diffusion.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "amended-worry",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "no display found. Using non-interactive Agg backend\n"
     ]
    }
   ],
   "source": [
    "import networkx as nx\n",
    "import ndlib.models.ModelConfig as mc\n",
    "import ndlib.models.epidemics as ep\n",
    "import pandas as pd\n",
    "from tqdm.notebook import tqdm\n",
    "\n",
    "import numpy as np\n",
    "#np.random.seed(0)\n",
    "\n",
    "from ndlib.viz.mpl.DiffusionTrend import DiffusionTrend\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "from myDiffusion import IndependentCascadesModel\n",
    "\n",
    "adj = pd.read_csv(\"adj.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "theoretical-logic",
   "metadata": {},
   "outputs": [],
   "source": [
    "# bc = pd.read_csv(f\"bc_{group}.csv\")\n",
    "# links\n",
    "# np.where(bc.betweenness_centrality>0)[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "retired-blackjack",
   "metadata": {},
   "source": [
    "# 1. Define initial active nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "rental-savannah",
   "metadata": {},
   "outputs": [],
   "source": [
    "infected_nodes = {}\n",
    "for group in [\"Liu Yu\", \"Mika\"]:\n",
    "    infected_nodes[group] = {}\n",
    "    links = adj[adj[group]].drop(group,axis=1).drop_duplicates(ignore_index=True)\n",
    "    g = nx.from_pandas_edgelist(links, source=\"m_node\", target=\"node\", create_using=nx.DiGraph)\n",
    "    \n",
    "    for i_group, deg in zip([\"degree\", \"in_degree\", \"out_degree\", \"random\"], [g.degree(), g.in_degree(), g.out_degree()]):\n",
    "        temp = sorted(deg, key=lambda x: x[1], reverse=True)\n",
    "        infected_nodes[group][i_group] = np.array(temp)[:,0]\n",
    "    \n",
    "    # Random\n",
    "    temp = list(g.nodes)\n",
    "    np.random.shuffle(temp)\n",
    "    infected_nodes[group][\"random\"] =  temp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "harmful-intention",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[53677 54146 55173 56045 54371 54149 53701 55534 55256 55565] Liu Yu degree\n",
      "[ 5988 54233 54019 54204 56215 35678 39478 48688 19380   776] Liu Yu in_degree\n",
      "[53677 54146 55173 56045 54371 54149 53701 55534 55256 54147] Liu Yu out_degree\n",
      "[3755, 46074, 46928, 3042, 12086, 15651, 39471, 44947, 37018, 29722] Liu Yu random\n",
      "[53677 57117 56232 54371 55891 54254 54247 54677 55173 56397] Mika degree\n",
      "[42643 50617 56397 56871 53837 54766 54113 54411 55364 54044] Mika in_degree\n",
      "[53677 57117 56232 54371 55891 54254 54247 54677 55173 55283] Mika out_degree\n",
      "[13413, 49942, 8281, 4773, 32096, 33088, 48150, 6640, 53813, 1389] Mika random\n"
     ]
    }
   ],
   "source": [
    "for group in [\"Liu Yu\", \"Mika\"]:\n",
    "    for i_group in [\"degree\", \"in_degree\", \"out_degree\", \"random\"]:\n",
    "        print(infected_nodes[group][i_group][:10], group,i_group)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "broadband-burning",
   "metadata": {},
   "source": [
    "# 2. Run diffusion experiments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ultimate-heaven",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 20/20 [00:01<00:00, 15.07it/s]\n",
      "100%|██████████| 20/20 [00:00<00:00, 23.19it/s]\n",
      "100%|██████████| 20/20 [00:00<00:00, 22.88it/s]\n",
      "100%|██████████| 20/20 [00:00<00:00, 26.39it/s]\n",
      "100%|██████████| 20/20 [00:00<00:00, 27.89it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.59it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  7.09it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  7.30it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  7.47it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  7.92it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 15.17it/s]\n",
      "100%|██████████| 20/20 [00:00<00:00, 22.46it/s]\n",
      "100%|██████████| 20/20 [00:00<00:00, 25.04it/s]\n",
      "100%|██████████| 20/20 [00:00<00:00, 28.18it/s]\n",
      "100%|██████████| 20/20 [00:00<00:00, 29.52it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.60it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.56it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.64it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  6.68it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  6.84it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 11.96it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 11.28it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 11.01it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 10.92it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 10.84it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 15.65it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 15.42it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 15.18it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 15.17it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 14.82it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 11.79it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 10.99it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 10.90it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 10.70it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 10.55it/s]\n",
      "100%|██████████| 20/20 [00:00<00:00, 28.99it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 16.13it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 16.18it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 16.02it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 14.77it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  8.53it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 14.40it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 14.72it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 14.73it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 16.22it/s]\n",
      "100%|██████████| 20/20 [00:04<00:00,  4.00it/s]\n",
      "100%|██████████| 20/20 [00:04<00:00,  4.16it/s]\n",
      "100%|██████████| 20/20 [00:04<00:00,  4.35it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  5.59it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  5.61it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  7.42it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 15.67it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 15.86it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 16.54it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 16.86it/s]\n",
      "100%|██████████| 20/20 [00:05<00:00,  3.87it/s]\n",
      "100%|██████████| 20/20 [00:05<00:00,  3.91it/s]\n",
      "100%|██████████| 20/20 [00:05<00:00,  3.91it/s]\n",
      "100%|██████████| 20/20 [00:04<00:00,  4.03it/s]\n",
      "100%|██████████| 20/20 [00:04<00:00,  4.01it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.66it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.23it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.45it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.36it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.35it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  8.50it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  8.30it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  8.07it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  8.15it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  8.39it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  6.89it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.58it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.39it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.27it/s]\n",
      "100%|██████████| 20/20 [00:03<00:00,  6.15it/s]\n",
      "100%|██████████| 20/20 [00:01<00:00, 16.94it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  8.44it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  8.37it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  8.55it/s]\n",
      "100%|██████████| 20/20 [00:02<00:00,  8.42it/s]\n"
     ]
    }
   ],
   "source": [
    "# Model selection\n",
    "all_trs = {}\n",
    "init_ns = [10, 100, 200, 500, 1000]\n",
    "threshold = 0.5\n",
    "ITERATIONS = 20\n",
    "\n",
    "for group in [\"Liu Yu\", \"Mika\"]:\n",
    "    all_trs[group] = {}\n",
    "    links = adj[adj[group]].drop(group,axis=1).drop_duplicates(ignore_index=True)\n",
    "    g = nx.from_pandas_edgelist(links, source=\"m_node\", target=\"node\", create_using=nx.DiGraph)\n",
    "    all_trs[group][\"n_nodes\"] = len(g.nodes())\n",
    "    \n",
    "    for m in [\"TM\", \"ICM\"]:\n",
    "        all_trs[group][m] = {}\n",
    "        for i_group in [\"degree\", \"in_degree\", \"out_degree\", \"random\"]:\n",
    "            trs = []\n",
    "            for init_n in init_ns:\n",
    "                config = mc.Configuration()\n",
    "                config.add_model_initial_configuration(\"Infected\", infected_nodes[group][i_group][:init_n])\n",
    "\n",
    "                if m == \"TM\":\n",
    "                    model = ep.ThresholdModel(g)\n",
    "                    for i in g.nodes():\n",
    "                        config.add_node_configuration(\"threshold\", i, threshold)\n",
    "\n",
    "                elif m == \"ICM\":\n",
    "                    model = IndependentCascadesModel(g)\n",
    "                    for e in g.edges():\n",
    "                        config.add_edge_configuration(\"threshold\", e, threshold)\n",
    "\n",
    "\n",
    "                model.set_initial_status(config)\n",
    "                # Simulation execution\n",
    "                iterations = model.iteration_bunch(ITERATIONS)\n",
    "\n",
    "                # Plot Trend\n",
    "                trends = model.build_trends(iterations)\n",
    "\n",
    "                trs.append(trends[0][\"trends\"][\"node_count\"][1])\n",
    "                #viz = DiffusionTrend(model, trends)\n",
    "                #viz.plot()\n",
    "                #plt.plot(trs[0])\n",
    "\n",
    "            all_trs[group][m][i_group] = trs"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "daily-brazil",
   "metadata": {},
   "source": [
    "# 3. Evaluate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "oriented-junction",
   "metadata": {},
   "outputs": [],
   "source": [
    "def inverse(x):\n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "religious-processor",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Liu Yu | TM  degree | N0 = 1000: #active node=27154 (0.96)| N0=10: 22669 (0.80)\n",
      "Liu Yu | TM  out_degree | N0 = 1000: #active node=27685 (0.98)| N0=10: 22669 (0.80)\n",
      "Liu Yu | TM  in_degree | N0 = 1000: #active node=5618 (0.20)| N0=10: 30 (0.00)\n",
      "Liu Yu | TM  random | N0 = 1000: #active node=1149 (0.04)| N0=10: 10 (0.00)\n",
      "Liu Yu | ICM  degree | N0 = 1000: #active node=27174 (0.96)| N0=10: 23609 (0.83)\n",
      "Liu Yu | ICM  out_degree | N0 = 1000: #active node=27692 (0.98)| N0=10: 23609 (0.83)\n",
      "Liu Yu | ICM  in_degree | N0 = 1000: #active node=14687 (0.52)| N0=10: 14685 (0.52)\n",
      "Liu Yu | ICM  random | N0 = 1000: #active node=15220 (0.54)| N0=10: 10 (0.00)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/envs/rapids/lib/python3.7/site-packages/ipykernel_launcher.py:28: UserWarning: This figure was using constrained_layout==True, but that is incompatible with subplots_adjust and or tight_layout: setting constrained_layout==False. \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mika | TM  degree | N0 = 1000: #active node=44142 (0.96)| N0=10: 38538 (0.84)\n",
      "Mika | TM  out_degree | N0 = 1000: #active node=44590 (0.97)| N0=10: 32298 (0.70)\n",
      "Mika | TM  in_degree | N0 = 1000: #active node=17810 (0.39)| N0=10: 747 (0.02)\n",
      "Mika | TM  random | N0 = 1000: #active node=1131 (0.02)| N0=10: 11 (0.00)\n",
      "Mika | ICM  degree | N0 = 1000: #active node=44196 (0.96)| N0=10: 40186 (0.88)\n",
      "Mika | ICM  out_degree | N0 = 1000: #active node=44614 (0.97)| N0=10: 40186 (0.88)\n",
      "Mika | ICM  in_degree | N0 = 1000: #active node=27388 (0.60)| N0=10: 27387 (0.60)\n",
      "Mika | ICM  random | N0 = 1000: #active node=27833 (0.61)| N0=10: 15 (0.00)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/envs/rapids/lib/python3.7/site-packages/ipykernel_launcher.py:28: UserWarning: This figure was using constrained_layout==True, but that is incompatible with subplots_adjust and or tight_layout: setting constrained_layout==False. \n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for group in [\"Liu Yu\", \"Mika\"]:\n",
    "    fig, axes = plt.subplots(1,2, constrained_layout=True, figsize=(10,3))\n",
    "    #fig.set_title(group)\n",
    "    N_NODES = all_trs[group][\"n_nodes\"]\n",
    "    def forward(x, N_NODES=N_NODES):\n",
    "        return (x/N_NODES)\n",
    "        \n",
    "    for p_i, m in enumerate([\"TM\", \"ICM\"]):\n",
    "        for i_group in [\"degree\", \"out_degree\", \"in_degree\", \"random\"]:\n",
    "        #for i_group in [\"degree\",  \"random\"]:    \n",
    "            _ = axes[p_i].plot(init_ns, np.array(all_trs[group][m][i_group])[:, -1], label=i_group)\n",
    "            temp = np.array(all_trs[group][m][i_group])[-1, -1]\n",
    "            temp2 = np.array(all_trs[group][m][i_group])[0, -1]\n",
    "            print(f\"{group} | {m}  {i_group} | N0 = 1000: #active node={temp} ({temp/N_NODES:.2f})| N0=10: {temp2} ({temp2/N_NODES:.2f})\")\n",
    "        secax = axes[p_i].secondary_yaxis(1, functions=(forward, inverse))\n",
    "        _ = axes[p_i].set_ylabel(\"Number of active nodes\")\n",
    "        _ = axes[p_i].set_xlabel(\"Number of initial active nodes ($N_0$)\")\n",
    "        _ = axes[p_i].set_xticks(init_ns)\n",
    "        _ = secax.set_ylabel(\"Influence rate\")\n",
    "        if p_i == 1:\n",
    "            _ = axes[p_i].legend()\n",
    "        _ = axes[p_i].set_title(m)\n",
    "        _ = axes[p_i].set_ylim(0, np.ceil(N_NODES/1000)*1000)\n",
    "        \n",
    "        \n",
    "    plt.tight_layout()\n",
    "    plt.savefig(f\"fig/diffusion_{group}.png\")"
   ]
  }
 ],
 "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": "3.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}