plot_component_experiment.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline \n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_now(ax1, vdsh_prec, edge_prec, node2hash_prec, title_str, min_ylim, max_ylim):\n",
    "    N = 5\n",
    "    ind = np.arange(N)\n",
    "    width = 0.20       \n",
    "\n",
    "    ax1.bar(ind, vdsh_prec, width, align='edge', label='Text-Only')\n",
    "    ax1.bar(ind + width, edge_prec, width, align='edge', label='Connection-Only')\n",
    "    ax1.bar(ind + width + width, node2hash_prec, width, align='edge', label='Both')\n",
    "\n",
    "    ax1.set_ylim(min_ylim, max_ylim)\n",
    "\n",
    "    ax1.set_ylabel('Prec@100')\n",
    "    ax1.title.set_text(title_str)\n",
    "\n",
    "    ax1.set_xlabel('Number of bits')\n",
    "    ax1.set_xticks(ind)\n",
    "    ax1.set_xticklabels(['8', '16', '32', '64', '128'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_df(dataset_name):\n",
    "    vdsh_df = pd.read_csv('vdsh_logs/vdsh.experiment.{}.txt'.format(dataset_name), sep='\\t', header=None, \n",
    "                             names=['dataset', 'bits', 'samples', 'Prec'])\n",
    "    edgeonly_df = pd.read_csv('edgeonly_logs/binary.Experiment.{}.txt'.format(dataset_name), sep='\\t', header=None, \n",
    "                             names=['dataset', 'bits', 'samples', 'Prec'])\n",
    "    node2hash_df = pd.read_csv('binary_logs/binary.Experiment.{}.txt'.format(dataset_name), sep='\\t', header=None, \n",
    "                                 names=['dataset', 'bits', 'samples', 'Prec'])\n",
    "    return vdsh_df, edgeonly_df, node2hash_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(8, 6), dpi=80)\n",
    "\n",
    "# Cora\n",
    "vdsh_df, edgeonly_df, node2hash_df = load_df('cora')\n",
    "\n",
    "ax1 = plt.subplot(231)\n",
    "plot_now(ax1, list(vdsh_df.Prec), list(edgeonly_df.Prec), list(node2hash_df.Prec), \"Cora\", 0.0, 1.0)\n",
    "\n",
    "# Cora\n",
    "vdsh_df, edgeonly_df, node2hash_df = load_df('citeseer')\n",
    "\n",
    "ax1 = plt.subplot(232)\n",
    "plot_now(ax1, list(vdsh_df.Prec), list(edgeonly_df.Prec), list(node2hash_df.Prec), \"Citeseer\", 0.0, 1.0)\n",
    "\n",
    "# Pubmed\n",
    "vdsh_df, edgeonly_df, node2hash_df = load_df('pubmed')\n",
    "\n",
    "ax1 = plt.subplot(233)\n",
    "plot_now(ax1, list(vdsh_df.Prec), list(edgeonly_df.Prec), list(node2hash_df.Prec), \"Pubmed\", 0.0, 1.0)\n",
    "\n",
    "# Dblp\n",
    "vdsh_df, edgeonly_df, node2hash_df = load_df('dblp')\n",
    "\n",
    "ax1 = plt.subplot(234)\n",
    "plot_now(ax1, list(vdsh_df.Prec), list(edgeonly_df.Prec), list(node2hash_df.Prec), \"Dblp\", 0.0, 1.0)\n",
    "ax1.legend(loc='lower right')\n",
    "\n",
    "# ng20\n",
    "vdsh_df, edgeonly_df, node2hash_df = load_df('ng20')\n",
    "\n",
    "ax1 = plt.subplot(235)\n",
    "plot_now(ax1, list(vdsh_df.Prec), list(edgeonly_df.Prec), list(node2hash_df.Prec), \"ng20\", 0.0, 1.0)\n",
    "\n",
    "# agnews\n",
    "vdsh_df, edgeonly_df, node2hash_df = load_df('agnews')\n",
    "\n",
    "ax1 = plt.subplot(236)\n",
    "plot_now(ax1, list(vdsh_df.Prec), list(edgeonly_df.Prec), list(node2hash_df.Prec), \"agnews\", 0.0, 1.0)\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig.savefig(\"component_experiment.pdf\", bbox_inches='tight')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Research2018",
   "language": "python",
   "name": "research"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}