Multi-Armed Bandits.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Multi-Armed Bandits"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook is designed to be read along with the [second article](https://medium.com/@alejandro.aristizabal24/understanding-reinforcement-learning-hands-on-part-2-multi-armed-bandits-526592072bdc) on a series on Reinforcement Learning. It is adviced to go through both resources at the same time, to cover both theory and code on the topic. The notebook will be focused on applying the theory, while the article is focused on explaining it."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preliminaries\n",
    "\n",
    "Let's start by loading all the required libraries for this notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import necessary libraries\n",
    "%matplotlib notebook\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.animation import FuncAnimation\n",
    "import gym\n",
    "from tqdm import tqdm\n",
    "import time"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scenario\n",
    "\n",
    "The Multi-Armed Bandit describes a situation in which an agent has only one state, and multiple actions to interact with the environment. Each action gives a random reward, centered on an unkown value. Our agent wants to maximize the reward received, which means it wants to find the action that yields a higher reward. Based on this, let's build the scenario.\n",
    "\n",
    "We're going to implement the multi-armed bandit environment using OpenAI's gym interface. Also, our implementation will be vectorized for the sake of optimization. This allows us to run multiple agents on multiple environments at the same time. The code below is based on [this implementation](https://github.com/diegoalejogm/openai-k-armed-bandits)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from gym import spaces\n",
    "from gym.utils import seeding\n",
    "\n",
    "class ArmedBanditsEnv(gym.Env):\n",
    "    \"\"\"\n",
    "    The famous k-Armed Bandit Environment, implemented for the gym interface.\n",
    "    Initialization requires an array of length equals to k, where each item is\n",
    "    a function which samples from a specified distribution.\n",
    "    \"\"\"\n",
    "    metadata = {'render.modes': ['human']}\n",
    "    \n",
    "    def __init__(self, mean, stddev):\n",
    "        assert len(mean.shape) == 2\n",
    "        assert len(stddev.shape) == 2\n",
    "        \n",
    "        super(ArmedBanditsEnv, self).__init__()\n",
    "        # Define action and observation space\n",
    "        self.num_bandits = mean.shape[1]\n",
    "        self.num_experiments = mean.shape[0]\n",
    "        self.action_space = spaces.Discrete(self.num_bandits)\n",
    "        \n",
    "        # Theres one state only in the k-armed bandits problem\n",
    "        self.observation_space = spaces.Discrete(1)\n",
    "        self.mean = mean\n",
    "        self.stddev = stddev\n",
    "        \n",
    "    def step(self, action):\n",
    "        # Sample from the specified bandit using it's reward distribution\n",
    "        assert (action < self.num_bandits).all()\n",
    "        \n",
    "        sampled_means = self.mean[np.arange(self.num_experiments),action]\n",
    "        sampled_stddevs = self.stddev[np.arange(self.num_experiments),action]\n",
    "        \n",
    "        reward = np.random.normal(loc=sampled_means, scale=sampled_stddevs, size=(self.num_experiments,))\n",
    "        \n",
    "        # Return a constant state of 0. Our environment has no terminal state\n",
    "        observation, done, info = 0, False, dict()\n",
    "        return observation, reward, done, info\n",
    "    \n",
    "    def reset(self):\n",
    "        return 0\n",
    "        \n",
    "    def render(self, mode='human', close=False):\n",
    "        pass\n",
    "    \n",
    "    def _seed(self, seed=None):\n",
    "        self.np_random, seed = seeding.np.random(seed)\n",
    "        return [seed]\n",
    "    \n",
    "    def close(self):\n",
    "        pass\n",
    "    \n",
    "    \n",
    "class ArmedBanditsGaussian(ArmedBanditsEnv):\n",
    "    def __init__(self, num_experiments=1, num_bandits=3):\n",
    "        self.means = np.random.normal(size=(num_experiments, num_bandits))\n",
    "        \n",
    "        ArmedBanditsEnv.__init__(self, self.means, np.ones((num_experiments, num_bandits)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we're creating two classes, `ArmedBanditsEnv` is the main class, while `ArmedBanditsGaussian` is an auxiliary class that allows us to easily create an environment with random mean rewards for each action. Our environment receives numpy arrays for the means and standard deviations for each action. The dimensions of this arrays are described as `num_experiments`x`num_bandits`. Taking a step requires a numpy vector of size `num_experiments`, where each value specifies which action to take for each experiment. The step functions returns, among other information, a vector of reward obtained for each experiment. Let's see this in action:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Bandit: 0  gave a reward of: 5.192030344836741\n",
      "Bandit: 1  gave a reward of: 0.924191944312297\n",
      "Bandit: 2  gave a reward of: -6.694986087601626\n",
      "Bandit: 3  gave a reward of: -10.344768246225584\n"
     ]
    }
   ],
   "source": [
    "means = np.array([[5, 1, 0, -10]]) # The mean for a four-armed bandit. Single experiment\n",
    "stdev = np.array([[1, 0.1, 5, 1]]) # The standard deviation for a four-armed bandit.\n",
    "\n",
    "env = ArmedBanditsEnv(means, stdev) # Create the environment\n",
    "\n",
    "for i in range(4):\n",
    "    action = np.array([[i]])\n",
    "    _, reward, _, _ = env.step(action)\n",
    "    print(\"Bandit:\", i, \" gave a reward of:\", reward[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluating our actions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In order to learn something from our interaction with the environment, we need to know how exactly we determine the value of our actions, as well as how to keep that value up to date. A simple strategy our agents can take is to calculate the expected return for each action. This can be done through experience by taking the average of previous rewards given by a determined action.\n",
    "\n",
    "For example, we can use the next list of rewards to calculate the expected return of the chosen action:\n",
    "\n",
    "<img src='assets/average.png'></img>\n",
    "\n",
    "Having to store all the previously seen rewards to calculate the value of an action is cumbersome, inefficient and unnecessary. We can derive another form of average, called *Incremental average update rule*, which only requires us to know the previous average $Q_{n-1}$ and the number `n`:\n",
    "\n",
    "<img src='assets/incremental-average.png'></img>\n",
    "\n",
    "Here's a small demonstration of this function at work"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average obtained from incremental update rule:  4.991666666666666\n",
      "Average obtained from basic average function:   4.991666666666666\n"
     ]
    }
   ],
   "source": [
    "def inc_avg(prev_avg, new_val, n):\n",
    "    return prev_avg + 1/n*(new_val - prev_avg)\n",
    "\n",
    "# Obtain the previous average\n",
    "vals = np.array([4.5, 5.04, 5.32, 4.8, 5.11])\n",
    "prev_avg = vals.mean()\n",
    "\n",
    "# Calculate a new average using the incremental average update function\n",
    "new_val = 5.18\n",
    "new_avg = inc_avg(prev_avg, new_val, 6)\n",
    "\n",
    "# Calculate the same average using all previous values for comparison\n",
    "avg = np.append(vals, new_val).mean()\n",
    "\n",
    "print(\"Average obtained from incremental update rule: \", new_avg)\n",
    "print(\"Average obtained from basic average function:  \", avg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This method will be used on the next strategies"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The Greedy Agent\n",
    "\n",
    "This strategy is focused on always choosing the best known action at the time. Every time the agent takes an action, it looks at the estimated values for each action, and chooses the one that has a greater score. If more than two values look best, then the agent selects arbitrarily among those best-valued actions. This is called breaking ties arbitrarily\n",
    "\n",
    "Taking the action with the greatest value is equivalent to using the **argmax** function. Although, we need to implement some changes so that the funciton breaks the ties the way we intend it to do:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def argmax(q_values):\n",
    "    \"\"\"\n",
    "    Takes in a matrix of n*k q_values and returns the index\n",
    "    of the item with the highest value for each row. \n",
    "    Breaks ties randomly.\n",
    "    returns: vector of size n, where each item is the index of\n",
    "    the highest value in q_values for each row.\n",
    "    \"\"\"\n",
    "    # Generate a mask of the max values for each row\n",
    "    mask = q_values == q_values.max(axis=1)[:, None]\n",
    "    # Generate noise to be added to the ties\n",
    "    r_noise = 1e-6*np.random.random(q_values.shape)\n",
    "    # Get the argmax of the noisy masked values\n",
    "    return np.argmax(r_noise*mask,axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we're generating a mask of all the values in the input that are equal to the maximum value. Then ,we generate some noise and multiply it with the mask. Taking the argmax of this new list will be equivalent to the desired argmax with ties broken arbitrarily. \n",
    "\n",
    "Now that we have specified the argmax function, we can declare our class for the Greedy Agent, which is capable of acting upon the environment, and updating its estimates after receiving a reward."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "class GreedyAgent:\n",
    "    def __init__(self, reward_estimates):\n",
    "        \"\"\"\n",
    "        Our agent takes as input the initial reward estimates.\n",
    "        This estimates will be updated incrementally after each \n",
    "        interaction with the environment.\n",
    "        \"\"\"\n",
    "        assert len(reward_estimates.shape) == 2\n",
    "        \n",
    "        self.num_bandits = reward_estimates.shape[1]\n",
    "        self.num_experiments = reward_estimates.shape[0]\n",
    "        self.reward_estimates = reward_estimates.astype(np.float64)\n",
    "        self.action_count = np.zeros(reward_estimates.shape)\n",
    "        \n",
    "    def get_action(self):\n",
    "        # Our agent is greedy, so there's no need for exploration.\n",
    "        # Our argmax will do just fine for this situation\n",
    "        action = argmax(self.reward_estimates)\n",
    "        \n",
    "        # Add a 1 to each action selected in the action count\n",
    "        self.action_count[np.arange(self.num_experiments), action] += 1\n",
    "        \n",
    "        return action\n",
    "    \n",
    "    def update_estimates(self, reward, action):\n",
    "        # rew is a matrix with the obtained rewards from our previuos\n",
    "        # action. Use this to update our estimates incrementally\n",
    "        n = self.action_count[np.arange(self.num_experiments), action]\n",
    "        \n",
    "        # Compute the difference between the received rewards vs the reward estimates\n",
    "        error = reward - self.reward_estimates[np.arange(self.num_experiments), action]\n",
    "        \n",
    "        # Update the reward difference incementally\n",
    "        self.reward_estimates[np.arange(self.num_experiments), action] += (1/n)*error"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The greedy agent contains a matrix for the estimates, as well as for the number of times each action has been taken. This is necessary for using the incremental average update rule, used inside the `update_estimates` function."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing the agent\n",
    "\n",
    "Let's see how the Greedy Agent behaves on the environment. For this, we're going to generate some animations where we're able to see how the agent estimates the values for each action, as well as the real values provided by the environment. Remember that the agent doesn't have access to that information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support. ' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            if (mpl.ratio != 1) {\n",
       "                fig.send_message(\"set_dpi_ratio\", {'dpi_ratio': mpl.ratio});\n",
       "            }\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        fig.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var backingStore = this.context.backingStorePixelRatio ||\n",
       "\tthis.context.webkitBackingStorePixelRatio ||\n",
       "\tthis.context.mozBackingStorePixelRatio ||\n",
       "\tthis.context.msBackingStorePixelRatio ||\n",
       "\tthis.context.oBackingStorePixelRatio ||\n",
       "\tthis.context.backingStorePixelRatio || 1;\n",
       "\n",
       "    mpl.ratio = (window.devicePixelRatio || 1) / backingStore;\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width * mpl.ratio);\n",
       "        canvas.attr('height', height * mpl.ratio);\n",
       "        canvas.attr('style', 'width: ' + width + 'px; height: ' + height + 'px;');\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>');\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option);\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'] / mpl.ratio;\n",
       "    var y0 = (fig.canvas.height - msg['y0']) / mpl.ratio;\n",
       "    var x1 = msg['x1'] / mpl.ratio;\n",
       "    var y1 = (fig.canvas.height - msg['y1']) / mpl.ratio;\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width / mpl.ratio, fig.canvas.height / mpl.ratio);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x * mpl.ratio;\n",
       "    var y = canvas_pos.y * mpl.ratio;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overridden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    var width = fig.canvas.width/mpl.ratio\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\" width=\"' + width + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var width = this.canvas.width/mpl.ratio\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\" width=\"' + width + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>');\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"998.8889153504084\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Initialize the environment of our multi-armed bandit problem\n",
    "num_experiments = 2\n",
    "num_bandits = 8\n",
    "num_steps = 100\n",
    "means = np.random.normal(size=(num_experiments, num_bandits))\n",
    "stdev = np.ones((num_experiments, num_bandits))\n",
    "\n",
    "env = ArmedBanditsEnv(means, stdev)\n",
    "\n",
    "# Initialize the agent\n",
    "agent = GreedyAgent(np.zeros((num_experiments,num_bandits)))\n",
    "\n",
    "# Code for plotting the interaction\n",
    "fig, axs = plt.subplots(1, num_experiments, figsize=(10, 4))\n",
    "x_pos = np.arange(num_bandits)\n",
    "\n",
    "def init():\n",
    "    for i in range(num_experiments):\n",
    "        init_ax(i)\n",
    "        \n",
    "    \n",
    "def init_ax(i):\n",
    "    ax = axs[i]\n",
    "    ax.clear()\n",
    "    ax.set_ylim(-4, 4)\n",
    "    ax.set_xlim(-0.5, num_bandits-.5)\n",
    "    ax.set_xlabel('Actions', fontsize=14)\n",
    "    ax.set_ylabel('Value', fontsize=14)\n",
    "    ax.set_title(label='Estimated Values vs. Real values', fontsize=15)\n",
    "    ax.plot(x_pos, env.mean[i], marker='D', linestyle='', alpha=0.8, color='r', label='Real Values')\n",
    "    ax.axhline(0, color='black', lw=1)\n",
    "    \n",
    "# Implement a step, which involves the agent acting upon the\n",
    "# environment and learning from the received reward.\n",
    "def step(g):\n",
    "    action = agent.get_action()\n",
    "    _, reward, _, _ = env.step(action)\n",
    "    agent.update_estimates(reward, action)\n",
    "    for i in range(num_experiments):\n",
    "        ax = axs[i]\n",
    "        # Plot the estimated values from the agent compared to the real values\n",
    "        estimates = agent.reward_estimates[i]\n",
    "        init_ax(i)\n",
    "        values = ax.bar(x_pos, estimates, align='center', color='blue', alpha=0.4, label='Estimated Values')\n",
    "        ax.legend()\n",
    "    \n",
    "anim = FuncAnimation(fig, func=step, frames=np.arange(num_steps), init_func=init, interval=10, repeat=False, blit=True)\n",
    "plt.show()\n",
    "\n",
    "# Uncomment the next line if you wish to store the animations as a gif\n",
    "# anim.save('./greedy-agent.gif', writer='imagemagick', fps=60)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "you may try the previous block of code multiple times to see multiple animations. In general, you may realize that most of the actions are not explored by the agent, and that it will mostly stick to the one action that gave it some positive reward. The agent is good at evading negative values, but it will only land on the optimal action by pure chance.\n",
    "\n",
    "We can plot the average behavior by doing more experiments. The next plot displays the percentage of times the agent chose the optimal action over an average of 10000 experiments."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1000/1000 [00:04<00:00, 214.90it/s]\n"
     ]
    },
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support. ' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            if (mpl.ratio != 1) {\n",
       "                fig.send_message(\"set_dpi_ratio\", {'dpi_ratio': mpl.ratio});\n",
       "            }\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        fig.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var backingStore = this.context.backingStorePixelRatio ||\n",
       "\tthis.context.webkitBackingStorePixelRatio ||\n",
       "\tthis.context.mozBackingStorePixelRatio ||\n",
       "\tthis.context.msBackingStorePixelRatio ||\n",
       "\tthis.context.oBackingStorePixelRatio ||\n",
       "\tthis.context.backingStorePixelRatio || 1;\n",
       "\n",
       "    mpl.ratio = (window.devicePixelRatio || 1) / backingStore;\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width * mpl.ratio);\n",
       "        canvas.attr('height', height * mpl.ratio);\n",
       "        canvas.attr('style', 'width: ' + width + 'px; height: ' + height + 'px;');\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>');\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option);\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'] / mpl.ratio;\n",
       "    var y0 = (fig.canvas.height - msg['y0']) / mpl.ratio;\n",
       "    var x1 = msg['x1'] / mpl.ratio;\n",
       "    var y1 = (fig.canvas.height - msg['y1']) / mpl.ratio;\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width / mpl.ratio, fig.canvas.height / mpl.ratio);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x * mpl.ratio;\n",
       "    var y = canvas_pos.y * mpl.ratio;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overridden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    var width = fig.canvas.width/mpl.ratio\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\" width=\"' + width + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var width = this.canvas.width/mpl.ratio\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\" width=\"' + width + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>');\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"958.8889142907702\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_experiments = 10000\n",
    "num_steps = 1000\n",
    "num_actions = 10\n",
    "\n",
    "# Initialize the environment\n",
    "env = ArmedBanditsGaussian(num_experiments, num_actions)\n",
    "# Initialize the agent\n",
    "agent = GreedyAgent(np.zeros((num_experiments, num_actions)))\n",
    "\n",
    "# Store the scores and averages for later plotting\n",
    "averages = np.zeros((num_steps))\n",
    "optimality = np.zeros((num_steps))\n",
    "scores = np.zeros((num_experiments, num_steps+1))\n",
    "\n",
    "#Store the optimal actions for later use\n",
    "optimal = np.argmax(env.mean, axis=1)\n",
    "\n",
    "for i in tqdm(range(num_steps)):\n",
    "    # Select an action to execute on the environment\n",
    "    action = agent.get_action()\n",
    "    _, reward, _, _ = env.step(action)\n",
    "    \n",
    "    # Update the agent estimates with the previously observed rewards\n",
    "    agent.update_estimates(reward, action)\n",
    "    \n",
    "    # Store the average cumulative score and optimality of the current step\n",
    "    scores[:,i+1] = scores[:,i] + reward\n",
    "    avg_score = np.mean(scores[:,i+1]/(i+1))\n",
    "    averages[i] = avg_score\n",
    "    \n",
    "    # Get optimal actions from the environment\n",
    "    current_optimality = np.mean(action == optimal)\n",
    "    optimality[i] = current_optimality\n",
    "    \n",
    "plt.figure(figsize=(12,6), dpi=80, facecolor='w', edgecolor='k')\n",
    "plt.plot([1.0 for _ in range(num_steps)], linestyle='--')\n",
    "plt.plot([0.0 for _ in range(num_steps)], linestyle='--')\n",
    "plt.plot(optimality)\n",
    "plt.legend([\"Best Possible\", \"Worst Possible\", \"Greedy\"])\n",
    "plt.title(\"Average Optimality of Greedy Agent\", fontsize=14)\n",
    "plt.ylabel(\"% of Optimal Action Taken\")\n",
    "plt.xlabel(\"Steps\")\n",
    "plt.show()\n",
    "greedy_scores = averages\n",
    "greedy_optimality = optimality"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The Greedy Agent chooses the best action on average ~40% of the times. Additionally, extra experience won't improve it's score, as it usually lands on a sub-optimal action on the first steps, and stays there for ever. The Greedy Agent is not a good strategy for finding the optimal action on this scenario."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The Epsilon-Greedy Agent\n",
    "\n",
    "Another strategy is the Epsilon Greedy Agent, which adds to the previous strategy. An Epsilon-Greedy Agent allows for some exploratory actions, by every once in a while choosing any action randomly, instead of always acting greedily. The probability of taking an exploratory action is defined by the parameter `epsilon`. An epsilon of `0` is a Greedy Agent. An epsilon of `0.2` means our agent takes a random action 20% of the time. An epsilon of `1` is an agent that behaves randomly. Let's build this new agent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "class EpsilonGreedyAgent(GreedyAgent):\n",
    "    def __init__(self, reward_estimates, epsilon):\n",
    "        GreedyAgent.__init__(self, reward_estimates)\n",
    "        # Store the epsilon value\n",
    "        assert epsilon >= 0 and epsilon <= 1\n",
    "        self.epsilon = epsilon\n",
    "    \n",
    "    def get_action(self):\n",
    "        # We need to redefine this function so that it takes an exploratory action with epsilon probability\n",
    "        \n",
    "        # One hot encoding: 0 if exploratory, 1 otherwise\n",
    "        action_type = (np.random.random_sample(self.num_experiments) > self.epsilon).astype(int)\n",
    "        # Generate both types of actions for every experiment\n",
    "        exploratory_action = np.random.randint(self.num_bandits, size=self.num_experiments)\n",
    "        greedy_action = argmax(self.reward_estimates)\n",
    "        # Use the one hot encoding to mask the actions for each experiment\n",
    "        action = greedy_action * action_type + exploratory_action * (1 - action_type)\n",
    "        \n",
    "        self.action_count[np.arange(self.num_experiments), action] += 1\n",
    "        \n",
    "        return action"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As may be seen, we're inheriting the `GreedyAgent` implementation, and only changing the `get_action` function. For taking an action, we're using a masking that defines which experiments will choose an action randomly and which will act greedily. Then, we apply that masking between random and greedy actions to obtain the action vector for all of our experiments."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing the Agent"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, we're going to use the same animation as before to observe how this new strategy behaves. Our agents will have a value of epsilon of `0.1`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support. ' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            if (mpl.ratio != 1) {\n",
       "                fig.send_message(\"set_dpi_ratio\", {'dpi_ratio': mpl.ratio});\n",
       "            }\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        fig.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var backingStore = this.context.backingStorePixelRatio ||\n",
       "\tthis.context.webkitBackingStorePixelRatio ||\n",
       "\tthis.context.mozBackingStorePixelRatio ||\n",
       "\tthis.context.msBackingStorePixelRatio ||\n",
       "\tthis.context.oBackingStorePixelRatio ||\n",
       "\tthis.context.backingStorePixelRatio || 1;\n",
       "\n",
       "    mpl.ratio = (window.devicePixelRatio || 1) / backingStore;\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width * mpl.ratio);\n",
       "        canvas.attr('height', height * mpl.ratio);\n",
       "        canvas.attr('style', 'width: ' + width + 'px; height: ' + height + 'px;');\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>');\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option);\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'] / mpl.ratio;\n",
       "    var y0 = (fig.canvas.height - msg['y0']) / mpl.ratio;\n",
       "    var x1 = msg['x1'] / mpl.ratio;\n",
       "    var y1 = (fig.canvas.height - msg['y1']) / mpl.ratio;\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width / mpl.ratio, fig.canvas.height / mpl.ratio);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x * mpl.ratio;\n",
       "    var y = canvas_pos.y * mpl.ratio;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overridden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    var width = fig.canvas.width/mpl.ratio\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\" width=\"' + width + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var width = this.canvas.width/mpl.ratio\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\" width=\"' + width + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>');\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"998.8889153504084\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Initialize the environment of our multi-armed bandit problem\n",
    "num_experiments = 2\n",
    "num_bandits = 8\n",
    "num_steps = 200\n",
    "means = np.random.normal(size=(num_experiments, num_bandits))\n",
    "stdev = np.ones((num_experiments, num_bandits))\n",
    "\n",
    "env = ArmedBanditsEnv(means, stdev)\n",
    "\n",
    "# Initialize the agent\n",
    "agent = EpsilonGreedyAgent(np.zeros((num_experiments,num_bandits)), 0.1)\n",
    "\n",
    "# Code for plotting the interaction\n",
    "fig, axs = plt.subplots(1, num_experiments, figsize=(10, 4))\n",
    "x_pos = np.arange(num_bandits)\n",
    "\n",
    "def init():\n",
    "    for i in range(num_experiments):\n",
    "        init_ax(i)\n",
    "        \n",
    "    \n",
    "def init_ax(i):\n",
    "    ax = axs[i]\n",
    "    ax.clear()\n",
    "    ax.set_ylim(-4, 4)\n",
    "    ax.set_xlim(-0.5, num_bandits-.5)\n",
    "    ax.set_xlabel('Actions', fontsize=14)\n",
    "    ax.set_ylabel('Value', fontsize=14)\n",
    "    ax.set_title(label='Estimated Values vs. Real values', fontsize=15)\n",
    "    ax.plot(x_pos, env.mean[i], marker='D', linestyle='', alpha=0.8, color='r', label='Real Values')\n",
    "    ax.axhline(0, color='black', lw=1)\n",
    "    \n",
    "# Implement a step, which involves the agent acting upon the\n",
    "# environment and learning from the received reward.\n",
    "def step(g):\n",
    "    action = agent.get_action()\n",
    "    _, reward, _, _ = env.step(action)\n",
    "    agent.update_estimates(reward, action)\n",
    "    for i in range(num_experiments):\n",
    "        ax = axs[i]\n",
    "        # Plot the estimated values from the agent compared to the real values\n",
    "        estimates = agent.reward_estimates[i]\n",
    "        init_ax(i)\n",
    "        values = ax.bar(x_pos, estimates, align='center', color='blue', alpha=0.4, label='Estimated Values')\n",
    "        ax.legend()\n",
    "    \n",
    "anim = FuncAnimation(fig, func=step, frames=np.arange(num_steps), init_func=init, interval=10, repeat=False, blit=True)\n",
    "plt.show()\n",
    "\n",
    "# Uncomment the next line if you wish to store the animations as a gif\n",
    "# anim.save('./epsilon-greedy-agent.gif', writer='imagemagick', fps=60)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Again, you may run the previous block multiple times to see different runs. As can be observed, the epsilon-greedy agent not only finds the optimal solution most of the times, but it is also capable of finding close estimated values for all the actions! How does our new strategy compares to the greedy agent on average?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_experiment(num_experiments=1000, num_steps=1000, num_actions=10,epsilon=0.1):\n",
    "    # Initialize the environment\n",
    "    env = ArmedBanditsGaussian(num_experiments, num_actions)\n",
    "    # Initialize the agent\n",
    "    agent = EpsilonGreedyAgent(np.zeros((num_experiments, num_actions)), epsilon)\n",
    "    \n",
    "    # Store the scores and averages for later plotting\n",
    "    averages = np.zeros((num_steps))\n",
    "    optimality = np.zeros((num_steps))\n",
    "    scores = np.zeros((num_experiments, num_steps+1))\n",
    "    \n",
    "    #Store the optimal actions for later use\n",
    "    optimal = np.argmax(env.mean, axis=1)\n",
    "    \n",
    "    for i in tqdm(range(num_steps)):\n",
    "        # Select an action to execute on the environment\n",
    "        action = agent.get_action()\n",
    "        _, reward, _, _ = env.step(action)\n",
    "        \n",
    "        # Update the agent estimates with the previously observed rewards\n",
    "        agent.update_estimates(reward, action)\n",
    "        \n",
    "        # Store the average cumulative score and optimality of the current step\n",
    "        scores[:,i+1] = scores[:,i] + reward\n",
    "        avg_score = np.mean(scores[:,i+1]/(i+1))\n",
    "        averages[i] = avg_score\n",
    "        \n",
    "        # Get optimal actions from the environment\n",
    "        current_optimality = np.mean(action == optimal)\n",
    "        optimality[i] = current_optimality\n",
    "    return optimality, averages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1000/1000 [00:05<00:00, 190.89it/s]\n"
     ]
    },
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support. ' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            if (mpl.ratio != 1) {\n",
       "                fig.send_message(\"set_dpi_ratio\", {'dpi_ratio': mpl.ratio});\n",
       "            }\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        fig.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var backingStore = this.context.backingStorePixelRatio ||\n",
       "\tthis.context.webkitBackingStorePixelRatio ||\n",
       "\tthis.context.mozBackingStorePixelRatio ||\n",
       "\tthis.context.msBackingStorePixelRatio ||\n",
       "\tthis.context.oBackingStorePixelRatio ||\n",
       "\tthis.context.backingStorePixelRatio || 1;\n",
       "\n",
       "    mpl.ratio = (window.devicePixelRatio || 1) / backingStore;\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width * mpl.ratio);\n",
       "        canvas.attr('height', height * mpl.ratio);\n",
       "        canvas.attr('style', 'width: ' + width + 'px; height: ' + height + 'px;');\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>');\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option);\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'] / mpl.ratio;\n",
       "    var y0 = (fig.canvas.height - msg['y0']) / mpl.ratio;\n",
       "    var x1 = msg['x1'] / mpl.ratio;\n",
       "    var y1 = (fig.canvas.height - msg['y1']) / mpl.ratio;\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width / mpl.ratio, fig.canvas.height / mpl.ratio);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x * mpl.ratio;\n",
       "    var y = canvas_pos.y * mpl.ratio;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overridden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    var width = fig.canvas.width/mpl.ratio\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\" width=\"' + width + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var width = this.canvas.width/mpl.ratio\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\" width=\"' + width + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>');\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"958.8889142907702\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_experiments = 10000\n",
    "num_steps = 1000\n",
    "num_actions = 10\n",
    "epsilon = 0.1\n",
    "\n",
    "balanced_eps_optimality, balanced_eps_scores = run_experiment(num_experiments, num_steps, num_actions, epsilon)\n",
    "\n",
    "plt.figure(figsize=(12,6), dpi=80, facecolor='w', edgecolor='k')\n",
    "plt.plot([1.0 for _ in range(num_steps)], linestyle='--')\n",
    "plt.plot([0.0 for _ in range(num_steps)], linestyle='--')\n",
    "plt.plot(balanced_eps_optimality)\n",
    "plt.plot(greedy_optimality)\n",
    "plt.legend([\"Best Possible\", \"Worst Possible\", \"0.1 epsilon\", \"Greedy\"])\n",
    "plt.title(\"Average Optimality of Greedy Agent vs Epsilon-Greedy Agent\")\n",
    "plt.ylabel(\"% of Optimal Action Taken\")\n",
    "plt.xlabel(\"Steps\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After 1000 steps, the Epsilon-Greedy Agent is capable of reaching an 80% average optimality. This means that on average, they choose the best possible action 80% of the times after that amount of iterations. We can also see how our agent is capable of improving from experience, contrary to the Greedy Agent."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Values for epsilon"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On the previous example, we used a value for epsilon of `0.1`. This means that the agent chose an exploratory action 10% of the times. What would have happened if we chose another action? Like with almost all hyper-parameters in Machine Learning, there's a range of values which behave better than others. Let's make an actual experiment using different values for epsilon"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support. ' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            if (mpl.ratio != 1) {\n",
       "                fig.send_message(\"set_dpi_ratio\", {'dpi_ratio': mpl.ratio});\n",
       "            }\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        fig.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var backingStore = this.context.backingStorePixelRatio ||\n",
       "\tthis.context.webkitBackingStorePixelRatio ||\n",
       "\tthis.context.mozBackingStorePixelRatio ||\n",
       "\tthis.context.msBackingStorePixelRatio ||\n",
       "\tthis.context.oBackingStorePixelRatio ||\n",
       "\tthis.context.backingStorePixelRatio || 1;\n",
       "\n",
       "    mpl.ratio = (window.devicePixelRatio || 1) / backingStore;\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width * mpl.ratio);\n",
       "        canvas.attr('height', height * mpl.ratio);\n",
       "        canvas.attr('style', 'width: ' + width + 'px; height: ' + height + 'px;');\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>');\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option);\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'] / mpl.ratio;\n",
       "    var y0 = (fig.canvas.height - msg['y0']) / mpl.ratio;\n",
       "    var x1 = msg['x1'] / mpl.ratio;\n",
       "    var y1 = (fig.canvas.height - msg['y1']) / mpl.ratio;\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width / mpl.ratio, fig.canvas.height / mpl.ratio);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x * mpl.ratio;\n",
       "    var y = canvas_pos.y * mpl.ratio;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overridden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    var width = fig.canvas.width/mpl.ratio\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\" width=\"' + width + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var width = this.canvas.width/mpl.ratio\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\" width=\"' + width + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>');\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        event.shiftKey = false;\n",
       "        // Send a \"J\" for go to next cell\n",
       "        event.which = 74;\n",
       "        event.keyCode = 74;\n",
       "        manager.command_mode();\n",
       "        manager.handle_keydown(event);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"958.8889142907702\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1000/1000 [00:00<00:00, 1266.63it/s]\n",
      "100%|██████████| 1000/1000 [00:00<00:00, 1257.61it/s]\n",
      "100%|██████████| 1000/1000 [00:00<00:00, 1296.82it/s]\n",
      "100%|██████████| 1000/1000 [00:00<00:00, 1288.89it/s]\n"
     ]
    }
   ],
   "source": [
    "num_experiments = 1000\n",
    "num_steps = 1000\n",
    "num_actions = 10\n",
    "epsilons = np.array([0.0, 0.01, 0.1, 0.4])\n",
    "\n",
    "plt.figure(figsize=(12,6), dpi=80, facecolor='w', edgecolor='k')\n",
    "plt.plot([1.6 for _ in range(num_steps)], linestyle='--')\n",
    "\n",
    "\n",
    "for epsilon in epsilons:\n",
    "    _, reward = run_experiment(num_experiments, num_steps, num_actions, epsilon)\n",
    "    \n",
    "    plt.plot(reward)\n",
    "    \n",
    "plt.legend([\"Best Possible\"] + epsilons.tolist())\n",
    "plt.title(\"Average reward for multiple values of epsilon\")\n",
    "plt.xlabel(\"Steps\")\n",
    "plt.ylabel(\"Average Reward\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, we're plotting the average amount of reward received by the agent. This is because plotting the optimality on this many tests gets too crowded and noisy. As can be seen, the agent with an epsilon of `0.1` was the one that received the most amount of reward. Values of epsilon that are too small take very few exploratory actions, and therefore take a long time to explore and find the most optimal value. On the other hand, values of epsilon too high will obstruct the capacity of the agent of acting optimally. Too much exploration gets in the way of exploitation."
   ]
  },
  {
   "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
}