4way_figures.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline\n",
    "import networkx as nx\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle as pkl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.patches as mpatches\n",
    "import matplotlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from subprocess import call"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/spiaggesi/anaconda3/envs/tf2/lib/python3.6/site-packages/sklearn/utils/deprecation.py:143: FutureWarning: The sklearn.metrics.ranking module is  deprecated in version 0.22 and will be removed in version 0.24. The corresponding classes / functions should instead be imported from sklearn.metrics. Anything that cannot be imported from sklearn.metrics is now part of the private API.\n",
      "  warnings.warn(message, FutureWarning)\n"
     ]
    }
   ],
   "source": [
    "from utils import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_list = ['contact-high-school', 'contact-primary-school', 'email-Enron']\n",
    "HASSE_TYPE = 'uniform'\n",
    "WORK_FOLDER = './'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "TUPLE_SIZE = 4"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# FIGURE A1 (a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x180 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, len(dataset_list), figsize=(2.*len(dataset_list), 2.5), facecolor='white', sharey=True, sharex=True)\n",
    "if len(dataset_list)==1:\n",
    "    ax=[ax]\n",
    "fig.tight_layout()\n",
    "fig.subplots_adjust(wspace=0.75)\n",
    "\n",
    "cmp = matplotlib.cm.get_cmap('magma_r')\n",
    "colors = [cmp(i) for i in np.linspace(0.2,0.8,4)][1:]\n",
    "\n",
    "\n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    load_path = WORK_FOLDER + 'processed-output/figures/%s/'%(dataset)\n",
    "\n",
    "    max_order = 1\n",
    "    \n",
    "    data_scores = []\n",
    "\n",
    "    for nlink in range(TUPLE_SIZE+1):\n",
    "\n",
    "        results_file = load_path+'best-aucpr-reconstruction-%dbounds-%dsimplex-max5000-%s-%s-simplexorder%d-maxorder%d.pkl'%\\\n",
    "                         (nlink, 0, 'cbow', HASSE_TYPE, TUPLE_SIZE-1, max_order)\n",
    "\n",
    "        pred_dict_list = pkl.load(open(results_file, 'rb'))\n",
    "\n",
    "        pred_scores = [average_precision(d['y_test'], d['y_pred'], pi0=0.5) \n",
    "                       if d['y_test'] is not None else None for d in pred_dict_list]   \n",
    "\n",
    "        data_scores.append(pred_scores)\n",
    "\n",
    "    ax[j].errorbar(x=np.arange(TUPLE_SIZE+1),\n",
    "               y=[np.mean(l) for l in data_scores], \n",
    "               yerr=[np.std(l) for l in data_scores],\n",
    "                color=colors[0], marker='D', linestyle = 'None', markersize=8, capsize=4)\n",
    "\n",
    "\n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    load_path = WORK_FOLDER + 'processed-output/figures/%s/'%(dataset)\n",
    "\n",
    "    max_order = 0\n",
    "    \n",
    "    data_scores = []\n",
    "\n",
    "    for nlink in range(TUPLE_SIZE+1):\n",
    "\n",
    "        results_file = load_path+'best-aucpr-reconstruction-%dbounds-%dsimplex-max5000-%s-%s-simplexorder%d-maxorder%d.pkl'%\\\n",
    "                        (nlink, 0, 'sg', HASSE_TYPE, TUPLE_SIZE-1, max_order)\n",
    "\n",
    "        pred_dict_list = pkl.load(open(results_file, 'rb'))\n",
    "\n",
    "        pred_scores = [average_precision(d['y_test'], d['y_pred'], pi0=0.5) \n",
    "                       if d['y_test'] is not None else None for d in pred_dict_list]   \n",
    "    \n",
    "        data_scores.append(pred_scores)\n",
    "\n",
    "    ax[j].errorbar(x=np.arange(TUPLE_SIZE+1),\n",
    "               y=[np.mean(l) for l in data_scores], \n",
    "               yerr=[np.std(l) for l in data_scores],\n",
    "                color=colors[1], marker='D', linestyle = 'None', markersize=8, capsize=4)\n",
    "    \n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    ax[j].set_xticks(list(range(TUPLE_SIZE+1)))\n",
    "\n",
    "    ax[j].text(-0.25,0.5, dataset, transform=ax[j].transAxes, rotation=90, fontsize=12, weight='bold', va='center')\n",
    "    \n",
    "    xmin, xmax = ax[j].get_xbound()\n",
    "    ax[j].set_xbound(xmin-0.05, xmax+0.05)\n",
    "\n",
    "    if j==0:\n",
    "        ax[j].spines['left'].set_position(('outward', 40))\n",
    "\n",
    "    if j>0:\n",
    "        ax[j].spines[\"left\"].set_visible(False)\n",
    "        ax[j].tick_params(axis='y', which='major', left=False)\n",
    "\n",
    "    ax[j].tick_params(axis='x', which='major', labelsize=14)\n",
    "    ax[j].tick_params(axis='y', which='major', labelsize=14)\n",
    "        \n",
    "    ax[j].grid()\n",
    "    ax[j].spines[\"top\"].set_visible(False)\n",
    "    ax[j].spines[\"right\"].set_visible(False)\n",
    "    ax[j].spines[\"bottom\"].set_visible(False)\n",
    "    \n",
    "ax = fig.add_subplot(111, frameon=False)\n",
    "plt.tick_params(labelcolor='none', top='off', bottom='off', left='off', right='off')\n",
    "plt.xticks([])\n",
    "plt.yticks([])\n",
    "\n",
    "ax.set_xlabel('# of training triangles in open 4-node configurations'+r\" $n_\\Delta(\\delta)$\", fontsize=16, weight='bold', labelpad=25)\n",
    "\n",
    "ax.set_ylabel('balanced AUC-PR', fontsize=16, weight='bold', labelpad=75)\n",
    "\n",
    "legend = ax.legend(*zip(*[(mpatches.Patch(color=c), l) \n",
    "                for c,l in zip(colors[:2],['simplex2vec -'+r\" $s_0(\\delta)$\", 'k-simplex2vec -'+r\" $s_0(\\delta)$\"])]),\n",
    "                bbox_to_anchor=(0.5,1), loc='lower center', ncol=2, fontsize=14, frameon=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# FIGURE A1 (b)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x180 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, len(dataset_list), figsize=(2.*len(dataset_list), 2.5), facecolor='white', sharey=True, sharex=True)\n",
    "if len(dataset_list)==1:\n",
    "    ax=[ax]\n",
    "fig.tight_layout()\n",
    "fig.subplots_adjust(wspace=0.75)\n",
    "\n",
    "cmp = matplotlib.cm.get_cmap('viridis')\n",
    "colors = ['C3']+[cmp(i) for i in np.linspace(0.2,0.8,4)][1:]\n",
    "\n",
    "\n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    load_path = WORK_FOLDER + 'processed-output/figures/%s/'%(dataset)\n",
    "\n",
    "    max_order = 1\n",
    "    \n",
    "    data_scores = []\n",
    "\n",
    "    for nlink in range(TUPLE_SIZE+1):\n",
    "\n",
    "        results_file = load_path+'best-aucpr-prediction-%dbounds-%dsimplex-max5000-%s-%s-simplexorder%d-maxorder%d.pkl'%\\\n",
    "                         (nlink, 0, 'cbow', HASSE_TYPE, TUPLE_SIZE-1, max_order)\n",
    "\n",
    "        pred_dict_list = pkl.load(open(results_file, 'rb'))\n",
    "\n",
    "        pred_scores = [average_precision(d['y_test'], d['y_pred'], pi0=0.5) \n",
    "                       if d['y_test'] is not None else None for d in pred_dict_list]   \n",
    "\n",
    "        data_scores.append(pred_scores)\n",
    "\n",
    "    ax[j].errorbar(x=np.arange(TUPLE_SIZE+1),\n",
    "               y=[np.mean(l) for l in data_scores], \n",
    "               yerr=[np.std(l) for l in data_scores],\n",
    "                color=colors[0], marker='D', linestyle = 'None', markersize=8, capsize=4)\n",
    "\n",
    "\n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    load_path = WORK_FOLDER + 'processed-output/figures/%s/'%(dataset)\n",
    "\n",
    "    max_order = 0\n",
    "    \n",
    "    data_scores = []\n",
    "\n",
    "    for nlink in range(TUPLE_SIZE+1):\n",
    "\n",
    "        results_file = load_path+'best-aucpr-prediction-%dbounds-%dsimplex-max5000-%s-%s-simplexorder%d-maxorder%d.pkl'%\\\n",
    "                        (nlink, 0, 'sg', HASSE_TYPE, TUPLE_SIZE-1, max_order)\n",
    "\n",
    "        pred_dict_list = pkl.load(open(results_file, 'rb'))\n",
    "\n",
    "        pred_scores = [average_precision(d['y_test'], d['y_pred'], pi0=0.5) \n",
    "                       if d['y_test'] is not None else None for d in pred_dict_list]   \n",
    "    \n",
    "        data_scores.append(pred_scores)\n",
    "\n",
    "    ax[j].errorbar(x=np.arange(TUPLE_SIZE+1),\n",
    "               y=[np.mean(l) for l in data_scores], \n",
    "               yerr=[np.std(l) for l in data_scores],\n",
    "                color=colors[1], marker='D', linestyle = 'None', markersize=8, capsize=4)\n",
    "    \n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    ax[j].set_xticks(list(range(TUPLE_SIZE+1)))\n",
    "\n",
    "    ax[j].text(-0.25,0.5, dataset, transform=ax[j].transAxes, rotation=90, fontsize=12, weight='bold', va='center')\n",
    "    \n",
    "    xmin, xmax = ax[j].get_xbound()\n",
    "    ax[j].set_xbound(xmin-0.05, xmax+0.05)\n",
    "\n",
    "    if j==0:\n",
    "        ax[j].spines['left'].set_position(('outward', 40))\n",
    "\n",
    "    if j>0:\n",
    "        ax[j].spines[\"left\"].set_visible(False)\n",
    "        ax[j].tick_params(axis='y', which='major', left=False)\n",
    "#         ax[j].set_yticks([])\n",
    "\n",
    "    ax[j].tick_params(axis='x', which='major', labelsize=14)\n",
    "    ax[j].tick_params(axis='y', which='major', labelsize=14)\n",
    "        \n",
    "    ax[j].grid()\n",
    "    ax[j].spines[\"top\"].set_visible(False)\n",
    "    ax[j].spines[\"right\"].set_visible(False)\n",
    "    ax[j].spines[\"bottom\"].set_visible(False)\n",
    "    \n",
    "ax = fig.add_subplot(111, frameon=False)\n",
    "plt.tick_params(labelcolor='none', top='off', bottom='off', left='off', right='off')\n",
    "plt.xticks([])\n",
    "plt.yticks([])\n",
    "\n",
    "ax.set_xlabel('# of training triangles in open 4-node configurations'+r\" $n_\\Delta(\\delta)$\", fontsize=16, weight='bold', labelpad=25)\n",
    "\n",
    "ax.set_ylabel('balanced AUC-PR', fontsize=16, weight='bold', labelpad=75)\n",
    "\n",
    "legend = ax.legend(*zip(*[(mpatches.Patch(color=c), l) \n",
    "                for c,l in zip(colors[:2],['simplex2vec -'+r\" $s_0(\\delta)$\", 'k-simplex2vec -'+r\" $s_0(\\delta)$\"])]),\n",
    "                bbox_to_anchor=(0.5,1), loc='lower center', ncol=2, fontsize=14, frameon=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# FIGURE A1 (c)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 324x180 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, len(dataset_list), figsize=(1.5*len(dataset_list), 2.5), facecolor='white', sharey=True, sharex=True)\n",
    "if len(dataset_list)==1:\n",
    "    ax=[ax]\n",
    "fig.tight_layout()\n",
    "fig.subplots_adjust(wspace=0.75)\n",
    "\n",
    "cmp = matplotlib.cm.get_cmap('magma_r')\n",
    "colors = [cmp(i) for i in np.linspace(0.2,0.8,4)][1:]\n",
    "\n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    load_path = WORK_FOLDER + 'processed-output/figures/%s/'%(dataset)\n",
    "    \n",
    "    data_scores =  []\n",
    "\n",
    "    for norder in range(TUPLE_SIZE-1):\n",
    "        \n",
    "        max_order = norder+1\n",
    "        \n",
    "        results_file = load_path+'best-aucpr-reconstruction-%dbounds-%dsimplex-max5000-%s-%s-simplexorder%d-maxorder%d.pkl'%\\\n",
    "                        (TUPLE_SIZE, norder, 'cbow', HASSE_TYPE, TUPLE_SIZE-1, max_order)\n",
    "\n",
    "        pred_dict_list = pkl.load(open(results_file, 'rb'))\n",
    "\n",
    "        pred_scores = [average_precision(d['y_test'], d['y_pred'], pi0=0.5) if d['y_test'] is not None else None for d in pred_dict_list]  \n",
    "\n",
    "        data_scores.append(pred_scores)\n",
    "        \n",
    "    ax[j].errorbar(x=np.arange(TUPLE_SIZE-1),\n",
    "           y=[np.mean(l) for l in data_scores], \n",
    "           yerr=[np.std(l) for l in data_scores],\n",
    "            color=colors[0], marker='D', linestyle = 'None', markersize=8, capsize=4)\n",
    "    \n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    load_path = WORK_FOLDER + 'processed-output/figures/%s/'%(dataset)\n",
    "    \n",
    "    data_scores =  []\n",
    "\n",
    "    for norder in range(TUPLE_SIZE-1):\n",
    "        \n",
    "        max_order = norder\n",
    "\n",
    "        results_file = load_path+'best-aucpr-reconstruction-%dbounds-%dsimplex-max5000-%s-%s-simplexorder%d-maxorder%d.pkl'%\\\n",
    "                        (TUPLE_SIZE, norder, 'sg', HASSE_TYPE, TUPLE_SIZE-1, max_order)\n",
    "\n",
    "        pred_dict_list = pkl.load(open(results_file, 'rb'))\n",
    "\n",
    "        pred_scores = [average_precision(d['y_test'], d['y_pred'], pi0=0.5) if d['y_test'] is not None else None for d in pred_dict_list]  \n",
    "\n",
    "        data_scores.append(pred_scores)\n",
    "    \n",
    "    ax[j].errorbar(x=np.arange(TUPLE_SIZE-1),\n",
    "           y=[np.mean(l) for l in data_scores], \n",
    "           yerr=[np.std(l) for l in data_scores],\n",
    "            color=colors[1], marker='D', linestyle = 'None', markersize=8, capsize=4)\n",
    "    \n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    ax[j].set_xticks(list(range(TUPLE_SIZE-1)))\n",
    "    ax[j].set_xticklabels([r\"$s_%d$\"%(norder) for norder in range(TUPLE_SIZE-1)], fontsize=14)\n",
    "\n",
    "    ax[j].text(-0.25,0.5, dataset, transform=ax[j].transAxes, rotation=90, fontsize=12, weight='bold', va='center')\n",
    "    \n",
    "    xmin, xmax = ax[j].get_xbound()\n",
    "    ax[j].set_xbound(xmin-0.05, xmax+0.05)\n",
    "\n",
    "    if j==0:\n",
    "        ax[j].spines['left'].set_position(('outward', 40))\n",
    "\n",
    "    if j>0:\n",
    "        ax[j].spines[\"left\"].set_visible(False)\n",
    "        ax[j].tick_params(axis='y', which='major', left=False)\n",
    "\n",
    "    ax[j].tick_params(axis='x', which='major', labelsize=18)\n",
    "    ax[j].tick_params(axis='y', which='major', labelsize=14)\n",
    "        \n",
    "    ax[j].grid()\n",
    "    ax[j].spines[\"top\"].set_visible(False)\n",
    "    ax[j].spines[\"right\"].set_visible(False)\n",
    "    ax[j].spines[\"bottom\"].set_visible(False)\n",
    "    \n",
    "ax = fig.add_subplot(111, frameon=False)\n",
    "# hide tick and tick label of the big axes\n",
    "plt.tick_params(labelcolor='none', top='off', bottom='off', left='off', right='off')\n",
    "plt.xticks([])\n",
    "plt.yticks([])\n",
    "\n",
    "ax.set_xlabel('similarity scores \\n ( 4-node configurations with'+r\" $n_\\Delta(\\delta)=4$\"+' )', fontsize=16, weight='bold', labelpad=30)\n",
    "    \n",
    "ax.set_ylabel('balanced AUC-PR', fontsize=16, weight='bold', labelpad=75)\n",
    "\n",
    "legend = ax.legend(*zip(*[(mpatches.Patch(color=c), l) \n",
    "                for c,l in zip(colors[:2],['simplex2vec embedding', 'k-simplex2vec embedding'])]),\n",
    "               bbox_to_anchor=(1.025, .8), loc = 'center left', ncol=1, fontsize=14, frameon=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# FIGURE A1 (d)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 324x180 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, len(dataset_list), figsize=(1.5*len(dataset_list), 2.5), facecolor='white', sharey=True, sharex=True)\n",
    "if len(dataset_list)==1:\n",
    "    ax=[ax]\n",
    "fig.tight_layout()\n",
    "fig.subplots_adjust(wspace=0.75)\n",
    "\n",
    "cmp = matplotlib.cm.get_cmap('viridis')\n",
    "colors = ['C3']+[cmp(i) for i in np.linspace(0.2,0.8,4)][1:]\n",
    "\n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    load_path = WORK_FOLDER + 'processed-output/figures/%s/'%(dataset)\n",
    "    \n",
    "    data_scores =  []\n",
    "\n",
    "    for norder in range(TUPLE_SIZE-1):\n",
    "        \n",
    "        max_order = norder+1\n",
    "        \n",
    "        results_file = load_path+'best-aucpr-prediction-%dbounds-%dsimplex-max5000-%s-%s-simplexorder%d-maxorder%d.pkl'%\\\n",
    "                        (TUPLE_SIZE, norder, 'cbow', HASSE_TYPE, TUPLE_SIZE-1, max_order)\n",
    "\n",
    "        pred_dict_list = pkl.load(open(results_file, 'rb'))\n",
    "\n",
    "        pred_scores = [average_precision(d['y_test'], d['y_pred'], pi0=0.5) if d['y_test'] is not None else None for d in pred_dict_list]  \n",
    "\n",
    "        data_scores.append(pred_scores)\n",
    "        \n",
    "    ax[j].errorbar(x=np.arange(TUPLE_SIZE-1),\n",
    "           y=[np.mean(l) for l in data_scores], \n",
    "           yerr=[np.std(l) for l in data_scores],\n",
    "            color=colors[0], marker='D', linestyle = 'None', markersize=8, capsize=4)\n",
    "    \n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    load_path = WORK_FOLDER + 'processed-output/figures/%s/'%(dataset)\n",
    "    \n",
    "    data_scores =  []\n",
    "\n",
    "    for norder in range(TUPLE_SIZE-1):\n",
    "        \n",
    "        max_order = norder\n",
    "\n",
    "        results_file = load_path+'best-aucpr-prediction-%dbounds-%dsimplex-max5000-%s-%s-simplexorder%d-maxorder%d.pkl'%\\\n",
    "                        (TUPLE_SIZE, norder, 'sg', HASSE_TYPE, TUPLE_SIZE-1, max_order)\n",
    "\n",
    "        pred_dict_list = pkl.load(open(results_file, 'rb'))\n",
    "\n",
    "        pred_scores = [average_precision(d['y_test'], d['y_pred'], pi0=0.5) if d['y_test'] is not None else None for d in pred_dict_list]  \n",
    "\n",
    "        data_scores.append(pred_scores)\n",
    "    \n",
    "    ax[j].errorbar(x=np.arange(TUPLE_SIZE-1),\n",
    "           y=[np.mean(l) for l in data_scores], \n",
    "           yerr=[np.std(l) for l in data_scores],\n",
    "            color=colors[1], marker='D', linestyle = 'None', markersize=8, capsize=4)\n",
    "    \n",
    "for j, dataset in enumerate(dataset_list):\n",
    "    \n",
    "    ax[j].set_xticks(list(range(TUPLE_SIZE-1)))\n",
    "    ax[j].set_xticklabels([r\"$s_%d$\"%(norder) for norder in range(TUPLE_SIZE-1)], fontsize=14)\n",
    "\n",
    "    ax[j].text(-0.25,0.5, dataset, transform=ax[j].transAxes, rotation=90, fontsize=12, weight='bold', va='center')\n",
    "    \n",
    "    xmin, xmax = ax[j].get_xbound()\n",
    "    ax[j].set_xbound(xmin-0.05, xmax+0.05)\n",
    "\n",
    "    if j==0:\n",
    "        ax[j].spines['left'].set_position(('outward', 40))\n",
    "\n",
    "    if j>0:\n",
    "        ax[j].spines[\"left\"].set_visible(False)\n",
    "        ax[j].tick_params(axis='y', which='major', left=False)\n",
    "\n",
    "    ax[j].tick_params(axis='x', which='major', labelsize=18)\n",
    "    ax[j].tick_params(axis='y', which='major', labelsize=14)\n",
    "        \n",
    "    ax[j].grid()\n",
    "    ax[j].spines[\"top\"].set_visible(False)\n",
    "    ax[j].spines[\"right\"].set_visible(False)\n",
    "    ax[j].spines[\"bottom\"].set_visible(False)\n",
    "    \n",
    "ax = fig.add_subplot(111, frameon=False)\n",
    "# hide tick and tick label of the big axes\n",
    "plt.tick_params(labelcolor='none', top='off', bottom='off', left='off', right='off')\n",
    "plt.xticks([])\n",
    "plt.yticks([])\n",
    "\n",
    "ax.set_xlabel('similarity scores \\n ( 4-node configurations with'+r\" $n_\\Delta(\\delta)=4$\"+' )', fontsize=16, weight='bold', labelpad=30)\n",
    "\n",
    "ax.set_ylabel('balanced AUC-PR', fontsize=16, weight='bold', labelpad=75)\n",
    "\n",
    "legend = ax.legend(*zip(*[(mpatches.Patch(color=c), l) \n",
    "                for c,l in zip(colors[:2],['simplex2vec embedding', 'k-simplex2vec embedding'])]),\n",
    "               bbox_to_anchor=(1.025, .8), loc = 'center left', ncol=1, fontsize=14, frameon=False)"
   ]
  },
  {
   "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": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}