Ranking protein-protein interfaces using Graph Convolution.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "6f137f93",
   "metadata": {},
   "source": [
    "# Demo for ranking protein-protein interfaces using Graph Convolution"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c6ee0e26",
   "metadata": {},
   "source": [
    "## Project setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b4218de3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Copying data and code from GitHub repository\n",
    "!git clone https://github.com/withai/Demo-Ranking-protein-protein-interfaces-using-GNN.git\n",
    "!mv Demo-Ranking-protein-protein-interfaces-using-GNN/* .\n",
    "!rm -r Demo-Ranking-protein-protein-interfaces-using-GNN"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ffbac141",
   "metadata": {},
   "source": [
    "## Import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "6d89e241",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import PyTorch\n",
    "import torch\n",
    "import torch.backends.cudnn as cudnn\n",
    "from torch.utils import data\n",
    "import torch.nn.functional as F\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from torch.utils import data\n",
    "import torch.nn.init as init\n",
    "\n",
    "# Import Neural Network layer\n",
    "from nn import Dense\n",
    "\n",
    "# Import utility functions\n",
    "from dataset import Dataset\n",
    "from sampler import CustomSampler\n",
    "from train import train\n",
    "from test import test\n",
    "\n",
    "# Import libraries for graphing\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from graph import show_results\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3cb4db1a",
   "metadata": {},
   "source": [
    "## Run computation on GPU or CPU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "5ca523ac",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "use_cuda = torch.cuda.is_available()\n",
    "device = torch.device(\"cuda:0\" if use_cuda else \"cpu\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "039ba253",
   "metadata": {},
   "source": [
    "## Define Graph Convolution layer"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "34e5920f",
   "metadata": {},
   "source": [
    "$$\n",
    "h^{(l)}_i = \\sigma \\left(\\mathbf{W}^{(l)}_c h^{(l-1)}_i +\n",
    "    \\frac{1}{|N_i^{(int)}|}\\sum_{j \\epsilon N_i^{(int)}} \\mathbf{W}^{(l)}_{int} h^{(l-1)}_j \\bigodot E_{ij} \\right.  \\left. + \\frac{1}{|N_i^{(prot)}|}\\sum_{j \\epsilon N_i^{(prot)}} \\mathbf{W}^{(l)}_{prot} h^{(l-1)}_j \\bigodot E_{ij} + b^{(l)} \\vphantom{\\frac{1}{|N_i^{(neigh)}|}}\\right)\n",
    "$$\n",
    "\n",
    "\n",
    "where for the layer $l$,  \n",
    "\n",
    "$\\mathbf{W}^{(l)}_c$ is the weight matrix for learning representations w.r.t central node,\n",
    "\n",
    "$\\mathbf{W}^{(l)}_{int}$ and $\\mathbf{W}^{(l)}_{prot}$ are the weight matrix for learning representations w.r.t the neighbor nodes in the interface graph and the protein graph respectively,\n",
    "\n",
    "$b^{(l)}$ is the bias vector, \n",
    "\n",
    "$h^{(l)}_i$ is the hidden node embedding of node $i$,\n",
    "\n",
    "$N_i^{(int)}$ is the number of atoms surrounding node $i$ in the interface graph, \n",
    "\n",
    "$N_i^{(prot)}$ is the number of atoms surrounding node $i$ in the protein graph, \n",
    "\n",
    "$\\sigma$  is the activation function, and\n",
    "\n",
    "$E_{ij}$ is the value in the edge matrix between node $i$ and node $j$."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0acd3872",
   "metadata": {},
   "source": [
    "### How does computing neighborhood signals look like in the code?\n",
    "\n",
    "Let's compute neighborhood signals in the interface graph\n",
    "\n",
    "$\n",
    "\\begin{align}\n",
    "\\frac{1}{|N_i^{(int)}|}\\sum_{j \\epsilon N_i^{(int)}} \\mathbf{W}^{(l)}_{int} h^{(l-1)}_j \\bigodot E_{ij} \n",
    "\\end{align}\n",
    "$"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "62c3f487",
   "metadata": {},
   "source": [
    "**Define all the variables**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "65a0759a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# node features: One-hot encoding of 11 different atom types\n",
    "node_feat = torch.tensor([\n",
    "                            [0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0], # atom 1\n",
    "                            [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1], # atom 2\n",
    "                            [0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0], # atom 3\n",
    "                            [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], # atom 4 \n",
    "                        ]).type(torch.float)\n",
    "\n",
    "# indices of neighbor atoms in interface graph\n",
    "int_indices = torch.tensor([\n",
    "                            [1, 2, -1],  # neighbors of atom 1\n",
    "                            [0, -1, -1], # neighbors of atom 2\n",
    "                            [1, -1, -1], # neighbors of atom 3\n",
    "                            [-1, -1, -1] # neighbors of atom 4\n",
    "                          ])\n",
    "\n",
    "# edge weights of neighbor atoms in interface graph\n",
    "int_edges = torch.tensor([\n",
    "                            [0.9, 0.2, -1], # weights of neighbors of atom 1\n",
    "                            [0.1, -1, -1],  # weights of neighbors of atom 2\n",
    "                            [0.5, -1, -1],  # weights of neighbors of atom 3\n",
    "                            [-1, -1, -1]    # weights of neighbors of atom 4\n",
    "                        ])\n",
    "\n",
    "# weight matrix\n",
    "num_of_features = 11\n",
    "filters = 16\n",
    "\n",
    "w_int = torch.randn(num_of_features, filters)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3f5c4bd2",
   "metadata": {},
   "source": [
    "**Step 1: Compute signals for all the nodes in the graph**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "83d73ead",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 16])\n",
      "tensor([[-0.3941, -1.2574, -0.6091,  1.0736, -0.9058, -1.7475,  0.9353, -0.2424,\n",
      "          0.6113,  0.5230,  0.2550, -0.2047, -1.3506,  0.5298, -0.4015,  0.0150],\n",
      "        [ 1.6465,  0.4131,  0.0321,  0.8855, -0.2912,  0.7838, -0.4488, -1.3473,\n",
      "         -1.0906,  1.0139,  0.4203,  0.1642,  1.2011, -0.7894,  0.6780,  0.8977],\n",
      "        [-0.9270, -0.7774, -1.2629,  1.7078,  0.5497,  1.8573, -0.0765,  0.0360,\n",
      "          2.3269, -2.7075, -1.7738,  0.1427, -0.0294, -0.2745,  0.2629, -0.5668],\n",
      "        [ 0.0925,  0.8056, -0.4443,  0.7264, -0.4592, -0.4714, -0.8235,  0.0877,\n",
      "          0.3983,  0.3654,  1.1938, -0.9651, -0.7113,  0.4504,  0.0540,  1.1967]])\n"
     ]
    }
   ],
   "source": [
    "node_Wn_int = torch.matmul(node_feat, w_int)\n",
    "\n",
    "print(node_Wn_int.shape)\n",
    "print(node_Wn_int)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4885a83b",
   "metadata": {},
   "source": [
    "**Step 2: Find all the neighbors of atoms that are participating in the interface graph**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "52a4f8d7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 3, 1])\n",
      "tensor([[[1.],\n",
      "         [1.],\n",
      "         [0.]],\n",
      "\n",
      "        [[1.],\n",
      "         [0.],\n",
      "         [0.]],\n",
      "\n",
      "        [[1.],\n",
      "         [0.],\n",
      "         [0.]],\n",
      "\n",
      "        [[0.],\n",
      "         [0.],\n",
      "         [0.]]])\n"
     ]
    }
   ],
   "source": [
    "int_part = torch.unsqueeze(int_indices > -1, 2).type(torch.float)\n",
    "\n",
    "print(int_part.shape)\n",
    "print(int_part)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da652e32",
   "metadata": {},
   "source": [
    "**Step 3: For each atom find the number of neighbors participating in the graph**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "70d29ea0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 1])\n",
      "tensor([[2],\n",
      "        [1],\n",
      "        [1],\n",
      "        [0]])\n"
     ]
    }
   ],
   "source": [
    "no_of_neighs_int = torch.sum(int_indices > -1, 1).unsqueeze(1)\n",
    "\n",
    "print(no_of_neighs_int.shape)\n",
    "print(no_of_neighs_int)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e5519efa",
   "metadata": {},
   "source": [
    "**Step 4a: Make sure we don't divide by zeros**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b7863a54",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 1])\n",
      "tensor([[2],\n",
      "        [1],\n",
      "        [1],\n",
      "        [1]])\n"
     ]
    }
   ],
   "source": [
    "no_of_neighs_int[no_of_neighs_int == 0] = 1\n",
    "\n",
    "print(no_of_neighs_int.shape)\n",
    "print(no_of_neighs_int)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "81a1d729",
   "metadata": {},
   "source": [
    "**Step 4b: Expand the dimension of edge weights to allow scalar multiplication**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "5c0c95b8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 3, 1])\n",
      "tensor([[[ 0.9000],\n",
      "         [ 0.2000],\n",
      "         [-1.0000]],\n",
      "\n",
      "        [[ 0.1000],\n",
      "         [-1.0000],\n",
      "         [-1.0000]],\n",
      "\n",
      "        [[ 0.5000],\n",
      "         [-1.0000],\n",
      "         [-1.0000]],\n",
      "\n",
      "        [[-1.0000],\n",
      "         [-1.0000],\n",
      "         [-1.0000]]])\n"
     ]
    }
   ],
   "source": [
    "int_edges = torch.unsqueeze(int_edges, 2)\n",
    "\n",
    "print(int_edges.shape)\n",
    "print(int_edges)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "19216100",
   "metadata": {},
   "source": [
    "**Step 5: Compute neighbor signals of all the nodes in the interface graph**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b33ab2ca",
   "metadata": {},
   "source": [
    "Step 5a: Grab all the neighbors of interface atoms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "7d71f3cb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 3, 16])\n",
      "tensor([[[ 1.6465,  0.4131,  0.0321,  0.8855, -0.2912,  0.7838, -0.4488,\n",
      "          -1.3473, -1.0906,  1.0139,  0.4203,  0.1642,  1.2011, -0.7894,\n",
      "           0.6780,  0.8977],\n",
      "         [-0.9270, -0.7774, -1.2629,  1.7078,  0.5497,  1.8573, -0.0765,\n",
      "           0.0360,  2.3269, -2.7075, -1.7738,  0.1427, -0.0294, -0.2745,\n",
      "           0.2629, -0.5668],\n",
      "         [ 0.0925,  0.8056, -0.4443,  0.7264, -0.4592, -0.4714, -0.8235,\n",
      "           0.0877,  0.3983,  0.3654,  1.1938, -0.9651, -0.7113,  0.4504,\n",
      "           0.0540,  1.1967]],\n",
      "\n",
      "        [[-0.3941, -1.2574, -0.6091,  1.0736, -0.9058, -1.7475,  0.9353,\n",
      "          -0.2424,  0.6113,  0.5230,  0.2550, -0.2047, -1.3506,  0.5298,\n",
      "          -0.4015,  0.0150],\n",
      "         [ 0.0925,  0.8056, -0.4443,  0.7264, -0.4592, -0.4714, -0.8235,\n",
      "           0.0877,  0.3983,  0.3654,  1.1938, -0.9651, -0.7113,  0.4504,\n",
      "           0.0540,  1.1967],\n",
      "         [ 0.0925,  0.8056, -0.4443,  0.7264, -0.4592, -0.4714, -0.8235,\n",
      "           0.0877,  0.3983,  0.3654,  1.1938, -0.9651, -0.7113,  0.4504,\n",
      "           0.0540,  1.1967]],\n",
      "\n",
      "        [[ 1.6465,  0.4131,  0.0321,  0.8855, -0.2912,  0.7838, -0.4488,\n",
      "          -1.3473, -1.0906,  1.0139,  0.4203,  0.1642,  1.2011, -0.7894,\n",
      "           0.6780,  0.8977],\n",
      "         [ 0.0925,  0.8056, -0.4443,  0.7264, -0.4592, -0.4714, -0.8235,\n",
      "           0.0877,  0.3983,  0.3654,  1.1938, -0.9651, -0.7113,  0.4504,\n",
      "           0.0540,  1.1967],\n",
      "         [ 0.0925,  0.8056, -0.4443,  0.7264, -0.4592, -0.4714, -0.8235,\n",
      "           0.0877,  0.3983,  0.3654,  1.1938, -0.9651, -0.7113,  0.4504,\n",
      "           0.0540,  1.1967]],\n",
      "\n",
      "        [[ 0.0925,  0.8056, -0.4443,  0.7264, -0.4592, -0.4714, -0.8235,\n",
      "           0.0877,  0.3983,  0.3654,  1.1938, -0.9651, -0.7113,  0.4504,\n",
      "           0.0540,  1.1967],\n",
      "         [ 0.0925,  0.8056, -0.4443,  0.7264, -0.4592, -0.4714, -0.8235,\n",
      "           0.0877,  0.3983,  0.3654,  1.1938, -0.9651, -0.7113,  0.4504,\n",
      "           0.0540,  1.1967],\n",
      "         [ 0.0925,  0.8056, -0.4443,  0.7264, -0.4592, -0.4714, -0.8235,\n",
      "           0.0877,  0.3983,  0.3654,  1.1938, -0.9651, -0.7113,  0.4504,\n",
      "           0.0540,  1.1967]]])\n"
     ]
    }
   ],
   "source": [
    "Zn_inter_step = node_Wn_int[int_indices]\n",
    "\n",
    "print(Zn_inter_step.shape)\n",
    "print(Zn_inter_step)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b4f89eba",
   "metadata": {},
   "source": [
    "Step 5b: Select all the participating neighbors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "a0b6a1a7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 3, 16])\n",
      "tensor([[[ 1.6465,  0.4131,  0.0321,  0.8855, -0.2912,  0.7838, -0.4488,\n",
      "          -1.3473, -1.0906,  1.0139,  0.4203,  0.1642,  1.2011, -0.7894,\n",
      "           0.6780,  0.8977],\n",
      "         [-0.9270, -0.7774, -1.2629,  1.7078,  0.5497,  1.8573, -0.0765,\n",
      "           0.0360,  2.3269, -2.7075, -1.7738,  0.1427, -0.0294, -0.2745,\n",
      "           0.2629, -0.5668],\n",
      "         [ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000, -0.0000, -0.0000,\n",
      "           0.0000,  0.0000,  0.0000,  0.0000, -0.0000, -0.0000,  0.0000,\n",
      "           0.0000,  0.0000]],\n",
      "\n",
      "        [[-0.3941, -1.2574, -0.6091,  1.0736, -0.9058, -1.7475,  0.9353,\n",
      "          -0.2424,  0.6113,  0.5230,  0.2550, -0.2047, -1.3506,  0.5298,\n",
      "          -0.4015,  0.0150],\n",
      "         [ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000, -0.0000, -0.0000,\n",
      "           0.0000,  0.0000,  0.0000,  0.0000, -0.0000, -0.0000,  0.0000,\n",
      "           0.0000,  0.0000],\n",
      "         [ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000, -0.0000, -0.0000,\n",
      "           0.0000,  0.0000,  0.0000,  0.0000, -0.0000, -0.0000,  0.0000,\n",
      "           0.0000,  0.0000]],\n",
      "\n",
      "        [[ 1.6465,  0.4131,  0.0321,  0.8855, -0.2912,  0.7838, -0.4488,\n",
      "          -1.3473, -1.0906,  1.0139,  0.4203,  0.1642,  1.2011, -0.7894,\n",
      "           0.6780,  0.8977],\n",
      "         [ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000, -0.0000, -0.0000,\n",
      "           0.0000,  0.0000,  0.0000,  0.0000, -0.0000, -0.0000,  0.0000,\n",
      "           0.0000,  0.0000],\n",
      "         [ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000, -0.0000, -0.0000,\n",
      "           0.0000,  0.0000,  0.0000,  0.0000, -0.0000, -0.0000,  0.0000,\n",
      "           0.0000,  0.0000]],\n",
      "\n",
      "        [[ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000, -0.0000, -0.0000,\n",
      "           0.0000,  0.0000,  0.0000,  0.0000, -0.0000, -0.0000,  0.0000,\n",
      "           0.0000,  0.0000],\n",
      "         [ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000, -0.0000, -0.0000,\n",
      "           0.0000,  0.0000,  0.0000,  0.0000, -0.0000, -0.0000,  0.0000,\n",
      "           0.0000,  0.0000],\n",
      "         [ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000, -0.0000, -0.0000,\n",
      "           0.0000,  0.0000,  0.0000,  0.0000, -0.0000, -0.0000,  0.0000,\n",
      "           0.0000,  0.0000]]])\n"
     ]
    }
   ],
   "source": [
    "Zn_inter_step = node_Wn_int[int_indices] * int_part\n",
    "\n",
    "print(Zn_inter_step.shape)\n",
    "print(Zn_inter_step)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3bfb58c3",
   "metadata": {},
   "source": [
    "Step 5c: Scale each of the participating neighbor by its edge weight\n",
    "\n",
    "$\\mathbf{W}^{(l)}_{int} h^{(l-1)}_j \\bigodot E_{ij}$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "6fad5be9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 3, 16])\n",
      "tensor([[[ 1.4818,  0.3718,  0.0289,  0.7970, -0.2621,  0.7054, -0.4039,\n",
      "          -1.2125, -0.9815,  0.9125,  0.3783,  0.1478,  1.0810, -0.7105,\n",
      "           0.6102,  0.8079],\n",
      "         [-0.1854, -0.1555, -0.2526,  0.3416,  0.1099,  0.3715, -0.0153,\n",
      "           0.0072,  0.4654, -0.5415, -0.3548,  0.0285, -0.0059, -0.0549,\n",
      "           0.0526, -0.1134],\n",
      "         [-0.0000, -0.0000,  0.0000, -0.0000,  0.0000,  0.0000,  0.0000,\n",
      "          -0.0000, -0.0000, -0.0000, -0.0000,  0.0000,  0.0000, -0.0000,\n",
      "          -0.0000, -0.0000]],\n",
      "\n",
      "        [[-0.0394, -0.1257, -0.0609,  0.1074, -0.0906, -0.1747,  0.0935,\n",
      "          -0.0242,  0.0611,  0.0523,  0.0255, -0.0205, -0.1351,  0.0530,\n",
      "          -0.0401,  0.0015],\n",
      "         [-0.0000, -0.0000,  0.0000, -0.0000,  0.0000,  0.0000,  0.0000,\n",
      "          -0.0000, -0.0000, -0.0000, -0.0000,  0.0000,  0.0000, -0.0000,\n",
      "          -0.0000, -0.0000],\n",
      "         [-0.0000, -0.0000,  0.0000, -0.0000,  0.0000,  0.0000,  0.0000,\n",
      "          -0.0000, -0.0000, -0.0000, -0.0000,  0.0000,  0.0000, -0.0000,\n",
      "          -0.0000, -0.0000]],\n",
      "\n",
      "        [[ 0.8232,  0.2066,  0.0161,  0.4428, -0.1456,  0.3919, -0.2244,\n",
      "          -0.6736, -0.5453,  0.5069,  0.2102,  0.0821,  0.6006, -0.3947,\n",
      "           0.3390,  0.4489],\n",
      "         [-0.0000, -0.0000,  0.0000, -0.0000,  0.0000,  0.0000,  0.0000,\n",
      "          -0.0000, -0.0000, -0.0000, -0.0000,  0.0000,  0.0000, -0.0000,\n",
      "          -0.0000, -0.0000],\n",
      "         [-0.0000, -0.0000,  0.0000, -0.0000,  0.0000,  0.0000,  0.0000,\n",
      "          -0.0000, -0.0000, -0.0000, -0.0000,  0.0000,  0.0000, -0.0000,\n",
      "          -0.0000, -0.0000]],\n",
      "\n",
      "        [[-0.0000, -0.0000,  0.0000, -0.0000,  0.0000,  0.0000,  0.0000,\n",
      "          -0.0000, -0.0000, -0.0000, -0.0000,  0.0000,  0.0000, -0.0000,\n",
      "          -0.0000, -0.0000],\n",
      "         [-0.0000, -0.0000,  0.0000, -0.0000,  0.0000,  0.0000,  0.0000,\n",
      "          -0.0000, -0.0000, -0.0000, -0.0000,  0.0000,  0.0000, -0.0000,\n",
      "          -0.0000, -0.0000],\n",
      "         [-0.0000, -0.0000,  0.0000, -0.0000,  0.0000,  0.0000,  0.0000,\n",
      "          -0.0000, -0.0000, -0.0000, -0.0000,  0.0000,  0.0000, -0.0000,\n",
      "          -0.0000, -0.0000]]])\n"
     ]
    }
   ],
   "source": [
    "Zn_inter_step = node_Wn_int[int_indices] * int_part * int_edges\n",
    "\n",
    "print(Zn_inter_step.shape)\n",
    "print(Zn_inter_step)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a42a587c",
   "metadata": {},
   "source": [
    "Step 5d: Sum all the neighbors of an atom\n",
    "\n",
    "$\\sum_{j \\epsilon N_i^{(int)}} \\mathbf{W}^{(l)}_{int} h^{(l-1)}_j \\bigodot E_{ij}$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "aac47740",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 16])\n",
      "tensor([[ 1.2964,  0.2163, -0.2237,  1.1385, -0.1521,  1.0769, -0.4192, -1.2053,\n",
      "         -0.5162,  0.3710,  0.0236,  0.1764,  1.0751, -0.7654,  0.6628,  0.6946],\n",
      "        [-0.0394, -0.1257, -0.0609,  0.1074, -0.0906, -0.1747,  0.0935, -0.0242,\n",
      "          0.0611,  0.0523,  0.0255, -0.0205, -0.1351,  0.0530, -0.0401,  0.0015],\n",
      "        [ 0.8232,  0.2066,  0.0161,  0.4428, -0.1456,  0.3919, -0.2244, -0.6736,\n",
      "         -0.5453,  0.5069,  0.2102,  0.0821,  0.6006, -0.3947,  0.3390,  0.4489],\n",
      "        [ 0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000,\n",
      "          0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000]])\n"
     ]
    }
   ],
   "source": [
    "Zn_inter_step = torch.sum(node_Wn_int[int_indices] * int_part * int_edges, 1)\n",
    "\n",
    "print(Zn_inter_step.shape)\n",
    "print(Zn_inter_step)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32b3bc37",
   "metadata": {},
   "source": [
    "Step 5e: Divide with the number of atoms in interface graph"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e4bfc250",
   "metadata": {},
   "source": [
    "$\\frac{1}{|N_i^{(int)}|}\\sum_{j \\epsilon N_i^{(int)}} \\mathbf{W}^{(l)}_{int} h^{(l-1)}_j \\bigodot E_{ij} $"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "77ec9a61",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 16])\n",
      "tensor([[ 0.6482,  0.1082, -0.1118,  0.5693, -0.0761,  0.5384, -0.2096, -0.6027,\n",
      "         -0.2581,  0.1855,  0.0118,  0.0882,  0.5376, -0.3827,  0.3314,  0.3473],\n",
      "        [-0.0394, -0.1257, -0.0609,  0.1074, -0.0906, -0.1747,  0.0935, -0.0242,\n",
      "          0.0611,  0.0523,  0.0255, -0.0205, -0.1351,  0.0530, -0.0401,  0.0015],\n",
      "        [ 0.8232,  0.2066,  0.0161,  0.4428, -0.1456,  0.3919, -0.2244, -0.6736,\n",
      "         -0.5453,  0.5069,  0.2102,  0.0821,  0.6006, -0.3947,  0.3390,  0.4489],\n",
      "        [ 0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000,\n",
      "          0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000,  0.0000]])\n"
     ]
    }
   ],
   "source": [
    "Zn_inter = torch.sum(node_Wn_int[int_indices] * int_part * int_edges, 1) / no_of_neighs_int\n",
    "\n",
    "print(Zn_inter.shape)\n",
    "print(Zn_inter)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b380a035",
   "metadata": {},
   "source": [
    "### GCN Layer code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "826c61bd",
   "metadata": {},
   "outputs": [],
   "source": [
    "class GCN_layer(nn.Module):\n",
    "\n",
    "    def __init__(self, v_feats, filters, dropout=0.1, bias=True, trainable=True, **kwargs):\n",
    "\n",
    "        super(GCN_layer, self).__init__()\n",
    "\n",
    "        self.v_feats = v_feats\n",
    "        self.filters = filters\n",
    "        self.dropout= dropout\n",
    "        self.bias = bias\n",
    "        self.trainable = trainable\n",
    "        \n",
    "        use_cuda = torch.cuda.is_available()\n",
    "        self.device = torch.device(\"cuda:0\" if use_cuda else \"cpu\")\n",
    "        \n",
    "        # Define non-linear function\n",
    "        self.non_lin = nn.LeakyReLU()\n",
    "\n",
    "        # Initialize dropout\n",
    "        self.drop_layer = nn.Dropout(p=self.dropout)\n",
    "       \n",
    "        # Initialize learnable parameters\n",
    "        self.Wvc = nn.Parameter(\n",
    "                    init.kaiming_uniform_(\n",
    "                        torch.randn(self.v_feats, self.filters, requires_grad=self.trainable))\n",
    "                            .type(torch.float).to(self.device))  # (v_dims, filters)\n",
    "        self.Wvn_int = nn.Parameter(\n",
    "                        init.kaiming_uniform_(\n",
    "                            torch.randn(self.v_feats, self.filters, requires_grad=self.trainable))\n",
    "                                .type(torch.float).to(self.device))  # (v_dims, filters)\n",
    "        self.Wvn_prot = nn.Parameter(\n",
    "                        init.kaiming_uniform_(\n",
    "                            torch.randn(self.v_feats, self.filters, requires_grad=self.trainable))\n",
    "                                .type(torch.float).to(self.device))  # (v_dims, filters)\n",
    "        self.bv = nn.Parameter(\n",
    "                        torch.zeros(self.filters, requires_grad=self.trainable)\n",
    "                            .type(torch.float).to(self.device))\n",
    "        \n",
    "        \n",
    "    def forward(self, x):\n",
    "\n",
    "        node_feat, prot_indices, int_indices, prot_edges, int_edges = x\n",
    "\n",
    "        # generate center node signals\n",
    "        Zc = torch.matmul(node_feat, self.Wvc)  # (n_verts, filters)\n",
    "\n",
    "        # create neighbor signals\n",
    "        v_Wvn_int = torch.matmul(node_feat, self.Wvn_int)  # (n_verts, filters)\n",
    "        v_Wvn_prot = torch.matmul(node_feat, self.Wvn_prot)  # (n_verts, filters)\n",
    "        \n",
    "        # Find participating neighbors of the interface and protein graph in Graph Convolution \n",
    "        int_part = torch.unsqueeze(int_indices > -1, 2).type(torch.float)\n",
    "        prot_part = torch.unsqueeze(prot_indices > -1, 2).type(torch.float)\n",
    "\n",
    "        # Find the number of neighbors per atom\n",
    "        no_of_neighs_int = torch.sum(int_indices > -1, 1).unsqueeze(1).to(self.device).type(torch.float)\n",
    "        no_of_neighs_prot = torch.sum(prot_indices > -1, 1).unsqueeze(1).to(self.device).type(torch.float)\n",
    "        no_of_neighs_int[no_of_neighs_int == 0] = 1\n",
    "        no_of_neighs_prot[no_of_neighs_prot == 0] = 1\n",
    "        \n",
    "        # Expand interface and neighborh edge matrices to allow scalar multiplication\n",
    "        if(len(int_edges.size()) != 3):\n",
    "            int_edges = torch.unsqueeze(int_edges, 2)\n",
    "            prot_edges = torch.unsqueeze(prot_edges, 2)\n",
    "\n",
    "        # Generate neighbor signals w.r.t the interface and protein graph\n",
    "        Zn_inter = torch.sum(v_Wvn_int[int_indices] * int_part * int_edges, 1) / no_of_neighs_int\n",
    "        Zn_prot = torch.sum(v_Wvn_prot[prot_indices] * prot_part * prot_edges, 1) / no_of_neighs_prot\n",
    "\n",
    "        # Combine central and neighbor signals\n",
    "        Z = Zc + Zn_inter + Zn_prot\n",
    "\n",
    "        # Add bias\n",
    "        if self.bias:\n",
    "            Z += self.bv\n",
    "\n",
    "        # Apply activation function\n",
    "        h = self.non_lin(Z)\n",
    "\n",
    "        # Apply dropout\n",
    "        if self.dropout:\n",
    "            h = self.drop_layer(h)\n",
    "\n",
    "\n",
    "        return h, prot_indices, int_indices, prot_edges, int_edges"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d029d4b4",
   "metadata": {},
   "source": [
    "## Define Graph Neural network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "f3f6b4de",
   "metadata": {},
   "outputs": [],
   "source": [
    "class GNN(torch.nn.Module):\n",
    "    \"\"\"\n",
    "        GCN + Sum unit pooling + L1-loss\n",
    "    \"\"\"\n",
    "    def __init__(self):\n",
    "        super(GNN, self).__init__()\n",
    "\n",
    "        # Define neural network layers\n",
    "        self.conv1 = GCN_layer(v_feats=11, filters=16, dropout=0.1)\n",
    "        self.conv2 = GCN_layer(v_feats=16, filters=32, dropout=0.1)\n",
    "        self.conv3 = GCN_layer(v_feats=32, filters=64, dropout=0.1)\n",
    "        self.dense = Dense(in_dims=64, out_dims=1, nonlin=\"linear\")\n",
    "\n",
    "        # Define optimizer\n",
    "        self.optimizer = optim.Adam(self.parameters(), lr=0.01, weight_decay=0.0001)\n",
    "\n",
    "    def forward(self, x):\n",
    "        \n",
    "        # GCN layers\n",
    "        x = self.conv1(x)\n",
    "        x = self.conv2(x)\n",
    "        x = self.conv3(x)\n",
    "        x = x[0]\n",
    "        # each row of x is now a feature vector.\n",
    "\n",
    "        # Graph pooling\n",
    "        x = torch.sum(x, 0).view(1, -1)\n",
    "        x = F.normalize(x)\n",
    "\n",
    "        # Dense layer\n",
    "        x = self.dense(x)\n",
    "        x = torch.squeeze(x, 1)\n",
    "\n",
    "        return x\n",
    "\n",
    "    def loss(self, output, target, reduction='mean'):\n",
    "        # Define loss function\n",
    "        return F.l1_loss(output, target, reduction=reduction)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "710e09f6",
   "metadata": {},
   "source": [
    "## Initialize Graph Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "c67b07c0",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = GNN().to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eec8f9e8",
   "metadata": {},
   "source": [
    "## Initialize PyTorch data loader and data sampler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "7fb132c1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No. of train complexes: 3\n",
      "No. of test complexes: 2\n"
     ]
    }
   ],
   "source": [
    "dataset_path = \"data\"\n",
    "\n",
    "# Train dataset generator - training\n",
    "train_dataset = Dataset(dataset_path, \"train\")\n",
    "\n",
    "train_params = {'sampler': CustomSampler(batch_complexes = 2,\n",
    "                                   decoys_per_complex = 50,\n",
    "                                   dataset_path = dataset_path,\n",
    "                                   dataset_cat = \"train\",\n",
    "                                   random_sample = True),\n",
    "          'num_workers': 0,\n",
    "          'collate_fn' : lambda x: x[0]}\n",
    "training_generator = data.DataLoader(train_dataset, **train_params)\n",
    "\n",
    "# Train dataset generator - testing\n",
    "train_params = {'sampler': CustomSampler(batch_complexes = 2,\n",
    "                                   decoys_per_complex = 300,\n",
    "                                   dataset_path = dataset_path,\n",
    "                                   dataset_cat = \"train\",\n",
    "                                   verbose=False),\n",
    "          'num_workers': 0,\n",
    "          'collate_fn' : lambda x: x[0]}\n",
    "test_generator_train_set = data.DataLoader(train_dataset, **train_params)\n",
    "\n",
    "# Test dataset generator\n",
    "test_dataset = Dataset(dataset_path, \"test\")\n",
    "test_params = {'sampler': CustomSampler(batch_complexes = 2,\n",
    "                                   decoys_per_complex = 300,\n",
    "                                   dataset_path = dataset_path,\n",
    "                                   dataset_cat = \"test\"),\n",
    "          'num_workers': 0,\n",
    "          'collate_fn' : lambda x: x[0]}\n",
    "test_generator_test_set = data.DataLoader(test_dataset, **test_params)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9819549a",
   "metadata": {},
   "source": [
    "## Train model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "e0d59a1f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1 loss : 1.7805060297250748\n",
      "Epoch 2 loss : 1.9482030868530273\n",
      "Epoch 3 loss : 0.4889952540397644\n",
      "Epoch 4 loss : 0.47706417739391327\n",
      "Epoch 5 loss : 0.21714823693037033\n",
      "Epoch 6 loss : 0.18489650636911392\n",
      "Epoch 7 loss : 0.22734376788139343\n",
      "Epoch 8 loss : 0.24544887244701385\n",
      "Epoch 9 loss : 0.11929459124803543\n",
      "Epoch 10 loss : 0.14776961505413055\n"
     ]
    }
   ],
   "source": [
    "epochs = 10\n",
    "mini_batches_per_epoch = 30\n",
    "\n",
    "loss = []\n",
    "\n",
    "for epoch in range(0, epochs):\n",
    "    \n",
    "    loss_ = train(model, device, training_generator, model.optimizer, epoch, mini_batches_per_epoch)\n",
    "    \n",
    "    loss.append(loss_)\n",
    "    \n",
    "    print(\"Epoch \" + str(epoch+1)+ \" loss : \" + str(loss_))\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4e40c945",
   "metadata": {},
   "source": [
    "## Plot loss function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "0bd7e019",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "epoch_list = list(range(1, epochs+1))\n",
    "epoch_losses = list(zip(epoch_list, loss))\n",
    "\n",
    "loss_df = pd.DataFrame(epoch_losses, columns=[\"Epoch\", \"Train loss\"])\n",
    "\n",
    "lines = loss_df.plot.line(x=\"Epoch\", y=\"Train loss\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d41fe96",
   "metadata": {},
   "source": [
    "## Test model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "374da1da",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores_dockq_train = test(model, device, test_generator_train_set, \"train\")\n",
    "scores_dockq_test = test(model, device, test_generator_test_set, \"test\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8e16ed3c",
   "metadata": {},
   "source": [
    "### Train complex 1 results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "4b25f5b3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_results(\"1bvn\", scores_dockq_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "b78a521c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_results(\"1f34\", scores_dockq_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "7a61ddf7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfEAAAEwCAYAAACnhkhsAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Il7ecAAAACXBIWXMAAAsTAAALEwEAmpwYAAAyfklEQVR4nO3de7xV5X3v+8+PmxjlqmBaEMVuqkEE1AUaUZoLQYypiqZHidWqQWMjtjuNScz2bE3cfb2SkzR1q7FRam7aWrUmWk6Dt2wleIktYNAI3gjxsjhal1wERFRYv/PHnBMHg3F55mWsOeda3/frtV4w5hyX33jG5fc8zxhzDHN3REREpP30a3YAIiIiUhslcRERkTalJC4iItKmlMRFRETalJK4iIhIm1ISFxERaVOFJnEzm2Nmz5vZGjO7POH7cWb2sJn9xsyeNrNPFxmPiIhIb2JF/U7czPoDLwCfAjqBZcA8d18dGWch8Bt3/4GZTQQWu/vBhQQkIiLSyxTZEp8OrHH3te7+HnA7cGpsHAeGlv8/DPj/CoxHRESkVxlQ4LzHAK9GhjuBY2LjfAN4wMwuBfYBZhUYj4iISK9SZBIPMQ/4ibt/z8w+CtxqZpPcvTs6kpldBFwEsM8++xx92GGHNSFUERGRnrdixYo33X1U0ndFJvF1wIGR4bHlz6I+D8wBcPdfm9lgYH/gjehI7r4QWAjQ0dHhy5cvLypmERGRlmJmL6d9V+Q18WXABDMbb2aDgLOARbFxXgE+CWBmHwEGA10FxiQiItJrFJbE3X0HsAC4H3gWuNPdV5nZ1WZ2Snm0LwMXmtlTwL8A57leqyYiIhKk0Gvi7r4YWBz77MrI/1cDM4qMQUREpLdq9o1tIiLSQO+//z6dnZ1s37692aFIlQYPHszYsWMZOHBg8DRK4iIivUhnZydDhgzh4IMPxsyaHY4EcnfWr19PZ2cn48ePD55Oz04XEelFtm/fzn777acE3mbMjP3226/qHhQlcRGRXkYJvD3Vst2UxEVEpKH69+/P1KlTOfzww5kyZQrf+9736O7uzp8wwcEHH8ybb765x+ednZ2ceuqpTJgwgUMOOYQFCxbw7rvv1ht621ESFxGRhtp7771ZuXIlq1at4sEHH+Tee+/lm9/8ZsPm7+6cfvrpnHbaabz44ou8+OKLvPPOO3z1q19t2DLS7Nixo/BlVENJvAUMHQpmpX9FRHqT0aNHs3DhQr7//e/j7mzfvp3zzz+fI444giOPPJKHH34YgJ07d3LZZZcxadIkJk+ezPXXX7/bfN555x1OOukk/vEf/5GHHnqIwYMHc/755wOllv8111zDLbfcwtatW3eb7rXXXmPmzJlMnTqVSZMm8cgjjwBw3333cdRRRzFlyhQ++clPArBhwwZOO+00Jk+ezLHHHsvTTz8NwDe+8Q3OOeccZsyYwTnnnENXVxdnnHEG06ZNY9q0aTz22GOFlmEW3Z3eArZs2f1fEZHe5JBDDmHnzp288cYb/NM//RNmxm9/+1uee+45Zs+ezQsvvMCPf/xjXnrpJVauXMmAAQPYsGHDrum3bt3KWWedxbnnnsu5557Lddddx9FHH73bMoYOHcrBBx/MmjVrmDp16q7Pb7vtNk488USuuOIKdu7cybZt2+jq6uLCCy9k6dKljB8/fteyrrrqKo488kjuueceHnroIc4991xWrlwJwOrVq3n00UfZe++9+dznPseXvvQljj/+eF555RVOPPFEnn322cLLMYmSeAsYMqSUwIcMaXYkItIXDR36wTlo8+Zil/Xoo49y6aWXAnDYYYdx0EEH8cILL/DLX/6Siy++mAEDSmlp5MiRu6Y59dRT+epXv8rZZ59d9fKmTZvGBRdcwPvvv89pp53G1KlTWbJkCTNnztz1U67Ksh599FF+9rOfAfCJT3yC9evXs7lcIKeccgp77703AL/85S9ZvXr1rmVs3ryZrVu3su+++1YdX73Und4CNm8G9+IPHhGRJEX3Bq5du5b+/fszevTomqafMWMG9913H5Wnck+cOJEVK1bsNs7mzZt5/fXXOfTQQ3f7fObMmSxdupQxY8Zw3nnnccstt9QUwz777LPr/93d3TzxxBOsXLmSlStXsm7duqYkcFASFxHp8yq9gEX0BnZ1dXHxxRezYMECzIwTTjiBf/7nfwbghRde4JVXXuHQQw/lU5/6FDfddNOuG8ei3elXX301I0aM4JJLLgHgk5/8JNu2bduVkHfu3MmXv/xlFixYsKu1XPHyyy9zwAEHcOGFFzJ//nyefPJJjj32WJYuXcrvf//73ZYVjW3JkiXsv//+DE24WWn27Nm7XbOvdLk3g5K4iEgf1+jewHfeeWfXT8xmzZrF7NmzueqqqwD44he/SHd3N0cccQRnnnkmP/nJT9hrr72YP38+48aNY/LkyUyZMoXbbrttt3lee+21u+5ANzPuvvtu7rrrLiZMmMB+++1Hv379uOKKK/aIZcmSJUyZMoUjjzySO+64g7/+679m1KhRLFy4kNNPP50pU6Zw5plnAqUb2FasWMHkyZO5/PLL+elPf5q4ftdddx3Lly9n8uTJTJw4kRtvvLExBVcDa7eXhul94iIi6Z599lk+8pGPNDuMHvX4448zb9487r77bo466qhmh1OXpO1nZivcvSNpfN3YJiIibe24447j5ZdfbnYYTaHudBERkTalJC4iItKmlMRFRETalJK4iIhIm1ISFxERaVNK4iIi0lCVV5FOmjSJP/uzP2Pbtm01z+u8887jrrvuAmD+/Pm7Pe40bsmSJTz++ONVLyPtdaftQElcREQaqvIq0meeeYZBgwbt8TCUWl/nefPNNzNx4sTU72tN4j2hqFeYKomLiEhhTjjhBNasWcOSJUs44YQTOOWUU5g4cSI7d+7kK1/5CtOmTWPy5MncdNNNQOld4QsWLODQQw9l1qxZvPHGG7vm9bGPfYzKw77irxJ96aWXuPHGG7nmmmuYOnUqjzzySOorQ9evX8/s2bM5/PDDmT9/PkkPPdu5cyfnnXcekyZN4ogjjuCaa64BYM2aNcyaNYspU6Zw1FFH8bvf/Q535ytf+cquce+44w6A4HWuhx72IiLS13V3Q1cXjB4NZg2b7Y4dO7j33nuZM2cOAE8++STPPPMM48ePZ+HChQwbNoxly5bx7rvvMmPGDGbPns1vfvMbnn/+eVavXs1//dd/MXHiRC644ILd5pv0KtGRI0dy8cUXs++++3LZZZcBpL4y9Jvf/CbHH388V155Jb/4xS/44Q9/uEfslRebPPPMMwBs2rQJgLPPPpvLL7+cuXPnsn37drq7u/n5z3/OypUreeqpp3jzzTeZNm0aM2fODF7nytvUalFoEjezOcC1QH/gZnf/duz7a4CPlwc/BIx29+FFxiQiIhHd3fDxj8Pjj8Nxx8HDD0O/+jppK89Oh1JL/POf/zyPP/4406dP35WwHnjgAZ5++uld17vfeustXnzxRZYuXcq8efPo378/f/iHf8gnPvGJPeb/xBNPJL5KNC7tlaFLly7l5z//OQAnn3wyI0aM2GPaQw45hLVr13LppZdy8sknM3v2bLZs2cK6deuYO3cuAIMHDwZKrzCtxHzAAQfwJ3/yJyxbtoyhQ4cGrXNLJnEz6w/cAHwK6ASWmdkid99Vou7+pcj4lwJHFhWPiIgk6OoqJfAdO0r/dnXBAQfUNcvKNfG46Os83Z3rr7+eE088cbdxFi9eXNeyoyqvDK0k22qMGDGCp556ivvvv58bb7yRO++8k2uvvbbq+YSscz2KvCY+HVjj7mvd/T3gduDUjPHnAf9SYDwiIhI3enSpBT5gQOnfGt/5Xa0TTzyRH/zgB7z//vtA6bWkb7/9NjNnzuSOO+5g586dvPbaazz88MN7TJv2KtEhQ4awJfJS9LRXhs6cOXPXW9LuvfdeNm7cuMcy3nzzTbq7uznjjDP427/9W5588kmGDBnC2LFjueeeewB499132bZtGyeccMKumLu6uli6dCnTp08PXud6FNmdPgZ4NTLcCRyTNKKZHQSMBx4qMB4REYkzK3WhF3BNPMv8+fN56aWXOOqoo3B3Ro0axT333MPcuXN56KGHmDhxIuPGjeOjH/3oHtNGXyXa3d3N6NGjefDBB/nTP/1TPvvZz/Jv//ZvXH/99Vx33XVccsklTJ48mR07djBz5kxuvPFGrrrqKubNm8fhhx/Occcdx7hx4/ZYxrp16zj//PPp7u4G4Fvf+hYAt956K1/4whe48sorGThwIP/6r//K3Llz+fWvf82UKVMwM77zne/w4Q9/mOeeey5onetR2KtIzeyzwBx3n18ePgc4xt0XJIz7NWCsu1+aMq+LgIsAxo0bd3RffVuNiEievvgq0t6k2leRFtmdvg44MDI8tvxZkrPI6Ep394Xu3uHuHaNGjWpgiCIiIu2ryCS+DJhgZuPNbBClRL0oPpKZHQaMAH5dYCwiIiK9TmFJ3N13AAuA+4FngTvdfZWZXW1mp0RGPQu43Yvq1xcREemlCv2duLsvBhbHPrsyNvyNImMQEelr3B3roRvUpHFqacvqsasiIr3I4MGDWb9+fU0JQZrH3Vm/fn3Vv2nXY1dFRHqRsWPH0tnZSVdXV7NDkSoNHjyYsWPHVjWNkriISC8ycODAuh7jKe1F3ekiIiJtSklcRESkTSmJi4iItCklcRERkTalJC4iItKmlMRFRETalJK4iIhIm1ISFxERaVNK4iIiIm1KSVxERKRNKYmLiIi0KSVxERGRNqUkLiIi0qaUxEVERNqUkriIiEibUhIXERFpU0riIiIibUpJXEREpE0piYuIiLSpQpO4mc0xs+fNbI2ZXZ4yzv9lZqvNbJWZ3VZkPCIiIr3JgKJmbGb9gRuATwGdwDIzW+TuqyPjTAC+Dsxw941mNrqoeERERHqbIlvi04E17r7W3d8DbgdOjY1zIXCDu28EcPc3CoxHRESkVykyiY8BXo0Md5Y/i/pj4I/N7DEze8LM5hQYj4iISK9SWHd6FcufAHwMGAssNbMj3H1TdCQzuwi4CGDcuHE9HKKIiEhrKrIlvg44MDI8tvxZVCewyN3fd/ffAy9QSuq7cfeF7t7h7h2jRo0qLGAREZF2UmQSXwZMMLPxZjYIOAtYFBvnHkqtcMxsf0rd62sLjElERKTXKCyJu/sOYAFwP/AscKe7rzKzq83slPJo9wPrzWw18DDwFXdfX1RMIiIivYm5e7NjqEpHR4cvX7682WGIiIj0CDNb4e4dSd/piW0iIiJtSklcRESkTSmJi4iItCklcRERkTalJC4iItKmlMRFRETalJK4iIhIm1ISFxERaVNK4iIiIm1KSVxERKRNKYmLiIi0KSVxERGRNqUkLiIi0qaUxEVERNqUkriIiEibUhIXERFpU0riIiIibUpJXEREpE0piYuIiLQpJXEREZE2pSQuIiLSppTERURE2pSSuIiISJsakPWlmS0KmMcGdz8vZfo5wLVAf+Bmd/927PvzgO8C68offd/dbw5YpoiISJ+XmcSBjwDzM7434IbEL8z6l7/7FNAJLDOzRe6+OjbqHe6+IDBeERERKctL4le4+6+yRjCzb6Z8NR1Y4+5ry+PdDpwKxJO4iIiI1CDzmri735k3g4xxxgCvRoY7y5/FnWFmT5vZXWZ2YNKMzOwiM1tuZsu7urryQhIREekTar6xzcwWNmD5/y9wsLtPBh4Efpo0krsvdPcOd+8YNWpUAxYrIiLS/vJubBuZ9hXw6Zx5rwOiLeuxfHADGwDuvj4yeDPwnZx5ioiISFneNfEu4GVKSbvCy8Ojc6ZdBkwws/GUkvdZwOeiI5jZH7j7a+XBU4BnA+MWERHp8/KS+Frgk+7+SvwLM3s1Yfxd3H2HmS0A7qf0E7MfufsqM7saWO7ui4C/MrNTgB3ABuC8GtZBRESkT8pL4v8bGAHskcQJ6Pp298XA4thnV0b+/3Xg67lRioiIyB4yk7i7J/4GvPzd9Y0PR0RERELpsasiIiJtSklcRESkTSmJi4iItKmakriZ/YGZ7dXoYERERCRcrS3xW4HnzOzvGhmMiIiIhMv7iVkid59lZgZMbHA8IiIiEig3iZeT9XQ+eHnJOuA/3d2BVQXGJiIiIhnynp0+G/gH4EU+eO75WOC/mdkX3f2BguMTERGRFHkt8WuBWe7+UvTD8vPQFwMfKSguERERyZF3Y9sASu8Bj1sHDGx8OCIiIhIqryX+I2CZmd0OVF54ciClN5L9sMjAREREJFves9O/ZWb3AKcCHy1/vA44291XFxybiIiIZMi9O93dn0Xv+RYREWk5mdfEzWxh3gxCxhEREZHGy2uJn2Zm2zO+N+DjDYxHREREAuUl8a8EzOORRgQiIiIi1cm7se2nPRWIiIiIVEevIhUREWlTSuIiIiJtqqokbmYfKioQERERqU5QEjez48xsNfBceXiKmf1DoZGJiIhIptCW+DXAicB6AHd/CpiZN5GZzTGz581sjZldnjHeGWbmZtYRGI+IiEifF9yd7u6vxj7amTW+mfUHbgBOAiYC88xsYsJ4Q4C/Bv4jNBYREREJT+KvmtlxgJvZQDO7jPxHsU4H1rj7Wnd/D7id0jPY4/4X8P8AWQ+VERERkZjQJH4xcAkwhtILUKaWh7OM4YM3n0HplaZjoiOY2VHAge7+i8A4REREpCz3BSjlbvFr3f3sRi7YzPoBfw+cFzDuRcBFAOPGjWtkGCIiIm0rtyXu7juBg8xsUJXzXkfp3eMVY8ufVQwBJgFLzOwl4FhgUdLNbe6+0N073L1j1KhRVYYhIiLSO+W2xMvWAo+Z2SLg7cqH7v73GdMsAyaY2XhKyfss4HORad8C9q8Mm9kS4DJ3Xx4cvYiISB8WmsR/V/7rR6kFncvdd5jZAuB+oD/wI3dfZWZXA8vdfVEtAYuIiEiJuXv4yGb7Arj71sIiytHR0eHLl6uxLiIifYOZrXD3xOeohD6xbZKZ/QZYBawysxVmdngjgxQREZHqhP7EbCHwN+5+kLsfBHwZ+MfiwhIREZE8oUl8H3d/uDLg7kuAfQqJSERERIIE351uZv8TuLU8/OeU7lgXERGRJgltiV8AjAJ+DvyM0k/DLigqKBEREckX1BJ3943AXxUci4iIiFQh9O70B81seGR4hJndX1hUIiIikiu0O31/d99UGSi3zEcXEpGIiIgECU3i3Wa2680jZnYQEP6UGBEREWm40LvTrwAeNbNfAQacQPmtYiIiItIcoTe23Vd+9/ex5Y/+u7u/WVxYIiIikif0xrYZwDvu/u/AcOB/lLvURUREpElCr4n/ANhmZlOAv6H0RrNbCotKREREcoUm8R1eet3ZqcAN7n4Dga8kFRERkWKE3ti2xcy+TulxqzPNrB8wsLiwREREJE9oS/xM4F3g8+7+OjAW+G5hUYmIiEiu0LvTXwf+PjL8CromLiIi0lShLXERERFpMUriIiIibSqoO93MPgT8t/Lg8+7+bnEhiYiISIjMlriZDTSz/w10Aj8GfgKsNbPLy99PLTg+ERERSZHXEv8e8CHgIHffAmBmQ4G/M7MfAHOA8cWGKCIiIknyrol/GriwksAB3H0z8JfAWcC8rInNbI6ZPW9mayqt99j3F5vZb81spZk9amYTa1gHERGRPikviXeXn9S2G3ffCXS5+xNpE5pZf+AG4CRgIjAvIUnf5u5HuPtU4DtEfsYmIiIi2fKS+GozOzf+oZn9OfBszrTTgTXuvtbd3wNup/TY1l3KrfqKfdA7ykVERILlXRO/BPi5mV0ArCh/1gHsDczNmXYM8GpkuBM4Jj6SmV1C6aUqg4BPBMQsIiIi5LTE3X2dux8DXA28VP672t2nu/u6RgTg7je4+x8BXwP+76RxzOwiM1tuZsu7uroasVgREZG2F/qwl/Hufn357/8AmNm3c6ZZBxwYGR5b/izN7cBpSV+4+0J373D3jlGjRgWGLCIi0ruFJvEzzOzsyoCZ3QDkZdNlwAQzG29mgyjdzb4oOoKZTYgMngy8GBiPiIhInxf6KtIzgEVm1k3pt+Gb3P3zWRO4+w4zWwDcD/QHfuTuq8zsamC5uy8CFpjZLOB9YCPwF7WuiIiISF9jCb8g++BLs5GRwSHAPcBjwJUA7r6hyOCSdHR0+PLly3t6sSIiIk1hZivcvSPpu7yW+ApKP/uyyL8nl/8cOKSBcYqIiEgVMpO4u+uRqiIiIi0q6MY2M7vEzIZHhkeY2RcLi0pERERyhd6dfqG7b6oMuPtG4MJCIhIREZEgoUm8v5lZZaD8XPRBxYQkIiIiIUJ/YnYfcIeZ3VQe/kL5MxEREWmS0CT+NUqJ+y/Lww8CNxcSkYiIiAQJSuLu3m1mPwQepfTTsufLryMVERGRJglK4mb2MeCnlF6AYsCBZvYX7r60sMhEREQkU2h3+veA2e7+PICZ/THwL8DRRQUmIiIi2ULvTh9YSeAA7v4CMLCYkERERCREaEt8uZndDPxTefhsQA8wFxERaaLQJP6XwCXAX5WHHwH+oZCIREREJEjo3envmtmtwK3u3lVwTCIiIhIg85q4lXzDzN4EngeeN7MuM7uyZ8ITERGRNHk3tn0JmAFMc/eR7j4SOAaYYWZfKjw6ERERSZWXxM8B5rn77ysfuPta4M+Bc4sMTERERLLlJfGB7v5m/MPydXH9xExERKSJ8pL4ezV+JyIiIgXLuzt9ipltTvjcgMEFxCMiIiKBMpO4u/fvqUBERESkOqGPXRUREZEWoyQuIiLSpgpN4mY2x8yeN7M1ZnZ5wvd/Y2arzexpM/s/ZnZQkfGIiIj0JoUlcTPrD9wAnARMBOaZ2cTYaL8BOtx9MnAX8J2i4hEREeltimyJTwfWuPtad38PuB04NTqCuz/s7tvKg08AYwuMR0REpFcpMomPAV6NDHeWP0vzeeDeAuMRERHpVUJfRVooM/tzoAP4k5TvLwIuAhg3blwPRiYiItK6imyJrwMOjAyPLX+2GzObBVwBnOLu7ybNyN0XunuHu3eMGjWqkGBFRETaTZFJfBkwwczGm9kg4CxgUXQEMzsSuIlSAn+jwFhERER6ncKSuLvvABYA9wPPAne6+yozu9rMTimP9l1gX+BfzWylmS1KmZ2IiIjEFHpN3N0XA4tjn10Z+f+sIpcvIiLSm+mJbSIiIm1KSVxERKRNKYmLiIi0KSVxERGRNqUkLiIi0qaUxEVERNqUkriIiEibUhIXERFpU0riIiIibUpJXEREpE0piYs0wdChYFb6V0SkVkriIk2wZcvu/4qI1EJJXKQJhgzZ/V8RkVooiYs0webN4F76t9Wp61+kdSmJi0gmdf2LtC4lcRHJpK5/kdY1oNkBiEhra4cuf5G+Si1xkQC6LiwirUhJXCSArgv3Taq8SatTEhcJ0I7XhZMSkJJSdVR5k7hWO4aUxKUpWu1AyNMqPwmrptySEpCSUnXasfImjRU/5lrtGFISl6ZotQOhJ9VTgamm3JISkJJSdVql8ibNEz/mWu0YKjSJm9kcM3vezNaY2eUJ3880syfNbIeZfbbIWKS1tNqB0JPqqcBUU25JCUhJSaQ68WOu1Y6hwpK4mfUHbgBOAiYC88xsYmy0V4DzgNuKikNaU6sdCD2pngpMkeXWbpc4RHpCq5+rimyJTwfWuPtad38PuB04NTqCu7/k7k8D3QXGIQ2gE3zjtOpJoS9f4mhVOu4kT5FJfAzwamS4s/yZlLXTAaoTfHuoZ5/qq5c4Wvk41HFXn7xt28rbPlRb3NhmZheZ2XIzW97V1dXscBomfoC28g7VV0/w7aaek36r9hAUrZUTZT3HXeV80krnlJ4+x+Vt27TvW/lcHFdkEl8HHBgZHlv+rGruvtDdO9y9Y9SoUQ0JrhXED9CsHa7ZO1VfPcG3G1W2qtfKZVbPcZf008JGqfV81IikWc24eds27ftWrtjFFZnElwETzGy8mQ0CzgIWFbi8thM/QLN2uGbtVEUdXFKM+D5V1DZph20dGmNvraAm/bQwrtHJODSmepJmNePmbdu071u5YhdXWBJ39x3AAuB+4FngTndfZWZXm9kpAGY2zcw6gT8DbjKzVUXF0w6ydrhm7VRFHVx9STMTaT3bJGv+Rc233umj37XS/tiMSk/lfJJ2TqmnjGo9HzUiaday7JDyj47TThU7c/dmx1CVjo4OX758eY8vt7LDDxnSHhu2UaIHOmSvf1YZ9fbyy1o/sw/+38jDLWS+9ZR71vyLmm+900e/GzKk+H0utBySYo5P29PHSLysmnVc9sR6h+xzRR2njWBmK9y9I+m7trixrRW0Uq2+J1VqpBVZ659Ve+3t5Ze1fkX1ooTMt54WRXS+8VZMI+abVx5praes6aPf9URrKnS/Too5Om0zeg7iZVXR7JvPilh+vPyTltGI47QZPS5K4oHa6RpJEepd/2aXX9EHV9b6FZVMik5S1VTgKkLKOTTutKSWNX29ZVLUfpIUV3Sfifd21St6Z3ra+qSVVZGViZDkWcTy4+sar0AljROiFZ6rriQeqJ2ukRSh3vUvqvxCT7pFH1yh69fIJNFTtf5qKmCNLOdmVPyK2k+StlV0n0lqFTfiGftpw1kaVe5J8SeVb94NvqHlUMtd6/FYqp1PKzxXXUm8D2qHO4tDxWvUaesVcnD1RLnkJYm0GEJOiD3ZiuwJPb3cpNZhntCTdt52T1rXpGlCt3E8nmruTm9UuSfFX8sloNCKVTXlFa845c0nTSs8V11JvA9qx+vTIddHs9YrenClzavWBFuNvJNYWgwhJ8RGVhBCNOq6YrMqlfEu53jrMG38tIRXbyUyZJrQSmv0zvSspFJtRbDW1u7QobXfwJZVdtF4QsorT3ycvOOlFXpolcR7iVoOrkb/drRIIddH620V1Zpg40JOrmkHfVoMSZ/HE0g8htD4a63UhXSNpqn3p2DxMq5ln01bXq3bP7QSmSSrggAfVDSiMWa1POPzyqsEV5aRtQ5Z2yot/ootW/LLL+8afl48SRWEeELP69pPii1ewYlO1wrnSSXxXqKaE2HeCaWnW+ohJ+B4rb6R866oNcHG1dOiT3tYC4Rts7ThrPhrvVu8nspgtIyi8wntLYifSEPLPCn+qKxWYjzh5SWcvHWJTp/VIk6rFCTFE09ISfNPqyjkifcCpVXEQso4dFlZFYeshFuRdNNeVss9LbZ42ebF0ZP6dBKvt3sxa7xGtWZD5lPLtbyQrj+ormZcq5BrWUnXCkPmk3RyrKj25pPQVmZe9181laRqWltJMUQltexCKghJsVRTsaiMX4k3KeZKXKG9BdVuu6R9It57k9dSztrvaun5iA5nXRpJqnRlxZMUR9aNXNHlRlW2VbwXoDKPrBgrZRKvbFTTlR5PnCHX+UM/S2u5V46Rai5/NPsXS306iafVHkNPslnjNao1GzKf+AkqpFKR1S2Wtj5FVHYguSWWFF9eiy2p8hH/Lq2sovMLvS6Y1J1bWWaluzOevNJaMyHXUuNlE1+HiryEHNKiSVrPtBN12rRpy85L0BVpLbqsVlGaeCKJ93DEl5cUd5akRJwkbRtn3aEd35abN+fHM2TInvPIOn6i19Cj6xRfTlKrPRp/UmUhrQIejyF6DT9tGdFpK2UeL/e0yxJpyTztxreky3Tx2Jp9PbwchLfV39FHH+2NMmRIdLcp/bnvOZw3/ZAh6fONf1dPbJV5xZebNW5F/LvoPJKmT/tLm2fSfNPGS4oxLYa08svaRmnzyVqnrO/yyipr+rx1zyrfeFlWM9+0eSSVddZ+FFKGcUnzD4m7lvIKmU/ScZM1TS1llrSf1lJ+afte1nLT9sksIdsvPu+k/SRpvbPKLzTetHNc0n4Qul5J86+mzLLWr2jAcvfknJj4YSv/NTKJu+cf0GnjZ23AeqePzyM+v7STdlbsWd9lnTSzYg45oSYtOymOvJNbVvnEhSSR0ASZt55pJ7asdU7bZvFKUNb0IckqrTxDKoGhf3nJqxHLyNoWIckzZDvm7Z9pZZe136WVV1b8WTHnyZp33javJyFVm9RCyjlrmnj8oZWLpHibkZBrkZXE+3R3OiQ/lSqrWyzrBpSKaqZPE51HfH5p17eSnlce/SlE3rIq/6+UR97PKEJu8Eq6LhaV122dF3Patd94HPFps+7sji8jK4bKsqLdkZs3p98wFO9mh9J4lc/z9o/o/JOu54ZcK40vI238ynyT1iMaC+y+ryV1BYdeh80Svy4b30+rmTZE6PGbdJkjGmPW9eh4F3FaDHk/r6rMK34uS4s7bdy0y0pJN+NVxo/PO36/Rb2Szq15v6ePxxGdR3T8Rv9ErFH3QlUlLbu36l+jW+IVoTWyvBZctdPXG298uaGtmDy1jltNLTuphZLXE5JUBtW0xqqJP76crNZVrWWTtG1qKcuQ5aWVW8g+WW0LMnT6vF6SvDJPW25oeWbNO69buJoyTxovr5cibXvkdQWHxp30eTXllDR+1jZJi7+aXsms2PP2m7yyyFNvL2w9UEs8X2iNLO0GlIq8VlDScmqpvaXVpNNa8NXWDKu5OzOqsguHTB9vsYWUf8jPWrKWlRdH9P/Rn+1Eyzou62ad0PLLunkm9IaprJZPZdroDUFJPRZJN/ZFp6ls37RembQY0x4+kndTaNbPwLJ+aZC0TZMkne5hz5sr4zfCJf0WuVI+aWWRdid8rXe9x+dbkXQspR1fWTfUZZVhreeH6DLT9qW4tGXFY0+6GS0+j6Tp4tLOxyG9qPWUS83Ssnur/hXVEq9XrTWwpJZnaK05vty0WmnoTRxF1VCrqXlnXesKudYdL8N6asV526WaZVbTWgm5tyGvJRw6XvyzkL+QFl6SWlq2WardvqH7Yd76N1LIdfToeKHXceu97ltNaz7v/BNfz0b1QhZ1HTtpWzcy/tpiUks8U/x6T/wvpBWbVgPLa2Un/YQq5Kdvadeckq53JtUcq/lda57Qh8dEh9N+ypV0jT/097bxmEJU0wtS7XqG/qwuOn7StdW0fSDrPoy8+xjiP7OKfh9yr0P8/9F9Mq1HIGk7hvTAJB2f8WvvWaLrmfXTrXh5pPUCRMdN+9ll6Lkj9N6YrGMzreekMn7ocZ3WyxWPNasnIO1+i6T5JMUe8nnI+lQ736T1Cb33qJp5FyItu7fqXxEt8WpbIXmtkngNOKv2n1XjzWvpZbWSQmv4WcsMiTVLSAxpNd688qhlXnnbJ2Rdq2n9p+0zaappkeXNK6QlltfiyNvHqi2PWntsqimPvPXMKrtq1jttPiH7Z9p8krZV3vbIW2be9GnnqaR4qz0Oa1n/rPIN6UlpxHbL6s2s5nzQSOgnZil27nTv7PTDedphZ907Zej4FVknoGpOaKEnrayKRV6s8fnGhXaLJi03r1IQUjnKmk+1iSlrXUPmEXLiTZI376R9IiRBh86r2v03q3xDkmvI9smrDOfFGLq9Q/eTvH2w2jJMK9e0inw1+1u156W8bVfP+tb7F3rM55V33nkqrazj0+Tts42mJJ5k50734493B+8G38xgn8Lj/jEe8GmDV5a+9/Cdyb26HT0vcYYuo9odK7S1mBd7/GDIOmCylp9U640KaXmmrUetJ7Gs7ZKX/JJiio5XTfLIO0GlLTu00hO6rePbJ2vbV1OGodNnyWsJ5ZVpXKP2lVrKNW9bJ61v0uehrcWQuJL2s7TpsoYbVWa1zjutHKppSGXtV9XsY7VQEk/y+uvu/fqlb4m99nJ//HGfs9cvfTTrHLp320BpG62aHb2W8WptRYXEV2tsSfPNijMv3tDyTFt22sFb74kmTVYLOemkENJiD1lOyDZO+qyaxJ50Qsoqw2pbcWnxZlV6suJJ2u5Z61JtGTfyL6Ril1RJr2W98tYtr7xD9r28ylm1+12tFZPQc1TWMVSRVs6hx2KjZCVxK33fPjo6Onz58uX1z8gdZs6ERx8NGn0Zk/kq32EDI3nqgU189OThDH5/E9sHDufXv9gEw4fDpt0/H2Gb6HxvJIP3Hsh/bp0E/folPozDPf1moMrmiX7vvvvPfkJujok+lKPaG9ei80l7oEW1D3bIi7uy3nnrGV92fLz47t2I9U97jnNS/FHxbZhWZkk3XUWllUnScGU+8WUnxZRVxknLz4o96adsSTfRpY2btc6w+/hJ8YSuS5q08oouK287JY0fjT1rP82LKWtbh8bTqHlVxs86RyXdRFjLMvPWJ2mfyNpPokKO2aRlV7Mf18rMVrh7R+J3fTaJA3R3Q2cnHHRQ7qjxUoqffz32XXyYvfaChx+GrVth+HA+ffwm1r03nJFsYgO7/7uekRjOmEEb2eilSkF0nBFs5tW9P0JX/z9gy9bdI4nfzRr9vNpEHk1aSfOMzjfpYIpOkzSPrFhDx006WYQcTCFJLb6OodOkLTctzmqmjY+TltDTEk+atP0jtAyjkpadtD9Uc6ILWVZ8vFqTE6TvQ1kn9WpVe9LP23/qiSfp+K02IRWZxIqSF3OrrJOSeIahQ2HblveYvddSFj8yDNavh5NO2jMJN0BeSceTf56VTOQ8fsRINvPwfw6HTZtg5Ehw59MzNu5RSYj3GkyZ/UGl4ellDhs37vpsA8N56oHyuJs3c8Csj/AGf0BSaVTTcownw6QWU62tqJCaf1qiz2oZ5LUaeuJAT3vkaFRI4klS6ykgpBKS1ftQzXKzWuJp4xWxLVrlpN7IeNISeEWRKaLVyhPSY+qJnoU0TUviZjYHuBboD9zs7t+Ofb8XcAtwNLAeONPdX8qaZ6OTeNIJrj/vcQJL2cBQfsrnmcIz9S+n7jnsKat3IG2rVltRiIpWGiq9BSPZyAaGM2bQJhY/Ony3ykG8h6FSMTj2MyN5570Ppk0bN947sc8gZ/B7G3ln0EieWDaQocdNYsvb/VIrB2lJONp6T5LVustr+UH4Samak0VW6zi0NZGk2pNN2rzyenqSKk8Q3gqKj5tXdiHrV00Cqaa3p1Gtu6zerUYmiaT9u5peokYuO2250Ph9OOSyWHT+SeeXRh5bWZqSxM2sP/AC8CmgE1gGzHP31ZFxvghMdveLzewsYK67n5k13yJa4lldjUY3B/Aao3mdkWwITjqVfx/+j2HsOOYY+mfEUNk3qtkSRVQK8oT0JOSNV8u6ps1nM3txGneznv13u09hV9mXeycmz/6gwlHpjUiraOzqrSj3aLBxY+r9Dwwfzsenp18OeeqBjbuNG/83Ou3bg0fzn28fAf36Vd1aDRm/US2epEpvVsUmK6a8uLO60Ku5HFLr8rMuVYTeo5K3XrVcrojPu5bKYN5naTHU0xuWNX69qokrbfvkPbY46VJeUb1ccc1K4h8FvuHuJ5aHvw7g7t+KjHN/eZxfm9kA4HVglGcE1egkXtGonSmqsiMPsErLfljV3dt7Xj8fxq1cwGRWAQnX3gOEJNla592TkuKr5bP4d7XGUO389uhJ2WsvuPtuPnrq/qmVhqR/Kz0bYwZt3NUbkjdNtJKSO27k308fX7pZM14pio9T2V93qxSZwYABMGn3mzxbtSWedNknbflZ46YttyKv4hWdb7wc4nGmJaW0G2TzhFY4spYZIi8Zpl1iqyWutO2TdV9AXs9SWpztnsQ/C8xx9/nl4XOAY9x9QWScZ8rjdJaHf1ce5820+RaVxNtKdzesWwfPPw/DhtV/cs6aZtgwuOACWLWqx1dTeqlyZYX99y+uwlFjJSXpvpLopaLEabZuZehnZrJlW/+ab1KrR70t8UbGWet6hcRaT5lVewmjluvhjYgzTdsncTO7CLgIYNy4cUe//PLLhcQsKZIqDQ04SdZ0Ut6wAb72NVi5sufWXyTEY4/B22/37PHQqONv//3hwx8u/TXqReDSMFlJfECBy10HHBgZHlv+LGmcznJ3+jBKN7jtxt0XAguh1BIvJFpJ168fHHhg6a8VzJoFr70Gr79eSurtcJLM+nfDBpg7F955p8BCk8LNmNHsCOp3xBHw3e+W9uFmHQ8h02zdWnrOR/+su436hiKT+DJggpmNp5SszwI+FxtnEfAXwK+BzwIPZV0PFwFKlYoxY0p/vcXmzfD006WfOLbCSbLR044YARdeqB6UVvfb38KcOc2OIty//zuMHt34fbqNKgmFJXF332FmC4D7Kf3E7EfuvsrMrqb0CLlFwA+BW81sDbCBUqIX6XsGDICjjmp2FMVasaK6HpR2qKQMHQof/7h6UZrlM58pdv7RSkJ8248eXeq56Nev2Bhy9PmHvYiI1GXHDnjmmdK/rdLdXM20GzbA6afDtm0FF1QvFL9Js6B7C5p1TVxEpPcbMACmTm12FPV5663sikgrXRMfOhQ+9jHYvr0HCibHu+/Cpz+95+dTp8KyZaV9o2BK4iIifV27VUS2bNnzHpJGVTQaUUlYubI0zw0bCk/kSuIiItJeir6HJKmSUO0vSrZsgeeeKz3cqEBK4iIiIlEhlYT4L0ri9xYMGwYTJxYfauFLEBER6W2SEn3l3oJ+/XY9XrjwMApfgoiISF/QhHsLmvsDNxEREamZkriIiEibUhIXERFpU0riIiIibUpJXEREpE0piYuIiLQpJXEREZE2pSQuIiLSppTERURE2lTbvU/czLqAl5sdh4iISA85yN1HJX3RdklcREREStSdLiIi0qaUxEVERNqU3mIm0geY2U7gt8BAYAdwC3CNu3c3NTARqYuSuEjf8I67TwUws9HAbcBQ4KpmBiUi9VF3ukgf4+5vABcBC6ykv5l918yWmdnTZvaFyrhm9jUz+62ZPWVm3y5/NtXMniiPe7eZjTCzPzKzJyPTTagMm9m3zWx1efy/6+n1FenN1BIX6YPcfa2Z9QdGA6cCb7n7NDPbC3jMzB4ADit/d4y7bzOzkeXJbwEudfdfmdnVwFXu/t/N7C0zm+ruK4HzgR+b2X7AXOAwd3czG96zayrSu6klLiKzgXPNbCXwH8B+wARgFvBjd98G4O4bzGwYMNzdf1We9qfAzPL/bwbOL1cOzqTUZf8WsB34oZmdDmzrmVUS6RuUxEX6IDM7BNgJvAEYpZb11PLfeHd/oIbZ/gw4CfgMsMLd17v7DmA6cFf58/saswYiAkriIn2OmY0CbgS+76WnPd0P/KWZDSx//8dmtg/wIKWW9YfKn49097eAjWZ2Qnl25wC/AnD37eV5/QD4cXmafYFh7r4Y+BIwpYdWU6RP0DVxkb5h73J3eeUnZrcCf1/+7mbgYOBJMzOgCzjN3e8zs6nAcjN7D1gM/A/gL4Aby8l9LaXr3xX/TOkaeKUlPwT4NzMbTKnF/zdFraBIX6THropIw5jZZZRa3v+z2bGI9AVqiYtIQ5jZ3cAfAZ9odiwifYVa4iIiIm1KN7aJiIi0KSVxERGRNqUkLiIi0qaUxEVERNqUkriIiEibUhIXERFpU/8/WurWhXgVbvkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_results(\"1pvh\", scores_dockq_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fba2cfd6",
   "metadata": {},
   "source": [
    "### Test complex 1 results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "8811f9ee",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_results(\"1b6c\", scores_dockq_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7418d0f4",
   "metadata": {},
   "source": [
    "### Test complex 2 results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "490d8de4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_results(\"1e6e\", scores_dockq_test)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}