data/molecules/prepare_molecules_ZINC_full.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Notebook for preparing and saving MOLECULAR (ZINC full) 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 full 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\n",
      "100 43.6M  100 43.6M    0     0  7898k      0  0:00:05  0:00:05 --:--:-- 10.3M\n",
      "Archive:  molecules_zinc_full.zip\n",
      "   creating: ../molecules/zinc_full/\n",
      "  inflating: ../__MACOSX/molecules/._zinc_full  \n",
      "  inflating: ../molecules/zinc_full/test.pickle  \n",
      "  inflating: ../__MACOSX/molecules/zinc_full/._test.pickle  \n",
      "  inflating: ../molecules/zinc_full/.DS_Store  \n",
      "  inflating: ../__MACOSX/molecules/zinc_full/._.DS_Store  \n",
      "  inflating: ../molecules/zinc_full/val.pickle  \n",
      "  inflating: ../__MACOSX/molecules/zinc_full/._val.pickle  \n",
      "  inflating: ../molecules/zinc_full/atom_dict.pickle  \n",
      "  inflating: ../__MACOSX/molecules/zinc_full/._atom_dict.pickle  \n",
      "  inflating: ../molecules/zinc_full/bond_dict.pickle  \n",
      "  inflating: ../__MACOSX/molecules/zinc_full/._bond_dict.pickle  \n",
      "  inflating: ../molecules/zinc_full/train.pickle  \n",
      "  inflating: ../__MACOSX/molecules/zinc_full/._train.pickle  \n"
     ]
    }
   ],
   "source": [
    "if not os.path.isfile('molecules_zinc_full.zip'):\n",
    "    print('downloading..')\n",
    "    !curl https://www.dropbox.com/s/grhitgnuuixoxwl/molecules_zinc_full.zip?dl=1 -o molecules_zinc_full.zip -J -L -k\n",
    "    !unzip molecules_zinc_full.zip -d ../\n",
    "    !rm -r ../__MACOSX/\n",
    "    # !tar -xvf molecules_zinc_full.zip -C ../\n",
    "else:\n",
    "    print('File already downloaded')"
   ]
  },
  {
   "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": [
      "/home/vijay/graphdeeplearning/benchmarking-gnns\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": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:root:The OGB package is out of date. Your version is 1.2.2, while the latest version is 1.2.3.\n"
     ]
    }
   ],
   "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 220011 graphs for the TRAIN set...\n",
      "preparing 24445 graphs for the VAL set...\n",
      "preparing 5000 graphs for the TEST set...\n",
      "Time taken: 1075.5591s\n"
     ]
    }
   ],
   "source": [
    "DATASET_NAME = 'ZINC-full'\n",
    "dataset = MoleculeDatasetDGL(DATASET_NAME)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "min/max : 6 38\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 : 6 38\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 : 8 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": [
      "220011\n",
      "24445\n",
      "5000\n",
      "(DGLGraph(num_nodes=33, num_edges=72,\n",
      "         ndata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}\n",
      "         edata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}), tensor([3.0464]))\n",
      "(DGLGraph(num_nodes=21, num_edges=44,\n",
      "         ndata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}\n",
      "         edata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}), tensor([2.0992]))\n",
      "(DGLGraph(num_nodes=24, num_edges=52,\n",
      "         ndata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}\n",
      "         edata_schemes={'feat': Scheme(shape=(), dtype=torch.int64)}), tensor([3.1382]))\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": "stderr",
     "output_type": "stream",
     "text": [
      "/home/vijay/miniconda3/envs/benchmark_gnn/lib/python3.7/site-packages/torch/storage.py:34: FutureWarning: pickle support for Storage will be removed in 1.5. Use `torch.save` instead\n",
      "  warnings.warn(\"pickle support for Storage will be removed in 1.5. Use `torch.save` instead\", FutureWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time (sec): 131.04525065422058\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "with open('data/molecules/ZINC-full.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) #131s"
   ]
  },
  {
   "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-full...\n",
      "train, test, val sizes : 220011 5000 24445\n",
      "[I] Finished loading.\n",
      "[I] Data load time: 154.6146s\n"
     ]
    }
   ],
   "source": [
    "DATASET_NAME = 'ZINC-full'\n",
    "dataset = LoadData(DATASET_NAME)\n",
    "trainset, valset, testset = dataset.train, dataset.val, dataset.test #154s"
   ]
  },
  {
   "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": []
  },
  {
   "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": 4
}