{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Use K-medoid algorithm to find the suitable human model representitives"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os, sys\n",
    "lib_path = os.path.abspath(os.path.join('..', '..', 'utils'))\n",
    "sys.path.append(lib_path)\n",
    "from GraphType import GraphStat\n",
    "import readCSV as reader\n",
    "from scipy import stats\n",
    "from ipywidgets import interact, fixed, interactive\n",
    "import ipywidgets as widgets\n",
    "from pyclustering.cluster.kmedoids import kmedoids\n",
    "from pyclustering.utils.metric import distance_metric, type_metric\n",
    "import random"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define a new distance metric"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def ks_value(dest1, dest2):\n",
    "    value, p = stats.ks_2samp(dest1, dest2)\n",
    "    return value\n",
    "\n",
    "\n",
    "ks_metric = distance_metric(type_metric.USER_DEFINED, func=ks_value)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read Human Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1253"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Progress Widge\n",
    "w = widgets.FloatProgress(\n",
    "    value=0,\n",
    "    min=0,\n",
    "    max=1.0,\n",
    "    step=0.1,\n",
    "    description='Loading Files...:',\n",
    "    bar_style='info',\n",
    "    orientation='horizontal'\n",
    ")\n",
    "\n",
    "\n",
    "humanFiles = reader.readmultiplefiles('../input/humanOutput/', 1300, False)\n",
    "modelToFileName = {}\n",
    "for name in humanFiles:\n",
    "    modelToFileName[GraphStat(name)] = name\n",
    "\n",
    "models = list(modelToFileName.keys())\n",
    "len(humanFiles)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find Representative by K-medroid for different dists on GraphStat"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Returns the index of the representative"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def findRep(graphStats, func):\n",
    "    out_ds = list(map(func, models))\n",
    "\n",
    "    #choose a random starting point\n",
    "    start_index = random.randint(0, len(out_ds))\n",
    "\n",
    "    # start with one initial metrid [start_index]\n",
    "    outdegree_kmedoid = kmedoids(out_ds, [start_index], metric=ks_metric)\n",
    "\n",
    "    outdegree_kmedoid.process()\n",
    "    centoids = outdegree_kmedoid.get_medoids()\n",
    "    return centoids[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find representative for out degree"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* the rep found is ../input/humanOutput\\R_20158_run_1.csv\n",
    "* the average distance between it and others is 0.05515988287586802"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "../input/humanOutput\\R_20158_run_1.csv\n",
      "../input/humanOutput\\R_20158_run_1.csv\n"
     ]
    }
   ],
   "source": [
    "od_rep_index = findRep(models, lambda m: m.out_d)\n",
    "print(list(modelToFileName.values())[od_rep_index])\n",
    "od_rep_model = models[od_rep_index]\n",
    "print(modelToFileName[od_rep_model])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.05515988287586802\n"
     ]
    }
   ],
   "source": [
    "total_distance = 0\n",
    "count = 0\n",
    "for model in models:\n",
    "    total_distance += ks_value(od_rep_model.out_d, model.out_d)\n",
    "print(total_distance / len(models))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find Representative for node activities"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* the rep found is ../input/humanOutput\\R_2016176_run_1.csv\n",
    "* the average distance between it and others is 0.05275267434589047"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "../input/humanOutput\\R_2016176_run_1.csv\n",
      "../input/humanOutput\\R_2016176_run_1.csv\n"
     ]
    }
   ],
   "source": [
    "total_distance = 0\n",
    "for model in models:\n",
    "    total_distance += ks_value(od_rep_model.mpc, model.mpc)\n",
    "print(total_distance / len(models))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.05275267434589047\n"
     ]
    }
   ],
   "source": [
    "total_distance = 0\n",
    "count = 0\n",
    "for model in models:\n",
    "    total_distance += ks_value(od_rep_model.na, model.na)\n",
    "print(total_distance / len(models))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find Representative for MPC"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* the rep found is ../input/humanOutput\\R_2015246_run_1.csv\n",
    "* the average distance between it and others is 0.08556632702185384"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "../input/humanOutput\\R_2015246_run_1.csv\n",
      "../input/humanOutput\\R_2015246_run_1.csv\n"
     ]
    }
   ],
   "source": [
    "mpc_rep_index = findRep(models, lambda m: m.mpc)\n",
    "print(list(modelToFileName.values())[mpc_rep_index])\n",
    "mpc_rep_model = models[mpc_rep_index]\n",
    "print(modelToFileName[mpc_rep_model])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.08556632702185384\n"
     ]
    }
   ],
   "source": [
    "total_distance = 0\n",
    "count = 0\n",
    "for model in models:\n",
    "    total_distance += ks_value(od_rep_model.mpc, model.mpc)\n",
    "print(total_distance / len(models))"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}