probabilistic_scheduler.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from datetime import datetime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "sns.set()\n",
    "sns.set_context(\"notebook\", font_scale=1.5, rc={\"lines.linewidth\": 2.5})\n",
    "plt.rcParams['figure.figsize'] = 16, 12"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from scheduler.scheduler import ProbabilisticTaskScheduler, TaskScheduler, Task, Slot, LikelihoodEstimator, UniformDistribution, Time, DiscreteDistribution, BetaDistribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "CATEGORY_PRIORS = dict(wellness=UniformDistribution(Time(hours=14), Time(hours=20)),\n",
    "              learning=UniformDistribution(Time(hours=10), Time(hours=14)),\n",
    "              work=UniformDistribution(Time(hours=11), Time(hours=15)),\n",
    "              personal=UniformDistribution(Time(hours=8), Time(hours=22)),\n",
    "              fun=UniformDistribution(Time(hours=12), Time(hours=22)),\n",
    "              chores=UniformDistribution(Time(hours=16), Time(hours=22)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for cat, dist in CATEGORY_PRIORS.items():\n",
    "    interval_starts = [s for s in range(0, 24 * 60, 15)]\n",
    "    freq = dict()\n",
    "    for s in interval_starts:\n",
    "        freq[s] = np.mean(BetaDistribution(np.random.randint(1, 100), np.random.randint(1, 50)).sample(1000))\n",
    "    d = DiscreteDistribution(freq, 15)\n",
    "    CATEGORY_PRIORS[cat] = d.joint(dist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class TaskCategoryLikelihoodEstimator(LikelihoodEstimator):\n",
    "\n",
    "    def likelihood(self, task, priors):\n",
    "        if not task.category in priors:\n",
    "            return UniformDistribution(Time(hours=0), Time(hours=24))\n",
    "        return priors[task.category]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "t1 = Task(\"Workout in the park\", \"wellness\", 60, Time(hours=11))\n",
    "t2 = Task(\"Read a book\", \"learning\", 30, Time(hours=10))\n",
    "scheduler = ProbabilisticTaskScheduler(TaskScheduler([t1, t2], 9, 16), TaskCategoryLikelihoodEstimator(CATEGORY_PRIORS))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "task_list = []\n",
    "for t in scheduler.tasks:\n",
    "    task_list.append(dict(name=t.name, context=t.category, duration=t.duration))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>context</th>\n",
       "      <th>duration</th>\n",
       "      <th>name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>wellness</td>\n",
       "      <td>60</td>\n",
       "      <td>Workout in the park</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>learning</td>\n",
       "      <td>30</td>\n",
       "      <td>Read a book</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    context  duration                 name\n",
       "0  wellness        60  Workout in the park\n",
       "1  learning        30          Read a book"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.DataFrame.from_dict(task_list)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "dummies = pd.get_dummies(df['context']).rename(columns=lambda x: 'context_' + str(x))\n",
    "df = pd.concat([df, dummies], axis=1)\n",
    "df.drop(['context'], inplace=True, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>duration</th>\n",
       "      <th>name</th>\n",
       "      <th>context_learning</th>\n",
       "      <th>context_wellness</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>60</td>\n",
       "      <td>Workout in the park</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>30</td>\n",
       "      <td>Read a book</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   duration                 name  context_learning  context_wellness\n",
       "0        60  Workout in the park                 0                 1\n",
       "1        30          Read a book                 1                 0"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "t3 = Task(\"Meditate\", \"wellness\", 15)\n",
    "slots = scheduler.propose_slots(t3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "probs = {s[1].start_minute: s[0] for s in slots}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.FacetGrid at 0x119b15b00>"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABGYAAAInCAYAAADanV8AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nc3mCgQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x119952b00>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data = pd.DataFrame(list(probs.items()), columns=[\"start_minute\", \"density\"])\n",
    "sns.lmplot(\"start_minute\", \"density\", data, fit_reg=False, size=8, aspect=2)"
   ]
  }
 ],
 "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": "3.5.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}