data/molecules/prepare_molecules.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Notebook for preparing and saving MOLECULAR graphs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import torch\n",
    "import pickle\n",
    "import time\n",
    "import os\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Download ZINC dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "downloading..\n",
      "  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current\n",
      "                                 Dload  Upload   Total   Spent    Left  Speed\n",
      "  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0\n",
      "  0     0    0     0    0     0      0      0 --:--:--  0:00:01 --:--:--     0\n",
      "100  366M  100  366M    0     0  1805k      0  0:03:27  0:03:27 --:--:-- 2863k:36  0:00:42  0:02:54  850k00:52  0:03:03  717k      0  0:03:55  0:01:26  0:02:29 1089k     0  0:04:04  0:01:56  0:02:08 2071k32k 0  0:03:45  0:03:00  0:00:45  534kk   0     0  1790k      0  0:03:29  0:03:25  0:00:04 2757k\n",
      "x molecules/atom_dict.pickle\n",
      "x molecules/bond_dict.pickle\n",
      "x molecules/test.pickle\n",
      "x molecules/train.pickle\n",
      "x molecules/val.pickle\n"
     ]
    }
   ],
   "source": [
    "if not os.path.isfile('molecules.zip'):\n",
    "    print('downloading..')\n",
    "    !curl https://www.dropbox.com/s/feo9qle74kg48gy/molecules.zip?dl=1 -o molecules.zip -J -L -k\n",
    "    !tar -xvf molecules.zip -C ../\n",
    "else:\n",
    "    print('File already downloaded')\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Convert to DGL format and save with pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/Users/xbresson/Documents/Dropbox/06_NTU_2017_now/03_my_codes/34_benchmark20/GITHUB_benchmark_project/benchmarking-gnn\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "os.chdir('../../') # go to root folder of the project\n",
    "print(os.getcwd())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "from data.molecules import MoleculeDatasetDGL \n",
    "\n",
    "from data.data import LoadData\n",
    "from torch.utils.data import DataLoader\n",
    "from data.molecules import MoleculeDataset\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "preparing 10000 graphs for the TRAIN set...\n",
      "preparing 1000 graphs for the VAL set...\n",
      "preparing 1000 graphs for the TEST set...\n",
      "Time taken: 90.7071s\n"
     ]
    }
   ],
   "source": [
    "DATASET_NAME = 'ZINC'\n",
    "dataset = MoleculeDatasetDGL(DATASET_NAME) \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "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"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "min/max : 9 37\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "min/max : 10 36\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAARgElEQVR4nO3de4zlZX3H8fenq3hBLSgD2XJxwaAWjV2bKWq8lBatCCpqA7JRC9W6kkqi0baCbeqlIaVWNE1asUshoFWQFC80UJVSKxrrZcAVF4EKuOrCdncUKFANleXbP+a36XE9w+yc3znMzjPvV3Iyv/P8Luf78LCfffY5v3MmVYUkqS2/tNQFSJLGz3CXpAYZ7pLUIMNdkhpkuEtSgwx3SWqQ4S5JDTLc1Zwkm5O8sOc1Tkny5T2lHmmxDHdJapDhrqYk+ShwCPDPSe5N8idJnp3kK0nuSvKtJEcNHH9KkluT3JPke0lek+RXgQ8Dz+mucVd37LFJvtMde1uSPxq4zkuTbOxe4ytJnjFfPQ/hfw6tYPHrB9SaJJuBP6iqf01yIHAd8Drgs8DRwMXAU4GfAFuB36iqm5KsBh5fVdcnOaW7xvMGrrsVOLGqvpRkX+DQqro2ya93134ZMAO8FngP8JSqum+wnoei/xI4c1f7XgtcUVVXVNUDVXUlcwF8bLf/AeDpSR5VVVur6voHudbPgCOSPK6q7qyqa7v2NwJ/X1Vfq6odVXUhcB/w7An1SVqQ4a7WPRE4oVsuuatbYnkesLqq/gd4NXAqsDXJ5Ume+iDX+l3m/lL4fpIvJnnOwGu8fZfXOBj4lYn1SlqA4a4WDa41/hD4aFXtM/DYu6rOAqiqz1XVi4DVwI3AuUOuQXfsN6rqeGB/4NPAJQOvceYur/HoqrpovmtJk2a4q0XbgMO67X8EXpbkxUlWJXlkkqOSHJTkgCQvT7I3c8so9wI7Bq5xUJK9AJLs1b3Z+stV9TPg7oFjzwVOTfKszNk7yXFJHjukHukhYbirRX8J/Fm3PPJq4HjgncAsc7PsP2bu//1fAt4O3A7cAfwm8IfdNf4NuB74ryQ/6tpeB2xOcjdzSzmvBaiqGebW3f8WuBO4GThlWD2Dd9hIk+TdMpLUIGfuktQgw12SGmS4S1KDDHdJatDDlroAgP3226/WrFmz1GVI0rJyzTXX/Kiqpobt2yPCfc2aNczMzCx1GZK0rCT5/nz7XJaRpAYZ7pLUIMNdkhpkuEtSgwx3SWqQ4S5JDTLcJalBhrskNchwl6QG7RGfUJX2ZGtOv3zkczefddwYK5F234Iz9yTnJ9meZNNA2yeSbOwem5Ns7NrXJPnpwL4PT7J4SdJwuzNzv4C5Xx/2kZ0NVfXqndtJzgb+e+D4W6pq7bgKlCQt3oLhXlVXJ1kzbF+SACcCvz3esiRJffR9Q/X5wLaq+u5A26FJvpnki0meP9+JSdYnmUkyMzs727MMSdKgvuG+Drho4PlW4JCqeibwNuDjSR437MSq2lBV01U1PTU19OuIJUkjGjnckzwMeBXwiZ1tVXVfVf24274GuAV4ct8iJUmL02fm/kLgxqrasrMhyVSSVd32YcDhwK39SpQkLdbu3Ap5EfAfwFOSbEnyhm7XSfz8kgzAC4DrknwL+Cfg1Kq6Y5wFS5IWtjt3y6ybp/2UIW2XApf2L0uS1IdfPyBJDTLcJalBhrskNchwl6QGGe6S1CDDXZIaZLhLUoMMd0lqkOEuSQ0y3CWpQYa7JDXIcJekBhnuktQgw12SGmS4S1KDDHdJapDhLkkNMtwlqUGGuyQ1yHCXpAYtGO5Jzk+yPcmmgbZ3J7ktycbucezAvjOS3JzkpiQvnlThkqT57c7M/QLgmCHtH6yqtd3jCoAkRwAnAU/rzvlQklXjKlaStHsWDPequhq4YzevdzxwcVXdV1XfA24GjuxRnyRpBH3W3E9Lcl23bLNv13Yg8MOBY7Z0bb8gyfokM0lmZmdne5QhSdrVqOF+DvAkYC2wFTi7a8+QY2vYBapqQ1VNV9X01NTUiGVIkoYZKdyraltV7aiqB4Bz+f+lly3AwQOHHgTc3q9ESdJijRTuSVYPPH0lsPNOmsuAk5I8IsmhwOHA1/uVKElarIctdECSi4CjgP2SbAHeBRyVZC1zSy6bgTcBVNX1SS4BvgPcD7y5qnZMpnRJ0nwWDPeqWjek+bwHOf5M4Mw+RUmS+vETqpLUIMNdkhpkuEtSgwx3SWqQ4S5JDTLcJalBhrskNchwl6QGGe6S1KAFP6EqDVpz+uUjn7v5rOPGWImkB+PMXZIaZLhLUoMMd0lqkOEuSQ0y3CWpQYa7JDXIcJekBhnuktQgw12SGmS4S1KDFgz3JOcn2Z5k00DbXye5Mcl1ST6VZJ+ufU2SnybZ2D0+PMniJUnD7c7M/QLgmF3argSeXlXPAP4TOGNg3y1VtbZ7nDqeMiVJi7FguFfV1cAdu7R9vqru755+FThoArVJkkY0jjX31wP/MvD80CTfTPLFJM+f76Qk65PMJJmZnZ0dQxmSpJ16hXuSPwXuBz7WNW0FDqmqZwJvAz6e5HHDzq2qDVU1XVXTU1NTfcqQJO1i5HBPcjLwUuA1VVUAVXVfVf24274GuAV48jgKlSTtvpHCPckxwDuAl1fVTwbap5Ks6rYPAw4Hbh1HoZKk3bfgb2JKchFwFLBfki3Au5i7O+YRwJVJAL7a3RnzAuC9Se4HdgCnVtUdQy8sSZqYBcO9qtYNaT5vnmMvBS7tW5QkqR8/oSpJDTLcJalBhrskNchwl6QGGe6S1CDDXZIaZLhLUoMMd0lqkOEuSQ0y3CWpQYa7JDXIcJekBhnuktQgw12SGmS4S1KDDHdJapDhLkkNMtwlqUGGuyQ1yHCXpAYtGO5Jzk+yPcmmgbbHJ7kyyXe7n/sO7Dsjyc1Jbkry4kkVLkma3+7M3C8Ajtml7XTgqqo6HLiqe06SI4CTgKd153woyaqxVStJ2i0LhntVXQ3csUvz8cCF3faFwCsG2i+uqvuq6nvAzcCRY6pVkrSbRl1zP6CqtgJ0P/fv2g8Efjhw3Jau7RckWZ9kJsnM7OzsiGVIkoYZ9xuqGdJWww6sqg1VNV1V01NTU2MuQ5JWtlHDfVuS1QDdz+1d+xbg4IHjDgJuH708SdIoRg33y4CTu+2Tgc8MtJ+U5BFJDgUOB77er0RJ0mI9bKEDklwEHAXsl2QL8C7gLOCSJG8AfgCcAFBV1ye5BPgOcD/w5qraMaHaJUnzWDDcq2rdPLuOnuf4M4Ez+xQlSerHT6hKUoMMd0lqkOEuSQ0y3CWpQYa7JDXIcJekBhnuktQgw12SGmS4S1KDDHdJapDhLkkNMtwlqUGGuyQ1aMFvhZTGZc3pl4987uazjhtjJVL7DHctC33+YgD/ctDKY7hLeyj/paM+XHOXpAYZ7pLUIMNdkhpkuEtSg0Z+QzXJU4BPDDQdBvw5sA/wRmC2a39nVV0xcoWSpEUbOdyr6iZgLUCSVcBtwKeA3wc+WFXvH0uFkhbNW0c1rmWZo4Fbqur7Y7qeJKmHcYX7ScBFA89PS3JdkvOT7DvshCTrk8wkmZmdnR12iCRpRL0/xJRkL+DlwBld0znAXwDV/TwbeP2u51XVBmADwPT0dPWtQ3owfZcppOVmHDP3lwDXVtU2gKraVlU7quoB4FzgyDG8hiRpEcYR7usYWJJJsnpg3yuBTWN4DUnSIvRalknyaOBFwJsGmt+XZC1zyzKbd9knSXoI9Ar3qvoJ8IRd2l7XqyJJUm9+QlWSGuRX/koT5F06WirO3CWpQYa7JDXIcJekBhnuktQgw12SGmS4S1KDDHdJapDhLkkNMtwlqUGGuyQ1yHCXpAYZ7pLUIMNdkhpkuEtSgwx3SWqQ4S5JDTLcJalBhrskNajXr9lLshm4B9gB3F9V00keD3wCWANsBk6sqjv7lSlJWoxxzNx/q6rWVtV09/x04KqqOhy4qnsuSXoITWJZ5njgwm77QuAVE3gNSdKD6BvuBXw+yTVJ1ndtB1TVVoDu5/7DTkyyPslMkpnZ2dmeZUiSBvVacweeW1W3J9kfuDLJjbt7YlVtADYATE9PV886JEkDes3cq+r27ud24FPAkcC2JKsBup/b+xYpSVqckcM9yd5JHrtzG/gdYBNwGXByd9jJwGf6FilJWpw+yzIHAJ9KsvM6H6+qzyb5BnBJkjcAPwBO6F+mJGkxRg73qroV+LUh7T8Gju5TlCSpHz+hKkkNMtwlqUGGuyQ1yHCXpAb1/RCTRrTm9MtHPnfzWceNsRJJLXLmLkkNcua+DPWZ9YMzf2klcOYuSQ0y3CWpQYa7JDXINfcVqO+avaQ9nzN3SWqQ4S5JDTLcJalBhrskNchwl6QGGe6S1CDDXZIaZLhLUoMMd0lq0MjhnuTgJF9IckOS65O8pWt/d5LbkmzsHseOr1xJ0u7o8/UD9wNvr6prkzwWuCbJld2+D1bV+/uXJ0kaxcjhXlVbga3d9j1JbgAOHFdhkqTRjeWLw5KsAZ4JfA14LnBakt8DZpib3d855Jz1wHqAQw45ZBxlSBoTfw3k8tf7DdUkjwEuBd5aVXcD5wBPAtYyN7M/e9h5VbWhqqaranpqaqpvGZKkAb3CPcnDmQv2j1XVJwGqaltV7aiqB4BzgSP7lylJWow+d8sEOA+4oao+MNC+euCwVwKbRi9PkjSKPmvuzwVeB3w7ycau7Z3AuiRrgQI2A2/qVaEkadH63C3zZSBDdl0xejmSpHHwE6qS1CB/h6qkPYa3YI6PM3dJapAzd0lj1Wf2rfFx5i5JDTLcJalBhrskNchwl6QG+YZqD75xJGlP5cxdkhpkuEtSgwx3SWqQ4S5JDTLcJalBTdwt45cNSdLPc+YuSQ1qYubeh/eqS2rRig93SeprT1wadllGkhrkzF3Sitfi8uzEwj3JMcDfAKuAf6iqsyb1WpLUYkD3MZFlmSSrgL8DXgIcAaxLcsQkXkuS9IsmteZ+JHBzVd1aVf8LXAwcP6HXkiTtYlLLMgcCPxx4vgV41uABSdYD67un9ya5aWD3fsCPJlTbnmSl9BNWTl9XSj9h5fR1ov3MX/U6/Ynz7ZhUuGdIW/3ck6oNwIahJyczVTU9icL2JCuln7By+rpS+gkrp6/LtZ+TWpbZAhw88Pwg4PYJvZYkaReTCvdvAIcnOTTJXsBJwGUTei1J0i4msixTVfcnOQ34HHO3Qp5fVdcv4hJDl2satFL6CSunryuln7By+ros+5mqWvgoSdKy4tcPSFKDDHdJatCSh3uS85NsT7JpoO3xSa5M8t3u575LWeM4zNPPdye5LcnG7nHsUtY4DkkOTvKFJDckuT7JW7r2Fsd0vr42Na5JHpnk60m+1fXzPV17i2M6X1+X3Zgu+Zp7khcA9wIfqaqnd23vA+6oqrOSnA7sW1XvWMo6+5qnn+8G7q2q9y9lbeOUZDWwuqquTfJY4BrgFcAptDem8/X1RBoa1yQB9q6qe5M8HPgy8BbgVbQ3pvP19RiW2Zgu+cy9qq4G7til+Xjgwm77Qub+wCxr8/SzOVW1taqu7bbvAW5g7hPLLY7pfH1tSs25t3v68O5RtDmm8/V12VnycJ/HAVW1Feb+AAH7L3E9k3Rakuu6ZZtl/8/aQUnWAM8EvkbjY7pLX6GxcU2yKslGYDtwZVU1O6bz9BWW2ZjuqeG+UpwDPAlYC2wFzl7acsYnyWOAS4G3VtXdS13PJA3pa3PjWlU7qmotc582PzLJ05e6pkmZp6/Lbkz31HDf1q1n7lzX3L7E9UxEVW3r/kd6ADiXuW/TXPa6tcpLgY9V1Se75ibHdFhfWx1XgKq6C/h35tagmxzTnQb7uhzHdE8N98uAk7vtk4HPLGEtE7PzD0bnlcCm+Y5dLro3pM4DbqiqDwzsam5M5+tra+OaZCrJPt32o4AXAjfS5pgO7etyHNM94W6Zi4CjmPtazW3Au4BPA5cAhwA/AE6oqmX9ZuQ8/TyKuX/mFbAZeNPONczlKsnzgC8B3wYe6JrfydxadGtjOl9f19HQuCZ5BnNvmK5ibkJ4SVW9N8kTaG9M5+vrR1lmY7rk4S5JGr89dVlGktSD4S5JDTLcJalBhrskNchwl6QGGe6S1CDDXZIa9H+nSoFf5nOUxQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "min/max : 11 37\n"
     ]
    }
   ],
   "source": [
    "def plot_histo_graphs(dataset, title):\n",
    "    # histogram of graph sizes\n",
    "    graph_sizes = []\n",
    "    for graph in dataset:\n",
    "        graph_sizes.append(graph[0].number_of_nodes())\n",
    "    plt.figure(1)\n",
    "    plt.hist(graph_sizes, bins=20)\n",
    "    plt.title(title)\n",
    "    plt.show()\n",
    "    graph_sizes = torch.Tensor(graph_sizes)\n",
    "    print('min/max :',graph_sizes.min().long().item(),graph_sizes.max().long().item())\n",
    "    \n",
    "plot_histo_graphs(dataset.train,'trainset')\n",
    "plot_histo_graphs(dataset.val,'valset')\n",
    "plot_histo_graphs(dataset.test,'testset')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000\n",
      "1000\n",
      "1000\n",
      "(DGLGraph(num_nodes=29, num_edges=64,\n",
      "         ndata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}\n",
      "         edata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}), tensor([0.8350]))\n",
      "(DGLGraph(num_nodes=35, num_edges=78,\n",
      "         ndata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}\n",
      "         edata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}), tensor([0.6299]))\n",
      "(DGLGraph(num_nodes=16, num_edges=34,\n",
      "         ndata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}\n",
      "         edata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}), tensor([1.9973]))\n"
     ]
    }
   ],
   "source": [
    "print(len(dataset.train))\n",
    "print(len(dataset.val))\n",
    "print(len(dataset.test))\n",
    "\n",
    "print(dataset.train[0])\n",
    "print(dataset.val[0])\n",
    "print(dataset.test[0])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_atom_type = 28\n",
    "num_bond_type = 4\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time (sec): 6.659376859664917\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "with open('data/molecules/ZINC.pkl','wb') as f:\n",
    "        pickle.dump([dataset.train,dataset.val,dataset.test,num_atom_type,num_bond_type],f)\n",
    "print('Time (sec):',time.time() - start)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test load function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[I] Loading dataset ZINC...\n",
      "train, test, val sizes : 10000 1000 1000\n",
      "[I] Finished loading.\n",
      "[I] Data load time: 8.5329s\n"
     ]
    }
   ],
   "source": [
    "DATASET_NAME = 'ZINC'\n",
    "dataset = LoadData(DATASET_NAME)\n",
    "trainset, valset, testset = dataset.train, dataset.val, dataset.test\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'data.molecules.MoleculeDataset'>\n"
     ]
    }
   ],
   "source": [
    "batch_size = 10\n",
    "collate = MoleculeDataset.collate\n",
    "print(MoleculeDataset)\n",
    "train_loader = DataLoader(trainset, batch_size=batch_size, shuffle=True, collate_fn=collate)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}