WEGL_RandomForest_ogbg_molhiv.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: ogb in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (1.1.1)\n",
      "Requirement already satisfied: tqdm>=4.29.0 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from ogb) (4.46.0)\n",
      "Requirement already satisfied: torch>=1.2.0 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from ogb) (1.4.0)\n",
      "Requirement already satisfied: pandas>=0.24.0 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from ogb) (1.0.3)\n",
      "Requirement already satisfied: numpy>=1.16.0 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from ogb) (1.18.1)\n",
      "Requirement already satisfied: scikit-learn>=0.20.0 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from ogb) (0.23.1)\n",
      "Requirement already satisfied: urllib3>=1.24.0 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from ogb) (1.25.9)\n",
      "Requirement already satisfied: six>=1.12.0 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from ogb) (1.14.0)\n",
      "Requirement already satisfied: python-dateutil>=2.6.1 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from pandas>=0.24.0->ogb) (2.8.1)\n",
      "Requirement already satisfied: pytz>=2017.2 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from pandas>=0.24.0->ogb) (2020.1)\n",
      "Requirement already satisfied: scipy>=0.19.1 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from scikit-learn>=0.20.0->ogb) (1.4.1)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from scikit-learn>=0.20.0->ogb) (2.0.0)\n",
      "Requirement already satisfied: joblib>=0.11 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from scikit-learn>=0.20.0->ogb) (0.15.1)\n",
      "Requirement already satisfied: POT in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (0.7.0)\n",
      "Requirement already satisfied: numpy>=1.16 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from POT) (1.18.1)\n",
      "Requirement already satisfied: cython>=0.23 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from POT) (0.29.19)\n",
      "Requirement already satisfied: scipy>=1.0 in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (from POT) (1.4.1)\n",
      "Requirement already satisfied: prettytable in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (0.7.2)\n",
      "Requirement already satisfied: tqdm in /home/user/miniconda/envs/py36/lib/python3.6/site-packages (4.46.0)\n"
     ]
    }
   ],
   "source": [
    "# Install (some of the) required packages\n",
    "\n",
    "!pip install ogb\n",
    "!pip install POT\n",
    "!pip install prettytable\n",
    "!pip install tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ogb version 1.1.1\n"
     ]
    }
   ],
   "source": [
    "# Import required modules\n",
    "\n",
    "import ogb; print('ogb version {}'.format(ogb.__version__)) # make sure the version is =>1.1.1.\n",
    "from ogb.graphproppred import PygGraphPropPredDataset\n",
    "from WEGL.WEGL import WEGL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set the random seed\n",
    "\n",
    "random_seed = 55"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# of graphs = 41127\n",
      "# of classes = 2\n",
      "# of node features = 9\n",
      "# of edge features = 3\n",
      "# of tasks = 1\n"
     ]
    }
   ],
   "source": [
    "# Load the dataset\n",
    "\n",
    "dataset = PygGraphPropPredDataset(name=\"ogbg-molhiv\")\n",
    "\n",
    "print('# of graphs = {0}\\n# of classes = {1}\\n# of node features = {2}\\n# of edge features = {3}'.\\\n",
    "         format(len(dataset), dataset.num_classes, dataset.num_node_features, dataset.num_edge_features))\n",
    "\n",
    "if isinstance(dataset, PygGraphPropPredDataset): # OGB datasets\n",
    "    print('# of tasks = {}'.format(dataset.num_tasks))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Specify the parameters\n",
    "\n",
    "# num_hidden_layers = range(3, 9)\n",
    "num_hidden_layers = [4]\n",
    "\n",
    "# node_embedding_sizes = [100, 300, 500]\n",
    "node_embedding_sizes = [300]\n",
    "\n",
    "# final_node_embeddings = ['concat', 'avg', 'final']\n",
    "final_node_embeddings = ['final']\n",
    "\n",
    "num_pca_components = 20\n",
    "num_experiments = 10\n",
    "classifiers = ['RF']\n",
    "device = 'cpu'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "****************************************************************************************************\n",
      "# diffusion layers = 4, node embedding size = 300, node embedding mode: final\n",
      "\n",
      "Now diffusing the train data ...\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a1fd7d635d564cc089a673db9172208b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=1029.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Now diffusing the valid data ...\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "516c1e2e03714e1a865f2f03de70e917",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=129.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Now diffusing the test data ...\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7066e2eda2064272a6141302d1ce983d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=129.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Now running PCA ...\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": [
      "Now running k-means for deriving the template ...\n",
      "\n",
      "Now deriving the final graph embeddings for the train data ...\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "84017f4afa3a4033ac0574ffda2503e9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=32901.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Now deriving the final graph embeddings for the valid data ...\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c1a775de37e443bab96a6e9b8b526646",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=4113.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Now deriving the final graph embeddings for the test data ...\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "25f489ec0fc94e87bcc8c15f43ddcd4a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=4113.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Now running the classifiers ...\n",
      "experiment 1/10 for RF completed ...\n",
      "experiment 2/10 for RF completed ...\n",
      "experiment 3/10 for RF completed ...\n",
      "experiment 4/10 for RF completed ...\n",
      "experiment 5/10 for RF completed ...\n",
      "experiment 6/10 for RF completed ...\n",
      "experiment 7/10 for RF completed ...\n",
      "experiment 8/10 for RF completed ...\n",
      "experiment 9/10 for RF completed ...\n",
      "experiment 10/10 for RF completed ...\n",
      "\n",
      "\n",
      "Final ROC-AUC(%) results for the ogbg-molhiv dataset with 'final' node embedding and one-hot 13-dim edge embedding\n",
      "+------------+--------------------+---------------------+-------------------+------------------+------------------+\n",
      "| Classifier | # Diffusion Layers | Node Embedding Size |       Train.      |       Val.       |       Test       |\n",
      "+------------+--------------------+---------------------+-------------------+------------------+------------------+\n",
      "|     RF     |         4          |         300         | 100.00 $\\pm$ 0.00 | 80.70 $\\pm$ 2.03 | 77.92 $\\pm$ 1.30 |\n",
      "+------------+--------------------+---------------------+-------------------+------------------+------------------+\n"
     ]
    }
   ],
   "source": [
    "# Run the algorithm\n",
    "\n",
    "for final_node_embedding in final_node_embeddings:\n",
    "    WEGL(dataset=dataset,\n",
    "         num_hidden_layers=num_hidden_layers,\n",
    "         node_embedding_sizes=node_embedding_sizes,\n",
    "         final_node_embedding=final_node_embedding,\n",
    "         num_pca_components=num_pca_components,\n",
    "         num_experiments=num_experiments,\n",
    "         classifiers=classifiers,\n",
    "         random_seed=random_seed,\n",
    "         device=device)"
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}