notebook/Learning the IK of the arm of a simulated Poppy using Explauto.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Learning the IK of the arm of a [Poppy robot](http://www.poppy-project.org) using Explauto"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Explauto comes with a [PoppyEnvironment](http://flowersteam.github.io/explauto/explauto.environment.html#explauto.environment.poppy.poppy_env.PoppyEnvironment) which links a [Poppy robot](http://www.poppy-project.org) to the explauto framework. \n",
    "\n",
    "Those robots, also developed in the [Flowers team](https://flowers.inria.fr), were designed as an open-source platform for the creation, use and sharing of interactive 3D printed robots."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![Alt text](https://flowers.inria.fr/poppy/wp-content/uploads/banners/walk_detail_hands-1100x380.jpg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The [PoppyEnvironment](http://flowersteam.github.io/explauto/explauto.environment.html#explauto.environment.poppy.poppy_env.PoppyEnvironment) works with:\n",
    "* a physical Poppy robot (either a [Humanoid](https://github.com/poppy-project/poppy-humanoid), a [Torso](https://github.com/poppy-project/poppy-torso) or an [Ergo Jr](https://github.com/poppy-project/poppy-ergo-jr))\n",
    "* or with their simulated equivalent in [V-REP](http://www.coppeliarobotics.com)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To illustrate how explauto can be used with poppy robots, we will setup a simple experiment where we will learn the inverse model of one of the arm of the Poppy Humanoid."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This problem is called [Inverse Kinematics](https://en.wikipedia.org/wiki/Inverse_kinematics) (IK) and consists in computing the joint parameters (angles q) that provide a desired position *p* of the end-effector in the cartesian space.\n",
    "\n",
    "![IK example](https://studywolf.files.wordpress.com/2013/08/rotated_robot_coordinate_frames1.png)\n",
    "(image source [studywolf blog](https://studywolf.wordpress.com))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To be more concrete, here it means to learn to position the hand effector of the robot in the 3D space."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will learn this kinematics in simulation, using the V-REP robot simulator, and then we will show how this can be used and tested on a real robot."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**In the following, we assume that you have installed the required libraries to control a Poppy in V-REP. If this is not the case, please refer to the [official documentation](http://poppy-project.github.io/poppy-docs/) and in particular this [notebook](https://github.com/poppy-project/poppy-humanoid/blob/master/software/samples/notebooks/Controlling%20a%20Poppy%20humanoid%20in%20V-REP%20using%20pypot.ipynb).** You should make sure that you have installed the following python packages:\n",
    "\n",
    "* [pypot>=2.10](https://github.com/poppy-project/pypot)\n",
    "* [poppy-creature>=1.8](https://github.com/poppy-project/poppy-creature)\n",
    "* [poppy-humanoid>=1.1](https://github.com/poppy-project/poppy-humanoid)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You should also have installed:\n",
    "\n",
    "* [explauto](https://github.com/flowersteam/explauto)>=1.2."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Learn Poppy IK in simulation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setup your Poppy Humanoid in V-REP"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, you will need to launch V-REP. Once it is started, you can use the following code to instantiate a Poppy Humanoid in the simulator."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from poppy.creatures import PoppyHumanoid\n",
    "\n",
    "poppy = PoppyHumanoid(simulator='vrep')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You should see something like: ![Poppy Humanoid in V-REP](http://poppy-project.github.io/poppy-docs/img/humanoid/vrep.png)\n",
    "\n",
    "*Note: if you have a connection error or see a popup in V-REP, please refer to this [documentation](http://poppy-project.github.io/poppy-docs/#test-your-installation) on how to solve the problem.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setup the PoppyEnvironment in Explauto"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Assuming that you have explauto installed and setup and that you are familiar with Explauto basic functionalities (otherwise you should start by following this [introduction notebook](https://github.com/flowersteam/explauto/blob/master/notebook/full_tutorial.ipynb)), we will now show you how Explauto can be used with the PoppyEnvironment and more precisely the Humanoid you just created in V-REP."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from explauto.environment.poppy import PoppyEnvironment"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To instantiate a PoppyEnvironment in Explauto, you will need to define a few arguments:\n",
    "* the **poppy robot** we want to use\n",
    "* which **motors** to use: here we will use the left arm of the robot which can be directly access via the dedicated alias *l_arm*\n",
    "* the **move duration**, meaning how long will take each motor command you sent to the robot\n",
    "* a **tracker** used to determine the tracked object position which will be used as the sensori output. When using a V-REP simulated robot you can directly use the robot itself as its own tracker and retrieve any part of the robot 3D position. That is what we will do here.\n",
    "* the **tracked object** - we will track the left forearm of the robot as we want to learn its inverse kinematics\n",
    "* and finally, the usual Explauto environment configuration: the motor bounds **(m_mins, m_maxs)** and sensori bounds **(s_mins, s_maxs)**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# We will use the left arm of the robot\n",
    "motors = poppy.l_arm\n",
    "# The move duration (in s)\n",
    "move_duration = 2.0\n",
    "\n",
    "# As explained above, we will use the simulated robot as its own tracker\n",
    "# Through V-REP API, we can directly retrieve the 3D position of any part of the robot\n",
    "tracker = poppy\n",
    "\n",
    "# You can use any object present in the V-REP scene here\n",
    "tracked_obj = 'l_forearm_visual'\n",
    "\n",
    "# The motors bounds for each one of the 4 motors in the arm\n",
    "m_mins = [-15, 0, -90, -90]\n",
    "m_maxs = [90, 90, 90, 0]\n",
    "\n",
    "# The sensori bounds of the more or less reachable 3D space\n",
    "s_mins = [-0.3, -0.3, 0.4]\n",
    "s_maxs = [0.35, 0.3, 0.8]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have our config, we can create our PoppyEnvironment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "env = PoppyEnvironment(poppy, \n",
    "                       motors, move_duration,\n",
    "                       poppy, tracked_obj,\n",
    "                       m_mins, m_maxs, s_mins, s_maxs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Run a few motor commands"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that our environment is setup, we can use it to run a few motor commands and check that we obtain the 3D position of the forearm. Using the same procedure as describe in the [introduction notebook](https://github.com/flowersteam/explauto/blob/master/notebook/full_tutorial.ipynb) we can generate a few random motor commands (within the chosen motor bounds) and send them to the environment. We will also store the reached 3D positions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "reached_point = []\n",
    "\n",
    "for m in env.random_motors(n=10):\n",
    "    reached_point.append(env.update(m))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You should see the left arm of your robot moves into some various positions. You can plot the reached positions and check that everything works well:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/pierrerouanet/.pyenv/versions/anaconda-2.4.0/lib/python2.7/site-packages/matplotlib/collections.py:590: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  if self._edgecolors == str('face'):\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<mpl_toolkits.mplot3d.art3d.Path3DCollection at 0x109258dd0>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAV0AAADtCAYAAAAcNaZ2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJztnXt4FPW9/9+z990kEG4CIeGiIJdyEUgAq3jhgAgWDqCt\n1MvxUaBUD17QR9GeitjqEVotWrFKrVapSunxhr8WAiJiLSYBEcWKAsrFEAgCIclm75f5/RE/w3cn\nM7szs7Ozu8m8nodHk+zsfGd35j2f+Vw5nudhYmJiYmIMlmwvwMTExKQjYYquiYmJiYGYomtiYmJi\nIKbompiYmBiIKbomJiYmBmKKromJiYmB2FL83cwnMzExMVEPJ/cH09I1MTExMRBTdE1MTEwMxBRd\nExMTEwMxRdfExMTEQEzRNTExMTEQU3RNTExMDMQUXRMTExMDMUXXxMTExEBM0TUxMTExEFN0TUxM\nTAzEFF0TExMTAzFF18TExMRATNE1SZtYLIZIJAJz3p6JSWpSdRkzMZGE53nwPI9IJIJwOIxoNAqO\na22sZLVaYbfbYbVaYbFYYLFYhL+ZmHR0TNE1UQUrtj6fDxaLBTabDRzHwWKxIBQKIRqNIhaLJWxn\nsVhgtVqFf6YYm3RUuBSPhObzogmARLGNx+MAAL/fj3g8jlgsBp7nBQHlOA52u10QVvF7sJhibNJO\nkT2BTdE1SQrP84jH44hGo4jH4+A4DvF4HKFQCMFgEFarFW63W7Bsw+GwIMDxeFz4fxJTElZWVNnX\nEaYYm+Q5puiaqENObIPBIMLhMBwOB4BWcbTb7YhGo4J7geM44e/0PuJ/PM8LQsr+I1Elq9gUY5M8\nRfaENH26JgnwPI9YLIZoNJpgsYZCIYTDYTidTnTu3BkWiwWBQKCNKNJ7EBzHCQIpfg0rwuS2kBJj\njuMSxJiCd+x+yZp2OByCMLPbmZjkCqbomgCQFlue5+H3+xGJRBLENhm0XSr0FuNoNIpgMJiQRQGg\njVVsirFJtjFFt4NDYuv3+wEADocjQWxdLhc8Hk9KsdULrWLMij0rquJAH/2XFWPWz2yKsUmmMUW3\ngyK2bKPRKAAIObculwsFBQUpRciogohUYhyJRAAgpWVMNw+prAsAphibZBxTdDsYJLCs4LCWo9vt\nRmFhoSKRyQUhIjGmoJvb7Qag3U1BYiyGRNhms0luZ2KiFFN0OwgktmTRchyHWCyGYDCIWCwmZCGQ\naOU7ySzjWCwmiDFlZwBnMyPEgkwCHovFEA6HE96PXm+z2UzL2EQRpui2c8RiC7T2SggEAojH43C5\nXCgsLEQwGNTNVZDLPRg4joPNlnjak5VMQky9JChVTi61jbYJBoMAALvdLuyDBNy0jE3EmKLbTpES\nW4rwkxvB4XCkJQJSmQr5KCokhuJgoVIxps8glWVsirEJYIpuu4PneYRCIUEYgFaxDQQCAACXyyUp\ntuTb1UJ7FQylYkw/+3w+RZaxnBizLgpTjNsvpui2E+hijkajaG5uhsfjAQBBbN1uN+x2u3kR64BY\njCkgabfb03JThEKhNvuRqr4zKn3PJDOYopvnUDCIjbhTni3HcarEVq0vVmkhREchXTeFlKjSjVS8\nH7GLgoo+THIfU3TzFCmxjUQiQmmuy+WCy+VSfCGaF2zmyKQYU1k0dXVjxViquZBJ9jFFN89gc06J\ncDiMYDAIi8WCgoIC+P1+ocet0ZjWr3JSiTGltlFam1THNqBVgF0ul/D/4lJuVozNJkHZxxTdPECq\nly1wVmytVisKCgqElCU2oq4UUyxzh2RiLFXwAbT67uU6tgFtxRiA8DpTjI3FFN0cRk5sqZetzWZD\nYWFhm7xTozCF2likCj6owMXpdCqqvpMTYxazfWZmMUU3B2GT7qPRqNCblhXboqIiWbHVKobpCGhH\nviBJ3LKFHh3bxKIajUaFfhaEKcb6YIpuDiFuHE4nPqUT2e12dOrUqc3FpQfmhZOfSPUzJowQY3oa\nczqdphgrxBTdHECul20kEkEkEgHHcarE1qjHftO9kJ/oKcYkwlar1bSMFWKKbhaRE9tgMIhQKCRE\nnAsLCzO+Fj0F1BTi/ESLGLNVj3KWsXjKR0cXY1N0s4CU2AKt03Vp/ljnzp2FxjRqMdoCZS+WjnLh\n5ArJ3At6kUyMw+Fwm9Q2ALKWMRsc7qhibIqugchZtoFAoM38MQCSfV1Nco+OatlTWhsV4wBtCz7E\nw021irG44COfxdgUXQPQOn9Mq8WqZTujtmmv5KsApIvY0taz+o4V43A43GZf4hzjfJl/Z4puBqH2\niqFQCKFQCAUFBarmj+WLqNHFlImsCpPkGOFe0INMinEwGBSmQH/22Wf4+uuvMW/evCwdaWpM0c0A\ncr1sW1paVM0f04qWNo1aBJ5OfK/Xi1gsJlxYFGzJ98fATNPU1ISqqo/h84UxcuRADBo0KNtLUk26\noq+HGLOlznV1dTh9+nS6h5VRzB5xOkI+KXYUeDweF6Y02Gw2FBcXw+12KzpRc9XSJQvD5/MJ/jyP\nxwO32y0UbMRiMYRCIfh8Pvh8PgQCAYRCoYQ+Ah0Zr9eLJ554DX//uwsffliGlSu34dNPP1P9Pu31\nc2R7DDscDuEcKygoSOicR3GPQCCA6dOnY/Xq1Xj//ffxyiuvYPfu3ULf4srKSgwZMgSDBg3CihUr\nJPe5bds2jB49GsOHD8dll10m/F7JtqqOLcWX1j6/UZ2Rmj9GUxpisRgcDgdCoRC6dOmi6n3j8Tga\nGxvRtWtXVduFQiFEIhFVqWZK9kXHSePanU4n/H4/unTpIgQ/6CZTUFAgbMM2b2Ebf8s9OmqxnChf\nNBsz3gKBAOx2u6py7OrqaqxZ40e/flcCALzeY3C53sYvfzlf1b5DoRA4jhOqFo2Gen9Q3w+joebx\nHo8HX3/9NdasWYP6+noAwBdffIG//OUvGDlyJAYPHowtW7agT58+qKiowNq1azF06FDhfRobG3HR\nRRdh06ZNKC0txalTp9C9e3fEYrGU28ogexKb7oU0YBuHE1Lzx2iag1qMfCxPtS9qG8mO+qHMi1Tv\nq7R5C1kt4ih1rrsotFib0WgMwFmhtFqdiETMbBWtWCwWnH/++fB4PLjlllswdepU4W9VVVUYOHAg\n+vfvDwCYO3cu1q9fnyCcr732Gq6++mqUlpYCALp37w4A2LFjR8pt1WKKrgakxDbV/LF0HgPV+s30\ndEtIiS2tJd1eDeLcTzV+vFyLVKtdx5Ahg+F2/xXffXcOnM5OaGjYhh//WP2FnO1AWq7t3+v1ori4\nOOE1dXV1KCsrE34uLS1FTU1NwmsOHDiASCSCyy+/HF6vF3feeSduvPFGRduqxRRdFUg1Dlcyf4zQ\nIp5GwQopuUf8fn/SIZZ6+5yVBFXoZkcpRGxzb3pNLomxHN27d8fixbOwYcNHaGkJYdq0gbj44gma\n3ivbopdNxNdUc3MzOnfunPAaJZ9PJBLBJ598gvfeew9+vx8XXnghJkyYkJHP1hRdBZDYBgIBRKNR\neDyeBLFNNRIn3eiukZYu64t2uVxwOp1ZFzBWjFm/KWsVU4COJmfkg4uitLQUP/vZT7K9jLTJpc+1\nqampTeykT58+qK2tFX6ura0V3AhEWVkZunfvDrfbDbfbjUsuuQSfffYZSktLU26rFjN7QQa6oCnH\nlqzbWCwGr9eLQCAAt9uNTp06KRplbmShgxbIVdLS0gK73Y7OnTurGvfDYpT1Qy4Ku90Ou90uNHP3\neDzCdyKXRUE+5GxbaumSa4/32d5/c3NzG/dCeXk5Dhw4gMOHDyMcDmPdunWYOXNmwmv+8z//E//6\n178Qi8Xg9/tRU1ODYcOGKdpWLaalK0KucTj5NgEIUxqy5WfVc1+sBQ9AdTczIFFkc8HqEc8TA9r6\ni9nUNT2zKEyMRSy6NC+OxWazYdWqVZg6dSpisRjmzZuHoUOHYvXq1QCAhQsXYsiQIbjyyisxcuRI\nWCwWLFiwAMOGDQMAyW3TwUwZ+x5xL1uCnT9mt9sRDofb+IyU0NTUhIKCAtVTHrRsF41G4fP5kq6T\nsiyoMs7lcqGxsVGyHDkZDQ0NKC4uFlpQUgqPEZ3RWLSmjImzKORS2pINefT5fHC73VlpZK4lXU1P\nWlpaMlrokwp6YnG5XOB5HtOnT8eHH36YCzdMM2VMDrHY0pclNX+MAjjZWKMaklm6YrFlL5hcLcbI\nJMk6aLG5xcmatmTzM8v24322kTr+XP88OqzoyoltMBiUnT+Wjiil49PVg2Ria9IWjuPaWI9SKW10\nEw4EApKBu/b8GefCDZoV3VxYjxI6nOjK9bJlxVZu/li2RDcdS5cGF0q1jtQDLdkV+YpcSltLSwuc\nTqcgyMkmLlB+cXsiV44nGAwKLSZzmQ4juqzYNjc3o6CgABaLRYhmK50/lm6Rg1GQb1WN2OrpXugo\nQgwgweXAwuYWK+0rq5Zsfs658B3TjQ1ojX9oibcYTbsXXbletsFgUIh0Ko3YiwsI1KD15FQrhJSr\nStvqbdmmItsXYS6htepO3Kjb/EzlYa9FqcKIXKTdiq6cG4FKWnmeVz1ZV48ih0xtRyPbQ6GQ0PxE\naTezdNaYzo2ovaDm2NX0okiW0paNTAkxufCds2tobGxsk6Obi7Q70aVOWJT4TkJCI3EcDgdsNpsw\nMlotWn2YmcoMEIstWbZaGuxoJV8CGLlMsiwKVozD4XCCVUzne7KUto6CaekajFR7RVZsWb9mS0uL\n4QExWqNe+xOLrVqrXc81stt25Is+E6RyUcRiMeGJTuwvZoU4E99LLnzf7BpMn65BSIltPB5POn8s\nX7IQCHY7Kk0OBoNJ/dFaezaopSPn+GYLElH67Klcm7WKa2tr8dJL/4eGhhZcfvkYXHnlFe0ypU0s\nuqZ7IYOkEttk88eyJbpqR+iw+6PgXyqxZTFFsGPA3visVitOnz6N+fP/B83NP4bNVoLq6r/A7w9g\n7twfJ7goALQRYjUuimxbuuLzu7m5Gf369cvSapSTd6JLYtvS0gKr1QqHwyEk/iudP5auVWa0pRuJ\nRFSltdH+1KKXtdqRcndzkQ8++ACNjZeiR4/rAQCh0LlYs+Ye3HjjdcJrlGRR5GrvYjFs9oLa6SzZ\nIO9El/Vf8XzrrC4S28LCQkUnhlark7Y1Ap5vnTYRCARgsVhkCzZSvYdJZqHPOFcEiYQ0sfS/baly\nsiwK1l8s1buY/mm9hvRCqsOY6dPNABaLJSEVzO12KxZbIpd9uqzY2mw2uFwuYail2v0ZgenLzR6s\n6DQ0NOCpp17Gnj1H4PHYYLN9gYaGPrDZShAOv4zbb5+m6D1ZMZbrXSwer+T3+7PSu1gsuqZPN0PQ\nFFr6krWU/eWi6JLYUpMdsmzD4bDm9C89G+WY5DZPPPEivvrqApSU3IuWlsPo2fMJDBxYjWAQmDx5\nEubMSa8HrFQWBZ2XVqs1pYvCCDGW6qWbi+Sd6FKDbbaxuFpySXTJRUINU8RNdrSSK4+7JpknHA5j\n797j6NPnF+A4Dp06nQefbyzmzh2E8ePHZ3TfHJe6MVCmeheLLd2WlhYUFRWlfUyZJu9El/2SspFr\nq5c1yIqtxWIR2kfquT8jLF2O44RetkCr1UMXnUlmIdGx2+3weKwIBI7D4ylBPB5DLHYcRUVjMr5/\nuewgrVV34hLoVPtnX8PzvC656pkm70SXyKboppv6pURsWbRa1pmGUvai0agwkp2ePgKBgNlLwCA4\njsMdd1yD3/72STQ2jkIsdgSTJnVNe8JBKrTcoFP1Ltbqosinm3zeia4eyfjZ2jYej6O5uRkAhGkQ\nqQRIq0BpXaeSbdjevBzHCVMT6ELx+/1wOBxC0FPKqsn1oZFKyZXUuPHjx2Hlyt44fPgwOnc+HyNG\njDBkXXrsQ4mLgnoXs1Yxnat+vx9Op1O39WSa7HfN0ADdAdNJ+zLqzkiWbUtLCwAIwyyVzlgzcq2p\n1hOPx+Hz+dDc3AyLxYLi4mLZ/rD0HdntdjidTrjdbhQUFKCgoCDp0EhKASSBNpFHLPhlZWWYOHEi\nRo4cmZXsAT1hMygcDoeQf19QUCD0TSFRfuaZZ1BWVoaDBw9i4cKFWLVqFT7//HMAQGVlJYYMGYJB\ngwZhxYoVbfazbds2dO7cGaNHj8bo0aPx61//Wvhb//79MXLkSIwePRrjxo3T7djyztIl9BCjTDau\n4Xk+YZily+WCz+dTPdBSK3qKtVxTHXY/So9JzqpJ5utrL1axSfqwLgp6srr33ntxzTXX4Pbbb8eI\nESPw73//Gy6XC8OGDcOiRYuwZcsW9OnTBxUVFZg5c2Ybt8ull16Kd955R3Jf27ZtQ9euXXU9hg4p\numwgTm/RFYut2+0WhNbn82laq5GWLvv0oKX0WOt+1fr6xEJsWsXGk23XChvIs1gsKCkpwaJFi4S/\nV1VVYeDAgejfvz8AYO7cuVi/fn0b0U11PetN3roXiFxK/YpEIvB6vfD7/XC5XOjUqZPwKK11n5ku\nxpCCxLaxsRHRaBSdOnVCYWGhKsHVq5yYfbz0eDwoKCiAy+USHi/pBkcphKFQSEjczyUhrqurw9/+\n9je89dZbOHPmjC7vmW3Ryzapmt3U1dWhrKxM+Lm0tBR1dXUJr+E4Dh999BFGjRqF6dOnY+/evQl/\nmzx5MsrLy/H888/rtu68tnTTqfHXU8zowo/H43C73QlCq8c+AWMuMLIsm5qaFJceSx1TJtcpZxVH\nIhEhsMdaxbnQAHzfvn245ZZfwOebDI7z47nnXserrz6F7t27G7oOvcm26ItFV1wCrGRtY8aMQW1t\nLTweDzZu3IhZs2Zh//79AIDt27ejd+/eOHnyJKZMmYIhQ4Zg4sSJaa87ry3dbBc5RCIRNDc3w+fz\nCS0knU6nrieiEdkLFOwja7GgoEBTr4dsQjdhsVXMBl0ikQj8fr8QtNPDKlYiPE8++RKCwdvQrds9\n6Nr1QZw4MQmvvfZ/mvandt8dBSnR7dOnD2pra4Wfa2trUVpamvCaoqIieDweAMC0adMQiUTQ0NAA\nAOjduzcAoEePHpg9ezZ27Nihy1rzUnSJbIqu1+uFz+cTAktKxNZI/6xSotGo4BKx2+2w2Wyagn3Z\nvPg/+ugjLFhwN2677T7U1NQkrMlqtbbJoGD97DQt2efzwe/3C5OTqbeHHjQ0tMDh6Cv8bLH0w+nT\nLbq8d7bIhfOYvelIlQCXl5fjwIEDOHz4MMLhMNatW4eZMxPLoU+cOCEcy44dO8DzPLp27Qq/3w+v\n1wsA8Pl82Lx5M0aMGKHLuvPHlJHAaNGNRqNCgMxutxvWaEeLGyVVEQcdSywWE1wi1EIyn/jwww9x\nww13IRi8H0AcW7feinXr/oAJEyZIvl5ptZR4LE46PQQmTy7HM888D7t9KeJxP3j+NVx66Y3pHHbO\nkCvuhebm5jZWrM1mw6pVqzB16lTEYjHMmzcPQ4cOxerVqwEACxcuxOuvv45nn30WNpsNHo8Hf/3r\nXwEA9fX1mDNnDoDWa+X666/HFVdcocu6uRQikP3bmQTUZczr9cLpdAqDGNVATXOUNMwhgYpGo3C7\n3fD7/ejSpYvqE07rehsbG1FUVKQqkBUMBgVXAQtb2OByuYSpAwAEF4Oa+nW/3w+Oay1FjcVi4DgO\ngUBAsJozzbXX/gwffDADDsdcAEA4vAZXXLEVa9Y8k/Z7SyXoi8tWOa61DNrj8cieD7FYDE899Rze\neOM92O123Hrrj3Httdekvb5QKCS4VIyGpk6Lzy+j4HkePp9P6J29YsUKXHTRRZg2TVk3NQOQFYe8\ntnQzXSAhFiiybAOBQM7708THRxeJeF5cvhOLxQE4md84vv9d+iixiskN4fP52ljFVDhitVpx993/\njbvv/m9d1mVyFtbSzYdeukCeim6mA2lisRVPojA6jSud40xW2KDnfoCzj3tG+q4XLPgxdux4CJGI\nBTwfh8Pxa8yf37bySE/YDAo6Vrfb3UaMpZp/t5epvdk2OsT7z5e2jkCeii6RruiKreRUYpvufo3e\nLhqNoqmpKeOFDdkMqkydOhWrV8fxhz+8DAC4445Hcfnllxu6BrrRpLKK9Z7aK9flqyMgPudM0TWI\nVMGiVNvSF0cR7HA4nHSgpdS2asm0QPH82ckTAFSlfmVbQLUybdo0TJkyBZFIBG63O9vLSYC1iqmb\nHOsrZnOKWatYTYvDbJBtSxdIDOLly9QIIE9FVy/3AjVwUevnTMdi1YKS/bF5qBaLBW63G+FwOK9y\nbTsKrFUsNRInFotJtjjMdSE2ErHoh0IhTVNkskFeX5FaxS8ejyMUCiEajcJms2kKKhntm022jmg0\nCr/fD+Bsy0gaTW8EuWD1tAekqu3IKpZqkUl/B2C4GGf7O5faf76cg3kpulotXTaCb7fbYbFYhGoU\nLfvXsl06DdDFkNjG43F4PJ6EogYtAq91G5PMkcxXTI3ilVjF7fl7yjeXWF6KLpDYeyEVUhF8qirT\num8jfbri/bGFDS6XS/fS43TJV9+wFrJ1nHT+22w2wUUhLvDIZItM09LVTt6KLpDackyWLpVOk2yj\nfbqEXN6wXmvsSGKpJ9m62MXCI+WeoNcpbZGZL1Yxe+zhcDgrBSJayWvRleujKhZbqXQpo0uI090n\nZSTkWmFDtoT6m2++wc6dO9GjRw/DU8TyDbKIWZJZxWIhlrKKs23pskg1u8ll8lZ02S+cTgCeV990\nW8vJo7dvVg7y24VCIVitVlVi256t1srKStx0022wWieB5/fi4osH4dVX9et32hGQs4pZIU7WIjPb\n5xabo9zU1IROnTpldT1qyFvRBc76tahBiRqxZX3CWkQ3kz5d8c2DfLaZtm7zRah/9rM7EQi8BeBi\nAGH8618TUFlZiSlTpmR7aYaSCWtTqt+wlFVMU5+lBNkIC5g99nzK0QXyWHRZgWhubtZUdZWNyrJk\nUF/bQCAAq9UqFDZQ8xoj1qgWo4U6Ho+jqek7AOO//40D8fhY1NfXG7aGjoaUVUytQMnokZrWK+5B\noSfsOWe6FwwiFArB6/WC53mhWbVajBYMuf2JCxsKCgqE6iX2NVrQYg2p3YaKTGKxWIKVRMEaPS84\ni8WCH/ygAnv3rkA8/j8A9gH4O8aOvV63faghl3ybRsJmT7BkskWm1BqA/CoBBvJYdGmcDAmVVrLZ\nQ0Fc2CDOtWW307KvTG9D5dPUk9dmswk3EPJHiwMzZPWkc7GtW/cCZs26AQcPPgqbzYYnnvgtRo0a\nhUgkovk985FsCr7cvpMVeLDNgMRlz2qtYrF7wbR0DcDhcCAajeZdFgLBFjYkm6um1/70RFxkYrPZ\n4HQ6EQ6HhSALz/NwOp0JVo9U5y0tpa1lZWXYtesDeL1eeDweWK1WQyvwTNSRrMBDiVUsdaNmRbe5\nuRklJSWGHlM65K3oEvkkurSd1+sVGqJnsrBBS6Aw2TZsgI/ynil3WI5kgRmpHgNSVrEcapqttzey\nfRPWw8pWahVL+YrZNeSbeyE3kj01wJa7ZrKRuV7bxWIx+Hw+AK1jRIqLixOmNmRinXpBncuampqE\nkewFBQWa3Tp0sSUbrR4Oh+Hz+RKGSLKPpSattDd/Mlm5NpsNDodDmGvHDhml6/3UqVMYMWIEtm3b\nhtdeew2vv/46Dhw4AJ7nUVlZiSFDhmDQoEFYsaJtf+Vt27ahc+fOGD16NEaPHo1HHnlE+FuqbdMl\n7y3ddHIGjRBd8cQGAKqtWyNvDuJtKMAHQDLApxeprB5xC0RxAr8pxMZitD+ZPT+sVitisRi6deuG\nt956C7/61a/gdrvxyiuv4JFHHsHHH3+MRYsWYcuWLejTpw8qKiowc+ZMDB06NOE9L730UrzzzjsJ\nv4vFYoq2TYe8Fd1sWrpEshOPLWxgS5DzZfCj1OBKucCJ+DNM5zsRvw9r+RDiBH56/ARaZ8MZOaGh\no2YvZBP6rq1WKwYOHIhwOIxly5ahe/fuAICqqioMHDgQ/fv3BwDMnTsX69evbyOcUtf+jh07FG2b\nDnnrXiCy5dOVg8S2sbER8Xi8zaO4kT0RtGzH87wwftputyseL0/7MwL28ZMmfDidTkFsqXUnuSfY\nsep63AxygWxnLmQb9thbWloSshfq6upQVlYm/FxaWoq6uro223/00UcYNWoUpk+fjr179yreNl3y\n1tIl0hXddPdL7yFX2KDXejN9orMpXhzH5VR/ByWQVWy32xMmNEhVUrHRca2jckxyp9FPLBZLuNaU\nrGvMmDGora2Fx+PBxo0bMWvWLOzfvz8j6xWTt6LLuheMtnTZbZUUNqS7T60nt5J9UZCMxqZbLBa4\nXK68Elw51PqJ9W5/2F7JtktFbOyI6dOnD2pra4Wfa2trUVpamvAaNvNl2rRpuO2229DQ0IDS0tKU\n26ZL3oouoK6nrtz26WwbiUQEH61cYYMeZCJ7QXyzIMu8ublZ1b7YtWX7YlSCnJ+Y56XbH+bqJN98\n+KyNhP0sysvLceDAARw+fBglJSVYt24d1q5dm/D6EydO4JxzzgHHcdixYwd4nkfXrl0VbZsueS26\nQHYsXfINBgIBeDyepIUNeu0TUH+hye2LCjN4ntf1ZpHPIsBxyUtapSb5kv+Ysify+fjVkO1jFVu6\n4rXYbDasWrUKU6dORSwWw7x58zB06FCsXr0aALBw4UK8/vrrePbZZ2Gz2eDxePDXv/416bZ6wqUQ\ngOx7zJMQiUQQjUbR2NiIrl27qt4+Go3C5/MpLiGMxWLw+/2IRqPCY7jang8tLS1C5zA1NDQ0oEuX\nLqpOdq/XC6fTKTR4Zpugy90smpub4Xa7FaeGxWIxeL1edO7cGeFwGBzHIRqNGj6Z16h9knuCrOJo\nNCpc+KmqqPSEiga0jJtKl1gshlAolJV9A619VziOg8PhQHNzM+bNm4fKysqsrCUJsl98u7B0Ae19\ncZVYnVITG6jQQct608lE0GLpsrnCFO2Xex+168t24YbRsO4J4GxfV5vNpnu5s5K1ZINsW7os+dZL\nF2gnoqtFkGjbZIIhLmwQR/RzXWwoo8Lv97cZV2SiH6yVyyJX7mxE68P2DN3MgPzrMAbkueimm8HA\nZiCwJ7229ju2AAAgAElEQVRcYYPUvrXuM5PbkdhGIhFYrVZVfYY7muWaSeSyJ5I1eVE6xTfbebq5\n4tPNtw5jQJ6LLqFXkYM4fSqZWBlZsMCuLxVs2S51AFPT2N0ks6Rb7pwr2RPZhL0O8m1qBJDnoqtX\nri754VIVNoi30yq6WuerJUOqbDdZ9690iUaj2LVrF5qbvejSpRhjx47N2L7aO2rLndmnLmocb6QQ\nZ9vSBRIbmJuWbhZIt8rL6/UqKmwQ79PIklK5Y4zH4/D7/ULkXjyWPRMlx7FYDE899Wd88okLFksp\nwuEtWLQohHHjyhW/R3shkwKUzE9MbolQKKSpLWY+I3Yv9OrVK8srUkeHFV16DOd5Hi6XS3GbxXT2\nmc52Yli/s9xY9kyJ3759+/Dpp0D//jcB4NDY+AO8/PKzqKgY224v9FyBdU9YLBY4nc6UfmK9y52z\nbemy+29ubsb555+ftbVoIa9FV4t7QTyxIRgMwmazZSzdTO/t1PidgcxYuqFQCBZLl+9fCzidneH1\nxtvMSDMxho5U7iw+N02fbpZQ8qjPFjawExvYtoBqMXo7qbLdZGTqQurfvz/c7kqcPPlvdOpUhqNH\nN+Pii/unXI+JfqSyNpWWO0tNZUhV7iz2K2cD1tI1RddA6IO3WCyyI8qlChvYk0mvzIdMbheNRoUZ\nYOR3Vvo+mXAvdOnSBfff/1OsWbMBJ096cdFF3XDLLXN134+J/qgtd5bKnsildDVTdLOElHCmKmxI\ntq3Wfeq9HXvDsFqtsNvtQkmv0n1lan39+/fH0qW3AWgtUfZ4PB1uGm97gXVPsG0xWauYbYvJNjjK\ndLlzKkz3QpYQnwji4YnJHoVyUXSlynYp91YtRmURhMPh7329FsHdk45FVF9fj5MnT6KoqAj9v+/i\nn4tky+rL9CO+uNyZ9kkBXOqxwbonjPATiz9vajqVT+S16LKBtHg8jmAwqDjAxL5HrhQ5JLthaLVa\ntWyjJhWO3B4UkAQguHp8Pp9kGlOqdX366R787W+fAxiAeHw/Lr30EKZNu1z1sSglHA7j9df/jt27\nD6FHjyLceOMM9O7dO2P7y1fou+O41mbxdH0Z2RZT6iaXbf+yWvJadAl69IlEIooCTCzpFCvo5Qum\nsl0qzpC6YaSbi6w3bH4w0OprJquHWh66XC7BX6jUKopEInjzzZ0455zr4XQWIRaL4sMP12LMmBPo\n2bNnRo7lT39aiy1brOja9XrU1dVi//7n8MQTd+dd0r1RiIUvlZ9YnE+sptw52b7zNRc8r0WX53mh\n6TbHcQnd4JWSbi6r1s5fRCan7Wbq8Y6scfKVNzU1IRgMJggnAMEHSJYRBWHYVCaxELcKuQ1OZ+t3\nabXawHGdM1ZdF41G8f77X6Cs7HFYLHYUFfVDXd0+HDhwAOXl5RnZpx5kO1c2FcnS2Fg/sVS5s1o/\ncS5/DlLkteiS0MbjcTQ3N2t+D60WK22r5UuPRCIIBoOIxWKKGolrtcj1ytPl+bOTJqhUmuM4oeyY\nMixojVarVfhs6LjYDBOyjtibTCwWg8vlQq9eFnz77U707Dkczc3H4XSeQNeuF2VEaFpTqjhEowE4\nHPbvbwh+M/0tCVq/h2R+YqVtMdl9R6PRvOwrkvdnFutX0nIyGF2ySvtqaWmRLNuVQ8s69RIo8aQJ\nch/QeshqoebsJMbsv3g8LtlTlr2RkEV8441X4c03t+LQoRp06+bB7NmXwmazwefz6d6f1mKx4Kc/\nvQwvvfQMHI6LEA5/i8GDm3SfFmAijZQQA/JtMYHW76ympgbfffddXrqA8l50AWMamae7LVu2C0BV\nu0WtaBVqVkzZvg4Oh0OwSIBWyzQYDAp9K9jjsdlskkn5YiFmBZQoKirCTTf9Z8Ka2O9VKmDD7kOt\nEM+YcSVKSnpg796D6NatCJMmLVI82SOb2QvZ2q8RyLknQqEQeJ7H/v378fLLL+PTTz9F3759ccEF\nF+C+++7DxRdfjMrKStx1112IxWKYP38+lixZIrmPnTt34sILL8S6detw9dVXA2hNhaRr0263Y8eO\nHbofW96LLlsgYXQWgpJtpcp2vV6voetUC90gWL8tCRqtIxgMCsEyJWXU5E5QI8SsC4c9dp7n2+yT\ngqlaGsBwHIfy8vKc9uHmGtkQfNqnzWbDjTfeiOHDh2PdunVYvHgxPv30U5xzzjmIxWJYtGgRtmzZ\ngj59+qCiogIzZ85s8+QSi8WwZMkSXHnllW32sW3bNk3jv5SS96JLZCP1K9m2rA9UXLZrVCaC2v3Q\nmqPRqOAvJ1cB/T0UCiESiQiz19K5+JQIMfmJxZkOYiGm92MtVKkGMB2lE1emyHYAj91/U1MTunTp\ngoEDB2LgwIEAgKqqKgwcOFDI7Z47dy7Wr1/fRnSffvppXHPNNdi5c6fkPjJJ3osum6ubjujq6ZqI\nRCIIBAKCD1Sqiixb/lk52EZAVqsVHo8nwW9L4+btdjsKCwszlhspJ8Ssfy8ajQpuBIvFIogrzSlj\nsVqtCTc7NnJOGRdSKWy5TrbFL1uw141UL926ujqUlZUJP5eWlqKmpqbNa9avX4+tW7di586dbdLf\nJk+eDKvVioULF2LBggW6H0Peiy6hZ96s1n1SUx22kbjU+2uN/GbC0hX7bS0WC3w+H4LBoJCBQJVm\nYr+tUUj598iCpbVZLJY2VrFUsA6gjAVpISYXSiZaIrYHckHsaf+NjY1tRFfJ2u666y4sX75c8olp\n+/bt6N27N06ePIkpU6ZgyJAhmDhxoq7r7/Ciy26r1dIVl+2mykgwOmNCCjawJ/bbOp1ORKNRIehH\ngkPjxrP9WE49KXieR0FBgSCgqSxiVjzFF5tYiMXBOnqNWIiz/T12NNjr1Ov14rzzzkv4e58+fVBb\nWyv8XFtbi9LS0oTX7Nq1C3PntjZoOnXqFDZu3Ai73Y6ZM2cKlYg9evTA7NmzsWPHDlN0xaTrXkhn\nW7aSTM203UxZrcnWyWZ40JptNlsbvy0rOC6XC3a7PcHPStYgiQ/7L9NCTJVNcj5luYi3lBCz/l2t\nQsy6XrQk9adDR8uakNq/lKVbXl6OAwcO4PDhwygpKcG6deuwdu3ahNccPHhQ+P+bb74ZM2bMwMyZ\nM4Wn1KKiIvh8PmzevBkPPfSQ7seQ96JLGCm6JFzhcBgWiyWn079YxNVv5A9V4rdNJmSZFmIK8AWD\nQdU+Zb2FmOM4QezpaQGAbJlztrtwtSfE57/X623TYcxms2HVqlWYOnUqYrEY5s2bh6FDh2L16tUA\ngIULF8q+f319PebMmQOg9fu8/vrrccUVV+h8FACX4kLO+Wcn1nKLx+MoKChQ/R7Nzc1wu92KSnBJ\nuFi/n9ouR9QIxuVyKd4mHo8L0Vo1nDlzBkVFRQgGg23ybclqiEajQlDJ7XZruoGIhYz+pSvEtDYA\nmtemdP1SKWwAJFPYgMSiEEphE1dYxWKxNkKcblEHz/Pw+XwoKCgwXMzpBqvm3NULOu7CwkIAwG23\n3YZf/vKXGDx4sOFrUYDsF9NuLN1kjcxTocSKZKftUtluKBTStE+j3AskAF6vVzbflm5WSvNtk60v\nlUWpxiKmrnHRaFRwc2RSYEhQLRaLZE9ZsRDT3202m/DEIMZms8HhcCQIsdS0Bq3VdR3Neha7NrQY\nIblAuxHdTLkXkk3bzUZusBKfGln/rCvBarUmCAZZvnrk28qhRYipyCUcDsPhcAg+52wgFmJyc1A3\nOCqHDofDAJDgmqDjkOo3QUIMoE2Zq95lznqTbZ8uSz6OXwfagehmKpAmFd0X+xGNjF4nO9HPnDmD\nr79ujdj27dtTePQrLCyEz+dDJBJJcCWEQiHYbLaM5tvKkUyIKQWMoKcLI4N1clCGCrmwxA1xpFwr\nlOkhLuqQEmK2ojJVX9qOiljwo9Gorl35jCLvRRdI7PildXvalvJSqWw3WUZCOpZuOj182RPvzJkz\nePvtXYjFzkMkEkF1dQ2uvrocvXr1Qjweh91uF8pjCbvdnlMnK8/zCR3XSNCMDNYlW1soFBIsb4/H\nI7k/EkX2c1UqxEBiBzYAbVw94vllAIS+F3o1CFdCNi1dcRYOkJ8ulnYhukD6li5ZWmqm7QLaSgb1\ntJC/+uowwuEBKC7uCbvdjuZmD/btO4oePXqA4zhBdAHA5XIlHCv9XixiRlm/qQRNTx+xFsjStlgs\nmp4KkgkxexxKOrABZ4WYUufIvaGl30Q+InXN5OPxtQvRTdfSJRGKRqOqGomnE3RKV6zJ79nU1Ayg\nE5NBwSEabRVT8tvKWWjiQBEFBlkXQCaEmOd5IStBjaDpHayTgw3iKc1qUYpUmbFUsE5OiOnv9LNc\nmTPbb0LPMmcKAGYLsaWbj7QL0QW09VCgst1oNCrk26qNHhv55dP+2LS1kSMH4ciRvWhocAPgEAh8\niYEDz0NLS0tKv61cxJ4VMVaIbTZb2tYktYOkjIl0BU1PIaYbWSgUMjSIR5+tkg5sBGVziPOIyX2h\npMxZLOpKjjVX3AvBYBButzsr60iXdiW6ShGX7TocDoTDYdUnk9HZCzzPC01pyALzeDy46qo4vvji\nMKLRGM47rxRdu3bVnNOaKWuSdSVkMmNC6zFwHIdIJAKO47LWY4KFFWL67OhmwLqIxD2JpVwTUkLM\nBuvypd9Eqmq0fKFdiC6bwUCPZVJQwEY8bZei+1r2a8R29LhLnbQKCwsT8m179uyJTp06CY120sm3\nlVuvViGmRjQ0LTgbGRPJjoGyOeizpBRBo0uc5WD9ykVFRYpdE1JFGKk6sEmVObOVdZSPnAtCnK/p\nYkA7EV2CTbthYavWrNa203bTdROofeRSuj9xJgVdJFL5tski65lAiRCTJQZAcE3QRZvtC1dcWkzV\nXUYH65KtLxAIpPQrK3VN0DEo7cDG/l08RJKCeNnoN0E3ncbGxjYlwPlCuxDdZLm6rP9TLkiWjsUq\nlcalB+y6yTr0+/1CdgX56mw2W048DgNnhZiqA3meh9PpFPrcUn8C9pGYbiRGFgHEYjGhZwKbosYe\nQ7ayJsQ3Ay1+ZTkhVtOBjS0T57iz/Sb8fr9QKCLXbyJTRR3sdWZaujkCK55sU+5U03bTTTfTs6SX\ngnvkKrDb7YK1QT9TW0OyHFtaWiSzDYy2JsWCIXYlUDN3sSVGQpzpYyD3ktpKPCOzJqgAQ3wzSBcl\nxyAWYnotna/RaFTYjhViNlgnLnPWU4jZa6apqckU3VyAvvyWlhahbJem06baLhtZCCzktw2FQnC5\nXCgoKEjw2wIQxFjci4AVMRI9NnjCXvyZEmK2MU0qyzvVIzH5WfUSYhIMameph185k1kTRrmJlByD\nuN8EuR5Yi5jFbrdntMyZXtvU1IRu3bqlc/hZo12ILhsEoIyE4uJiTRenURkMtD8Agt/W4XCgU6dO\nABIrlFJVRLEiRjPCxL7VTBVC6NWYJlNCnEnrUeoYtGRNkAWZC24i8TGQm8tmswmFNnKtMCmmIiXE\n7HUiJ8SpquvY69Pr9WLAgAEZ/jQyQ7sQ3Wg0iqamJlgsFrhcLtWtFtPxzWp1LwBnZ6mxflv2hE0n\n6i+uhhIXQsgJsdLMB9Y60+p7TEU6QsxxnNDzONMpaqmOQUqIKV+ZxIvSAXMla4IN5JF7Tvx3sWuC\nnbBBa5cTYrl+E1KDSEmI2eszXzuMAe1EdGkCgtbUL8DYnFsSO7/f38ZvSy4SelTXyzqjG0uqQggK\n1CW7+OlRnYKTRlpnUkIstiTJvcJxrWXQ7MWd7awJAAmuDo/HkyBMYj+30VkTwFljINnNVO5mIpU5\nAaTuwAZICzFb5gy0XjPPP/88Tp8+nRPfpRZyf+ypAuhCNDogpna7eDwOn8+H5uZmQbDEKWCBQAB+\nvx8Oh0Oym5We0IXjcDiEtpWdOnUShJ5cB83NzfB6vfD5fMJ/nU5nTjwOAxBuJKw163a74Xa7BYu3\npaVFWDsF09ipGUZAOcCBQABut1sQXKDtd1FUVJTwXVDGBX0Xfr8foVBImFun5/qCwSA8Ho/w+SmF\nvalTXKKoqAiFhYXCd0M3d3rSYjMkAAgBORJZq9UKp9MpdM6LRCKora1FVVUVpk+fjnPPPVeYBlFZ\nWYkhQ4Zg0KBBWLFihew6d+7cCZvNhjfeeEP4ndJt9SDvJ0cAieNzQqEQioqKVL+HmukRLEqmQLD5\ntg6HAy6XC16vNyHQFYlEBL+tkuCfkZD40iww+h3ldGYzYwJoG4iS+/zkAkR6+7ml1sdmdVDjIa3v\nRel34mIIrRaxnutTuj8lFrE4UByNRoXv9ic/+QnWrl2L06dPo66uDpdeeikGDx6MLVu2oE+fPqio\nqMDatWsxdOjQhH3HYjFMmTIFHo8HN998M66++mrEYjFF26qkfU+OSJanq+Y9MmHpijuXUdCPnbjL\nFhCQNZBNfx4LXYziiqhMZhuoQY2rQ6ufOx0hZnOC9XgySOYj1uKaMDLQyB6D2NVFaxELMVnCPM/j\n448/xjnnnIM9e/bgiy++gNPpxODBgzF48GBUVVVh4MCB6N+/PwBg7ty5WL9+fRvhfPrpp3HNNddg\n586dwu927NihaFu9aBeiC+jbU1ctUtuJx/vQ4zrb+Yn+nx4z5SLcZE0aXUCQrDGNkiAXm7rGHoce\nqWvk64tEIpqzJuT83GIhjkajCWKnRIjp6caoXhOpqtKkhJjW6HQ6Da1mlEN8U4xGo/D5fMIT1Vtv\nvYVNmzbh5MmTqKiowNKlS7F06VJ06dIFdXV1KCsrE96rtLQUNTU1Ce9fV1eH9evXY+vWrdi5c6dw\nvEq21ZN2I7pAbli6bDMdt9stmW9LUWGxWMhZL0ZakumIRaoglx6WpB4VW6mOQUnAkU2ZEudCp9uH\nV6/jkBNiOp/ovCX/di5kTdA6yZ1FBss//vEPfP755/jzn/+MsWPHYvfu3di1a5eQqaRkvXfddReW\nL18uXLPZCq62G9HNlqVL7gKx3zZVvm0qsUhlvZDwAPo8CmeigABIPVFBjRDr/aiuFDX5twRN58i2\n9SiGPnPyfdPvciFrAkgc0VRUVITm5mbcd999sFgs2Lx5s5AmNnnyZEyePFnYrk+fPqitrRV+rq2t\nRWlpacJ779q1C3PnzgUAnDp1Chs3boTdble0rZ60i0AacHY09JkzZ9ClSxfVJwiV1qrN8Q0Ggwll\nrBTxlcu3dblculo+qYJD7CO9HCRmPM8LXcqMhKwONjjEulfI9UJPB9nKuZWD1k59CaxWa8L3ki0B\nY2FvWKnafkoFuTItxKx1S+fgtm3bsGzZMvziF7/ArFmzku4rGo1i8ODBeO+991BSUoJx48YlDYbd\nfPPNmDFjBubMmaN6W4W070Aakc4JIBZKJbBtAQsLC4WKnUzm24qRCw6RgFECvtQFAxgzFTgV9JRC\nvRnoOMiCDIVCwrrIMs6VR2E2ECWV4peJHg1qYDM7lH7HWn3EWo9D7I4JBAJYsmQJTp8+jQ0bNqBH\njx4p38Nms2HVqlWYOnUqYrEY5s2bh6FDh2L16tUAIKSVqdk2U7QbS5dE5syZM0mHScpBQZnCwsKU\nr2X9tg6HA9FoFIWFhYLY0l073dJYvUhVPUSPwtlK+ZKCFTOyfOSOIxsBR6VpanLbshVYmbIk1Vi3\nWkjXImbjBy6XCzabDTU1NXjggQdw55134rrrrsuZ81EDHcPSBTJb5EBiSk3QqXF4KBQSOn1RDqXT\n6TRs3EsqWJ8k20mKLMtYLCZ0ZGMvFLa5iVEkC+SlSpcyKuCYrm+ZPY5MdF4Ti1mmbvrpWMQAEqzb\ncDiMhx56CPv378dbb72FPn366L7eXKHdWLp00TU1NUnWiqeCSh8pAMZCEXOqjRf7bdniDPbkVuNX\nzTRKGtOw0W0jiwcI9jHT7XZr3o+cBZbuDYX1OxrxBKPkOMRCTDfQdD9DPREfB1XRBQIBPPjgg+jX\nrx/eeecd/PznP8dtt92WE2vWgY5j6bK122qQs3QpQEJBNjm/Lc/zgk9PjV/VyMfgVClWalO+xKlS\nWmFvCHpM3011HFoyP+imbOTIISW50KxlD7SKLj0h5Ip40XFwHCdUNbrdbsEA2Lx5M3iex3333YeN\nGzfiH//4R7aXnFHajeimW5WWKt+WeoRSojwgn28rFxhi/XhGNO7WozGNVKBOr8CQ2C+aSXeM1tQ1\njuOEAgm9x7FrQUqI2W51NptNCD5m8txSg1Qw76uvvsLixYsxe/Zs/PGPf4TVakUwGMTRo0eTvldl\nZSXuuusuxGIxzJ8/H0uWLEn4+6uvvorf/OY34HkeRUVFePbZZzFy5EhF2xpFu3Ev0BwnJb0Q5LZv\nampCcXGx4Ld1Op0J/WkJciWk2ydB6rELSP9xXq8et0qRCwwlu+jZG0ImgjxaYKvRotFoQsCRhC4X\nMiYIcZqV+IYi7tHAVgca5bOnJjoAhJHpzzzzDDZu3IjnnntOVZZALJa6R0JVVRWGDRuGzp07o7Ky\nEsuWLUN1dbWibXWm47gX0i3nbWpqEhKz5fJtrVarbnX0qR7nybJmL/hk0WCjLEfxcSQLDIkfg0mk\nc/ExmM4fqjijgZXZSvmSgy0ikHJ3kGWvxcWihxCzlYP0PR86dAh33HEHJk2ahC1btqh+alDSI+HC\nCy8U/n/8+PGC5Wx0f4VktBvRTce9QH5bAJL9bcnVQH7dTBYPqH2cJzGmLAqLxZITLRfFNxS6IdBN\ny2azCU8MRgXqUpEsc0KcMZHKxZKpYZvJrNtU6FkdmAz2eqGb1gsvvIC//vWveOaZZzB69GjlB8yg\ntkfCCy+8gOnTp2vaNpO0G9El1BQ50KMPncDRaFSofkrltzWKVGlS5M8Dzl5UtH6j073kYFOsCgsL\nE45FzrLXM1CnBDX9ElJ9J+mmfMmhpLm4WvTsvMZat+R6O3bsGO644w5ccMEFeP/99wV3nRbUHO/7\n77+PF198Edu3b1e9baZpN6KrxtJl822dTic6d+4s+Lu8Xq9wYcRiMcmpttmGLnry15FVRj48vfsy\naIW1yuSqoTIZqFO7RirC0PK+ajMN1Agx66M3Ys4bx6nvvGaxWIQpwGTdrl27Fs8//zxWrlyJH/7w\nh2mvTWmPhD179mDBggWorKwUejUY3V8hGe0mkAacbSgTDAZl823D4bDgC5Prk0AWD/u3XHoEZhvT\nJOvlIK7eMsoXqWaNSt9PbaBOzRqNaNzN7lecC83zvKTPnlKsjF6jEtjvJBKJCEHglStX4osvvsDJ\nkydRUlKCZ555RrdCByU9Er799ltMmjQJr7zyCiZMmKBqW53pGIE0ukMrybcV97cV+6HkAhDUayEb\ngRR6TFfqW05lRbLVQnpNgKC8ZD2bYqsN1LGuCSkXSzYad7PHojQXmnA4HDnX5Ic+V8rwKCgogMVi\nwcCBA/Hvf/8b/fr1w4kTJzBkyBCsWbMGs2fPln2vVKlcX331FW6++Wbs3r0bN9xwQ0KPhGnTpgm5\nv71798bIkSNx5swZ3HrrrQBau73t2LHD8P4KyWhXli499ni9XhQXFwNI9NtScQOJj7hPgtKGIJmw\nvFLtL1ONaZJVPYmFONX7GNW0W45UHdeocXyujkUCEgOOFIyjY2FvPtmuchRb4I2Njbj33nvhcrmw\ncuVKdO7cGcDZtpFszjqLklSukydP4siRI3j77bfRpUsX3HPPPcLfBgwYgF27dqFr166ZPWD1dBxL\nl4RUzm/LWhFaR4insrzkfKpq/YVsYCJTlVBqq9CkLnh2pE82/d/JgkJUMACcDbaymRO5IL6sBS4O\nOIqfUrJV5Ujlu+RftlqteO+99/DrX/8aS5cuxY9+9CPJUng5lKRy9ejRAz169JCtVNOaIpot2pXo\nEpRvS6PZKShG0EmrZ3oVK15sQQX7+KtkvDm7Ror4G/0IrKZdJP2NCklyQbwIth8BVZRRmXYu5d2K\no/5So3PkMiaMDDqSi46uq5aWFvzP//wPfD4fNm7ciO7du6t+z3RTuTiOw+TJk2G1WrFw4UIsWLBA\n9RqMpl2JLrkWAKT020rN/dKbZD5VuVJgigLnSltIoG1ZM7kSQqFQQk19LpWeAskLCKTEi24qRpRo\nE2zFlloDIJkQ63ksUhke27dvxy9/+UvcfffduPbaazV/Lul+ntu3b0fv3r1x8uRJTJkyBUOGDMHE\niRPTes9M0+5E1+VywefzCdYtfam50qxbzi0hNRmYLh6taUyZgC3flXoEZoNb2UpbU9tAJ1XerdpA\nnRKk+hHo8R1rOZZkQiwenxMMBrF06VIcOXIE69evR+/evdNab7qpXLT/Hj16YPbs2dixY4cpukZC\nY81tNptkvm0uVGqJYf2iJGS0Zjm3RKaqnZKhpJ+D3mXNamEf09MtINDi61Z6U2FdR0ack6lyiOkz\no1x1Or8o55sC0Lt27cK9996Ln/3sZ1i5cqUuN8/y8nIcOHAAhw8fRklJCdatW4e1a9dKvlbsu/X7\n/YjFYigqKoLP58PmzZvx0EMPpb2mTNOushduueUWHD9+HGPGjEFhYSE+//xzPPbYY/B4PLL5kNks\nelDamEZJtkSmmpeIhSzdXFGxH5L+pTv9QZxOZ8TNVVw0QN+N3E0lFzI8ksHm3UYiEQDA5s2b8Ze/\n/AVOpxPHjx/H6tWrUVFRoet+N27cKKSMzZs3Dw888EDCmJ36+npUVFSgubkZFosFRUVF2Lt3L777\n7jvMmTMHQKtFfv311+OBBx7QdW1pIPvFtivR5XkeH330EW6//XYcPXoUl1xyCerq6jBo0CBUVFRg\nwoQJOO+88wCcHe+TDQtS3JhGSwBKScPxdN0SrEWWyU5g6TQdzzUhS3ZToZxoGk2TS4ILyI/PefLJ\nJ8FxHAKBAHbt2oWbbroJK1euTPpeanJvH3300YQ0sFxpwZgmHUN0AWDTpk3Yt28fbr31VqHh+L59\n+1BVVYXq6mrs3bsXTqcTY8aMQUVFBcaNG4fi4mLJi521uvSCTa9yuVy6CZm4LaHYglRzU8lkXrBS\nlDuhKOQAABcMSURBVNxUAAifZa5MSRBDn2U4HBZ8y5nO69ZCLJY4cSIej+Ppp5/Gli1b8Nxzz2Hw\n4MHC8fh8vqSzBNPJvVWybZ7QMfJ0AWDq1KmYOnWq8LPVasWwYcMwbNgwzJs3DzzPo6WlBR9//DGq\nqqrw2muv4cSJE+jbty/Ky8sxfvx4/OAHPwDHcbpOSqDMiVgsllaNvxxyNfOpItmsBSku381mzm0y\nnyqtkef5BJ94LuXcAokl5UVFRQmfpVTQkfWpsseS6bxb8Ty1r7/+GnfddRemTp2Kd999N+E7oLhD\nMtLJvc2lFoyZot2Jbio4jkNRUREuv/xyXH755QBaL+YjR46gqqoKb7zxBpYuXQqe5zFy5EiUl5dj\nwoQJ6Nmzp9AonSxIJaWz7Ektl4OZyWOVi2SLG+OwObe5MCFBCqrOikQiQk8HuZxbvcqatcA+KaQT\ndNQaqFMK6z4iIX3++efx+uuv4w9/+IMwcUEt6eTe5lILxkzR4URXCovFggEDBmDAgAG47rrrBKHc\nvXs3qqur8dBDD+HIkSPo3r07KioqMH78eFxwwQVC7blcahSd1FJ5otlCfLHH43HhpkCWVS5kS4hJ\n1i9BaXqUmrJmraRqLp6MZBV1dCx6FD9IpasdPXoUt99+O8aNG4etW7fKlu0qIZ3zJFeeUjKJKboS\ncBwHl8uFCy+8UOhEz/M8Tpw4gerqamzbtg2PP/44AoEAhgwZIrglBgwYAJ7ncebMGcEPSgJHF1Au\nnVSsf5l9/FXrlsgk4qBjqicFPcqata5Ta3PxZMci5zJKVoWW7CYpLsbgOA6vvPIKXnrpJTz55JMY\nP3582utOJ/c2l1owZgpTdBXCcRx69eqFWbNmYdasWQBarZovvvgCVVVV+P3vf48vv/wSjY2NOHny\nJFasWIGpU6fC5XJJClcmgnRKSVU8oMYtoadwidErn1VNWbMWC5KdFJzpEUmpih/kbpLkmmGt2xMn\nTmDx4sU499xzsXXrVmGGWbqkk3urZtt8pd1lL2SLQ4cOYeLEibjkkkswa9YsfPnll6ipqUFDQwMG\nDBggpKwNHjxYKH4giysThQJS6JGqxr4XK1zpZEtIvbc4uGOERa22cxzb/CXX/OBiNwv1u62pqcEH\nH3wAp9OJyspKrFy5Epdddpnun6/W3NvCwkLJbfOQjpMyli3i8Tg+++yzNvOf4vE4vvnmGyFl7fPP\nP4fVasWoUaME/3D37t0THhkzkUpkxPRdsVtCLgUvmVuC9Ymm2/w8XcQWJDutmdZqt9uFZvi5hriR\njt1ux86dO/HUU09h3759OHXqFBwOBxYvXox777036XspyZ294447sHHjRng8Hrz00kvCtdC/f390\n6tQJVqtV6G/bATBFN1fgeR5+vx+7du1CdXU1duzYgbq6OvTq1UvIGx45ciRsNpsuRQ9Gj2MXI2UN\nA23dEmxf41yzGlkocEo3R7ZXhhFPK0oRBx4tFgs2bdqExx57DA8//DCmTZsGADhy5AhCoZCQhyuF\nktzZDRs2YNWqVdiwYQNqampw5513orq6GgByuedtJsnfPN32VtnCca1jvS+55BJccsklAFqF6ejR\no6iursbGjRvx6KOPIhwOY/jw4ULKWmlpqZBpwGYXyJXN6tmHIN3jtdvtKf2pQKvv1el0Culr2RYu\nFvHnSUGoVIEtrWXN6UA+Zgo8er1ePPDAA4hEIti0aVOC+FE+bDKU5M6+8847uOmmmwC0jj5vbGzE\niRMn0LNnTwD51/M2k+S06MZiMSxatCjhDjtz5syEL7tbt254+umn8fbbb6veNlfgOA5lZWUoKyvD\nj3/8YwCtDdb37NmD6upqrFixAt988w2Ki4sxduxYjB8/HmPHjoXD4ZBMi6LCDhJ4I/oQKIVtE0mi\ny3Gc0IM4WdDRiGwJKZKlq2nt6pWJ7A96qonFYkL/iQ8//BAPPvgg7rvvPlxzzTWa9qckd1bqNXV1\ndejZsyc4Lv963maSnBbdjlzZ4nA4UF5ejvLycixatAg8z+P06dOoqalBVVUVVq1ahebmZqGvxPjx\n41FSUoLq6mqMHTtWsKzEPW5zwXoU54lKpYGx1nCm0ryUrDNVc3EpknX1ytS0ZnZ8TmFhIQKBAJYt\nW4Zjx47h73//u2BxakHpZyxnzf7rX/9CSUlJXvW8zSQ5LbpmZctZOI5D9+7dcdVVV+Gqq64CAKGv\nxEcffYRf/OIXqKqqQnl5OcaNGyf8t7i4WBhNk61cWxa2NDZZ8YBSt0SmHuNZ61aPp4VU+cNaCx/E\n43NsNht27NiBJUuW4L//+79xww03pB2MVJI7K37N0aNHhSnAJSUlAPKr520myWnRNStbkmO1tvaV\nOHjwIOrr67FhwwaMGTOmTV+JsrIyIUg3fPhwcBxnWK4toaQ0NhmsW4J9z2SP8VpyodUWY6RDqmnN\nqcqaxeNzwuEwHnnkEfz73//G//3f/6Fv3766rFNJ7uzMmTOxatUqzJ07F9XV1SguLkbPnj3ztudt\nJslp0TUrW5Qxffp0TJ06Vbh45fpKvPnmm3jooYeEvhJjx47FhAkT0KtXrwR/oN7WI1s8oGc5tNrq\ns1QNi/QqxtCKGv8wBfG++eYbnHPOOWhoaMDdd9+N66+/HsuXL9c11U5ufDmbdzt9+nRs2LABAwcO\nREFBAf785z8DAOrr69v0vL3iiit0W1s+ktMpY9FoFIMHD8Z7772HkpISjBs3TrbN27Jly1BUVCRk\nL6jZtiMh7itRXV2d0Fdi3LhxGD16NJxOZ5tcW7W9C9hHdOqsZjRKmqbTlIRc6cmbDLJuyfpdtmwZ\nXn31VQQCAVx00UWYNGkSfvKTnwh9o02yhuwJlP0OLElg77DDhg3DtddeK9xh6S5bX1+PsrIyrFy5\nEo888gj69u2LlpYW2W2VUllZiSFDhmDQoEFYsWJFm79/9dVXuPDCC+FyufDEE08k/K1///4YOXIk\nRo8ejXHjxqX3IegM21di8eLFWLduHaqqqvCnP/0Jo0aNwgcffIAbbrgBV111Fe677z68+eabOH78\nuGBFh8NheL1eNDc3w+/3IxQKIRqNJgRRSNhbWlqExi/ZEFw6Xqu1dSad2+1GYWEhOnXqJBSIRCIR\n+Hw+hEIhodF4NBoV8m9zBfLd+v1+4TiOHDmCzz77DEuWLMGePXswf/58nDp1CqdOnUr5fqnOb6C1\n2GHQoEEYNWoUdu/erWpbE3ly2tLNFuk0YQbaRzI421eiuroa+/fvR0FBAcaOHYuKigpUVFSgqKgo\nQaTIAqY5aNmybpUgNSVBbA0D6fdR1gM2+Oh2u8HzPFavXo3169fj2WefxfDhw1W9XzrFDu2oyXim\nyd/iiGyQTqoake/J4DabDaNGjcKoUaPw85//HDzPo6mpCTt27EBVVRVeeOGFhL4SF1xwAT777DOU\nl5fj/PPPFyrvjEzxUgo7JYH1Mavp5mVE0YNUD4ojR47gjjvuwMUXX4ytW7dqqtzTWuxQX1+PQ4cO\n5XUqZi5giq4E6aabtcdkcI7jUFxcjCuuuEIIhFBfiRdffBFz5szBgAEDsGnTJgwePFjIHe7Ro4ch\nKV5KUNNIR23Rg969esU3BgB4+eWX8corr+Cpp55KazhkOsUOx44da1epmNnAFF0J0hWB7du3o3fv\n3u0+GdxisaBv37744IMPsGbNGsyYMSOhr8T999+PY8eOoVevXkLe8KhRo2C1WhNEy4gJzUrzg5Oh\nd7aEFFI3hvr6etx5550YOnQotm7dCpfLpXrt4uNQuhYT/TFFV4J008169+4NoGMkgzudTmzfvl24\nkJP1laisrMT//u//JvSVGDduHPr16yf0ldBjMgJLJpqLs6Sba8siHp/DcZwwOufxxx/HxRdfrMtT\ngdZih9LSUkQikQ6Tipkpcjp7IVuwyeDhcBjr1q3DzJkzJV8rtgb8fj+8Xi8ACMngI0aMULX/VNHh\nV199FaNGjcLIkSNx0UUXYc+ePYq3zQSpJjlQT4knnngC77//Pt5//33Mnz8foVAIv/3tbzFt2jRc\nd911+P3vf4+dO3cKHbxIhJqbm9HS0oJAICBU1imxwqLRKLxeL3ieR2FhoSGdy5JlS1CQ0e/3C8dE\nk4KDwSB8Pp/QTKehoQE33XQTampqsGXLFkycOFE3N4yS83vmzJlYs2YNACQUO6i5NkykMbMXZNDa\nhPm7775rkwyupgmzkuhwVVUVhg0bhs6dO6OyshLLli3L68iyuK/Ezp07hb4SNAqJgnNKMgtyubk4\nwTZMD4fDAIB9+/bhd7/7HUpKSrBt2zb85je/wYwZMzLi8051fgPAokWLUFlZKRQ7jBkzRnZbkzaY\n/XTzhaqqKjz88MOorKwEACxfvhwAcP/990u+/syZMxgxYgSOHj2qettchvpKUMra3r174XQ6MWbM\nGKGIo0uXLm0KHsiatNvtWW+Cngy2mY7T6YTdbse3336LRx99FAcPHkQ4HMaXX36JyZMn45133lH8\nvg0NDbj22mtx5MgR9O/fH3/7299QXFzc5nVybU+XLVuGP/3pT+jRowcA4LHHHsOVV16pz0F3LMyU\nsXxBbebECy+8gOnTp2vaNpehvhLDhg3DvHnzwPM8vF4vPv74Y1RXV+O1115DfX09+vbti4qKCgwd\nOhT//Oc/ce2116Jv376Ca8GIIJ1aqFKP53kUFBTAYrFg27ZtWLZsGR544AHMnj0bHMchGAzi6NGj\nqt57+fLlmDJlCu677z6sWLECy5cvF26+RLK2pxzH4e6778bdd9+t5yGbMJiim2OoeZR8//338eKL\nL2L79u2qt803OI5Dp06dMGnSJEyaNAnA2b4STz75JH71q19h7Nix2LVrF4YOHSq4JUpKSmSDdEaP\nlhe3inQ6nfD7/XjwwQdx+vRpbNiwQbAwAcDlcmHgwIGq9vHOO+/ggw8+AADcdNNNuOyyy9qIbqo8\nXTNrIbOYoptjKM2c2LNnDxYsWIDKykp06dJF1bbtBYvFgq5du+KTTz7Bu+++iwkTJiT0lXj44YcT\n+kpUVFRgzJgxQpDOyAnN4laRFosF1dXVeOCBB3DnnXfiuuuu00X82WkNPXv2xIkTJ9q8JtUT0dNP\nP401a9agvLwcTzzxhKR7wkQ7pujmGEra6H377beYM2cOXnnllQRLqCOMrxbTuXNnfPjhh8LP1Ffi\nwgsvBNBqtdXX16O6uhr//Oc/8bvf/Q5+vx9DhgwRfMPnnnuu0NIxE+W/4vE5oVAIjz76KPbv34+3\n3npL6DurlClTpqC+vr7N7x999NGEn+V6JSc7lltvvRVLly4FADz44IO455578MILL6han0lyTNHN\nMZS00fvVr36FM2fO4NZbbwUAYcKq3LZqSDVX7tVXX8VvfvMb8DyPoqIiPPvssxg5ciSA3Jz6ynEc\nevfujdmzZ2P27NkAEvtKPP3009i/fz88Hg/Gjh2LcePGoaKiAp06dVKdZytGPD7HZrPh008/xT33\n3IObb74Zv/3tbzVZ1e+++67s33r27In6+nr06tULx48fxznnnNPmNcmeiNjXz58/HzNmzFC9PpPk\nmNkLJgLppKsB+dvoR9xXoqamJqGvxLhx4zB06FBYLJY2o9jlgnTs+ByXy4VoNIrHH38c1dXVeO65\n5zLWevG+++5Dt27dsGTJEixfvhyNjY1tfLrJ2p4eP35cKO5ZuXIldu7ciddeey0ja23nmCljJqlJ\nJ10NaBXdjz/+GN26dTNmwRkkHo/j66+/FkR4z549sFqtuOCCCxL6SlCXNTZIRz2IqUDiyy+/xF13\n3YU5c+bgjjvuyGhz9IaGBvzkJz/Bt99+m5AyduzYMSxYsEBo0CSXa/tf//Vf+PTTT8FxHAYMGIDV\nq1enNV+tA2OKrklqXn/9dWzatAnPP/88AOCVV15BTU0Nnn76acnXP/7449i/fz/++Mc/AgDOPfdc\ndO7cuV01+iGoaxr1laipqUFdXR169eolBOlisRhOnDiBK6+8Eo2NjSgvL8egQYNw6tQp3Hvvvbjm\nmmuEeWEm7R4zT9ckNemkqwHtu9EPjbOX6iuxbds2LFmyBN988w0uueQSVFVVoV+/fhg3bhyGDRuG\nHj16YPPmzXjsscdw8OBBuN1uxftVWuxwyy234B//+AfOOeccfP7556q3NzGO7GeKm+QMatPV3nnn\nHSFdDZBu9NOeob4SX3/9NUaMGIEjR47gzTffxPz581FfX4/Fixdj1apVeOihh/D3v/8dx44dUyW4\nwNlih/379+M//uM/2vhniZtvvllwC2nZ3sRAaLS1zD+TDkQkEuHPPfdc/tChQ3woFOJHjRrF7927\nN+E1R44c4c877zy+qqoq4fc+n49vbm7meZ7nW1pa+B/+8If8pk2bDFt7NolGoxl778GDB/P19fU8\nz/P88ePH+cGDB8u+9tChQ/zw4cM1b2+iK7K6alq6JgJKZtKx6WrsDLj6+npMnDgRF1xwAcaPH48f\n/ehHmqa+puqStn79eowaNQqjR49GRUVFgnsjW7O7MhkYU1LskMntTTJAMkXOyv3BpMMSjUb58847\njz906BAfDoclLe2Wlhbh//fs2cMPGTJE8ba5yuTJk/nhw4e3+bd+/Xq+uLg44bVdunSRfR8pS1fN\n9ia6IqurZiDNJGdQMruroKBA+P+WlhYhN1bJtrlKusUOyUh3exP9Md0LJjmD3FwuMW+//TaGDh2K\nH/3oR3jxxRdVbZtvzJw5Ey+//DKA1hlps2bNMnR7E/0xRdckZ1CasjZr1ix8+eWXePvtt/HLX/4y\nw6vKLvfffz/effddnH/++di6datQqHLs2DFcddVVwut++tOf4oc//CH279+PsrIy/PnPf066vUn2\nMN0LJjmD2i5pEydOxMGDB9HQ0IDS0tJ22WGta9eu2LJlS5vfl5SUCNVlAGQbG8ltb5I9TEvXJGdQ\nMn/rm2++Efq9fvLJJwiHw+jatWtezO5qaGjAlClTcP755+OKK65AY2Oj5OtuueUW9OzZs81svWXL\nlqG0tBSjR4/G6NGjJfNyTXIfU3RNcgYlKWtvvPEGRowYgdGjR2PRokVYt25d0m1ziXQLHWiqw+7d\nu7F7925zjE6eYvZeMDERkaq95fr167F06VKh3eOTTz6Jiy66CEDy9pZDhgzBBx98IGQUXHbZZfjq\nq68k13D48GHMmDEjoaT34YcfRmFhIe65554MHLWJzsgHKJLlk2Uht83EJKukkyvM8zzfv39//vTp\n05LvzebMxuPxNjm0LFI5t8uWLeP79evHjxw5kr/lllv4M2fOqDo2E0MxK9JMTJTA5vva7XYh35dF\nLlcYaK3Mu/jiizFixIiEf+KJvnJTHZJx66234tChQ/j000/Ru3dv0+LNU8zsBRMTBqUTld9++208\n8MAD+O6777Bhwwbh971794bT6ZRsb5luoYI51aF9kMqna2LSoeA47moAV/I8v+D7n28AMJ7n+dtl\nXj8RwFKe56d8/3NvnuePcxzXA8C7AG7nef7D7//2GwCneZ5fwXHc/QCKeZ6XTJzlOK4/gP/H8/wI\n5ne9eZ4//v3/LwZQwfP8dbocuIlhmO4FE5NE6gCUMT+XATgq9+LvBfVcjuO6fv/z8e//exLAWwDG\nMS9fDmAKx3H7AUz6/mdwHFfCcZyQdMtx3FoAHwE4n+O4Wo7jbv7+Tys4jtvDcdxnAC4FsDitIzXJ\nCqala2LCwHGcDcA+AP8B4BiAHQB+yvP8l8xrzgNwkOd5nuO4MQDW8zxfxnGcB4CV53kvx3EFADYD\neJjn+c3GH4lJrmL6dE1MGHiej3IctwjAJgBWAC/wPP8lx3ELv//7agBXA/gvjuMiAAIArv1+814A\n3vw+QGYD8KopuCZiTEvXxMTExEBMn66JiYmJgZiia2JiYmIg/x/OuqNxVFDcgAAAAABJRU5ErkJg\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x109133210>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%pylab inline\n",
    "\n",
    "ax = axes(projection='3d')\n",
    "x, y, z = array(reached_point).T\n",
    "ax.scatter3D(x, y, z)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setup and run an Explauto experiment with Poppy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we want to run the real experiment and really start learning the robot IK."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Setup the models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook, we will use a very simple setup: **random goal babling** as the interest model and **nearest neighbor** for the sensori-motor model. Of course, the idea is to show you how to setup a simple experiment with Poppy so you can freely design your own experiment with more complex and more interesting models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from explauto import InterestModel, SensorimotorModel\n",
    "\n",
    "sm_model = SensorimotorModel.from_configuration(env.conf, \n",
    "                                                'nearest_neighbor', \n",
    "                                                'default')\n",
    "\n",
    "im_model = InterestModel.from_configuration(env.conf, \n",
    "                                            env.conf.s_dims,\n",
    "                                            'random')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Setup the experiment"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The experiment with a Poppy (real or simulated) can be defined exactly as any other experiment in Explauto:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from explauto import Agent, Experiment\n",
    "\n",
    "agent = Agent(env.conf, sm_model, im_model)\n",
    "xp = Experiment(env, agent)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define the testcases"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will define a simple (and probably not so interesting for our task?) testcases. We will define a \"dense\" 3D grid of 3x3x3 points."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<mpl_toolkits.mplot3d.art3d.Path3DCollection at 0x1093dab90>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAV0AAADtCAYAAAAcNaZ2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x109133e90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n = 3j\n",
    "X, Y, Z = numpy.mgrid[s_mins[0] + 0.1: s_maxs[0]:n - 0.1, \n",
    "                      s_mins[1] + 0.1: s_maxs[1]:n - 0.1, \n",
    "                      s_mins[2] + 0.1: s_maxs[2]:n - 0.1]\n",
    "\n",
    "dense_testcase = numpy.column_stack((X.flat, Y.flat, Z.flat))\n",
    "\n",
    "ax = axes(projection='3d')\n",
    "ax.scatter3D(X, Y, Z)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For time saving purposes, we will run a rather short experiment (100 runs) and the results should thus be interpreted in that regard. Once again, feel free to play with the parameters of the experiment to make a more interesting one!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "xp.evaluate_at([1, 10, 25, 50, 100], dense_testcase)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Run!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have everything setup and ready, we can finally run the experiment.\n",
    "\n",
    "**Note that it will take about 5-10 minutes depending on how fast the simulation can run on your computer! So maybe you should first try with a shorter experiment.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "xp.run()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plotting results and interpretation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can of course use the classical plot functionalities of Explauto:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEZCAYAAAB4hzlwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x109219550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = axes()\n",
    "xp.log.plot_learning_curve(ax)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can also plot the \"chosen\" vs \"reached\" point (here projected on the x, y plan)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x109570110>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEKCAYAAADpfBXhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nyEP+f57OoSuJMGIdAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x109505910>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = axes()\n",
    "title(('Random goal babbling'))\n",
    "\n",
    "xp.log.scatter_plot(ax, (('sensori', [0, 1]),))\n",
    "xp.log.scatter_plot(ax, (('choice', [0, 1]),), marker='.', color='red')\n",
    "\n",
    "\n",
    "legend(['reached hand positions', 'chosen goals'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### Save the trained models so they can be reused later on"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can save explauto models (both interest or sensorimotor) and reload them later. This is particularly useful when:\n",
    "\n",
    "* you run a very long experiment and want to regularly save the results so you can restore it\n",
    "* want to transfer trained models to another environment (of course they need to be close enough so the models correspond)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, we will show you how to save the models of the experiment you just ran in simulation so they can be re-used on the real robot. To do that, we will use the [pickle](https://docs.python.org/2/library/pickle.html) python module."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "with open('simu-training.im', 'w') as f: \n",
    "    pickle.dump(im_model, f)\n",
    "    \n",
    "with open('simu-training.sm', 'w') as f:\n",
    "    pickle.dump(sm_model, f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The code above will save the trained models to your hard drive. You can store this file anywhere you want. But you will have to upload them to the real robot so they can be reused on this platform."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Reuse the learnt IK on the real robot"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Even if this section is written in the same notebook for consistency, this part is intended to be run on a real robot. Thus, you should start again this notebook from this point on the Jupyter server hosted on the robot. As pointed above, you should also upload your saved models on the robot, so they can be reused in the rest of this tutorial.**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A key feature of the PoppyEnvironment in Explauto (and of the Poppy ecosystem in general) is that it let you seamlessly switch from a simulated robot to a real one. This will let us reuse the IK we just learnt in simulation to test it on a real robot."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To do that, we will use the possibility to save and load models in Explauto. The idea is to train the models in simulation (what you did above) and create a new environment (based on the real robot) that will use the pretrained models.\n",
    "\n",
    "*Note: you will need to upload the saved models to the robot hard drive (using the upload features of the Notebook for instance, see the [dedicated section](http://docs.poppy-project.org/en/from-simulation-to-real-robot/using-jupyter.html) of the documentation for more details).*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load the previously saved models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To load the sensorimotor model we will simply use the pickle module once again:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "from explauto import SensorimotorModel\n",
    "\n",
    "with open('simu-training.sm') as f:\n",
    "    sm_model = pickle.load(f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*We assumed here that you upload the trained models in the same folder than the notebook. Of course, you can freely modify the path and filename to match the correct path.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we only need the sensorimotor model as we will not continue exploring but only exploit previous data. Though, the interest model could be loaded in exactly the same way. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can check that what you obtain are really the trained model. And that you can use it to perform inverse prediction. The inverse prediction here, $f: s \\rightarrow m$, corresponds here to computing the joints position to need a cartesian position: i.e. the inverse kinematics.\n",
    "\n",
    "For instance:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "sm_model.inverse_prediction([0.0, 0.0, 0.6])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Will give you the joint angles to send to the motor of the left arm to reach the (0, 0, 0.6) 3D position."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create a PoppyEnvironment based on the real robot"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will then create a new PoppyEnvironment, this time using a real robot. Of course to do that, you need to be connected on the real PoppyHumanoid. Otherwise, you will just get an error saying that pypot did not manage to \"ping\" the motors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from poppy.creatures import PoppyHumanoid\n",
    "\n",
    "real_poppy = PoppyHumanoid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You should then put the robot in a stable position, and make sure the motor of its left arm are stiff. **This will actually make the robot move to its standing position!**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "real_poppy.stand_position.start()\n",
    "\n",
    "for m in real_poppy.motors:\n",
    "    m.compliant = False"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Then, the environment can be created exactly like the one using V-REP. Note that we will not use tracker in this case and thus we will not be able to update the models as the sensori part will not give any information. We will directly use the bounds stored in the models to make sure we are coherent."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from explauto.environment.poppy import PoppyEnvironment\n",
    "\n",
    "real_env = PoppyEnvironment(real_poppy, \n",
    "                            real_poppy.l_arm, 2.0,\n",
    "                            None, None,\n",
    "                            sm_model.conf.m_mins, sm_model.conf.m_maxs, sm_model.conf.s_mins, sm_model.conf.s_maxs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Reuse of the trained models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As shown above, we can use our sensorimotor model to compute the inverse kinematics and send the results as a motor command to our environment (it will make the robot move):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "m_ag = sm_model.inverse_prediction([0.1, -0.3, 0.6])\n",
    "\n",
    "m_env = real_env.compute_motor_command(m_ag)\n",
    "real_env.compute_sensori_effect(m_env)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This can be easily wrapped in a function:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def goto(pos, duration=2.0):\n",
    "    m_ag = sm_model.inverse_prediction(pos)\n",
    "    m_env = real_env.compute_motor_command(m_ag)\n",
    "    \n",
    "    real_env.move_duration = duration\n",
    "    real_env.compute_sensori_effect(m_env)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Note that we directly change the move duration defined in the environment to get a finer control over the speed of the movement.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can then use this function to evaluate how accurate is the trained IK. It can also be used to follow a trajectory for instance a circle:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy\n",
    "\n",
    "r = .1\n",
    "x0, y0, z0 = (0.1, -0.3, 0.6)\n",
    "\n",
    "goto((x0, y0, z0), 1.)\n",
    "\n",
    "for alpha in numpy.arange(0, 2 * numpy.pi, .5):\n",
    "    x = r * numpy.cos(alpha) + x0\n",
    "    z = r * numpy.sin(alpha) + z0\n",
    "    \n",
    "    goto((x, y0, z), .5)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}