multiplicative_weight_update.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# import matplotlib as mpl\n",
    "# mpl.use('Agg')\n",
    "%matplotlib inline\n",
    "\n",
    "import networkx as nx\n",
    "import numpy as np\n",
    "import random\n",
    "import pandas as pd\n",
    "from matplotlib import pyplot as plt\n",
    "from itertools import combinations\n",
    "from tqdm import tqdm\n",
    "from joblib import Parallel, delayed\n",
    "\n",
    "from graph_generator import grid_2d, add_p_and_delta\n",
    "from ic import make_partial_cascade\n",
    "from plot_utils import plot_snapshot, plot_query_process\n",
    "from utils import infeciton_time2weight\n",
    "from mwu import main_routine as mwu, MAX_MU, MAX_ADV, RANDOM\n",
    "from synthetic_data import (load_data_by_gtype, GRID, KRONECKER_HIER, KRONECKER_PERI, KRONECKER_RAND,\n",
    "                            PL_TREE, ER, BARABASI)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEACAYAAABVtcpZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1ec49c2ba8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "g = load_data_by_gtype('barabasi', '2-6')[0]\n",
    "pos = nx.spring_layout(g)\n",
    "nx.draw_networkx(g, pos, with_labels=False, node_size=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEACAYAAABcXmojAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1ec5c046d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "g = load_data_by_gtype('kr-hier', '2-6')[0]\n",
    "# pos = nx.graphviz(g)\n",
    "from networkx.drawing.nx_agraph import graphviz_layout\n",
    "pos = graphviz_layout(g)\n",
    "nx.draw_networkx(g, pos=pos, with_labels=False, node_size=100)\n",
    "# nx.draw_networkx(g, pos, with_labels=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "ename": "FileNotFoundError",
     "evalue": "[Errno 2] No such file or directory: 'data/barabasi/graph.gpkl'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mFileNotFoundError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-50-c0e710fb46a0>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mgtype\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mBARABASI\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mg\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0ms2n_probas\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnode2id\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mid2node\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mload_data_by_gtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgtype\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0mu\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mv\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mg\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0medges_iter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0minfp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mg\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mu\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mv\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'p'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'infp = {}'\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/cloud-user/code/source_finding/synthetic_data.py\u001b[0m in \u001b[0;36mload_data_by_gtype\u001b[0;34m(gtype)\u001b[0m\n\u001b[1;32m     34\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     35\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mextract_larges_CC\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 36\u001b[0;31m     \u001b[0mnodes\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconnected_components\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     37\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mg\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msubgraph\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnodes\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     38\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<decorator-gen-295>\u001b[0m in \u001b[0;36mread_gpickle\u001b[0;34m(path)\u001b[0m\n",
      "\u001b[0;32m/home/cloud-user/code/source_finding/venv/lib/python3.5/site-packages/networkx/utils/decorators.py\u001b[0m in \u001b[0;36m_open_file\u001b[0;34m(func, *args, **kwargs)\u001b[0m\n\u001b[1;32m    196\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mis_string_like\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    197\u001b[0m             \u001b[0mext\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msplitext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 198\u001b[0;31m             \u001b[0mfobj\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_dispatch_dict\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mext\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmode\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmode\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    199\u001b[0m             \u001b[0mclose_fobj\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    200\u001b[0m         \u001b[0;32melif\u001b[0m \u001b[0mhasattr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'read'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mFileNotFoundError\u001b[0m: [Errno 2] No such file or directory: 'data/barabasi/graph.gpkl'"
     ]
    }
   ],
   "source": [
    "gtype = BARABASI\n",
    "g, s2n_probas, node2id, id2node = load_data_by_gtype(gtype)\n",
    "u, v = next(g.edges_iter())\n",
    "infp = g[u][v]['p']\n",
    "print('infp = {}'.format(infp))\n",
    "if gtype == GRID:\n",
    "    pos = {n: np.array(n) for n in g.nodes()}\n",
    "else:\n",
    "    pos = nx.spring_layout(g)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "source, obs_nodes, infection_times, tree = make_partial_cascade(g, 0.05, sampling_method='late_nodes')\n",
    "\n",
    "node2weight = infeciton_time2weight(infection_times)\n",
    "\n",
    "plot_snapshot(g, pos, node2weight, \n",
    "              query_node=(3, 3),\n",
    "              source_node=source,\n",
    "              queried_nodes=obs_nodes, \n",
    "              max_node_size=500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "%%timeit -r 1 -n1\n",
    "# demo of how mu(source increases)\n",
    "\n",
    "source, obs_nodes, infection_times, tree = make_partial_cascade(\n",
    "    g, fraction=0.05, sampling_method='late_nodes')\n",
    "query_count, mu_list, query_list = mwu(g, node2id, id2node,\n",
    "                                source, obs_nodes, infection_times, \n",
    "                                s2n_probas,\n",
    "                                epsilon=0.7,\n",
    "                                query_selection_method=MAX_MU,\n",
    "                                check_neighbor_threshold=0.1,\n",
    "                                save_log=True,\n",
    "                                debug=False)\n",
    "print(query_count)\n",
    "\n",
    "mu_of_source = [mu[source] for mu in mu_list]\n",
    "plt.clf()\n",
    "fig, ax = plt.subplots(1, 1, figsize=(5, 5))\n",
    "ax.plot(mu_of_source)\n",
    "ax.plot([0.1]*100, 'r')\n",
    "ax.set_xlabel('iteration')\n",
    "ax.set_ylabel('mu(source)')\n",
    "\n",
    "idx = np.nonzero(np.array(mu_of_source) > 0.1)[0]\n",
    "if len(idx) > 0:\n",
    "    half_value = idx[0]\n",
    "    ax.plot([half_value]*20, np.linspace(0, 1.0, 20), 'r')\n",
    "# fig.savefig('figs/{}_mu_source_vs_iteration.pdf'.format(gtype))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "source, obs_nodes, infection_times, tree = make_partial_cascade(g, 0.05, sampling_method='late_nodes')\n",
    "query_count, mu_list, query_list = mwu(g, node2id, id2node,\n",
    "                                source, obs_nodes, infection_times, \n",
    "                                s2n_probas,\n",
    "                                epsilon=0.7,\n",
    "                                query_selection_method=MAX_MU,\n",
    "                                check_neighbor_threshold=0.1,\n",
    "                                max_iter=10,\n",
    "                                save_log=True,\n",
    "                                debug=False)\n",
    "print(query_count)\n",
    "fig, ax = plot_query_process(g, source, obs_nodes, infection_times,\n",
    "                           mu_list, query_list,\n",
    "                           check_neighbor_threshold=0.1,\n",
    "                           pos=pos)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true,
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "if False:\n",
    "    print('|V|={}, |E|={}'.format(g.number_of_nodes(), g.number_of_edges()))\n",
    "    if True:\n",
    "        while True:\n",
    "            source, obs_nodes, infection_times, tree = make_partial_cascade(\n",
    "                g, fraction=0.05, sampling_method='late_nodes')    \n",
    "            query_count, mu_of_source = mwu(g, source, obs_nodes, infection_times, \n",
    "                                            mean_times_by_source,\n",
    "                                            epsilon=0.7,\n",
    "                                            query_selection_method=MAX_MU,\n",
    "                                            inspect_mu=False,\n",
    "                                            debug=False, \n",
    "                                            plot=False)            \n",
    "            if query_count > 50:\n",
    "                print(query_count)\n",
    "                break\n",
    "    else:\n",
    "        d = pkl.load(open('bad_examples/{}/1.pkl'.format(gtype), 'rb'))\n",
    "        source, obs_nodes, infection_times = d['source'], d['obs_nodes'], d['infection_times']\n",
    "\n",
    "    query_count, mu_of_source, fig = mwu(g, source, obs_nodes, infection_times, \n",
    "                                        times_by_source,\n",
    "                                        source2nodeid_counter,\n",
    "                                        epsilon=0.7,\n",
    "                                        query_selection_method=MAX_MU,\n",
    "                                        inspect_mu=False,\n",
    "                                        debug=False, \n",
    "                                        plot=True,\n",
    "                                        max_plots=12)\n",
    "    print(query_count)\n",
    "    fig.savefig('figs/query_process_{}'.format(gtype))\n",
    "\n",
    "    plt.clf()\n",
    "    fig, ax = plt.subplots(1, 1, figsize=(5, 5))\n",
    "    ax.plot(mu_of_source)\n",
    "    ax.plot([0.1]*100, 'r')\n",
    "    ax.set_xlabel('iteration')\n",
    "    ax.set_ylabel('mu(source)')\n",
    "\n",
    "    idx = np.nonzero(np.array(mu_of_source) > 0.1)[0]\n",
    "    if len(idx) > 0:\n",
    "        half_value = idx[0]\n",
    "        ax.plot([half_value]*20, np.linspace(0, 1.0, 20), 'r')\n",
    "    fig.savefig('figs/mu_vs_time_{}'.format(gtype))    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# dump bad example\n",
    "if False:\n",
    "    import os    \n",
    "    data = {\n",
    "        'source': source, 'obs_nodes': obs_nodes, 'infection_times': infection_times, \n",
    "        'tree': tree,\n",
    "        'epsilon': 0.7,\n",
    "        'query_selection_method': MAX_MU\n",
    "    }\n",
    "    if not os.path.exists('bad_examples/{}'.format(gtype)):\n",
    "        os.makedirs('bad_examples/{}'.format(gtype))\n",
    "    pkl.dump(data, open('bad_examples/{}/unsual_tree.pkl'.format(gtype), 'wb'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "def experiment_multiple_rounds(rounds, g,\n",
    "                               node2id, id2node,\n",
    "                               s2n_probas,\n",
    "                               infp, fraction, epsilon, sampling_method, \n",
    "                               query_selection_method,\n",
    "                               check_neighbor_threshold,\n",
    "                               max_iter=float('inf'),\n",
    "                               seed=None):\n",
    "    np.random.seed(seed)\n",
    "    random.seed(seed)\n",
    "    results = []\n",
    "    for i in tqdm(range(rounds)):\n",
    "        source, obs_nodes, infection_times, tree = make_partial_cascade(g, fraction, sampling_method=sampling_method)\n",
    "        r = mwu(g, node2id, id2node, \n",
    "                source, obs_nodes, infection_times, \n",
    "                s2n_probas,\n",
    "                epsilon,\n",
    "                query_selection_method=query_selection_method,\n",
    "                debug=False,\n",
    "                max_iter=max_iter,\n",
    "                save_log=False)\n",
    "        results.append(r)\n",
    "    return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "rounds = 100\n",
    "infp = 0.7\n",
    "fraction = 0.05\n",
    "epsilon = infp\n",
    "check_neighbor_threshold = 0.01\n",
    "sampling_method = 'late_nodes'\n",
    "params = {\n",
    "    'infp': infp,\n",
    "    'fraction': fraction,\n",
    "    'sampling_method': sampling_method\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "results_summary_by_query_method = {}\n",
    "results_by_query_method = {}\n",
    "# methods = [CENTROID, MAX_MU,  RANDOM]\n",
    "methods = [MAX_MU,  RANDOM]\n",
    "for method in methods:\n",
    "    cnts = experiment_multiple_rounds(rounds, g, \n",
    "                                      node2id, id2node,\n",
    "                                      s2n_probas,\n",
    "                                      infp, fraction, epsilon, sampling_method,\n",
    "                                      query_selection_method=method,\n",
    "                                      check_neighbor_threshold=check_neighbor_threshold,\n",
    "                                      max_iter=g.number_of_nodes(),\n",
    "                                      seed=123456)\n",
    "    s = pd.Series(list(filter(lambda c: c != False, cnts)))\n",
    "    results_summary_by_query_method[method] = s.describe().to_dict()\n",
    "    results_by_query_method[method] = s.as_matrix()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "from baselines import baseline_dog_tracker\n",
    "\n",
    "baseline_cnts = []\n",
    "for i in range(rounds):\n",
    "    source, obs_nodes, infection_times, tree = make_partial_cascade(g, fraction, sampling_method=sampling_method)\n",
    "    c = baseline_dog_tracker(g, obs_nodes, infection_times)\n",
    "    baseline_cnts.append(c)\n",
    "results_summary_by_query_method['dog'] = pd.Series(baseline_cnts).describe().to_dict()\n",
    "results_by_query_method['dog'] = np.array(baseline_cnts)\n",
    "result = pd.DataFrame.from_dict(results_summary_by_query_method)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "print('number of nodes: {}'.format(g.number_of_nodes()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "import os\n",
    "d = 'outputs/{}'.format(gtype)\n",
    "if not os.path.exists(d):\n",
    "    os.makedirs(d)\n",
    "result.to_pickle('{}/performance_comparison.pkl'.format(d))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEKCAYAAAACS67iAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/uQ9AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fa338a93eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, 1)\n",
    "data = list(results_by_query_method.values())\n",
    "ax.boxplot(data)\n",
    "ax.legend(list(results_by_query_method.keys()))\n",
    "ax.set_title('{}'.format(gtype))\n",
    "ax.set_ylim(0, g.number_of_nodes())\n",
    "fig.savefig('figs/boxplot-{}'.format(gtype))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{lrrrr}\n",
      "\\toprule\n",
      "{} &    dog &  max\\_adversarial &  max\\_mu &  random \\\\\n",
      "\\midrule\n",
      "25\\%   &  18.00 &            14.00 &   10.00 &   16.75 \\\\\n",
      "50\\%   &  24.00 &            22.00 &   15.00 &   28.50 \\\\\n",
      "75\\%   &  32.00 &            43.00 &   23.75 &   43.00 \\\\\n",
      "count & 100.00 &           100.00 &   98.00 &  100.00 \\\\\n",
      "max   &  54.00 &            86.00 &   93.00 &   92.00 \\\\\n",
      "mean  &  25.07 &            30.44 &   21.02 &   31.71 \\\\\n",
      "min   &  11.00 &             7.00 &    5.00 &    8.00 \\\\\n",
      "std   &   9.13 &            19.62 &   17.95 &   19.05 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(result.to_latex(float_format='%.2f'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.exit(-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "if False:\n",
    "    rounds = 500\n",
    "    infp = 0.7\n",
    "    fraction = 0.02\n",
    "    epsilon = 0.8\n",
    "    sampling_method = 'late_nodes'\n",
    "    check_neighbor_threshold = 0.1\n",
    "    \n",
    "    # calling it outside \n",
    "    result = experiment_multiple_rounds(rounds, g, \n",
    "                                        mean_times_by_source,\n",
    "                                        infp, fraction, epsilon, sampling_method,\n",
    "                                        query_selection_method=MAX_MU,\n",
    "                                        check_neighbor_threshold=check_neighbor_threshold)\n",
    "    cnts, _ = zip(*result)\n",
    "    print(np.mean(cnts))  # 10.252\n",
    "    print(np.median(cnts)) # 8.0\n",
    "\n",
    "    # calling it using parallel\n",
    "    epsilons = [epsilon]\n",
    "    results = Parallel(n_jobs=-1)(\n",
    "        delayed(experiment_multiple_rounds)(rounds, g, \n",
    "                                            mean_times_by_source,\n",
    "                                            infp, fraction, epsilon, sampling_method,\n",
    "                                            query_selection_method='max_mu',\n",
    "                                            check_neighbor_threshold=check_neighbor_threshold)\n",
    "        for eps in epsilons)\n",
    "\n",
    "    cnts, _ = zip(*results[0])\n",
    "    print(np.mean(cnts))  # 14.208\n",
    "    print(np.median(cnts))  # 11.0    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "epsilons = np.linspace(0.2, 0.9, 8)\n",
    "# epsilons = [0.8]\n",
    "results = []\n",
    "\n",
    "for eps in tqdm(epsilons):\n",
    "    result = experiment_multiple_rounds(rounds, g, \n",
    "                                        mean_times_by_source,\n",
    "                                        infp, fraction, epsilon, sampling_method,\n",
    "                                        query_selection_method='max_mu',\n",
    "                                        check_neighbor_threshold=check_neighbor_threshold)\n",
    "    results.append(result)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "list_cnts = [list(zip(*result))[0] for result in results]\n",
    "means = list(map(np.mean, list_cnts))\n",
    "medians = list(map(np.median, list_cnts))\n",
    "df = pd.DataFrame.from_dict({'eps': epsilons,\n",
    "                             'mean': means, \n",
    "                             'median': medians})\n",
    "df    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "from baselines import baseline_dog_tracker\n",
    "\n",
    "baseline_cnts = []\n",
    "for i in range(rounds):\n",
    "    source, obs_nodes, infection_times, tree = make_partial_cascade(g, fraction, sampling_method=sampling_method)\n",
    "    c = baseline_dog_tracker(g, obs_nodes, infection_times)\n",
    "    baseline_cnts.append(c)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "pd.Series(baseline_cnts).describe().to_dict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "print('mean', np.mean(baseline_cnts))\n",
    "print('median', np.median(baseline_cnts))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "fig, axes = plt.subplots(1, 2, figsize=(10, 5))\n",
    "ax0, ax1 = axes\n",
    "df.plot(x='eps', y=['mean'], ax=ax0)\n",
    "ax0.plot([np.mean(baseline_cnts)] * len(epsilons))\n",
    "ax0.legend(['multiplicative (max mu strategy)', 'baseline'])\n",
    "ax0.set_ylim(0, 100)\n",
    "ax0.set_title('mean')\n",
    "ax0.set_ylabel('query count')\n",
    "\n",
    "df.plot(x='eps', y=['median'], ax=axes[1])\n",
    "ax1.plot([np.median(baseline_cnts)] * len(epsilons))\n",
    "ax1.legend(['multiplicative (max mu strategy)', 'baseline'])\n",
    "ax1.set_ylim(0, 100)\n",
    "ax1.set_title('median')\n",
    "ax1.set_ylabel('query count')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "fig.savefig('figs/grid-{}-{}.png'.format(sampling_method, fraction))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "df.to_pickle('outputs/mwu-{}-{}.pkl'.format(sampling_method, fraction))\n",
    "import pickle\n",
    "pd.Series(baseline_cnts).to_pickle('outputs/dog-{}-{}.pkl'.format(sampling_method, fraction))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# summarize result\n",
    "eps = 0.8\n",
    "sampling_methods = ['uniform', 'late_nodes']\n",
    "fraction = 0.02\n",
    "for s in sampling_methods:\n",
    "    s1 = pd.read_pickle('outputs/mwu-{}-{}.pkl'.format(s, fraction))\n",
    "    s2_cnts = pd.read_pickle('outputs/dog-{}-{}.pkl'.format(s, fraction))\n",
    "    print('method', s)\n",
    "    print('mean', s1[s1['eps'] == 0.8].iloc[0]['mean'], np.mean(s2_cnts))\n",
    "    print('median', s1[s1['eps'] == 0.8].iloc[0]['median'], np.median(s2_cnts))    \n",
    "    print()"
   ]
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}