          aupr.append(auc(r, p))\n",
        "          count += 1\n",
        "        \n",
        "        conf_mat = np.zeros(shape=(2, 2))\n",
        "        conf_mat[0][0] = np.array(TN).sum()\n",
        "        conf_mat[0][1] = np.array(FP).sum()\n",
        "        conf_mat[1][0] = np.array(FN).sum()\n",
        "        conf_mat[1][1] = np.array(TP).sum()\n",
        "        output_file.write(\"\\n-------------- Evaluation Metrics for All Folds -----------------\")\n",
        "        output_file.write(f\"\\nAccuracy score: {np.array(accuracy).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nPrecision score: {np.array(precision).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRecall score: {np.array(recall).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nF1 score: {np.array(f1).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRoc_auc: {np.array(roc_auc).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nAUPR: {np.array(aupr).mean().round(decimals=3)}\")\n",
        "        output_file.write(\"\\nConfusion matrix for all folds:\")\n",
        "        output_file.write(f\"\\n{conf_mat}\")\n",
        "\n",
        "        summary.write(f\"AVG_ANN_{fused_sim}_{filename},\")\n",
        "        summary.write(f\"{np.array(accuracy).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(precision).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(recall).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(f1).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(roc_auc).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(aupr).mean().round(decimals=3)}\\n\")\n",
        "      output_file.close()\n",
        "    summary.close()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EkDyDtL_CRWi"
      },
      "source": [
        "AdaBoost Classifier - MAX"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6eoMiQjHCIF3"
      },
      "source": [
        "fused_sims = ['cosine_sim', 'Pearson_corr_sim']\n",
        "for fused_sim in fused_sims:\n",
        "  for filename in filenames:\n",
        "    with open(f\"../Output/nr/summary.csv\", \"a\") as summary:\n",
        "      with open(f\"../Output/nr/MAX_AdaBoost_{fused_sim}_{filename}_results.txt\", \"w\") as output_file:\n",
        "        accuracy = []\n",
        "        precision = []\n",
        "        recall = []\n",
        "        f1 = []\n",
        "        roc_auc = []\n",
        "        aupr = []\n",
        "        TP = []\n",
        "        TN = []\n",
        "        FP = []\n",
        "        FN = []\n",
        "        count = 1\n",
        "        for train_index, test_index in skf.split(test_edges_max, target_label_max):\n",
        "          output_file.write(\"\\n---------------------------------------------\")\n",
        "          output_file.write(f\"\\n------------------- Fold {count} ------------------\")\n",
        "          train_fold_edges, train_fold_labels = test_edges_max[train_index], target_label_max[test_index]\n",
        "          test_fold_edges, test_fold_labels = test_edges_max[test_index], target_label_max[test_index]\n",
        "          \n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G1)\n",
        "          DTI_train = DTIs.transpose()\n",
        "          DTI_train = mask_test_edges(test_fold_edges, test_fold_labels, DTI_train, len(all_drugs))\n",
        "\n",
        "          sumDDD1, maxDDD1 = DDD_TTT_sim(fused_DD_similarities_max)\n",
        "          sumTTT1, maxTTT1 = DDD_TTT_sim(fused_TT_similarities_max)\n",
        "\n",
        "          sumDDT1, maxDDT1 = calculate_drug_similarities_and_DTIs_meta_path(fused_DD_similarities_max, DTI_train, 2)\n",
        "          sumDTT1, maxDTT1 = calculate_DTIS_and_target_similarities_meta_path(fused_TT_similarities_max, DTI_train, 2)\n",
        "\n",
        "          sumDDDT1, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD1, DTI_train, 3)\n",
        "          _, maxDDDT1 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD1, DTI_train, 3)\n",
        "\n",
        "          sumDTTT1, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT1, DTI_train, 3)\n",
        "          _, maxDTTT1 = calculate_DTIS_and_target_similarities_meta_path(maxTTT1, DTI_train, 3)\n",
        "\n",
        "          sumDTDT1, maxDTDT1 = metaPath_DTDT(DTI_train)\n",
        "          sumDDTT1, maxDDTT1 = metaPath_DDTT(DTI_train, fused_DD_similarities_max, fused_TT_similarities_max)\n",
        "\n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G2)\n",
        "          DTI_new = np.genfromtxt(f\"embeddings/nr/{filename}/MAX/{fused_sim}/DTsim_fold{count}.csv\", delimiter=',')\n",
        "          DD_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/MAX/{fused_sim}/DDsim_fold{count}.csv\", delimiter=',')\n",
        "          TT_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/MAX/{fused_sim}/TTsim_fold{count}.csv\", delimiter=',')\n",
        "\n",
        "          DTI_train_new = mask_test_edges(test_fold_edges, test_fold_labels, DTI_new, len(all_drugs))\n",
        "\n",
        "          sumDDD2, maxDDD2 = DDD_TTT_sim(DD_similarities_new)\n",
        "          sumTTT2, maxTTT2 = DDD_TTT_sim(TT_similarities_new)\n",
        "\n",
        "          sumDDT2, maxDDT2 = calculate_drug_similarities_and_DTIs_meta_path(DD_similarities_new, DTI_train_new, 2)\n",
        "          sumDTT2, maxDTT2 = calculate_DTIS_and_target_similarities_meta_path(TT_similarities_new, DTI_train_new, 2)\n",
        "\n",
        "          sumDDDT2, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD2, DTI_train_new, 3)\n",
        "          _, maxDDDT2 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTTT2, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT2, DTI_train_new, 3)\n",
        "          _, maxDTTT2 = calculate_DTIS_and_target_similarities_meta_path(maxTTT2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTDT2, maxDTDT2 = metaPath_DTDT(DTI_train_new)\n",
        "          sumDDTT2, maxDDTT2 = metaPath_DDTT(DTI_train_new, DD_similarities_new, TT_similarities_new)\n",
        "\n",
        "          features = []\n",
        "          class_labels = []\n",
        "          for pair, label in zip(test_edges_max, target_label_max):\n",
        "            features.append([ # path scores from G1\n",
        "                              sumDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT1[pair[0]][pair[1] - len(all_drugs)], sumDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT1[pair[0]][pair[1] - len(all_drugs)], maxDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              # path scores from G2\n",
        "                              sumDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT2[pair[0]][pair[1] - len(all_drugs)], sumDTDT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT2[pair[0]][pair[1] - len(all_drugs)], maxDTDT2[pair[0]][pair[1] - len(all_drugs)]])\n",
        "            class_labels.append(label)\n",
        "\n",
        "          X = np.asarray(features)\n",
        "          y = np.array(class_labels)\n",
        "          X_train, y_train = X[train_index], y[train_index]\n",
        "          X_test, y_test = X[test_index], y[test_index]\n",
        "          scaler = MaxAbsScaler()\n",
        "          scaler.fit(X_train)\n",
        "          X_train = scaler.transform(X_train)\n",
        "          X_test = scaler.transform(X_test)\n",
        "\n",
        "          ada_boost.fit(X_train, y_train)\n",
        "          predictions = ada_boost.predict(X_test)\n",
        "          predicted_probabilities = ada_boost.predict_proba(X_test)[:, 1]\n",
        "\n",
        "          cm = confusion_matrix(y_test, predictions)\n",
        "          TN.append(cm[0][0])\n",
        "          FP.append(cm[0][1])\n",
        "          FN.append(cm[1][0])\n",
        "          TP.append(cm[1][1])\n",
        "          output_file.write(\"\\nConfusion Matrix for this fold\")\n",
        "          output_file.write(f\"\\n{cm}\")\n",
        "          output_file.write(f\"\\nAccuracy score: {accuracy_score(y_test, predictions)}\")\n",
        "          accuracy.append(accuracy_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nPrecision score: {precision_score(y_test, predictions)}\")\n",
        "          precision.append(precision_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nRecall score: {recall_score(y_test, predictions)}\")\n",
        "          recall.append(recall_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nF1 score: {f1_score(y_test, predictions)}\")\n",
        "          f1.append(f1_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nArea ROC: {roc_auc_score(y_test, predicted_probabilities)}\")\n",
        "          roc_auc.append(roc_auc_score(y_test, predicted_probabilities))\n",
        "          p, r, _ = precision_recall_curve(y_test, predicted_probabilities, pos_label=1)\n",
        "          output_file.write(f\"\\nAUPR: {auc(r, p)}\")\n",
        "          aupr.append(auc(r, p))\n",
        "          count += 1\n",
        "        \n",
        "        conf_mat = np.zeros(shape=(2, 2))\n",
        "        conf_mat[0][0] = np.array(TN).sum()\n",
        "        conf_mat[0][1] = np.array(FP).sum()\n",
        "        conf_mat[1][0] = np.array(FN).sum()\n",
        "        conf_mat[1][1] = np.array(TP).sum()\n",
        "        output_file.write(\"\\n-------------- Evaluation Metrics for All Folds -----------------\")\n",
        "        output_file.write(f\"\\nAccuracy score: {np.array(accuracy).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nPrecision score: {np.array(precision).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRecall score: {np.array(recall).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nF1 score: {np.array(f1).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRoc_auc: {np.array(roc_auc).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nAUPR: {np.array(aupr).mean().round(decimals=3)}\")\n",
        "        output_file.write(\"\\nConfusion matrix for all folds:\")\n",
        "        output_file.write(f\"\\n{conf_mat}\")\n",
        "\n",
        "        summary.write(f\"MAX_AdaBoost_{fused_sim}_{filename},\")\n",
        "        summary.write(f\"{np.array(accuracy).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(precision).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(recall).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(f1).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(roc_auc).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(aupr).mean().round(decimals=3)}\\n\")\n",
        "      output_file.close()\n",
        "    summary.close()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TId4I7A0g24o"
      },
      "source": [
        "XGBoost Classifier - MAX"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8_rJBvDLg57j"
      },
      "source": [
        "fused_sims = ['cosine_sim', 'Pearson_corr_sim']\n",
        "for fused_sim in fused_sims:\n",
        "  for filename in filenames:\n",
        "    with open(f\"../Output/nr/summary.csv\", \"a\") as summary:\n",
        "      with open(f\"../Output/nr/MAX_XGBoost_{fused_sim}_{filename}_results.txt\", \"w\") as output_file:\n",
        "        accuracy = []\n",
        "        precision = []\n",
        "        recall = []\n",
        "        f1 = []\n",
        "        roc_auc = []\n",
        "        aupr = []\n",
        "        TP = []\n",
        "        TN = []\n",
        "        FP = []\n",
        "        FN = []\n",
        "        count = 1\n",
        "        for train_index, test_index in skf.split(test_edges_max, target_label_max):\n",
        "          output_file.write(\"\\n---------------------------------------------\")\n",
        "          output_file.write(f\"\\n------------------- Fold {count} ------------------\")\n",
        "          train_fold_edges, train_fold_labels = test_edges_max[train_index], target_label_max[test_index]\n",
        "          test_fold_edges, test_fold_labels = test_edges_max[test_index], target_label_max[test_index]\n",
        "          \n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G1)\n",
        "          DTI_train = DTIs.transpose()\n",
        "          DTI_train = mask_test_edges(test_fold_edges, test_fold_labels, DTI_train, len(all_drugs))\n",
        "\n",
        "          sumDDD1, maxDDD1 = DDD_TTT_sim(fused_DD_similarities_max)\n",
        "          sumTTT1, maxTTT1 = DDD_TTT_sim(fused_TT_similarities_max)\n",
        "\n",
        "          sumDDT1, maxDDT1 = calculate_drug_similarities_and_DTIs_meta_path(fused_DD_similarities_max, DTI_train, 2)\n",
        "          sumDTT1, maxDTT1 = calculate_DTIS_and_target_similarities_meta_path(fused_TT_similarities_max, DTI_train, 2)\n",
        "\n",
        "          sumDDDT1, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD1, DTI_train, 3)\n",
        "          _, maxDDDT1 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD1, DTI_train, 3)\n",
        "\n",
        "          sumDTTT1, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT1, DTI_train, 3)\n",
        "          _, maxDTTT1 = calculate_DTIS_and_target_similarities_meta_path(maxTTT1, DTI_train, 3)\n",
        "\n",
        "          sumDTDT1, maxDTDT1 = metaPath_DTDT(DTI_train)\n",
        "          sumDDTT1, maxDDTT1 = metaPath_DDTT(DTI_train, fused_DD_similarities_max, fused_TT_similarities_max)\n",
        "\n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G2)\n",
        "          DTI_new = np.genfromtxt(f\"embeddings/nr/{filename}/MAX/{fused_sim}/DTsim_fold{count}.csv\", delimiter=',')\n",
        "          DD_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/MAX/{fused_sim}/DDsim_fold{count}.csv\", delimiter=',')\n",
        "          TT_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/MAX/{fused_sim}/TTsim_fold{count}.csv\", delimiter=',')\n",
        "\n",
        "          DTI_train_new = mask_test_edges(test_fold_edges, test_fold_labels, DTI_new, len(all_drugs))\n",
        "\n",
        "          sumDDD2, maxDDD2 = DDD_TTT_sim(DD_similarities_new)\n",
        "          sumTTT2, maxTTT2 = DDD_TTT_sim(TT_similarities_new)\n",
        "\n",
        "          sumDDT2, maxDDT2 = calculate_drug_similarities_and_DTIs_meta_path(DD_similarities_new, DTI_train_new, 2)\n",
        "          sumDTT2, maxDTT2 = calculate_DTIS_and_target_similarities_meta_path(TT_similarities_new, DTI_train_new, 2)\n",
        "\n",
        "          sumDDDT2, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD2, DTI_train_new, 3)\n",
        "          _, maxDDDT2 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTTT2, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT2, DTI_train_new, 3)\n",
        "          _, maxDTTT2 = calculate_DTIS_and_target_similarities_meta_path(maxTTT2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTDT2, maxDTDT2 = metaPath_DTDT(DTI_train_new)\n",
        "          sumDDTT2, maxDDTT2 = metaPath_DDTT(DTI_train_new, DD_similarities_new, TT_similarities_new)\n",
        "\n",
        "          features = []\n",
        "          class_labels = []\n",
        "          for pair, label in zip(test_edges_max, target_label_max):\n",
        "            features.append([ # path scores from G1\n",
        "                              sumDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT1[pair[0]][pair[1] - len(all_drugs)], sumDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT1[pair[0]][pair[1] - len(all_drugs)], maxDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              # path scores from G2\n",
        "                              sumDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT2[pair[0]][pair[1] - len(all_drugs)], sumDTDT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT2[pair[0]][pair[1] - len(all_drugs)], maxDTDT2[pair[0]][pair[1] - len(all_drugs)]])\n",
        "            class_labels.append(label)\n",
        "\n",
        "          X = np.asarray(features)\n",
        "          y = np.array(class_labels)\n",
        "          X_train, y_train = X[train_index], y[train_index]\n",
        "          X_test, y_test = X[test_index], y[test_index]\n",
        "          scaler = MaxAbsScaler()\n",
        "          scaler.fit(X_train)\n",
        "          X_train = scaler.transform(X_train)\n",
        "          X_test = scaler.transform(X_test)\n",
        "\n",
        "          xgb.fit(X_train, y_train)\n",
        "          predictions = xgb.predict(X_test)\n",
        "          predicted_probabilities = xgb.predict_proba(X_test)[:, 1]\n",
        "\n",
        "          cm = confusion_matrix(y_test, predictions)\n",
        "          TN.append(cm[0][0])\n",
        "          FP.append(cm[0][1])\n",
        "          FN.append(cm[1][0])\n",
        "          TP.append(cm[1][1])\n",
        "          output_file.write(\"\\nConfusion Matrix for this fold\")\n",
        "          output_file.write(f\"\\n{cm}\")\n",
        "          output_file.write(f\"\\nAccuracy score: {accuracy_score(y_test, predictions)}\")\n",
        "          accuracy.append(accuracy_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nPrecision score: {precision_score(y_test, predictions)}\")\n",
        "          precision.append(precision_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nRecall score: {recall_score(y_test, predictions)}\")\n",
        "          recall.append(recall_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nF1 score: {f1_score(y_test, predictions)}\")\n",
        "          f1.append(f1_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nArea ROC: {roc_auc_score(y_test, predicted_probabilities)}\")\n",
        "          roc_auc.append(roc_auc_score(y_test, predicted_probabilities))\n",
        "          p, r, _ = precision_recall_curve(y_test, predicted_probabilities, pos_label=1)\n",
        "          output_file.write(f\"\\nAUPR: {auc(r, p)}\")\n",
        "          aupr.append(auc(r, p))\n",
        "          count += 1\n",
        "        \n",
        "        conf_mat = np.zeros(shape=(2, 2))\n",
        "        conf_mat[0][0] = np.array(TN).sum()\n",
        "        conf_mat[0][1] = np.array(FP).sum()\n",
        "        conf_mat[1][0] = np.array(FN).sum()\n",
        "        conf_mat[1][1] = np.array(TP).sum()\n",
        "        output_file.write(\"\\n-------------- Evaluation Metrics for All Folds -----------------\")\n",
        "        output_file.write(f\"\\nAccuracy score: {np.array(accuracy).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nPrecision score: {np.array(precision).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRecall score: {np.array(recall).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nF1 score: {np.array(f1).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRoc_auc: {np.array(roc_auc).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nAUPR: {np.array(aupr).mean().round(decimals=3)}\")\n",
        "        output_file.write(\"\\nConfusion matrix for all folds:\")\n",
        "        output_file.write(f\"\\n{conf_mat}\")\n",
        "\n",
        "        summary.write(f\"MAX_XGBoost_{fused_sim}_{filename},\")\n",
        "        summary.write(f\"{np.array(accuracy).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(precision).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(recall).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(f1).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(roc_auc).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(aupr).mean().round(decimals=3)}\\n\")\n",
        "      output_file.close()\n",
        "    summary.close()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2l3mbSFmngAC"
      },
      "source": [
        "Artificial Neural Network Classifier - MAX"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i0GyFUhUnmc3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9df400af-7289-45c4-d643-f526fe0fe539"
      },
      "source": [
        "fused_sims = ['cosine_sim', 'Pearson_corr_sim']\n",
        "for fused_sim in fused_sims:\n",
        "  for filename in filenames:\n",
        "    with open(f\"../Output/nr/summary.csv\", \"a\") as summary:\n",
        "      with open(f\"../Output/nr/MAX_ANN_{fused_sim}_{filename}_results.txt\", \"w\") as output_file:\n",
        "        accuracy = []\n",
        "        precision = []\n",
        "        recall = []\n",
        "        f1 = []\n",
        "        roc_auc = []\n",
        "        aupr = []\n",
        "        TP = []\n",
        "        TN = []\n",
        "        FP = []\n",
        "        FN = []\n",
        "        count = 1\n",
        "        for train_index, test_index in skf.split(test_edges_max, target_label_max):\n",
        "          output_file.write(\"\\n---------------------------------------------\")\n",
        "          output_file.write(f\"\\n------------------- Fold {count} ------------------\")\n",
        "          train_fold_edges, train_fold_labels = test_edges_max[train_index], target_label_max[test_index]\n",
        "          test_fold_edges, test_fold_labels = test_edges_max[test_index], target_label_max[test_index]\n",
        "          \n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G1)\n",
        "          DTI_train = DTIs.transpose()\n",
        "          DTI_train = mask_test_edges(test_fold_edges, test_fold_labels, DTI_train, len(all_drugs))\n",
        "\n",
        "          sumDDD1, maxDDD1 = DDD_TTT_sim(fused_DD_similarities_max)\n",
        "          sumTTT1, maxTTT1 = DDD_TTT_sim(fused_TT_similarities_max)\n",
        "\n",
        "          sumDDT1, maxDDT1 = calculate_drug_similarities_and_DTIs_meta_path(fused_DD_similarities_max, DTI_train, 2)\n",
        "          sumDTT1, maxDTT1 = calculate_DTIS_and_target_similarities_meta_path(fused_TT_similarities_max, DTI_train, 2)\n",
        "\n",
        "          sumDDDT1, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD1, DTI_train, 3)\n",
        "          _, maxDDDT1 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD1, DTI_train, 3)\n",
        "\n",
        "          sumDTTT1, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT1, DTI_train, 3)\n",
        "          _, maxDTTT1 = calculate_DTIS_and_target_similarities_meta_path(maxTTT1, DTI_train, 3)\n",
        "\n",
        "          sumDTDT1, maxDTDT1 = metaPath_DTDT(DTI_train)\n",
        "          sumDDTT1, maxDDTT1 = metaPath_DDTT(DTI_train, fused_DD_similarities_max, fused_TT_similarities_max)\n",
        "\n",
        "          # Generate all features from the matrix multiplication of each path structure\n",
        "          # list for each feature (Graph G2)\n",
        "          DTI_new = np.genfromtxt(f\"embeddings/nr/{filename}/MAX/{fused_sim}/DTsim_fold{count}.csv\", delimiter=',')\n",
        "          DD_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/MAX/{fused_sim}/DDsim_fold{count}.csv\", delimiter=',')\n",
        "          TT_similarities_new = np.genfromtxt(f\"embeddings/nr/{filename}/MAX/{fused_sim}/TTsim_fold{count}.csv\", delimiter=',')\n",
        "\n",
        "          DTI_train_new = mask_test_edges(test_fold_edges, test_fold_labels, DTI_new, len(all_drugs))\n",
        "\n",
        "          sumDDD2, maxDDD2 = DDD_TTT_sim(DD_similarities_new)\n",
        "          sumTTT2, maxTTT2 = DDD_TTT_sim(TT_similarities_new)\n",
        "\n",
        "          sumDDT2, maxDDT2 = calculate_drug_similarities_and_DTIs_meta_path(DD_similarities_new, DTI_train_new, 2)\n",
        "          sumDTT2, maxDTT2 = calculate_DTIS_and_target_similarities_meta_path(TT_similarities_new, DTI_train_new, 2)\n",
        "\n",
        "          sumDDDT2, _ = calculate_drug_similarities_and_DTIs_meta_path(sumDDD2, DTI_train_new, 3)\n",
        "          _, maxDDDT2 = calculate_drug_similarities_and_DTIs_meta_path(maxDDD2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTTT2, _ = calculate_DTIS_and_target_similarities_meta_path(sumTTT2, DTI_train_new, 3)\n",
        "          _, maxDTTT2 = calculate_DTIS_and_target_similarities_meta_path(maxTTT2, DTI_train_new, 3)\n",
        "\n",
        "          sumDTDT2, maxDTDT2 = metaPath_DTDT(DTI_train_new)\n",
        "          sumDDTT2, maxDDTT2 = metaPath_DDTT(DTI_train_new, DD_similarities_new, TT_similarities_new)\n",
        "\n",
        "          features = []\n",
        "          class_labels = []\n",
        "          for pair, label in zip(test_edges_max, target_label_max):\n",
        "            features.append([ # path scores from G1\n",
        "                              sumDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT1[pair[0]][pair[1] - len(all_drugs)], sumDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT1[pair[0]][pair[1] - len(all_drugs)], sumDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT1[pair[0]][pair[1] - len(all_drugs)], maxDTTT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT1[pair[0]][pair[1] - len(all_drugs)], maxDTDT1[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              # path scores from G2\n",
        "                              sumDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDDT2[pair[0]][pair[1] - len(all_drugs)], sumDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              sumDDTT2[pair[0]][pair[1] - len(all_drugs)], sumDTDT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDDT2[pair[0]][pair[1] - len(all_drugs)], maxDTTT2[pair[0]][pair[1] - len(all_drugs)],\n",
        "                              maxDDTT2[pair[0]][pair[1] - len(all_drugs)], maxDTDT2[pair[0]][pair[1] - len(all_drugs)]])\n",
        "            class_labels.append(label)\n",
        "\n",
        "          X = np.asarray(features)\n",
        "          y = np.array(class_labels)\n",
        "          X_train, y_train = X[train_index], y[train_index]\n",
        "          X_test, y_test = X[test_index], y[test_index]\n",
        "          scaler = MaxAbsScaler()\n",
        "          scaler.fit(X_train)\n",
        "          X_train = scaler.transform(X_train)\n",
        "          X_test = scaler.transform(X_test)\n",
        "\n",
        "          ann.fit(X_train, y_train)\n",
        "          predictions = ann.predict(X_test)\n",
        "          predicted_probabilities = ann.predict_proba(X_test)[:, 1]\n",
        "\n",
        "          cm = confusion_matrix(y_test, predictions)\n",
        "          TN.append(cm[0][0])\n",
        "          FP.append(cm[0][1])\n",
        "          FN.append(cm[1][0])\n",
        "          TP.append(cm[1][1])\n",
        "          output_file.write(\"\\nConfusion Matrix for this fold\")\n",
        "          output_file.write(f\"\\n{cm}\")\n",
        "          output_file.write(f\"\\nAccuracy score: {accuracy_score(y_test, predictions)}\")\n",
        "          accuracy.append(accuracy_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nPrecision score: {precision_score(y_test, predictions)}\")\n",
        "          precision.append(precision_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nRecall score: {recall_score(y_test, predictions)}\")\n",
        "          recall.append(recall_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nF1 score: {f1_score(y_test, predictions)}\")\n",
        "          f1.append(f1_score(y_test, predictions))\n",
        "          output_file.write(f\"\\nArea ROC: {roc_auc_score(y_test, predicted_probabilities)}\")\n",
        "          roc_auc.append(roc_auc_score(y_test, predicted_probabilities))\n",
        "          p, r, _ = precision_recall_curve(y_test, predicted_probabilities, pos_label=1)\n",
        "          output_file.write(f\"\\nAUPR: {auc(r, p)}\")\n",
        "          aupr.append(auc(r, p))\n",
        "          count += 1\n",
        "        \n",
        "        conf_mat = np.zeros(shape=(2, 2))\n",
        "        conf_mat[0][0] = np.array(TN).sum()\n",
        "        conf_mat[0][1] = np.array(FP).sum()\n",
        "        conf_mat[1][0] = np.array(FN).sum()\n",
        "        conf_mat[1][1] = np.array(TP).sum()\n",
        "        output_file.write(\"\\n-------------- Evaluation Metrics for All Folds -----------------\")\n",
        "        output_file.write(f\"\\nAccuracy score: {np.array(accuracy).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nPrecision score: {np.array(precision).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRecall score: {np.array(recall).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nF1 score: {np.array(f1).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nRoc_auc: {np.array(roc_auc).mean().round(decimals=3)}\")\n",
        "        output_file.write(f\"\\nAUPR: {np.array(aupr).mean().round(decimals=3)}\")\n",
        "        output_file.write(\"\\nConfusion matrix for all folds:\")\n",
        "        output_file.write(f\"\\n{conf_mat}\")\n",
        "\n",
        "        summary.write(f\"MAX_ANN_{fused_sim}_{filename},\")\n",
        "        summary.write(f\"{np.array(accuracy).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(precision).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(recall).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(f1).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(roc_auc).mean().round(decimals=3)},\")\n",
        "        summary.write(f\"{np.array(aupr).mean().round(decimals=3)}\\n\")\n",
        "      output_file.close()\n",
        "    summary.close()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WHFNEq_DeLr5"
      },
      "source": [
        "# Results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f4IhGKwievFF"
      },
      "source": [
        "%cd /content/drive/My\\ Drive/DTI/Output"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BR6-_a3geIkI"
      },
      "source": [
        "df = pd.read_csv('nr/summary.csv')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "Q9oS1YVzezuS",
        "outputId": "7045ce99-9f35-4504-aafa-7e3ffa7c04c5"
      },
      "source": [
        "display(df)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "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>model</th>\n",
              "      <th>accuracy</th>\n",
              "      <th>precision</th>\n",
              "      <th>recall</th>\n",
              "      <th>f1</th>\n",
              "      <th>auc</th>\n",
              "      <th>aupr</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>SNF_AdaBoost_cosine_sim_hope</td>\n",
              "      <td>0.783</td>\n",
              "      <td>0.983</td>\n",
              "      <td>0.578</td>\n",
              "      <td>0.720</td>\n",
              "      <td>0.783</td>\n",
              "      <td>0.886</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>SNF_AdaBoost_cosine_sim_lap</td>\n",
              "      <td>0.861</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.744</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.861</td>\n",
              "      <td>0.924</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>SNF_AdaBoost_cosine_sim_lle</td>\n",
              "      <td>0.850</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.722</td>\n",
              "      <td>0.822</td>\n",
              "      <td>0.850</td>\n",
              "      <td>0.918</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>SNF_AdaBoost_Pearson_corr_sim_hope</td>\n",
              "      <td>0.772</td>\n",
              "      <td>0.986</td>\n",
              "      <td>0.556</td>\n",
              "      <td>0.704</td>\n",
              "      <td>0.772</td>\n",
              "      <td>0.882</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>SNF_AdaBoost_Pearson_corr_sim_lap</td>\n",
              "      <td>0.906</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.811</td>\n",
              "      <td>0.892</td>\n",
              "      <td>0.906</td>\n",
              "      <td>0.953</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>SNF_AdaBoost_Pearson_corr_sim_lle</td>\n",
              "      <td>0.861</td>\n",
              "      <td>0.950</td>\n",
              "      <td>0.767</td>\n",
              "      <td>0.841</td>\n",
              "      <td>0.861</td>\n",
              "      <td>0.916</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>SNF_hope_cosine_sim_XGBoost</td>\n",
              "      <td>0.739</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.478</td>\n",
              "      <td>0.638</td>\n",
              "      <td>0.907</td>\n",
              "      <td>0.923</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>SNF_lap_cosine_sim_XGBoost</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.700</td>\n",
              "      <td>0.809</td>\n",
              "      <td>0.949</td>\n",
              "      <td>0.959</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>SNF_lle_cosine_sim_XGBoost</td>\n",
              "      <td>0.867</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.756</td>\n",
              "      <td>0.847</td>\n",
              "      <td>0.904</td>\n",
              "      <td>0.912</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>SNF_hope_Pearson_corr_sim_XGBoost</td>\n",
              "      <td>0.733</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.467</td>\n",
              "      <td>0.628</td>\n",
              "      <td>0.904</td>\n",
              "      <td>0.920</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>SNF_lap_Pearson_corr_sim_XGBoost</td>\n",
              "      <td>0.861</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.722</td>\n",
              "      <td>0.834</td>\n",
              "      <td>0.960</td>\n",
              "      <td>0.972</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>SNF_lle_Pearson_corr_sim_XGBoost</td>\n",
              "      <td>0.844</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.711</td>\n",
              "      <td>0.810</td>\n",
              "      <td>0.953</td>\n",
              "      <td>0.951</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>SNF_ANN_cosine_sim_hope</td>\n",
              "      <td>0.633</td>\n",
              "      <td>0.645</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.651</td>\n",
              "      <td>0.744</td>\n",
              "      <td>0.825</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>SNF_ANN_cosine_sim_lap</td>\n",
              "      <td>0.494</td>\n",
              "      <td>0.497</td>\n",
              "      <td>0.989</td>\n",
              "      <td>0.662</td>\n",
              "      <td>0.944</td>\n",
              "      <td>0.962</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>SNF_ANN_cosine_sim_lle</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.500</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.949</td>\n",
              "      <td>0.969</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>SNF_ANN_Pearson_corr_sim_hope</td>\n",
              "      <td>0.883</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.767</td>\n",
              "      <td>0.861</td>\n",
              "      <td>0.946</td>\n",
              "      <td>0.962</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>SNF_ANN_Pearson_corr_sim_lap</td>\n",
              "      <td>0.933</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.867</td>\n",
              "      <td>0.925</td>\n",
              "      <td>0.944</td>\n",
              "      <td>0.967</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>SNF_ANN_Pearson_corr_sim_lle</td>\n",
              "      <td>0.922</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.844</td>\n",
              "      <td>0.913</td>\n",
              "      <td>0.946</td>\n",
              "      <td>0.967</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>AVG_AdaBoost_cosine_sim_hope</td>\n",
              "      <td>0.817</td>\n",
              "      <td>0.990</td>\n",
              "      <td>0.644</td>\n",
              "      <td>0.761</td>\n",
              "      <td>0.817</td>\n",
              "      <td>0.906</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>AVG_AdaBoost_cosine_sim_lap</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.986</td>\n",
              "      <td>0.689</td>\n",
              "      <td>0.805</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.915</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>AVG_AdaBoost_cosine_sim_lle</td>\n",
              "      <td>0.878</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.756</td>\n",
              "      <td>0.856</td>\n",
              "      <td>0.878</td>\n",
              "      <td>0.939</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>AVG_AdaBoost_Pearson_corr_sim_hope</td>\n",
              "      <td>0.806</td>\n",
              "      <td>0.980</td>\n",
              "      <td>0.633</td>\n",
              "      <td>0.746</td>\n",
              "      <td>0.806</td>\n",
              "      <td>0.898</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>AVG_AdaBoost_Pearson_corr_sim_lap</td>\n",
              "      <td>0.867</td>\n",
              "      <td>0.965</td>\n",
              "      <td>0.767</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.867</td>\n",
              "      <td>0.924</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>AVG_AdaBoost_Pearson_corr_sim_lle</td>\n",
              "      <td>0.889</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.778</td>\n",
              "      <td>0.868</td>\n",
              "      <td>0.889</td>\n",
              "      <td>0.944</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24</th>\n",
              "      <td>AVG_XGBoost_cosine_sim_hope</td>\n",
              "      <td>0.744</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.489</td>\n",
              "      <td>0.648</td>\n",
              "      <td>0.915</td>\n",
              "      <td>0.933</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>AVG_XGBoost_cosine_sim_lap</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.971</td>\n",
              "      <td>0.700</td>\n",
              "      <td>0.807</td>\n",
              "      <td>0.946</td>\n",
              "      <td>0.961</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>AVG_XGBoost_cosine_sim_lle</td>\n",
              "      <td>0.872</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.744</td>\n",
              "      <td>0.850</td>\n",
              "      <td>0.965</td>\n",
              "      <td>0.970</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27</th>\n",
              "      <td>AVG_XGBoost_Pearson_corr_sim_hope</td>\n",
              "      <td>0.744</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.489</td>\n",
              "      <td>0.648</td>\n",
              "      <td>0.937</td>\n",
              "      <td>0.944</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28</th>\n",
              "      <td>AVG_XGBoost_Pearson_corr_sim_lap</td>\n",
              "      <td>0.828</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.678</td>\n",
              "      <td>0.790</td>\n",
              "      <td>0.961</td>\n",
              "      <td>0.965</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29</th>\n",
              "      <td>AVG_XGBoost_Pearson_corr_sim_lle</td>\n",
              "      <td>0.839</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.678</td>\n",
              "      <td>0.800</td>\n",
              "      <td>0.978</td>\n",
              "      <td>0.981</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>AVG_ANN_cosine_sim_hope</td>\n",
              "      <td>0.606</td>\n",
              "      <td>0.597</td>\n",
              "      <td>0.889</td>\n",
              "      <td>0.697</td>\n",
              "      <td>0.891</td>\n",
              "      <td>0.928</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>31</th>\n",
              "      <td>AVG_ANN_cosine_sim_lap</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.500</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.934</td>\n",
              "      <td>0.961</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>32</th>\n",
              "      <td>AVG_ANN_cosine_sim_lle</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.500</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.965</td>\n",
              "      <td>0.979</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>33</th>\n",
              "      <td>AVG_ANN_Pearson_corr_sim_hope</td>\n",
              "      <td>0.900</td>\n",
              "      <td>0.989</td>\n",
              "      <td>0.811</td>\n",
              "      <td>0.887</td>\n",
              "      <td>0.965</td>\n",
              "      <td>0.973</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>34</th>\n",
              "      <td>AVG_ANN_Pearson_corr_sim_lap</td>\n",
              "      <td>0.922</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.844</td>\n",
              "      <td>0.911</td>\n",
              "      <td>0.940</td>\n",
              "      <td>0.963</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>35</th>\n",
              "      <td>AVG_ANN_Pearson_corr_sim_lle</td>\n",
              "      <td>0.922</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.844</td>\n",
              "      <td>0.911</td>\n",
              "      <td>0.938</td>\n",
              "      <td>0.959</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>36</th>\n",
              "      <td>MAX_AdaBoost_cosine_sim_hope</td>\n",
              "      <td>0.806</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.611</td>\n",
              "      <td>0.749</td>\n",
              "      <td>0.806</td>\n",
              "      <td>0.903</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>37</th>\n",
              "      <td>MAX_AdaBoost_cosine_sim_lap</td>\n",
              "      <td>0.794</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.589</td>\n",
              "      <td>0.734</td>\n",
              "      <td>0.794</td>\n",
              "      <td>0.897</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>38</th>\n",
              "      <td>MAX_AdaBoost_cosine_sim_lle</td>\n",
              "      <td>0.778</td>\n",
              "      <td>0.983</td>\n",
              "      <td>0.567</td>\n",
              "      <td>0.715</td>\n",
              "      <td>0.778</td>\n",
              "      <td>0.883</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>39</th>\n",
              "      <td>MAX_AdaBoost_Pearson_corr_sim_hope</td>\n",
              "      <td>0.761</td>\n",
              "      <td>0.986</td>\n",
              "      <td>0.533</td>\n",
              "      <td>0.684</td>\n",
              "      <td>0.761</td>\n",
              "      <td>0.876</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>40</th>\n",
              "      <td>MAX_AdaBoost_Pearson_corr_sim_lap</td>\n",
              "      <td>0.794</td>\n",
              "      <td>0.983</td>\n",
              "      <td>0.600</td>\n",
              "      <td>0.738</td>\n",
              "      <td>0.794</td>\n",
              "      <td>0.892</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>41</th>\n",
              "      <td>MAX_AdaBoost_Pearson_corr_sim_lle</td>\n",
              "      <td>0.767</td>\n",
              "      <td>0.966</td>\n",
              "      <td>0.556</td>\n",
              "      <td>0.701</td>\n",
              "      <td>0.767</td>\n",
              "      <td>0.872</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>42</th>\n",
              "      <td>MAX_XGBoost_cosine_sim_hope</td>\n",
              "      <td>0.744</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.489</td>\n",
              "      <td>0.648</td>\n",
              "      <td>0.900</td>\n",
              "      <td>0.923</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>43</th>\n",
              "      <td>MAX_XGBoost_cosine_sim_lap</td>\n",
              "      <td>0.789</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.578</td>\n",
              "      <td>0.724</td>\n",
              "      <td>0.901</td>\n",
              "      <td>0.925</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44</th>\n",
              "      <td>MAX_XGBoost_cosine_sim_lle</td>\n",
              "      <td>0.756</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.511</td>\n",
              "      <td>0.672</td>\n",
              "      <td>0.917</td>\n",
              "      <td>0.934</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>45</th>\n",
              "      <td>MAX_XGBoost_Pearson_corr_sim_hope</td>\n",
              "      <td>0.739</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.478</td>\n",
              "      <td>0.638</td>\n",
              "      <td>0.932</td>\n",
              "      <td>0.943</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>46</th>\n",
              "      <td>MAX_XGBoost_Pearson_corr_sim_lap</td>\n",
              "      <td>0.767</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.533</td>\n",
              "      <td>0.688</td>\n",
              "      <td>0.920</td>\n",
              "      <td>0.935</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>47</th>\n",
              "      <td>MAX_XGBoost_Pearson_corr_sim_lle</td>\n",
              "      <td>0.750</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.664</td>\n",
              "      <td>0.924</td>\n",
              "      <td>0.936</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>48</th>\n",
              "      <td>MAX_ANN_cosine_sim_hope</td>\n",
              "      <td>0.606</td>\n",
              "      <td>0.590</td>\n",
              "      <td>0.989</td>\n",
              "      <td>0.727</td>\n",
              "      <td>0.954</td>\n",
              "      <td>0.970</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>49</th>\n",
              "      <td>MAX_ANN_cosine_sim_lap</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.500</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.931</td>\n",
              "      <td>0.950</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50</th>\n",
              "      <td>MAX_ANN_cosine_sim_lle</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.500</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.948</td>\n",
              "      <td>0.966</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>51</th>\n",
              "      <td>MAX_ANN_Pearson_corr_sim_hope</td>\n",
              "      <td>0.939</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.878</td>\n",
              "      <td>0.932</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.983</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>52</th>\n",
              "      <td>MAX_ANN_Pearson_corr_sim_lap</td>\n",
              "      <td>0.894</td>\n",
              "      <td>0.986</td>\n",
              "      <td>0.800</td>\n",
              "      <td>0.880</td>\n",
              "      <td>0.973</td>\n",
              "      <td>0.980</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>53</th>\n",
              "      <td>MAX_ANN_Pearson_corr_sim_lle</td>\n",
              "      <td>0.900</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.800</td>\n",
              "      <td>0.885</td>\n",
              "      <td>0.968</td>\n",
              "      <td>0.976</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                 model  accuracy  ...    auc   aupr\n",
              "0         SNF_AdaBoost_cosine_sim_hope     0.783  ...  0.783  0.886\n",
              "1          SNF_AdaBoost_cosine_sim_lap     0.861  ...  0.861  0.924\n",
              "2          SNF_AdaBoost_cosine_sim_lle     0.850  ...  0.850  0.918\n",
              "3   SNF_AdaBoost_Pearson_corr_sim_hope     0.772  ...  0.772  0.882\n",
              "4    SNF_AdaBoost_Pearson_corr_sim_lap     0.906  ...  0.906  0.953\n",
              "5    SNF_AdaBoost_Pearson_corr_sim_lle     0.861  ...  0.861  0.916\n",
              "6          SNF_hope_cosine_sim_XGBoost     0.739  ...  0.907  0.923\n",
              "7           SNF_lap_cosine_sim_XGBoost     0.839  ...  0.949  0.959\n",
              "8           SNF_lle_cosine_sim_XGBoost     0.867  ...  0.904  0.912\n",
              "9    SNF_hope_Pearson_corr_sim_XGBoost     0.733  ...  0.904  0.920\n",
              "10    SNF_lap_Pearson_corr_sim_XGBoost     0.861  ...  0.960  0.972\n",
              "11    SNF_lle_Pearson_corr_sim_XGBoost     0.844  ...  0.953  0.951\n",
              "12             SNF_ANN_cosine_sim_hope     0.633  ...  0.744  0.825\n",
              "13              SNF_ANN_cosine_sim_lap     0.494  ...  0.944  0.962\n",
              "14              SNF_ANN_cosine_sim_lle     0.500  ...  0.949  0.969\n",
              "15       SNF_ANN_Pearson_corr_sim_hope     0.883  ...  0.946  0.962\n",
              "16        SNF_ANN_Pearson_corr_sim_lap     0.933  ...  0.944  0.967\n",
              "17        SNF_ANN_Pearson_corr_sim_lle     0.922  ...  0.946  0.967\n",
              "18        AVG_AdaBoost_cosine_sim_hope     0.817  ...  0.817  0.906\n",
              "19         AVG_AdaBoost_cosine_sim_lap     0.839  ...  0.839  0.915\n",
              "20         AVG_AdaBoost_cosine_sim_lle     0.878  ...  0.878  0.939\n",
              "21  AVG_AdaBoost_Pearson_corr_sim_hope     0.806  ...  0.806  0.898\n",
              "22   AVG_AdaBoost_Pearson_corr_sim_lap     0.867  ...  0.867  0.924\n",
              "23   AVG_AdaBoost_Pearson_corr_sim_lle     0.889  ...  0.889  0.944\n",
              "24         AVG_XGBoost_cosine_sim_hope     0.744  ...  0.915  0.933\n",
              "25          AVG_XGBoost_cosine_sim_lap     0.839  ...  0.946  0.961\n",
              "26          AVG_XGBoost_cosine_sim_lle     0.872  ...  0.965  0.970\n",
              "27   AVG_XGBoost_Pearson_corr_sim_hope     0.744  ...  0.937  0.944\n",
              "28    AVG_XGBoost_Pearson_corr_sim_lap     0.828  ...  0.961  0.965\n",
              "29    AVG_XGBoost_Pearson_corr_sim_lle     0.839  ...  0.978  0.981\n",
              "30             AVG_ANN_cosine_sim_hope     0.606  ...  0.891  0.928\n",
              "31              AVG_ANN_cosine_sim_lap     0.500  ...  0.934  0.961\n",
              "32              AVG_ANN_cosine_sim_lle     0.500  ...  0.965  0.979\n",
              "33       AVG_ANN_Pearson_corr_sim_hope     0.900  ...  0.965  0.973\n",
              "34        AVG_ANN_Pearson_corr_sim_lap     0.922  ...  0.940  0.963\n",
              "35        AVG_ANN_Pearson_corr_sim_lle     0.922  ...  0.938  0.959\n",
              "36        MAX_AdaBoost_cosine_sim_hope     0.806  ...  0.806  0.903\n",
              "37         MAX_AdaBoost_cosine_sim_lap     0.794  ...  0.794  0.897\n",
              "38         MAX_AdaBoost_cosine_sim_lle     0.778  ...  0.778  0.883\n",
              "39  MAX_AdaBoost_Pearson_corr_sim_hope     0.761  ...  0.761  0.876\n",
              "40   MAX_AdaBoost_Pearson_corr_sim_lap     0.794  ...  0.794  0.892\n",
              "41   MAX_AdaBoost_Pearson_corr_sim_lle     0.767  ...  0.767  0.872\n",
              "42         MAX_XGBoost_cosine_sim_hope     0.744  ...  0.900  0.923\n",
              "43          MAX_XGBoost_cosine_sim_lap     0.789  ...  0.901  0.925\n",
              "44          MAX_XGBoost_cosine_sim_lle     0.756  ...  0.917  0.934\n",
              "45   MAX_XGBoost_Pearson_corr_sim_hope     0.739  ...  0.932  0.943\n",
              "46    MAX_XGBoost_Pearson_corr_sim_lap     0.767  ...  0.920  0.935\n",
              "47    MAX_XGBoost_Pearson_corr_sim_lle     0.750  ...  0.924  0.936\n",
              "48             MAX_ANN_cosine_sim_hope     0.606  ...  0.954  0.970\n",
              "49              MAX_ANN_cosine_sim_lap     0.500  ...  0.931  0.950\n",
              "50              MAX_ANN_cosine_sim_lle     0.500  ...  0.948  0.966\n",
              "51       MAX_ANN_Pearson_corr_sim_hope     0.939  ...  0.975  0.983\n",
              "52        MAX_ANN_Pearson_corr_sim_lap     0.894  ...  0.973  0.980\n",
              "53        MAX_ANN_Pearson_corr_sim_lle     0.900  ...  0.968  0.976\n",
              "\n",
              "[54 rows x 7 columns]"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "RufO5aGLe-C3",
        "outputId": "33e07c5b-7729-4d58-c373-848db190e7cc"
      },
      "source": [
        "sorted_df_by_f1 = df.sort_values(['f1'], ascending=False)\n",
        "display(sorted_df_by_f1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "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>model</th>\n",
              "      <th>accuracy</th>\n",
              "      <th>precision</th>\n",
              "      <th>recall</th>\n",
              "      <th>f1</th>\n",
              "      <th>auc</th>\n",
              "      <th>aupr</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>51</th>\n",
              "      <td>MAX_ANN_Pearson_corr_sim_hope</td>\n",
              "      <td>0.939</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.878</td>\n",
              "      <td>0.932</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.983</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>SNF_ANN_Pearson_corr_sim_lap</td>\n",
              "      <td>0.933</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.867</td>\n",
              "      <td>0.925</td>\n",
              "      <td>0.944</td>\n",
              "      <td>0.967</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>SNF_ANN_Pearson_corr_sim_lle</td>\n",
              "      <td>0.922</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.844</td>\n",
              "      <td>0.913</td>\n",
              "      <td>0.946</td>\n",
              "      <td>0.967</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>35</th>\n",
              "      <td>AVG_ANN_Pearson_corr_sim_lle</td>\n",
              "      <td>0.922</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.844</td>\n",
              "      <td>0.911</td>\n",
              "      <td>0.938</td>\n",
              "      <td>0.959</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>34</th>\n",
              "      <td>AVG_ANN_Pearson_corr_sim_lap</td>\n",
              "      <td>0.922</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.844</td>\n",
              "      <td>0.911</td>\n",
              "      <td>0.940</td>\n",
              "      <td>0.963</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>SNF_AdaBoost_Pearson_corr_sim_lap</td>\n",
              "      <td>0.906</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.811</td>\n",
              "      <td>0.892</td>\n",
              "      <td>0.906</td>\n",
              "      <td>0.953</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>33</th>\n",
              "      <td>AVG_ANN_Pearson_corr_sim_hope</td>\n",
              "      <td>0.900</td>\n",
              "      <td>0.989</td>\n",
              "      <td>0.811</td>\n",
              "      <td>0.887</td>\n",
              "      <td>0.965</td>\n",
              "      <td>0.973</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>53</th>\n",
              "      <td>MAX_ANN_Pearson_corr_sim_lle</td>\n",
              "      <td>0.900</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.800</td>\n",
              "      <td>0.885</td>\n",
              "      <td>0.968</td>\n",
              "      <td>0.976</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>52</th>\n",
              "      <td>MAX_ANN_Pearson_corr_sim_lap</td>\n",
              "      <td>0.894</td>\n",
              "      <td>0.986</td>\n",
              "      <td>0.800</td>\n",
              "      <td>0.880</td>\n",
              "      <td>0.973</td>\n",
              "      <td>0.980</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>AVG_AdaBoost_Pearson_corr_sim_lle</td>\n",
              "      <td>0.889</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.778</td>\n",
              "      <td>0.868</td>\n",
              "      <td>0.889</td>\n",
              "      <td>0.944</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>SNF_ANN_Pearson_corr_sim_hope</td>\n",
              "      <td>0.883</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.767</td>\n",
              "      <td>0.861</td>\n",
              "      <td>0.946</td>\n",
              "      <td>0.962</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>AVG_AdaBoost_cosine_sim_lle</td>\n",
              "      <td>0.878</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.756</td>\n",
              "      <td>0.856</td>\n",
              "      <td>0.878</td>\n",
              "      <td>0.939</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>AVG_XGBoost_cosine_sim_lle</td>\n",
              "      <td>0.872</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.744</td>\n",
              "      <td>0.850</td>\n",
              "      <td>0.965</td>\n",
              "      <td>0.970</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>SNF_lle_cosine_sim_XGBoost</td>\n",
              "      <td>0.867</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.756</td>\n",
              "      <td>0.847</td>\n",
              "      <td>0.904</td>\n",
              "      <td>0.912</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>SNF_AdaBoost_Pearson_corr_sim_lle</td>\n",
              "      <td>0.861</td>\n",
              "      <td>0.950</td>\n",
              "      <td>0.767</td>\n",
              "      <td>0.841</td>\n",
              "      <td>0.861</td>\n",
              "      <td>0.916</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>AVG_AdaBoost_Pearson_corr_sim_lap</td>\n",
              "      <td>0.867</td>\n",
              "      <td>0.965</td>\n",
              "      <td>0.767</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.867</td>\n",
              "      <td>0.924</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>SNF_AdaBoost_cosine_sim_lap</td>\n",
              "      <td>0.861</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.744</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.861</td>\n",
              "      <td>0.924</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>SNF_lap_Pearson_corr_sim_XGBoost</td>\n",
              "      <td>0.861</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.722</td>\n",
              "      <td>0.834</td>\n",
              "      <td>0.960</td>\n",
              "      <td>0.972</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>SNF_AdaBoost_cosine_sim_lle</td>\n",
              "      <td>0.850</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.722</td>\n",
              "      <td>0.822</td>\n",
              "      <td>0.850</td>\n",
              "      <td>0.918</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>SNF_lle_Pearson_corr_sim_XGBoost</td>\n",
              "      <td>0.844</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.711</td>\n",
              "      <td>0.810</td>\n",
              "      <td>0.953</td>\n",
              "      <td>0.951</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>SNF_lap_cosine_sim_XGBoost</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.700</td>\n",
              "      <td>0.809</td>\n",
              "      <td>0.949</td>\n",
              "      <td>0.959</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>AVG_XGBoost_cosine_sim_lap</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.971</td>\n",
              "      <td>0.700</td>\n",
              "      <td>0.807</td>\n",
              "      <td>0.946</td>\n",
              "      <td>0.961</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>AVG_AdaBoost_cosine_sim_lap</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.986</td>\n",
              "      <td>0.689</td>\n",
              "      <td>0.805</td>\n",
              "      <td>0.839</td>\n",
              "      <td>0.915</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29</th>\n",
              "      <td>AVG_XGBoost_Pearson_corr_sim_lle</td>\n",
              "      <td>0.839</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.678</td>\n",
              "      <td>0.800</td>\n",
              "      <td>0.978</td>\n",
              "      <td>0.981</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28</th>\n",
              "      <td>AVG_XGBoost_Pearson_corr_sim_lap</td>\n",
              "      <td>0.828</td>\n",
              "      <td>0.975</td>\n",
              "      <td>0.678</td>\n",
              "      <td>0.790</td>\n",
              "      <td>0.961</td>\n",
              "      <td>0.965</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>AVG_AdaBoost_cosine_sim_hope</td>\n",
              "      <td>0.817</td>\n",
              "      <td>0.990</td>\n",
              "      <td>0.644</td>\n",
              "      <td>0.761</td>\n",
              "      <td>0.817</td>\n",
              "      <td>0.906</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>36</th>\n",
              "      <td>MAX_AdaBoost_cosine_sim_hope</td>\n",
              "      <td>0.806</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.611</td>\n",
              "      <td>0.749</td>\n",
              "      <td>0.806</td>\n",
              "      <td>0.903</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>AVG_AdaBoost_Pearson_corr_sim_hope</td>\n",
              "      <td>0.806</td>\n",
              "      <td>0.980</td>\n",
              "      <td>0.633</td>\n",
              "      <td>0.746</td>\n",
              "      <td>0.806</td>\n",
              "      <td>0.898</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>40</th>\n",
              "      <td>MAX_AdaBoost_Pearson_corr_sim_lap</td>\n",
              "      <td>0.794</td>\n",
              "      <td>0.983</td>\n",
              "      <td>0.600</td>\n",
              "      <td>0.738</td>\n",
              "      <td>0.794</td>\n",
              "      <td>0.892</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>37</th>\n",
              "      <td>MAX_AdaBoost_cosine_sim_lap</td>\n",
              "      <td>0.794</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.589</td>\n",
              "      <td>0.734</td>\n",
              "      <td>0.794</td>\n",
              "      <td>0.897</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>48</th>\n",
              "      <td>MAX_ANN_cosine_sim_hope</td>\n",
              "      <td>0.606</td>\n",
              "      <td>0.590</td>\n",
              "      <td>0.989</td>\n",
              "      <td>0.727</td>\n",
              "      <td>0.954</td>\n",
              "      <td>0.970</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>43</th>\n",
              "      <td>MAX_XGBoost_cosine_sim_lap</td>\n",
              "      <td>0.789</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.578</td>\n",
              "      <td>0.724</td>\n",
              "      <td>0.901</td>\n",
              "      <td>0.925</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>SNF_AdaBoost_cosine_sim_hope</td>\n",
              "      <td>0.783</td>\n",
              "      <td>0.983</td>\n",
              "      <td>0.578</td>\n",
              "      <td>0.720</td>\n",
              "      <td>0.783</td>\n",
              "      <td>0.886</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>38</th>\n",
              "      <td>MAX_AdaBoost_cosine_sim_lle</td>\n",
              "      <td>0.778</td>\n",
              "      <td>0.983</td>\n",
              "      <td>0.567</td>\n",
              "      <td>0.715</td>\n",
              "      <td>0.778</td>\n",
              "      <td>0.883</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>SNF_AdaBoost_Pearson_corr_sim_hope</td>\n",
              "      <td>0.772</td>\n",
              "      <td>0.986</td>\n",
              "      <td>0.556</td>\n",
              "      <td>0.704</td>\n",
              "      <td>0.772</td>\n",
              "      <td>0.882</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>41</th>\n",
              "      <td>MAX_AdaBoost_Pearson_corr_sim_lle</td>\n",
              "      <td>0.767</td>\n",
              "      <td>0.966</td>\n",
              "      <td>0.556</td>\n",
              "      <td>0.701</td>\n",
              "      <td>0.767</td>\n",
              "      <td>0.872</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>AVG_ANN_cosine_sim_hope</td>\n",
              "      <td>0.606</td>\n",
              "      <td>0.597</td>\n",
              "      <td>0.889</td>\n",
              "      <td>0.697</td>\n",
              "      <td>0.891</td>\n",
              "      <td>0.928</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>46</th>\n",
              "      <td>MAX_XGBoost_Pearson_corr_sim_lap</td>\n",
              "      <td>0.767</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.533</td>\n",
              "      <td>0.688</td>\n",
              "      <td>0.920</td>\n",
              "      <td>0.935</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>39</th>\n",
              "      <td>MAX_AdaBoost_Pearson_corr_sim_hope</td>\n",
              "      <td>0.761</td>\n",
              "      <td>0.986</td>\n",
              "      <td>0.533</td>\n",
              "      <td>0.684</td>\n",
              "      <td>0.761</td>\n",
              "      <td>0.876</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44</th>\n",
              "      <td>MAX_XGBoost_cosine_sim_lle</td>\n",
              "      <td>0.756</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.511</td>\n",
              "      <td>0.672</td>\n",
              "      <td>0.917</td>\n",
              "      <td>0.934</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>32</th>\n",
              "      <td>AVG_ANN_cosine_sim_lle</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.500</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.965</td>\n",
              "      <td>0.979</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>SNF_ANN_cosine_sim_lle</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.500</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.949</td>\n",
              "      <td>0.969</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50</th>\n",
              "      <td>MAX_ANN_cosine_sim_lle</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.500</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.948</td>\n",
              "      <td>0.966</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>31</th>\n",
              "      <td>AVG_ANN_cosine_sim_lap</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.500</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.934</td>\n",
              "      <td>0.961</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>49</th>\n",
              "      <td>MAX_ANN_cosine_sim_lap</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.500</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.931</td>\n",
              "      <td>0.950</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>47</th>\n",
              "      <td>MAX_XGBoost_Pearson_corr_sim_lle</td>\n",
              "      <td>0.750</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.664</td>\n",
              "      <td>0.924</td>\n",
              "      <td>0.936</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>SNF_ANN_cosine_sim_lap</td>\n",
              "      <td>0.494</td>\n",
              "      <td>0.497</td>\n",
              "      <td>0.989</td>\n",
              "      <td>0.662</td>\n",
              "      <td>0.944</td>\n",
              "      <td>0.962</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>SNF_ANN_cosine_sim_hope</td>\n",
              "      <td>0.633</td>\n",
              "      <td>0.645</td>\n",
              "      <td>0.667</td>\n",
              "      <td>0.651</td>\n",
              "      <td>0.744</td>\n",
              "      <td>0.825</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>42</th>\n",
              "      <td>MAX_XGBoost_cosine_sim_hope</td>\n",
              "      <td>0.744</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.489</td>\n",
              "      <td>0.648</td>\n",
              "      <td>0.900</td>\n",
              "      <td>0.923</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24</th>\n",
              "      <td>AVG_XGBoost_cosine_sim_hope</td>\n",
              "      <td>0.744</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.489</td>\n",
              "      <td>0.648</td>\n",
              "      <td>0.915</td>\n",
              "      <td>0.933</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27</th>\n",
              "      <td>AVG_XGBoost_Pearson_corr_sim_hope</td>\n",
              "      <td>0.744</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.489</td>\n",
              "      <td>0.648</td>\n",
              "      <td>0.937</td>\n",
              "      <td>0.944</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>45</th>\n",
              "      <td>MAX_XGBoost_Pearson_corr_sim_hope</td>\n",
              "      <td>0.739</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.478</td>\n",
              "      <td>0.638</td>\n",
              "      <td>0.932</td>\n",
              "      <td>0.943</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>SNF_hope_cosine_sim_XGBoost</td>\n",
              "      <td>0.739</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.478</td>\n",
              "      <td>0.638</td>\n",
              "      <td>0.907</td>\n",
              "      <td>0.923</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>SNF_hope_Pearson_corr_sim_XGBoost</td>\n",
              "      <td>0.733</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.467</td>\n",
              "      <td>0.628</td>\n",
              "      <td>0.904</td>\n",
              "      <td>0.920</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                 model  accuracy  ...    auc   aupr\n",
              "51       MAX_ANN_Pearson_corr_sim_hope     0.939  ...  0.975  0.983\n",
              "16        SNF_ANN_Pearson_corr_sim_lap     0.933  ...  0.944  0.967\n",
              "17        SNF_ANN_Pearson_corr_sim_lle     0.922  ...  0.946  0.967\n",
              "35        AVG_ANN_Pearson_corr_sim_lle     0.922  ...  0.938  0.959\n",
              "34        AVG_ANN_Pearson_corr_sim_lap     0.922  ...  0.940  0.963\n",
              "4    SNF_AdaBoost_Pearson_corr_sim_lap     0.906  ...  0.906  0.953\n",
              "33       AVG_ANN_Pearson_corr_sim_hope     0.900  ...  0.965  0.973\n",
              "53        MAX_ANN_Pearson_corr_sim_lle     0.900  ...  0.968  0.976\n",
              "52        MAX_ANN_Pearson_corr_sim_lap     0.894  ...  0.973  0.980\n",
              "23   AVG_AdaBoost_Pearson_corr_sim_lle     0.889  ...  0.889  0.944\n",
              "15       SNF_ANN_Pearson_corr_sim_hope     0.883  ...  0.946  0.962\n",
              "20         AVG_AdaBoost_cosine_sim_lle     0.878  ...  0.878  0.939\n",
              "26          AVG_XGBoost_cosine_sim_lle     0.872  ...  0.965  0.970\n",
              "8           SNF_lle_cosine_sim_XGBoost     0.867  ...  0.904  0.912\n",
              "5    SNF_AdaBoost_Pearson_corr_sim_lle     0.861  ...  0.861  0.916\n",
              "22   AVG_AdaBoost_Pearson_corr_sim_lap     0.867  ...  0.867  0.924\n",
              "1          SNF_AdaBoost_cosine_sim_lap     0.861  ...  0.861  0.924\n",
              "10    SNF_lap_Pearson_corr_sim_XGBoost     0.861  ...  0.960  0.972\n",
              "2          SNF_AdaBoost_cosine_sim_lle     0.850  ...  0.850  0.918\n",
              "11    SNF_lle_Pearson_corr_sim_XGBoost     0.844  ...  0.953  0.951\n",
              "7           SNF_lap_cosine_sim_XGBoost     0.839  ...  0.949  0.959\n",
              "25          AVG_XGBoost_cosine_sim_lap     0.839  ...  0.946  0.961\n",
              "19         AVG_AdaBoost_cosine_sim_lap     0.839  ...  0.839  0.915\n",
              "29    AVG_XGBoost_Pearson_corr_sim_lle     0.839  ...  0.978  0.981\n",
              "28    AVG_XGBoost_Pearson_corr_sim_lap     0.828  ...  0.961  0.965\n",
              "18        AVG_AdaBoost_cosine_sim_hope     0.817  ...  0.817  0.906\n",
              "36        MAX_AdaBoost_cosine_sim_hope     0.806  ...  0.806  0.903\n",
              "21  AVG_AdaBoost_Pearson_corr_sim_hope     0.806  ...  0.806  0.898\n",
              "40   MAX_AdaBoost_Pearson_corr_sim_lap     0.794  ...  0.794  0.892\n",
              "37         MAX_AdaBoost_cosine_sim_lap     0.794  ...  0.794  0.897\n",
              "48             MAX_ANN_cosine_sim_hope     0.606  ...  0.954  0.970\n",
              "43          MAX_XGBoost_cosine_sim_lap     0.789  ...  0.901  0.925\n",
              "0         SNF_AdaBoost_cosine_sim_hope     0.783  ...  0.783  0.886\n",
              "38         MAX_AdaBoost_cosine_sim_lle     0.778  ...  0.778  0.883\n",
              "3   SNF_AdaBoost_Pearson_corr_sim_hope     0.772  ...  0.772  0.882\n",
              "41   MAX_AdaBoost_Pearson_corr_sim_lle     0.767  ...  0.767  0.872\n",
              "30             AVG_ANN_cosine_sim_hope     0.606  ...  0.891  0.928\n",
              "46    MAX_XGBoost_Pearson_corr_sim_lap     0.767  ...  0.920  0.935\n",
              "39  MAX_AdaBoost_Pearson_corr_sim_hope     0.761  ...  0.761  0.876\n",
              "44          MAX_XGBoost_cosine_sim_lle     0.756  ...  0.917  0.934\n",
              "32              AVG_ANN_cosine_sim_lle     0.500  ...  0.965  0.979\n",
              "14              SNF_ANN_cosine_sim_lle     0.500  ...  0.949  0.969\n",
              "50              MAX_ANN_cosine_sim_lle     0.500  ...  0.948  0.966\n",
              "31              AVG_ANN_cosine_sim_lap     0.500  ...  0.934  0.961\n",
              "49              MAX_ANN_cosine_sim_lap     0.500  ...  0.931  0.950\n",
              "47    MAX_XGBoost_Pearson_corr_sim_lle     0.750  ...  0.924  0.936\n",
              "13              SNF_ANN_cosine_sim_lap     0.494  ...  0.944  0.962\n",
              "12             SNF_ANN_cosine_sim_hope     0.633  ...  0.744  0.825\n",
              "42         MAX_XGBoost_cosine_sim_hope     0.744  ...  0.900  0.923\n",
              "24         AVG_XGBoost_cosine_sim_hope     0.744  ...  0.915  0.933\n",
              "27   AVG_XGBoost_Pearson_corr_sim_hope     0.744  ...  0.937  0.944\n",
              "45   MAX_XGBoost_Pearson_corr_sim_hope     0.739  ...  0.932  0.943\n",
              "6          SNF_hope_cosine_sim_XGBoost     0.739  ...  0.907  0.923\n",
              "9    SNF_hope_Pearson_corr_sim_XGBoost     0.733  ...  0.904  0.920\n",
              "\n",
              "[54 rows x 7 columns]"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 457
        },
        "id": "FDKsByTQlilJ",
        "outputId": "135783a4-09dd-4aec-b12f-5f3cb8e5a652"
      },
      "source": [
        "df_first_5 = sorted_df_by_f1.head(5)\n",
        "df_first_5 = df_first_5[['model', 'f1']]\n",
        "df_first_5.plot(x='model', y='f1', kind='bar', xlabel='Model', ylabel='F1-score')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f7b90718d10>"
            ]
          },
          "metadata": {},
          "execution_count": 16
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}