{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Boson Bunching\n",
    "\n",
    "A prominent example of boson bunching is the Hong-Ou-Mandel effect, where the bunching of two photons arises from a destructive quantum interference between the trajectories where they both either cross a beam splitter or are reflected. This effect takes its roots in the indistinguishability of identical photons. You can read more about it [[1]](#references).\n",
    "More generally one can send in $n$ photons into an optical interferometer via $m$ modes and one then measures in which modes the photons are ending up.\n",
    "For the example discussed here there are 7 photons sent into 7 modes and we are interested in the probability that all 7 photons end up in the first two modes. The interferometer is described by an unitary matrix and the incoming photons can be fully indistinguishable, fully distinguishable or partially distinguishable.\n",
    "\n",
    "## Introduction\n",
    "\n",
    "The paper [[2]](#references) shows that, unlike in the Hong-Ou-Mandel effect, bunching gets maximized with partially distinguishable photons. This disproofs the following conjecture:\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "### Conjecture 1\n",
    "(Generalized Bunching). Consider any input state of classically correlated photons. For any linear interferometer\n",
    "$\\hat{U}$ and any nontrivial subset $\\mathcal{K}$ of output modes, the probability that all photons are found in $\\mathcal{K}$ is maximal if the photons are (perfectly) indistinguishable. \n",
    "\n",
    "Evidence for this conjecture was given [[3]](#references)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The probability for such a case is given by:\n",
    "$$\n",
    "P_n(S) = perm(H \\odot S^T)\n",
    "$$\n",
    "where $n$ is the number of input photons, $S$ the distinguishability  matrix and $H$ a matrix which depends on $\\hat{U}$ and $\\mathcal{K}$. The symbol $\\odot$ is the elementwise multiplication.\n",
    "\n",
    "The distinguishability  matrix $S$ is defined by:\n",
    "$$ S_{ij} = <i|j> $$ \n",
    "where $<i|$ is the photon in the $i^{th}$ input mode $S_{ij} = 1$ if $\\text{photon}_i$ and $\\text{photon}_j$ are indistinguishable and $0$ if $\\text{photon}_i$ and $\\text{photon}_j$ are fully distinguishable. So for fully indistinguishable photons, $S$ is the matrix with ones in all entries.\n",
    "\n",
    "The matrix $H$ is defined by: \n",
    "$$\n",
    "H_{ij} = \\sum_{l\\in\\mathcal{K}} U^*_{li} U_{lj}\n",
    "$$\n",
    "with $i,j \\in \\{1, ... , n\\}$ where $n$ is the number of input photons.\n",
    "If we now compare the probability for fully indistinguishable photons to partially distinguishable photons we can rewrite conjecture 1 like this:\n",
    "$$\n",
    "perm(H) ≥ perm(H \\odot S^T)\n",
    "$$\n",
    "\n",
    "This version of the conjecture has been proven wrong with a 7 by 7 counterexample by [[4]](#references).\n",
    "\n",
    "In this notebook we are constructing an optical circuit based on this counter example and input the partially distinguishable photons to show that this indeed amplifies the bunching probability.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setting up the required packages."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Perceval\n",
    "from perceval import BasicState, Matrix, Simulator, SLOSBackend, Unitary\n",
    "# Other\n",
    "import numpy as np\n",
    "import math\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Constructing the unitary matrix\n",
    "\n",
    "The counter example to the conjecture is $H = A$ and $S^T = A^T$ with\n",
    "$$A = M^{\\dagger} M \\text{ with }\n",
    "M= \\frac{1}{\\sqrt{2}} \\begin{pmatrix}\n",
    "\\sqrt{2} & 0 & 1 & 1 & 1 & 1 & 1\\\\\n",
    "0 & \\sqrt{2} & 1 & \\omega & \\omega^2 & \\omega^3 & \\omega^4\n",
    "\\end{pmatrix} \n",
    "$$\n",
    "and $\\omega = exp(\\frac{2*\\pi *i}{5})$, the $5^{th}$ root of unity.\n",
    "\n",
    "Thus we have to construct the unitary circuit fom the definition of $H$. There are infinitly many such unitaries and we therefore generate a random one which returns $H=A$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# number of input photons and modes (works for n ≥ 4) See [2] for the generalisation of the above to n photons.\n",
    "n = 7\n",
    "# definition of w as the (n-2)th root of unity\n",
    "w = np.exp((2*math.pi*1j)/(n-2))\n",
    "\n",
    "\n",
    "def create_vectors(n: int, w: complex) -> tuple:\n",
    "    # first two orthonormal vectors from which the Unitary matrix for the circuit is built\n",
    "    v1 = np.zeros(n, dtype=np.cdouble)\n",
    "    v1[0] = 1\n",
    "    v1[1] = 0\n",
    "    for i in range(2, n):\n",
    "        v1[i] = 1/math.sqrt(2)\n",
    "\n",
    "    v2 = np.zeros(n, dtype=np.cdouble)\n",
    "    v2[0] = 0\n",
    "    v2[1] = 1\n",
    "    for i in range(2, n):\n",
    "        v2[i] = np.conj(w**(i-2))/math.sqrt(2)\n",
    "\n",
    "    return v1, v2  # these are the complex conjugate of the first two rows of the matrix M for n=7\n",
    "\n",
    "\n",
    "def normalize(vector: np.ndarray) -> np.ndarray:\n",
    "    return vector / np.linalg.norm(vector)\n",
    "\n",
    "\n",
    "def make_unitary(n: int, w: complex) -> np.matrix:\n",
    "    vector1, vector2 = create_vectors(n, w)\n",
    "    orthonormal_basis = [normalize(vector1), normalize(vector2)]\n",
    "    for _ in range(n-2):\n",
    "        # Create a random n-dimensional complex vector\n",
    "        random_vector = np.random.rand(n) + 1j*np.random.rand(n)\n",
    "        # Subtract projections onto existing basis vectors to make it orthogonal\n",
    "        for basis_vector in orthonormal_basis:\n",
    "            random_vector -= np.vdot(basis_vector,\n",
    "                                     random_vector) * basis_vector\n",
    "            # Normalize the orthogonal vector to make it orthonormal\n",
    "        orthonormal_basis.append(normalize(random_vector))\n",
    "        # Define the unitary matrix from the calculated vectors\n",
    "    mat = []\n",
    "    for i in range(n):\n",
    "        mat.append(orthonormal_basis[i])\n",
    "    unitary_matrix = np.matrix(mat)\n",
    "\n",
    "    return unitary_matrix\n",
    "\n",
    "\n",
    "unitary_matrix = make_unitary(n, w)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Building the Circuit\n",
    "\n",
    "Now that we have the unitary matrix we can use directly use it as a Perceval circuit. This gives a 7 modes interferometer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert Unitary matrix into a Perceval circuit\n",
    "circuit = Unitary(Matrix(unitary_matrix))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inputting photons of different degrees of distinguishability\n",
    "\n",
    "Now that we have the circuit which represents the correct unitary matrix and therefore the correct $H$ from the conjecture, we need to feed in the right photons following the distinguishably matrix $S=A$. In the [[2]](#references) the input photons are described via polarization. In this simulation it is not specified how the photons are distinguishable. The first two photons are fully distinguishable. The other five photons are each a linear combination of the first two photons and form an equally spaced star shape when drawn on the Bloch sphere. \n",
    "\n",
    "\n",
    "\n",
    "![Blochsphere](../_static/img/Blochsphere.png)\n",
    "\n",
    "Here one can see the two fully distinguishable photons with red arrows and the five linear combinations in the equator plane."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_inputs(n: int, w: complex) -> tuple:\n",
    "    # make a list of fully distinguishable photons of the form |{i}> that means the photon has a noise tag \"i\"\n",
    "    states_dist = []\n",
    "    for i in range(1, n+1):\n",
    "        x = \"|{\" + str(i) + \"}>\"\n",
    "        states_dist.append(BasicState(x))\n",
    "    # make a list of all input photons which are superpositions of |{1}> and |{2}>\n",
    "    # start with the pure states |{1}> and |{2}>\n",
    "    states_par_dist = [states_dist[0], states_dist[1]]\n",
    "    for i in range(2, n):\n",
    "        # add the states |{1}> + w^(i-2)|{2}>\n",
    "        x = states_par_dist[1] * w**(i-2) + states_par_dist[0]\n",
    "        states_par_dist.append(x)\n",
    "    # Input states\n",
    "    # initialise the variables\n",
    "    indistinguishable_photons = []\n",
    "    partially_distinguishable_photons = 1\n",
    "    distinguishable_photons = 1\n",
    "    # fill the states\n",
    "    for i in range(n):\n",
    "        indistinguishable_photons.append(1)  # gives the state |1,1, ... , 1>\n",
    "        partially_distinguishable_photons = partially_distinguishable_photons * \\\n",
    "            states_par_dist[i]\n",
    "        distinguishable_photons = distinguishable_photons * \\\n",
    "            states_dist[i]  # gives the state |{1},{2}, ... ,{n}>\n",
    "    return BasicState(indistinguishable_photons), partially_distinguishable_photons, distinguishable_photons\n",
    "\n",
    "\n",
    "# indistinguishable photons, partially distinguishable photons, fully distinguishable photons respectively\n",
    "indistinguishable_photons, partially_distinguishable_photons, distinguishable_photons = create_inputs(n, w)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To see if photon bunching gets amplified for partially distiguishable photons we need to compare it to the case of indistinguishable photons and for interest we are also comparing it to fully distinguishable photons. We are expecting that the probability for partially distinguishable photons will be highest followed by the indistinguishable photons and fully distinguishable photons having the lowest probability."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# simulating boson sampling\n",
    "simulator = Simulator(SLOSBackend())\n",
    "simulator.set_circuit(circuit)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The probability for all 7 indistinguishable photons ending up in the first two modes is: 0.699 %\n",
      "The probability for all 7 distinguishable photons ending up in the first two modes is: 0.016 %\n",
      "The probability for all 7 partially distinguishable photons ending up in the first two modes is: 0.751 %\n"
     ]
    }
   ],
   "source": [
    "def calc_prob(distinguishable_type, n: int, simulator) -> tuple:\n",
    "    def reverse(lst):\n",
    "        new_lst = lst[::-1]\n",
    "        return new_lst\n",
    "    # initialize the probabilities\n",
    "    probability_photons = 0\n",
    "    probability_distribution_photons = []\n",
    "    # summing over all cases where all photons end up in only two modes\n",
    "    # range over half the distribution because of symmetry\n",
    "    for i in range(math.ceil((n+1)/2)):\n",
    "        probability = simulator.probability(distinguishable_type, BasicState([i, n-i] + [0]*(n-2)))\n",
    "        probability_distribution_photons.append(probability)\n",
    "        probability_photons += probability\n",
    "    x = []\n",
    "    for i in range(math.ceil((n+1)/2), n+1):\n",
    "        x.append(probability_distribution_photons[i-math.ceil((n+1)/2)])\n",
    "        probability_photons += probability_distribution_photons[i-math.ceil((n+1)/2)]\n",
    "    probability_distribution_photons = probability_distribution_photons + reverse(x)\n",
    "    for i in range(n+1):\n",
    "        if probability_photons != 0:\n",
    "            probability_distribution_photons[i] = probability_distribution_photons[i] / probability_photons\n",
    "\n",
    "    return probability_distribution_photons, probability_photons\n",
    "\n",
    "\n",
    "# Save the Probabilities so we don't have to run the CalcProb function multiple times\n",
    "probability_distribution_indistinguishable, probability_indistinguishable = calc_prob(indistinguishable_photons, n, simulator)\n",
    "print(f\"The probability for all {n} indistinguishable photons ending up in the first two modes is: {probability_indistinguishable*100:.3f} %\")\n",
    "\n",
    "probability_distribution_distinguishable, probability_distinguishable = calc_prob(distinguishable_photons, n, simulator)  # high runtime\n",
    "print(f\"The probability for all {n} distinguishable photons ending up in the first two modes is: {probability_distinguishable*100:.3f} %\")\n",
    "\n",
    "probability_distribution_partially_distinguishable, probability_partially_distinguishable = calc_prob(partially_distinguishable_photons, n, simulator)  # high runtime\n",
    "print(f\"The probability for all {n} partially distinguishable photons ending up in the first two modes is: {probability_partially_distinguishable*100:.3f} %\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the results above we can clearly see that the implemented partial distinguishablity amplified the bunching probability from 0.699% to 0,751%. This however does not proof that our set up gives the maximum bunching probability. As it only disproofs the conjecture with a counter example."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plotting the results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first plot shows two-mode bunching probabilities for the three different scenarios. Here we can see the results match our expectations.\n",
    "\n",
    "The second plot shows the photon-number probability distribution for the first two output modes. The distributions are normalized, i.e., the probabilities are conditioned on events where all 7 photons end up in the first two modes (two-mode bunching events). This plot is interesting because it shows that the shape of the distribution of indistinguishable photons is very different to the shape of the distributions of partially and fully distinguishable photons."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkgAAAGzCAYAAADUo+joAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjAsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvlHJYcgAAAAlwSFlzAAAPYQAAD2EBqD+naQAAVmRJREFUeJzt3QmcTfX/x/EPwzC2IbKvSSFCZEt/1U8/lRb9qp+0kJS2nwgJhZR+ipQsJS20ifwq9fOTkrQhslUqUlmzkyWZGcv9P97fOrd7z9wZd8aMOzNez8fjGvfc7z33e8/5nnM/57udfIFAIGAAAAAIyv/XfwEAACAESAAAAD4ESAAAAD4ESAAAAD4ESAAAAD4ESAAAAD4ESAAAAD4ESAAAAD4ESAAAAD4ESIiJfPny2YMPPmg5yXnnnWf16tWzE8GkSZPcPli7dm266W666SYrVqyY5WT6Hv/617+y7Dvndjqu9D137NiRZeusXr26XXrppUdN9/HHH7vP1t/QMqT3Z/b413u1juPpt99+s1tuucXKly/v8tqzZ0/LC06kc1xWIECKkZ9++sluu+02O+WUU6xw4cJWokQJO+ecc+ypp56yAwcOxDp7edamTZvciXn58uUxy8PkyZNt1KhRmX7/77//7r5D6I8QkJvNnz/flendu3dbTvDvf//bBdR33HGHvfLKK3bjjTdm22cpYFcQltbj1ltvtZzq9zx+LioQ6wyciP73v//ZNddcY4UKFbJOnTq5iD4lJcU+//xzu/fee+3bb7+1CRMmWF6mILBAgQIxCZCGDBnirkobNmxosQqQVqxYkemrUp2U9B28K0JERz9y1157rTvukD3+7//+zx3b8fHxGTr+FSCpTKumqGTJkmFpV61aZfnzH99r+Y8++siaN29ugwcPzvbPOvnkk10Q5jdr1ix77bXX7O9//7vlVL/n8XMRAdJxtmbNGneSrlatmjsIK1SoEHztrrvush9//NEFUHnRkSNHXCCoGjM9gOMpLi7OPXKbQ4cOuWPnaEFHTqBAJppjOyPHfywC2m3btlndunWPyz4sWrSo3XDDDamWqwZLLQuXXXZZluUDGUMT23E2fPhw1779wgsvhAVHnlNPPdV69OgRdmA9/PDDVrNmTXeiUM3HgAEDLDk5OWIfAVV1NmnSxBISEqx+/frBqs+33nrLPdeJqXHjxrZs2bKIfU1+/vlna9u2rTtoK1asaA899JAFAoGwtI8//ri1bNnSSpcu7T5H6/vPf/6TZt8QXQWdccYZLv+6KorUB8HrN6EA0buKTExMtC5durirFP/V5913321lypSx4sWL2+WXX26//PLLUfs1aFucffbZ7v9ar1eFrRNRqO+++87OP/98K1KkiFWqVMntMz9tf11dan/pe1WpUsX69u2bar/46SpLAfC6deuCnx/aP0Mn5q5du1q5cuXcvmrQoIG99NJLYdXxuuIUXbl56/C+99dff+22n9d0qz4UN998s+3cudOOxdHKRaS+J15+/dvYK2vaZ+3bt3f/13fq06ePHT58OOz9+lFRs7NXdpXuoosussWLF6fK4/Tp011trPaHyptX1tLrg+QdN6q9bdq0qfsMbbuXX3451fq1bVu3bu3KfOXKlW3o0KE2ceLEDPXlOtp29LaXjjE1w3rHvcqk6KLq3HPPde/XMXLFFVfY999/H/Ez1Qfpn//8p/uR1bGq80pSUlJYGuX/ggsusLJly7rPUVDwzDPPpPk9PvjgA1fzqu2ktDqvhEqrHPiFlln9Vc251KhRI1imvW0aqQ+SmuJUA6vjTvnWcfjYY4+58hJqypQp7vyk84S2g8qRylNavPzrQlbHqT8vRzs+o9mH0di8ebPNnTvX/vGPfxw1mPTyPHXqVPfboGNe5UPnxQ0bNkR8TzTnuG3HeC6Ktrxm5Nw/e/Zsa9WqlUuj4+n000933zm7UIN0nP33v/91J2AFGNFQR0EVyquvvtp69+5tCxcutGHDhrlC9vbbb4elVQG77rrrXN8mXZHoANXVx/jx410huvPOO106vV8nTn/VtX6c9OOjqmUdMPqBURCgIE0nco9OMDr4rr/+elcjpJOQmgxnzJhh7dq1C8uTDpA33njDBUoKaPydNf2UL50klcelS5fa888/707eOvl5dBBpnWoyUV4/+eSTVJ8bSZ06ddz3GDRokHXr1s0duBK6L3799Ve3DXRiUl4U+N13333uxHrxxRe7NDoJ6/vrR1Xr0Xq/+eYbe/LJJ+2HH35wP9Rpuf/++23Pnj22ceNGl168TtAK/BRAaT9qe2k7TJs2zX1f/SDoB04nJP2AqW/ElVde6fIpZ555ZvAEoh9hnVx0ovSaa/X3iy++cCeijIq2XGR0nQoUmjVr5srphx9+aCNHjnQ/JvpuHp2gFdho2+tY0Gd+9tln7rvoQsCjfaEfa5Vx/RiOHj3arrrqKlu/fr0LDtKj7a3jS5/VuXNne/HFF9021w+rAi1RMKcfFG2//v37uxO+ymZGajcysh0VuCiYUfnSZ5x00kluG2k76PyhHxWVlzFjxri+izpW/MeWyq+W6VjS9tI2UfkODf5UlvQdVZ7V5KXzk7ahyrhqtEOtXr3aOnToYLfffrvbTsqjjnt9jwsvvNAyS2VYx83rr7/ujgmdJ8T78fXTj6YCVe0TneuqVq3qmui0XxRYeP37dCx07NjR/va3vwXPHzpvzps3L+wiNJSOZTV33XPPPS4I1jnXy0s0x+fR9mG0dE7VPtA5NlqPPPKIK586Xym40XZo06aN62+poD4j57isOBd9mInymt65X+cwXcxo/TpetE2VP+3PbBPAcbNnzx5dKgauuOKKqNIvX77cpb/lllvClvfp08ct/+ijj4LLqlWr5pbNnz8/uOz99993yxISEgLr1q0LLn/22Wfd8rlz5waXde7c2S3r3r17cNmRI0cC7dq1C8THxwe2b98eXP7777+H5SclJSVQr169wAUXXBC2XOvLnz9/4Ntvv0313fTa4MGDg8/1fy27+eabw9JdeeWVgdKlSwefL1myxKXr2bNnWLqbbrop1Toj+fLLL126iRMnpnqtdevW7rWXX345uCw5OTlQvnz5wFVXXRVc9sorr7jv9dlnn4W9f/z48e798+bNSzcP2qbaX36jRo1y73/11VfDtm2LFi0CxYoVC+zdu9ct075I67v69428/vrrLv2nn34aXKbvr2Vr1qxJN6/RlguVJX+ZEq3fv729dT700ENhaRs1ahRo3Lhx8LnKt9LdfffdqfKlPHiURnn58ccfg8u++uort3zMmDHpfmfvuAndNtu2bQsUKlQo0Lt37+Ayff98+fIFli1bFly2c+fOwEknnZSl29HbXiVKlHD5CNWwYcNA2bJl3eeGfk+VxU6dOqU6li6//PKw9995551uud6TXnlp27Zt4JRTTglb5m2nN998M+x8VqFCBbffPJHKgb67v7z7y++IESPS3I56r9bhefjhhwNFixYN/PDDD2Hp+vXrF4iLiwusX7/ePe/Ro4fbjocOHQpklD5T+yYzx2d6+zBaOg60bQ8fPnzUtN42r1SpUjAP8sYbb7jlTz31VIbPcVlxLmqYwfJ6tHP/k08+6dKF/hZlN5rYjqO9e/e6v7rCjcbMmTPd3169eoUt965q/H2VVOXdokWL4HNdnYuq0HWV5V+umga/0OHSXhOZaol0NeDxX42oRkS1MYr6/XSll5G2fF2dhtJ61TzkbTuv2cSrDfN0797dsoJqc0L7A6jPgJpeQreVrqR0pVm7dm3XjOE9tJ1FVeOZof2tWh9d9XoKFizomhPVLKuasqMJ3Te6elW+VGMhkfZPtKIpFxkVaV+Hbuc333zTfVakjrL+mjBdKav2yaOrTDWpRCrjfiqfXm2i6MpYVfeh71W507EV2rFfNQIZucLPyHZU7VdoDYpqRlQToCv40JoIfU/V3njnilD+GiDvGAlNG1pedByrvOiY1XfX81BqElRNgUfbV4NM1Fy/ZcsWO150/Gl/lSpVKuz4UxlQLd2nn37q0qkZZv/+/a4mKStk9Pj078NoqTZtyZIlrq9qRjqna1+E/raoVlTdOPxlI5pz3LGeizZnorwe7dzvdd5/5513UjWlZhcCpONIJxTZt29fVOnVT0UHiNrXQ6ngqrDo9VChQZCoHVfUTh9puYKbUPosVYeGOu2009zf0D4WakrTj67apVX4vapW/wlVVGWaEf7voJNgaF69beJfr38bZZaq1f0/vspD6LZSU4Oqe/W9Qx/etlL1dmbou9WqVSvVSVHBmPf60ezatctVf6vfgH78lC9vW0XaP9GItlxkhNefKL3trKkw9KMcTdOEv9xEWt+xvFfbPlIZy0i5y8h29Jdvb98rcPNT+VCAoGAglMpSKAWQykPoZ6l5QoGF10dE+8Tr0+EvL/qu/mPjWMtBZuj4U8DqP/70PUKPP11EKX9q5tFxrb54/n5p2Xl8ZvTc51GfTclo8O3f39pX2mf+fRPNOe5Yz0XrMlFej3buV/OumufU1K7zmwJIdbXIzmCJPkjHOUDSCV9DvDMi2n4jaY3QSWu5v/N1NNT/Q/0VNJz36aefdlcourJQe7uGr/uFXqFGIyvzmhnRfL4OSLXXP/HEExHT+gPS40nt+OqPoU6vqu3Q1aLyqz4H2XkiSauM+jtde7J6NNmxlJtYl7lIMnrcZGYfKQBV/xzVhKosq9yqNkFX9+oLdLyu0jNK+VIthAZFROIFbeq/olqM999/39577z330HlKNS3+jtU5aR/qPKrAQn3gskNOLO/R5EvbU7WDqqFX64mCXXVMV829Bg9kxwhVAqTjTJ3M1Gl2wYIFYc1hkWgqAJ0MdMXkRe6ydetW11FOr2clfZaqWb0TjFfdK16HOjV76OpfJ53QDqo68RwP3jbRKJPQKyZ11otGZjop++lK/KuvvnI/LplZX1rv0XfTSCl9v9Art5UrVwZfT+/9utKaM2eOG1GijugelZ/sLhfe1Z5/or9oar3S284qZ6oVy0gH1+ygbR+pjEVb7qLdjul9vmhghZ/Khzo2qxYolPZ7aC2G8qo8eJ+lDtkadfnuu++GXb2n1USs9+vHKrT8RZv/o8nIcaRyoWYer8YoPQr4NFBFD3131So9++yzNnDgwAzXOkd7fB4LDcLRds7M4Af/ca59pXV5naYz4ljPRdUyUV6jobzovKuHgnpN6KmBLyqz0ZSHDH9elq8R6dJVjwqGqgkV6Pjpqs4bhnrJJZe4v/5Zl72ai2hGbmXU2LFjww4wPVcNkQqkKErXQRFaM6Aq3PRGbmUljXwS1V6F0uiIaHgH5bHM2KtaGo2gee6551K9ppEa/qrjSHmI1Nyl/a2+HLoq8miEk76baoLUN0Q0NDfSd/CuoPxXgscya3e05UInRH2+1//D499PGaE+HPosbyK6WF7tqtzpoiZ0BnYFbl5zSFZtx7SoplY1gqr5CN3vqo3W1bN3rgg1bty4iMeIN1IpUnlRuUzrYkeTrIaOnFXfEI2IU77U7H8sMnJc6vjTvlDw7Kf365gR/9QW+nH1goWjTccRSbTH57HwauE1GjmjtC9Cu29odJr6Ann7OyOO9VxUIRPl9Wh0vPl5fQIzsz+jQQ3ScaarHx0Eak9VrVDoTNpqGvGGUormndBwWtU4qZCpUC5atMgVOs0fo2HHWUk1Q6q21GeqI7eqpFWVqT4JXn8RBWUK0NRko4NY7f06EetqTFcc2U3Vzvrh1I++ToDeMH/vSvZoV6La/uproakP1KFRJ2Z914z0F9D0Amr7VqdCXbmoXVwBo66MtFwn7tAh6JG+g0486nyveZl0wtEVroYD6+pW+1+dNHVVrpOc+ono+3odMFXVrI7FWodqI1S7ojKkh5o+NYT84MGDbn4TnYxU25bd5UL92jTkWydQ7QNtZ/VVy2x/LFH51rbW8HRdHXvNhGrm1WvR3H8tKy9sXn31Vde0o87O3jB/1bzoxB1NDUg02zE9I0aMcD92qnnWlATesGlt+0jzf2m/qzlc200BhfKvY1bnFdEMzV4Ni4bLq1ZGQb+apvTD6qeyps/98ssvXR8QTYegi7ysqD32mpNUG6C+JQoala9ItQxqPlatl2rjvekYdFGiqTZ0vOiCTTUUugjVvlETjPrdqDZT20s/qqE18tGK9vjMLJ1DdEzrnBY64CBaOg9ojiBN8aH9ojzpvJyZW5VkxbloRAbL69GoVk0XYPoN0gWZzi26ANO+1ffOFsdtvBzCaIjqrbfeGqhevbob5lu8ePHAOeec44YlJyUlBdMdPHgwMGTIkECNGjUCBQsWDFSpUiXQv3//sDRpDUsV7eK77rorbJk3DFVDaz0aRquhsz/99FPg73//e6BIkSKBcuXKuSGY/qGmL7zwQqBWrVpuKHTt2rXd8GlvqObRPvtow/z9QzgjDc3ev3+/W6+GWGvIafv27QOrVq1y6R599NHA0bzzzjuBunXrBgoUKBA2BF1DYM8444xU6SMNU9aQ18cee8yl13YoVaqUG5qrfaXhz+n57bffAtddd12gZMmS7vND171169ZAly5dAmXKlHHlon79+hGnJNB0Dvo8pQndlhs3bnTDY7XuxMTEwDXXXBPYtGlTqu2dkWH+0ZYL7TsNFVYabY/bbrstsGLFiojD/LVOv0hlSEO0VU5VzvRdTz755MDFF1/spns4WjnzDw9Pa5h/pONGZUGPUBrif+6557r9Xbly5cCwYcMCo0ePduvcsmVLlmzHSMdmqA8//NCdJzR1h4aRX3bZZYHvvvsu4nbU8quvvtqdW7Q//vWvfwUOHDgQlvbdd98NnHnmmYHChQu7c5HK9IsvvpjmdtLUIUrvHfvTpk0LW19mh/l7w/c1VF3DwEM/378fZd++fe48eOqpp7pyoeOlZcuWgccff9wdm/Kf//zHbWsNNVeaqlWrujK5efPmdPdV6Pf1i+b4PNo+TMusWbPc+1SmMsLb5prOQ9tE31flQ/kPnd4lo+e4rcd4LspoeT3auX/OnDluipyKFSu6z9Lfjh07ppruISvl0z/ZE3ohN9GVgq4QdBWZG6npo1GjRu4qOaOjP4DM0mzOutLWcZNeJ9Hcfnwh59JM2qpRVeuDhvYj69AHCbmOqmr9VO2rPgZqYgKOR7lTE69mXVb1fm68xxuA9NEHCbmO+tioXVxXTbo9gjeEV+3msRxij7xNfSl0+wX1X1EfD91PUR2VNSIKQN5DgIRcR/dO0+y4uomvmizUUVad/tTBE8guGnmjZjINmlCn7LPOOssFSdRaAnkTfZAAAAB86IMEAADgQ4AEAADgQx+kTNKEdZpZVhNmZcXtKwAAQPZTzyLNOq57o/pvyBuKACmTFBwxYgoAgNxpw4YNbibutBAgZZI31bo2cIkSJWKdHQAAEAVNz6EKjqPdHoYAKZO8ZjUFRwRIAADkLkfrHkMnbQAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAJ8C/gUAMK777FhnATF215gLY50FIKaoQQIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAPAhQAIAAMiJAdK4ceOsevXqVrhwYWvWrJktWrQo3fTTpk2z2rVru/T169e3mTNnhr0eCARs0KBBVqFCBUtISLA2bdrY6tWrg69//PHHli9fvoiPL7/8Mtu+JwAAyB1iHiBNnTrVevXqZYMHD7alS5dagwYNrG3btrZt27aI6efPn28dO3a0rl272rJly6x9+/busWLFimCa4cOH2+jRo238+PG2cOFCK1q0qFtnUlKSe71ly5a2efPmsMctt9xiNWrUsCZNmhy37w4AAHKmfAFVt8SQaozOPvtsGzt2rHt+5MgRq1KlinXv3t369euXKn2HDh1s//79NmPGjOCy5s2bW8OGDV1ApK9TsWJF6927t/Xp08e9vmfPHitXrpxNmjTJrr322lTrPHjwoFWqVMl95sCBA6PK9969ey0xMdGtu0SJEsewBYCcZ1z32bHOAmLsrjEXxjoLQLaI9vc7pjVIKSkptmTJEtcEFsxQ/vzu+YIFCyK+R8tD04tqh7z0a9assS1btoSl0YZQIJbWOt99913buXOndenSJc28Jicnu40a+gAAAHlTgVh++I4dO+zw4cOudieUnq9cuTLiexT8REqv5d7r3rK00vi98MILLsiqXLlymnkdNmyYDRkyxI6HjvPWH5fPQc71+jlVY50FADihxbwPUqxt3LjR3n//fdenKT39+/d31XHeY8OGDcctjwAA4AQKkMqUKWNxcXG2devWsOV6Xr58+Yjv0fL00nt/o13nxIkTrXTp0nb55Zenm9dChQq5tsrQBwAAyJtiGiDFx8db48aNbc6cOcFl6qSt5y1atIj4Hi0PTS+zZ88OptdINAVCoWnUX0ij2fzrVIduBUidOnWyggULZvG3AwAAuVVM+yCJhvh37tzZDa9v2rSpjRo1yo1S8zpMK3jRCDP1AZIePXpY69atbeTIkdauXTubMmWKLV682CZMmOBe11xGPXv2tKFDh1qtWrVcwKSRaRrZpukAQn300UeuU7eG+AMAAOSYAEnD9rdv3+4mdlQnag3XnzVrVrCT9fr1693INo/mMJo8ebI98MADNmDAABcETZ8+3erVqxdM07dvXxdkdevWzXbv3m2tWrVy69TEkv7O2VqfJp0EAADIMfMg5VbZOQ8So9gQ61FszIME5kFCXpUr5kECAADIiQiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAfAiQAAAAclqANG7cOKtevboVLlzYmjVrZosWLUo3/bRp06x27douff369W3mzJlhrwcCARs0aJBVqFDBEhISrE2bNrZ69epU6/nf//7nPk9pSpUqZe3bt8/y7wYAAHKnmAZIU6dOtV69etngwYNt6dKl1qBBA2vbtq1t27YtYvr58+dbx44drWvXrrZs2TIX1OixYsWKYJrhw4fb6NGjbfz48bZw4UIrWrSoW2dSUlIwzZtvvmk33nijdenSxb766iubN2+eXXfddcflOwMAgJwvX0BVLjGiGpyzzz7bxo4d654fOXLEqlSpYt27d7d+/fqlSt+hQwfbv3+/zZgxI7isefPm1rBhQxcQ6atUrFjRevfubX369HGv79mzx8qVK2eTJk2ya6+91g4dOuRqrIYMGeICrczau3evJSYmuvWXKFHCslLHeeuzdH3IfV4/p2pMP39c99kx/XzE3l1jLox1FoBsEe3vd8xqkFJSUmzJkiWuCSyYmfz53fMFCxZEfI+Wh6YX1Q556desWWNbtmwJS6ONoEDMS6Oaql9++cV9VqNGjVxT3MUXXxxWCxVJcnKy26ihDwAAkDfFLEDasWOHHT582NXuhNJzBTmRaHl66b2/6aX5+eef3d8HH3zQHnjgAVcbpT5I5513nu3atSvN/A4bNswFW95DNV0AACBvinkn7eNNzXhy//3321VXXWWNGze2iRMnWr58+VwH8LT079/fVcd5jw0bNhzHXAMAgBMiQCpTpozFxcXZ1q1bw5brefny5SO+R8vTS+/9TS+NmtSkbt26wdcLFSpkp5xyiq1fn3bfH6VRW2XoAwAA5E0xC5Di4+Nd7c2cOXPCanf0vEWLFhHfo+Wh6WX27NnB9DVq1HCBUGga9RXSaDYvjT5Twc6qVauCaQ4ePGhr1661atWqZfn3BAAAuU+BWH64hvh37tzZmjRpYk2bNrVRo0a5UWoafi+dOnWySpUquf4/0qNHD2vdurWNHDnS2rVrZ1OmTLHFixfbhAkT3OtqJuvZs6cNHTrUatWq5QKmgQMHupFt3jxHqvm5/fbb3dQC6kekoGjEiBHutWuuuSZm2wIAAOQcMQ2QNGx/+/btbmJHdaLWcP1Zs2YFO1mryUujzTwtW7a0yZMnu87VAwYMcEHQ9OnTrV69esE0ffv2dUFWt27dbPfu3daqVSu3Tk0s6VFAVKBAATcX0oEDB9wot48++sh11gYAAIjpPEi5GfMgITsxDxJijXmQkFfl+HmQAAAAcioCJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAgJwYII0bN86qV69uhQsXtmbNmtmiRYvSTT9t2jSrXbu2S1+/fn2bOXNm2OuBQMAGDRpkFSpUsISEBGvTpo2tXr06LI0+L1++fGGPRx99NFu+HwAAyF1iHiBNnTrVevXqZYMHD7alS5dagwYNrG3btrZt27aI6efPn28dO3a0rl272rJly6x9+/busWLFimCa4cOH2+jRo238+PG2cOFCK1q0qFtnUlJS2Loeeugh27x5c/DRvXv3bP++AAAg54t5gPTEE0/Yrbfeal26dLG6deu6oKZIkSL24osvRkz/1FNP2UUXXWT33nuv1alTxx5++GE766yzbOzYscHao1GjRtkDDzxgV1xxhZ155pn28ssv26ZNm2z69Olh6ypevLiVL18++FAgBQAAENMAKSUlxZYsWeKawIIZyp/fPV+wYEHE92h5aHpR7ZCXfs2aNbZly5awNImJia7pzr9ONamVLl3aGjVqZCNGjLBDhw6lmdfk5GTbu3dv2AMAAORNBWL54Tt27LDDhw9buXLlwpbr+cqVKyO+R8FPpPRa7r3uLUsrjdx9992u5umkk05yzXb9+/d3zWyq0Ypk2LBhNmTIkEx+UwAAkJvENECKJfV78qgZLj4+3m677TYXCBUqVChVegVQoe9RDVKVKlWOW34BAMAJ0sRWpkwZi4uLs61bt4Yt13P1CYpEy9NL7/3NyDpFTXBqYlu7dm3E1xU0lShRIuwBAADyppgGSKq1ady4sc2ZMye47MiRI+55ixYtIr5Hy0PTy+zZs4Ppa9So4QKh0DSq7dFotrTWKcuXL3f9n8qWLZsF3wwAAORmMW9iU7NV586drUmTJta0aVM3Am3//v1uVJt06tTJKlWq5Jq+pEePHta6dWsbOXKktWvXzqZMmWKLFy+2CRMmuNc1n1HPnj1t6NChVqtWLRcwDRw40CpWrOimAxB11lbAdP7557uRbHp+zz332A033GClSpWK4dYAAAA5QcwDpA4dOtj27dvdxI7qRN2wYUObNWtWsJP1+vXrXc2Op2XLljZ58mQ3jH/AgAEuCNLw/Xr16gXT9O3b1wVZ3bp1s927d1urVq3cOjWxpNdcpsDqwQcfdKPTFEQpQArtYwQAAE5c+QKaOAgZpmY7TR+wZ8+eLO+P1HHe+ixdH3Kf18+pGtPPH9d9dkw/H7F315gLY50FIKa/3zGfKBIAACCnIUACAADwIUACAADwIUACAADwIUACAADwIUACAADwIUACAADwIUACAADIigBJs1QDAADkVZkKkHQbkJtvvtk+//zzrM8RAABAbgyQXn31Vdu1a5ddcMEFdtppp9mjjz5qmzZtyvrcAQAA5JYAqX379u4Gsb/88ovdfvvt7uax1apVs0svvdTeeustO3ToUNbnFAAAIDd00j755JOtV69e9vXXX9sTTzxhH374oV199dVWsWJFGzRokP3+++9Zl1MAAIDjpMCxvHnr1q320ksv2aRJk2zdunUuOOratatt3LjRHnvsMfviiy/sgw8+yLrcAgAA5NQASc1oEydOtPfff9/q1q1rd955p91www1WsmTJYJqWLVtanTp1sjKvAAAAOTdA6tKli1177bU2b948O/vssyOmUTPb/ffff6z5AwAAyB0B0ubNm61IkSLppklISLDBgwdnNl8AAAC5q5N28eLFbdu2bamW79y50+Li4rIiXwAAALkrQAoEAhGXJycnW3x8/LHmCQAAIPc0sY0ePdr9zZcvnz3//PNWrFix4GuHDx+2Tz/91GrXrp31uQQAAMipAdKTTz4ZrEEaP358WHOaao6qV6/ulgMAAJwwAdKaNWvc3/PPP98N9S9VqlR25QsAACB3jWKbO3du1ucEAAAgtwVIuqXIww8/bEWLFnX/T49uOwIAAJDnA6Rly5bZwYMHg/9PizpwAwAAnBABUmizGk1sAAAgL8vUPEgAAAB5WdQ1SP/4xz+iXqlGuAEAAOT5ACkxMTF7cwIAAJDbAqSJEydmb04AAAByCPogAQAAZLYG6ayzzrI5c+a42bMbNWqU7nD+pUuXRrtaAACA3BsgXXHFFVaoUCH3//bt22dnngAAAHJHgDR48OCI/wcAAMhrMnUvNs/ixYvt+++/d/+vW7euNW7cOKvyBQAAkLsCpI0bN1rHjh1t3rx5VrJkSbds9+7d1rJlS5syZYpVrlw5q/MJAACQs0ex3XLLLe6+bKo92rVrl3vo/0eOHHGvZdS4ceOsevXqVrhwYWvWrJktWrQo3fTTpk2z2rVru/T169e3mTNnhr0eCARs0KBBVqFCBUtISLA2bdrY6tWrI64rOTnZGjZs6DqdL1++PMN5BwAAeU+mAqRPPvnEnnnmGTv99NODy/T/MWPG2KeffpqhdU2dOtV69erl+jVp9FuDBg2sbdu2tm3btojp58+f72qvunbt6m6aqw7jeqxYsSKYZvjw4TZ69GgbP368LVy40IoWLerWmZSUlGp9ffv2tYoVK2YozwAAIG/LVIBUpUoVV4Pkd/jw4QwHG0888YTdeuut1qVLF9ePSUFNkSJF7MUXX4yY/qmnnrKLLrrI7r33XqtTp449/PDDbgqCsWPHBmuPRo0aZQ888IAbeXfmmWfayy+/bJs2bbLp06eHreu9996zDz74wB5//PEM5RkAAORtmQqQRowYYd27d3edtD36f48ePTIUbKSkpNiSJUtcE1gwQ/nzu+cLFiyI+B4tD00vqh3y0q9Zs8a2bNkSlka3SVHTXeg6t27d6gKzV155xQVkR6OmuL1794Y9AADACd5JWxNEhk4OuX//fhd0FCjwxyoOHTrk/n/zzTdHPU/Sjh07XK1TuXLlwpbr+cqVKyO+R8FPpPRa7r3uLUsrjWqZbrrpJrv99tutSZMmtnbt2qPmddiwYTZkyJCovhcAADhBAiQ1W+UV6iu1b98+69+/f9TvUVr1lfKoBklNjQAA4AQOkDp37pzlH16mTBmLi4tzzV2h9Lx8+fIR36Pl6aX3/mqZRrGFptFoNfnoo49cc5s3M7hHtUnXX3+9vfTSS6k+V2n96QEcf/XOrWyN/lbdipSIt52//Gaf/melbVuXdpN3zYZlrdmlp1rxkwrbnu2/24J3frR13+0IS9P0kppWt2UlK5RQwDav2W2fTF3p0krFU0vZlT2aRFz3tBELbdv6vW7dnYacm+r1/4xcZFvX7nH/P6l8UWvarqadXKWElSidYJ+9ucq+/nj9MW4NADlyokjRyDD1JQpVokSJqN4bHx/vJpfUPd68ZjlNFaDn//rXvyK+p0WLFu71nj17BpfNnj3bLZcaNWq4IElpvIBItT0azXbHHXe45xrhNnTo0OD71YFb/Zg0ok7NhgByplPPKmetrjzdPp76vW1dt8canFfVLrvzLJv88Dw78FvqgSPlayTa32+qb1/890dbu2KH1WpS3i6+tYG9MfwL27V5v0vTqE11O7N1FZvz6re2d+cBa9aupl12ZyN7/ZEFdvjQEduyZrdNHPBJ2HqbXlrTKp92kguOQr0zZont2vxb8HnS/r/yVCA+zvbuOGA/Lttqrf7x1whgAHkoQFL/o/vuu8/eeOMN27lzZ6rX1a8oWmq2Uu2Uam+aNm3qmvK0fo1qk06dOlmlSpVcHyBRR/DWrVvbyJEjrV27dm5iSnUQnzBhgntd/aQUPCkAqlWrlguYBg4c6EbXeUFY1apVw/JQrFgx97dmzZpMcgnkYA3Pr2bfLthoKxducs8VKFU7o4zVaVHJls5O3ZfwzPOq2vrvd9qyOevc80X/+8mqnH6S1f+/qvbJ1D/uAqAga/H7a2zNN9vd8w9f+da6/Pv/rMaZJ9uPS7fakcMB+33fXxeB+fPnsxr1y9o3n6Su/UnanxKWNpSCKS+ganF5rSzZHgByWICkuYPmzp3r5kK68cYb3USPv/zyiz377LP26KOPZmhdHTp0sO3bt7uJHdWJWrU+s2bNCnayXr9+vRvZ5tFs3ZMnT3bD+AcMGOCCIA3fr1evXlj+FGR169bNzfDdqlUrt05NLAkgd8ofl89OrlLclsxe89fCgNnGVbusfPXEiO/R8uVzwwOZDSt3Wo0zy7r/q6mraGIh27jqrwu9lKRDtnXtXitfo6QLkPyq1z/ZChctaN//GaSFuqRbQ4srGGe7t+23ZR+us7Ur/gi6AJwgAdJ///tfN7fQeeed52p6zj33XDv11FOtWrVq9tprr7l+PBmh5rS0mtQ+/vjjVMuuueYa90iLapEeeugh94iGZvHWyDYAOVfhovGWPy6//b43vIZGNTalyhWN+J4iJQrZgX2p0xcpHv/n63/89df6HNiXHHzNr26LSrbh+522f3dycNnB5MP2+VurbMvPu02nklMalrVLbm1gM5/7iiAJOJECJN1a5JRTTgn2N9JzUU2N188HAPKaoiULWZU6pe39F78OW66+Rl+F1FSpKU01U43aVCNAAk6kiSIVHGlCRtE90dQXyatZ8m5eCwBZSf17jhw+kqpmR7VBv+/9qzYnlJYnFI+Q/s8aI682yqtR8iQUL5SqpkrqNKvogqG1f/ZXSo9GryWWOfoktADyUICkZrWvvvrK/b9fv36uD5L699xzzz3uFiAAkNXUWXr7hn1u9FhQPnPPt/w5lN5Py8PSm1nl00vbljV/pNeotf17kt0yT8HCcVauegk3es2vdvOKtmrRJjty5OhN8mUqF08zcAOQR5vYFAh5dEuP77//3t1oVv2QdO8zAMgOy+eus7/dcMYfI8LW7XUj0AoUirPvv/ijw/TfbjzD9Q3SsH7RPEPtezSxhhdUs7XfbrdaZ5W3slVL2MdTvguu86uP11vjtjVs97bf/xjmf2lNFzSt+Tq8lkiBlmqEvlvwS6p8nd60wh8B3MY/RqnVbFDW6jSvZHMnfxfWyVxzIUlcgfxWLLGQlalUzPVf2rPjQDZtMQAxmwfJ6+SsBwBkJ40qSygW7+YqKlK8kO34ZZ/NeHppsCN28VKFXSdpj2qKZk/6xk0U2fzSU2339t/tvee+Cs6BJMs+XGsF4+Ps/I51LF4TRf682/779DI3B1KoOi0qutd2b/1jAkm/Jm1rWPGTEtxcbkrzwcSv7afl24Kvq09Sh35/zNfmzb+kxy+rd9n00UuycjMByAL5ApkcvqWJGJ988klXeyR16tRx8w/5bySbV2nySd0Ed8+ePVFPjBmtjvOYXfdE9/o54XN1HW/jus+O6ecj9u4ac2GsswDE9Pc7U32Qnn76abvooousePHibuJGPfQhl1xyieuPBAAAcMI1sf373/92tUehcxfdfffdds4557jX7rrrrqzMI3KQC8sXs8sqlbDE+Dhbvz/FJv38q/30W+SZg6VZ6QS7pmpJO7lwAdty4KC9vm63Lf81KSzN1VUT7YJyxaxoXD5btS/FXvxpl21JOhSWplGpwvaPKolWtUhBSwmYfb8nyZ5Y+df9tM5ILGT/rFrSqhQtaMmHA/bptv02dd1uC20kaV66iF1RuYRVSChgew8esQ+27LMZv+zLwq0DAMgrMlWDpNmpVYPk9/e//91VWSFval6miN1Yo5S9uWGPDVi+2dbtP2j9zihrJQpGLka1isdb99PL2MfbfrP+yzfb4l0HrHftk61ykYLBNJdVKm4XVShuL/y0ywZ+vdWSDx9x6yyY76/1NC2dYHfWKm2fbNtv9y3fYg9+vcXm/3kjUVHQdF/dsvbV7gPWf/kWG71qhzU+KcE6Vv9ryokGJQvbXaeVtjlbf7O+yzbbxJ932SUVi9vfy/9xmxkAAI45QLr88svt7bffTrX8nXfesUsvvTQzq0Qu0K5icfto628uUPnlwCEX1KQcPmLnlY0cZFxcsbh99WuSq6XZdOCQTVu/x9bsT7G2Ff5Kf3HFEvb2hj22ZNcBW//7QXt69U4rFR9nTUoXCRbQTjVK2Wtrd9uHW35zNUv67C92/hUgtShTxNVmvbVhr21NOmTf7022yWt3u+CncNwfkda5ZYu6AE3r2JZ82Jb9mmTvbNxrl1fO2v5jAIATrIlt9OjRwf/XrVvXHnnkEXcbkBYt/hiV8cUXX9i8efOsd+/e2ZNTxJTijBrF4l1Q4VHv/hV7klxNUSS1iheymb4mrK9/TbImpRPc/8sWinPBkNbhOXA4YD/tS3bvXbDjd/eZpQsVcJ81rEF517S3bn+KC5g2/v7HndIL5s9nB33z0qQcOWLxcfmtRtF4FzAVyJfPUlKlCbh1lykUZzuSo7/BMgAg74s6QFKfo1ClSpWy7777zj08mkX7xRdfdDeSRd5SomCcxeXLZ3sOhgcSe1KOWMXEv5rMQpUsGJc6/cHDbrko2PljHRHSxP9RuVm28B9F9Koqifbq2l9te9Iha1ephA2qV9buWbrZ9h86Yl/tTnK1VS3LFHFBVcn4ONdfSRSAyde7D7jmwTO2FbLv9iRbucIFrF3FEsE0BEgAgEwFSN6tRYDjyeuKNH3jHlu084/J9Mav3mnjzq7kOl2rT9E3u5NcjVLXmifZnaeVdrVJb2/Ya3USCwc7aX+0db+VK1zQ+tY52eLy57MDh47YrM377OqqJS2KSZEBACeYY54o0ptGKV++kF61yHP2HjxshwMBS/yz9seTGJ/fdvtqgDy7Dx5Onb5gnFseWnOkmqTdB4+EpVm7/4/mMy/tL7//NartUMBsW9IhK13or3XP3LTPPVQb9NuhI3ZyoTjXSVvpPBpBN2XdblfDpO9TL7GwWx6aBgCATHfSlpdfftnq169vCQkJ7qFbjLzyyits1TzqcMBszW8pwaBCFBKfkVjYVv85i7Hf6n3JdkbJv9JL/ZJK/8f9qdRZ+teUvwIVSYjLZzWLFwqm0Weqr5CG5of2hzq5UAHbkZw6sNH6VIPU8uSi7nW9P1TgzzT6Pkrzw95k2+ebMRkAgEzVID3xxBM2cOBANw+S5j6Szz//3G6//XbbsWNH2L3akHf8b9M+u6NWafv5txT78bdk1++nUFx++2Tbb+51vfZryiGbsu6PqR7e27TPBtUr50a/Lfv1gLUoU9ROKRZvz/20K7jO9zbttfZVEt3oNNXkXFM10QUwi/8cpaZO23O2qCks0XamHHZBz6WV/ug7tHDHXyPZLq30x4g5VWieXTrBrqhUwp5atcMFRFK8QH5rpvto7Ulynbpbly1mzUsn2EMr/roVBAAAxxQgjRkzxp555hnr1KlT2ND/M844wx588EECpDzqix2/W4kC+V2wUvLP0WSPfrvN9vzZPKbRYKF3rlHN0tgfdtg/q5W0DtVKuokiR67cHhx9Jv/9ZZ8Lsm6peZIVKZDfVu1Ndus8GNIvSP2LVONzV63SLrj56bdkG7pim+3Xwj81LJlg7SsnuvmT1v1+0B7/frvrvB3q3JOL2vV/zo2kvCk4Sm+SSwDAiStT92IrXLiwrVixwk499dSw5atXr3bNbklJ4T9MeRH3YkN24l5siDXuxYa8KlvvxabA6I033ki1fOrUqVarVq3MrBIAACB3N7ENGTLEOnToYJ9++mmwD5ImiZwzZ07EwAkAACA3yVQN0lVXXWWLFi2yMmXK2PTp091D/9eyK6+8MutzCQAAkJNrkA4ePGi33XabG8X26quvZk+uAAAAclMNUsGCBe3NN9/MntwAAADk1ia29u3bu2Y1AACAvChTnbQ1Uu2hhx5yHbMbN25sRYsWDXv97rvvzqr8AQAA5I4A6YUXXrCSJUvakiVL3COU7slGgAQAAE64AGnNmjXB/3OzWgAAkNdk+ma1qkWqV6+em1VbD/3/+eefz9rcAQAA5JYapEGDBrkb1nbv3t1atGjhli1YsMDdg239+vWufxIAAMAJFSDpRrXPPfecdezYMexmtWeeeaYLmgiQAADACdfEpskimzRpkmq5RrQdOnQoK/IFAACQuwKkG2+80dUi+U2YMMGuv/76rMgXAABA7mpi8zppf/DBB9a8eXP3fOHCha7/UadOnaxXr17BdOqrBAAAkOcDpBUrVthZZ53l/v/TTz+5v7pZrR56zcPQfwAAcMIESHPnzs36nAAAAOT2eZAAAADyqhwRII0bN86qV6/uJpxs1qyZLVq0KN3006ZNs9q1a7v09evXt5kzZ4a9rtm9NVdThQoVLCEhwdq0aWOrV68OS6NpCapWrerWoXTqeL5p06Zs+X4AACB3iXmANHXqVNepe/DgwbZ06VJr0KCBtW3b1rZt2xYx/fz58938S127drVly5ZZ+/bt3SO079Pw4cNt9OjRNn78eNd5XDfT1TqTkpKCac4//3x74403bNWqVfbmm2+6vlRXX331cfnOAAAgZ8sX8G6mFiOqMTr77LNt7Nix7vmRI0esSpUqbsLJfv36pUrfoUMH279/v82YMSO4TCPpGjZs6AIifZ2KFSta7969rU+fPu71PXv2WLly5WzSpEl27bXXRszHu+++6wKt5ORkK1iw4FHzvXfvXktMTHTrLlGihGWljvPWZ+n6kPu8fk7VmH7+uO6zY/r5iL27xlwY6ywA2SLa3++Y1iClpKTYkiVLXBNYMEP587vnunVJJFoeml5UO+Sl1410t2zZEpZGG0KBWFrr3LVrl7322mvWsmXLNIMjBU7aqKEPAACQN8U0QNqxY4cdPnzY1e6E0nMFOZFoeXrpvb/RrPO+++5zzW+lS5d2czi98847aeZ12LBhLtDyHqrlAgAAeVPM+yDF0r333uv6MWnCy7i4ODfJZVotjv3793fVcd5jw4YNxz2/AAAgh8+knRU0saQCk61bt4Yt1/Py5ctHfI+Wp5fe+6tlGp0Wmkb9lPyfr8dpp51mderUcbVCX3zxhbVo0SLV5xYqVMg9AABA3hfTGqT4+Hh3g9s5c+YEl6mTtp5HClJEy0PTy+zZs4Ppa9So4YKk0DTqL6TRbGmt0/tcr68RAAA4scW0Bkk0xL9z587WpEkTa9q0qY0aNcqNUuvSpYt7Xc1elSpVcn2ApEePHta6dWsbOXKktWvXzqZMmWKLFy92N8r1bm/Ss2dPGzp0qNWqVcsFTAMHDnQj2zRKTRQsffnll9aqVSsrVaqUG+KvNDVr1kw3iAIAACeGmAdIGra/fft2N7GjOlGrGWzWrFnBTtbqPK2RbR6NNJs8ebI98MADNmDAABcETZ8+3erVqxdM07dvXxdkdevWzXbv3u0CIa1Tk0JKkSJF7K233nJzLymdmuIuuugit06a0QAAQMznQcqtmAcJ2Yl5kBBrzIOEvCpXzIMEAACQExEgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA+BAgAQAA5MQAady4cVa9enUrXLiwNWvWzBYtWpRu+mnTplnt2rVd+vr169vMmTPDXg8EAjZo0CCrUKGCJSQkWJs2bWz16tXB19euXWtdu3a1GjVquNdr1qxpgwcPtpSUlGz7jgAAIPeIeYA0depU69WrlwtQli5dag0aNLC2bdvatm3bIqafP3++dezY0QU4y5Yts/bt27vHihUrgmmGDx9uo0ePtvHjx9vChQutaNGibp1JSUnu9ZUrV9qRI0fs2WeftW+//daefPJJl3bAgAHH7XsDAICcK19A1S0xpBqjs88+28aOHeueK3CpUqWKde/e3fr165cqfYcOHWz//v02Y8aM4LLmzZtbw4YNXZCjr1OxYkXr3bu39enTx72+Z88eK1eunE2aNMmuvfbaiPkYMWKEPfPMM/bzzz9Hle+9e/daYmKiW3eJEiUsK3Wctz5L14fc5/Vzqsb088d1nx3Tz0fs3TXmwlhnAcgW0f5+x7QGSU1aS5YscU1gwQzlz++eL1iwIOJ7tDw0vah2yEu/Zs0a27JlS1gabQgFYmmtU7ShTjrppDRfT05Odhs19AEAAPKmmAZIO3bssMOHD7vanVB6riAnEi1PL733NyPr/PHHH23MmDF22223pZnXYcOGuUDLe6iWCwAA5E0x74MUa7/88otddNFFds0119itt96aZrr+/fu7WibvsWHDhuOaTwAAcIIESGXKlLG4uDjbunVr2HI9L1++fMT3aHl66b2/0axz06ZNdv7551vLli1twoQJ6ea1UKFCrq0y9AEAAPKmmAZI8fHx1rhxY5szZ05wmTpp63mLFi0ivkfLQ9PL7Nmzg+k1dF+BUGga9RfSaLbQdarm6LzzznOfP3HiRNf3CQAAQArEejNoiH/nzp2tSZMm1rRpUxs1apQbpdalSxf3eqdOnaxSpUquD5D06NHDWrdubSNHjrR27drZlClTbPHixcEaoHz58lnPnj1t6NChVqtWLRcwDRw40I1s03QAocFRtWrV7PHHH7ft27cH85NWzRUAADhxxDxA0rB9BSia2FGdqDVcf9asWcFO1uvXrw+r3VFz2OTJk+2BBx5w8xYpCJo+fbrVq1cvmKZv374uyOrWrZvt3r3bWrVq5dapiSW9Gid1zNajcuXKYfmJ8awHAAAgB4j5PEi5FfMgITsxDxJijXmQkFflinmQAAAAciICJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAAB8CJAAAgJwWII0bN86qV69uhQsXtmbNmtmiRYvSTT9t2jSrXbu2S1+/fn2bOXNm2OuBQMAGDRpkFSpUsISEBGvTpo2tXr06LM0jjzxiLVu2tCJFiljJkiWz5XsBAIDcK6YB0tSpU61Xr142ePBgW7p0qTVo0MDatm1r27Zti5h+/vz51rFjR+vatastW7bM2rdv7x4rVqwIphk+fLiNHj3axo8fbwsXLrSiRYu6dSYlJQXTpKSk2DXXXGN33HHHcfmeAAAgd8kXUJVLjKjG6Oyzz7axY8e650eOHLEqVapY9+7drV+/fqnSd+jQwfbv328zZswILmvevLk1bNjQBUT6KhUrVrTevXtbnz593Ot79uyxcuXK2aRJk+zaa68NW5+W9ezZ03bv3n3UvCYnJ7uHZ+/evS6vWn+JEiUsK3Wctz5L14fc5/Vzqsb088d1nx3Tz0fs3TXmwlhnAcgW+v1OTEw86u93zGqQVIuzZMkS1wQWzEz+/O75ggULIr5Hy0PTi2qHvPRr1qyxLVu2hKXRRlAgltY6ozVs2DC3Lu+h4AgAAORNMQuQduzYYYcPH3a1O6H0XEFOJFqeXnrvb0bWGa3+/fu7aNN7bNiw4ZjWBwAAcq4Csc5AblGoUCH3AAAAeV/MapDKlCljcXFxtnXr1rDlel6+fPmI79Hy9NJ7fzOyTgAAgBwTIMXHx1vjxo1tzpw5wWXqpK3nLVq0iPgeLQ9NL7Nnzw6mr1GjhguEQtOoM5ZGs6W1TgAAgBzVxKYh/p07d7YmTZpY06ZNbdSoUW6UWpcuXdzrnTp1skqVKrkO0tKjRw9r3bq1jRw50tq1a2dTpkyxxYsX24QJE9zr+fLlc6PShg4darVq1XIB08CBA93INk0H4Fm/fr3t2rXL/VU/qOXLl7vlp556qhUrViwm2wIAAOQcMQ2QNGx/+/btbmJHdaLWcP1Zs2YFO1krgNHINo8md5w8ebI98MADNmDAABcETZ8+3erVqxdM07dvXxdkdevWzQ3fb9WqlVunJpb06PNeeuml4PNGjRq5v3PnzrXzzjvvOH17AACQU8V0HqQTYR6FzGAeJDAPEmKNeZCQV+X4eZAAAAByKgIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAAHwIkAAAyYNy4cVa9enUrXLiwNWvWzBYtWpRu+mnTplnt2rVd+vr169vMmTPDXg8EAjZo0CCrUKGCJSQkWJs2bWz16tVhaXbt2mXXX3+9lShRwkqWLGldu3a13377Lfh6UlKS3XTTTW79BQoUsPbt20fMS3Jyst1///1WrVo1K1SokPseL7744jFtj7yKAAkAgChNnTrVevXqZYMHD7alS5dagwYNrG3btrZt27aI6efPn28dO3Z0Ac2yZctc4KLHihUrgmmGDx9uo0ePtvHjx9vChQutaNGibp0KejwKjr799lubPXu2zZgxwz799FPr1q1b8PXDhw+74Oruu+92AVZa/vnPf9qcOXPshRdesFWrVtnrr79up59+epZtn7wkX0ChKzJs7969lpiYaHv27HERfVbqOG99lq4Puc/r51SN6eeP6z47pp+P2LtrzIWxzkKOpBqjs88+28aOHeueHzlyxKpUqWLdu3e3fv36pUrfoUMH279/vwtqPM2bN7eGDRu6gEg/wRUrVrTevXtbnz593Ov6XSlXrpxNmjTJrr32Wvv++++tbt269uWXX1qTJk1cmlmzZtkll1xiGzdudO8PpZqk3bt32/Tp08OW6z1a388//2wnnXSSnaj2Rvn7TQ0SAABRSElJsSVLloTV0OTPn989X7BgQcT3aLm/Rke1Q176NWvW2JYtW8LS6MdbgZiXRn/VrOYFR6L0+mzVOEXr3XffdetQjVWlSpXstNNOc0HZgQMHMrAVThwFYp0BAABygx07drimLNXuhNLzlStXRnyPgp9I6bXce91bll6asmXLhr2ufkaqBfLSREM1R59//rnrC/X222+773PnnXfazp07beLEiVGv50RBgAQAwAlAzYH58uWz1157zdVSyRNPPGFXX321Pf30064PE/5CExsAAFEoU6aMxcXF2datW8OW63n58uUjvkfL00vv/T1aGn8n8EOHDrmRbWl9biQaJaemNS84kjp16rh+UOrLhHAESAAARCE+Pt4aN27sRoGF1sroeYsWLSK+R8tD04tGonnpa9So4YKc0DTqRKy+RV4a/VWna/V/8nz00Ufus9VXKVrnnHOObdq0KWx6gB9++MH1ZapcuXLU6zlRECABABAlDfF/7rnn7KWXXnKjy+644w43Sq1Lly7u9U6dOln//v2D6Xv06OFGj40cOdL1U3rwwQdt8eLF9q9//cu9riavnj172tChQ10n6m+++catQyPTvLmMVMtz0UUX2a233urmXJo3b557v0akhY5g++6772z58uWuZkkjtPR/PTzXXXedlS5d2uVVaTVVwL333ms333wzzWsR0AcJAIAoadj+9u3b3cSO6iCt4foKgLxO1uvXr3c1Mp6WLVva5MmT7YEHHrABAwZYrVq13PD7evXqBdP07dvXBVma10g1Ra1atXLrVGdqj/oNKSj629/+5tZ/1VVXubmTQmnY/7p164LPGzVq5P56s/kUK1bM1V5pSgKNZlOwpHmRFJwhNeZByiTmQUJ2Yh4kxBrzIOFE//2mBgkAkOO0aH1xrLOAGFvwyXsx/Xz6IAEAAPgQIAEAAOTEACkn3hkZAACcuGIeIOXUOyMDAIATV8xHseWGOyNHwig2ZCdGseFEH8VGJ20syKZO2rliFJt3Z+TQSbWiuTOyapxCqXZI80pEc2dkBUhHuzPylVdemepzk5OT3cOjDett6Kx2cP++LF8ncpfsKFcZcSBlf0w/H7EX6zKoW2ngxLY3m8qgt96j1Q/FNEDKTXdGHjZsmA0ZMiTVctV2AVntzVhnACe8eyfEOgc40SWG3DMuO+zbty/dz2AepCipliu05kpNgerorZlINVU8sja6V+C5YcOGLG++BKJBGUSsUQazj2qOFBwdrTtNgbx8Z2SNYgtNo35Kmb0zcqFChdwjlJrpkH10UuDEgFiiDCLWKIOxq52K6Si23H5nZAAAkDfFvIlNzVadO3d2HaabNm1qo0aNSnVn5EqVKrk+QN6dkVu3bu3ujNyuXTubMmWKuzPyhAkTUt0ZWTcFVMA0cODANO+MrJFvBw8ejHhnZAAAcGKKeYCUk++MjNhQU6bmxfI3aQLHC2UQsUYZjL2Yz4MEAACQ08R8Jm0AAICchgAJAADAhwAJAADAhwAJAADAhwAplznvvPPcNAaZtXbtWjcVwvLly93zjz/+2D3XaL9Y5SmzHnzwweDkn1mZNtptF0lWbM/cRt/XuxdiZvbFTTfdFJyCI6tQzjMvL5bz0H1XvXp1N51MZumm57GYJDia/ZKZtMda5qsf4/bMyWI+zB8Z89Zbb1nBggWzbH2aNmHz5s1RzSqqk+L5559vv/76a9gJIqvzFK0+ffpY9+7dj/vnnqj046tAyH/SVfkpVaqU5WSUc3i+/PJLK1q0aFRp9eOvACE0SNDUNJdccokdb7rtiMqw7kCB44MAKZfRDXWzejbztG6vEqs8RatYsWLugeylmUB0U+m0HGv5OR4o5/CcfPLJx/T+hIQE9zjedFuu3HCs5SU0seUy/qrif//733bzzTdb8eLFrWrVqsEZxT2LFi2yRo0auUkyNVv5smXL0q0qX7dunV122WWuRkBXWWeccYbNnDnTVdnqqlr0mt6jphF/nqLN1/z5811TgJcv1UyEVglHqsb20qTVnKDvotnYlW+995xzznHfJ9Qrr7zi8qeaBM2crhsWejSZqCYV1Xt1E+JLL73Ufvrpp1T7YOXKla5GQnnXBKWffPJJuvvs888/t3PPPdedVHUVePfdd7uJTLOD9oUmQNVD31FXm5pJPnS6M20DbXPtG51wr7vuurB7E3pl4r333nO3AtJEda+++qoNGTLEvvrqK/eaHtpHkZrY7rvvPjvttNOsSJEidsopp7jP12z10Xj55Zfdtk9OTg5bria4G2+8Mc33Uc5PrHKeHn2m7sCgoFL349RdF9JqEtJxoe2rfadyrjspKN/e/tZ2veeee4JlPtI+8/ZPettc/7/++uvdPlOennzyyVTlKVJTtT7HO878zWaq4dQ6FfBpm2vS5IkTJ4a9/+eff3blWcdigwYNbMGCBcHXdu7caR07dnR3qtDr9evXt9dffz3V9jx06FC65xM/HWO33HKLy5fuIXfBBRe480ZuRICUy+ng934Q7rzzTrvjjjts1apV7rXffvvNnfzq1q3r7junA1nV9em566673I/Tp59+at9884099thj7kSjE96bb77p0mj9qup96qmnMpUv3RtPP046IJcuXWoPP/yw+1E9FjqI9SOq29B8/fXX7kSgmdRDf2j0I6AT0IwZM9xDJ/xHH3007MSqW9/o1jW6l59mWL/yyivdPfpC3Xvvvda7d2/33XRfP30XnWwi0WfqtjaaqV35mjp1qvsh0Qknu7z00ktWoEABFzRoHz3xxBP2/PPPB19XsKJtrpOWtodOvF4QEKpfv35u+3z//fd24YUXuu+sQEL7Xg81NUSiYEEn9e+++859/nPPPed+EKJxzTXXuNqqd999N7hMwdv//vc/F4hEQjk/Mct5WpRvfed33nnHPvjgAxdQavtHon2tsvnss8/a6tWr3XbT/vKaVCtXrmwPPfRQsMyn5WjbXNt73rx5rlzr3qGfffZZmnmKlgIVHWO6kNEx+swzz6Rqfrv//vvdsaCgShctCohUhiQpKcldAOnYWrFihStHugjReSMj55NIx7COWeVLx+NZZ53l7lihm8HnOppJG7lH69atAz169HD/r1atWuCGG24IvnbkyJFA2bJlA88884x7/uyzzwZKly4dOHDgQDCNXtNuX7ZsmXs+d+5c9/zXX391z+vXrx948MEHI362P22kPEWTL/315+u5554Ly9fEiRMDiYmJYZ/z9ttvuzSewYMHBxo0aOD+v3PnTvfaxx9/HDHvSlukSJHA3r17g8vuvffeQLNmzQJp2b59u1vnN998456vWbPGPX/00UeDaQ4ePBioXLly4LHHHou4jbp27Rro1q1b2Ho/++yzQP78+cO+f1bRvqhTp47b5p777rvPLUvLl19+6fK8b9++sO8wffr0sHSh2zuU0mrfpGXEiBGBxo0bp7mezp07B6644org8zvuuCNw8cUXB5+PHDkycMopp4R9p1CU8xOvnKdFZTg+Pj7wxhtvBJdpmyUkJISdN5988slg2TrttNMCKSkpEdcXmtbj32dH2+ZaXrBgwcC0adOCr+/evdu9J7Q8RTqO9Dn6vND94pWdyy67LNClS5eI+fbSPv/888Fl3377rVv2/fffp7n92rVrF+jdu3eGzifVQraR9nmJEiUCSUlJYeutWbOmO05zG2qQcrkzzzwz+H9dRarJxGsu0VWFXg+9B52uBNOj6mXd6FfV9roPkK4GszpfusL250tNBsfaP0S1IG3btnVXurrS8V/xqfpbtRseVXWHNi3pClJXWGoWUtWw0nv3AwwVug11ZaUaBG3rSFRLo9oUrx+JHsqjrtbXrFlj2aF58+ZhNQrKr76b149IV3XaRmpW0PZQbUSk76nvlRmqPVD50T7X99V9E/3rTo9uIq0r/19++cU91/bTvg39TqEo5ydmOU+rJiclJcWaNWsWts1OP/30NGs7Dhw44LaFyt3bb78drGHJiPS2uZq5VGsbuu/VXJVWnqKl2krdrF3Ne7r/qJpz0yufypN4+dL5QLWaqjHTNtI+e//991OVg+ZHOZ/4y4FqdNV0G1oWVAYiNePmdARIuZx/VI0Ksr+qPCPUdqwDWlWtanrQSXHMmDHHPV+q9ve3cx+tH4va39XkoH4T+pFWlfIXX3wRdZ70g6NqYDUJLVy40D1EJ9zM0snitttuc1Xc3kMnEZ1gatasacebmlf0w6UfRt2wWSN69KMQ6XtGO9InlLa/+kVolI+aGtQ8o2r+jGxD9SVSfwn1R1Iw9+2330ZsAjwWlPO8Xc6jpSZVBbJPP/2068ejZtL/+7//i7rPXFaeh/WejJSFiy++ONhHatOmTa4Zy9+0HJovL8jx8jVixAgXYKvZd+7cuW6f6dxwrOWgQoUKYeVAD21jNX3mNgRIeVidOnXclbHamj2hJ9L0Thq33367a4NXHwSdSL2RQJLeiKZo6MpJP0qhHXH1Qx1KHfzUsTG0k2c0c3rox7V///7uakodSydPnhxVntS3Qgexajt0otG2UyfISEK3oa429SOu9JGo/V39BE499dRUD297ZjXvBy80v+rAqVEw6nir76r+EepQW7t27bDahfQov0fb99ru1apVc0GRgg59rr8DcbQBjGokFAy0adPGlcm0UM5PzHIeiYIxBQWhx4C+3w8//JDmexQYKWgcPXq066+k4FP7LdoyfzSqnVKeQvf9nj17UuVJZSG0NlDB5e+//57uuvWezp07u0EU6njuHySQHvWJuuKKK+yGG25wFyTKZ6TttDCd80mkcrBlyxZX4+gvB7lxegICpDxMo5N01aCqY528NErn8ccfT/c9GlWhalZViaoToa4svJOifvi0PtUMbN++3V0tZDZfuopRp0BV2evzvHx5VzmqItfIigEDBriqWf0AeKM5IlF+9YOhk5t+kNVEoxNMWid0P41YUrWwTjA//vijffTRR65jZSTjxo1ztS4KNtTZVyfgtDoQ6+pMP2LqrKofPuVJnUezs/OqqsiVd/0QalSKakZ69OjhXlOzmk76WqYaFHUaVTV7tM0I2s76Hjt27Eg10kx04tTnq+pf+00/Ol4NVUbLyMaNG13Qkta2DU1LOT/xynkkas7p2rWrq63Qd1PnY9U+qqYuEm3rF154waXT8aBAQwGTyoBX5tWRX829KvOZoaY3BTHKk8qZakSVR+UptOlKo73Gjh3ral3VgV7Be3rzbg0aNMhtY+1HrVPlNdpy4B2r6jCu/abyqRrArVu3Zuh84qeLGTXBaSCByqYGgGj9umDSd8ptCJDyMJ0s/vvf/7qrIV1xqpBqtE56dLWkk6EONI1KUfW9qp9Fw0E11Fujm8qVK5fpk5+ad5QvnUjVfq586WAXr7+G2sR1stKPnTf8VKOT0qIfGZ3INYpGedaPkr6HDvpo6GSlH3VdJeuKXNXWqoKORLUveuiqSyN1FGSkdXWkPgAa0aIrM9XYaD/ou2o4cXbREGf1q1CfB20Dncy0PbwrTv0oTJs2zY360vc4WjDh0bZVmdCwYa0n0pDgyy+/3G07lQ3tW50cNdomo9RHQ5+nMny0WbYp5ydmOU+Lvo/yoFoh/WBrSgON1opEw+gVhKsvmr7Dhx9+6PaZgkjRCDb9yKtm6ljmT9LILwUOGm2pPOnzVPZC+6dpRKRqNZV3BddqLtP+TosudBQsK99qFlSNjvZttFSLqBofNatpygH1n4t0rHVK53zip4BPZVn56dKliyujmvJAwbyOpdwmn3pqxzoTgPrD6IBS1XMsJmHLK3Si049xXpj6X01AmlZAtVB5BeUcoiZVBeIKilSbhJyJmbQRE+qAqzZvnSTUmVNV9P/85z/50YBrylFfED28Wp3cinIOUbOZav5UC6PgWDVToj5AyLkIkBAT6sinKnj91agHDbd95JFHYp0t5ABqnlGQpGayYx0KHWuUc3jUlK1+PGoaU5OfJovMjR2XTyQ0sQEAAPjQSRsAAMCHAAkAAMCHAAkAAMCHAAkAAMCHAAkAAMCHAAkAAMCHAAkAAMCHAAkAAMDC/T8lr5Q0EkNi2QAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot(n: int, probability_distribution_indistinguishable, probability_indistinguishable, probability_distribution_partially_distinguishable, probability_partially_distinguishable, probability_distribution_distinguishable, probability_distinguishable):\n",
    "    x = []\n",
    "    for i in range(n+1):\n",
    "        # creates a list of all possible tuple such that a+b=n for tuple (a,b)\n",
    "        x.append(\"({},{})\".format(i, n-i))\n",
    "    output_modes = x\n",
    "    colours = ['#58c4e1', '#946cba', '#383e48']  # light blue, purple and gray\n",
    "\n",
    "    def add_labels(x, y):\n",
    "        for i in range(len(x)):\n",
    "            if y[i] > 0.0005:\n",
    "                plt.text(i, y[i]/2, round(y[i], 5), ha='center', color='white')\n",
    "            else:\n",
    "                plt.text(i, 1.25*y[i], round(y[i], 5), ha='center')\n",
    "    cases = [\"indistinguishable\",\n",
    "             \"partially distinguishable\", \"distinguishable\"]\n",
    "    probabilities = [probability_indistinguishable,\n",
    "                     probability_partially_distinguishable, probability_distinguishable]\n",
    "    plt.bar(cases, probabilities, color=colours)\n",
    "    plt.ylabel('probability')\n",
    "    plt.title(\"Comparing the total bunching probabilities for {} photons\".format(n))\n",
    "    add_labels(cases, probabilities)\n",
    "    plt.show()\n",
    "\n",
    "    plt.scatter(output_modes, probability_distribution_indistinguishable,\n",
    "                label='indistinguishable', color=colours[0])\n",
    "    plt.scatter(output_modes, probability_distribution_partially_distinguishable,\n",
    "                label='partially distinguishable', color=colours[1])\n",
    "    plt.scatter(output_modes, probability_distribution_distinguishable,\n",
    "                label='distinguishable', color=colours[2])\n",
    "    plt.ylabel('probability')\n",
    "    plt.xlabel('number of photons in first two modes')\n",
    "    plt.title(\"Normalized bunching distributions\")\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "\n",
    "\n",
    "plot(n, probability_distribution_indistinguishable, probability_indistinguishable, probability_distribution_partially_distinguishable,\n",
    "     probability_partially_distinguishable, probability_distribution_distinguishable, probability_distinguishable)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To see the that this set up only works for at least 7 photons we can compare the cases of 4,5,6 and 7 photons. You can see this if you run the above for n = 4,5 and 6. Running it for 8 or higher values for n will take a long time.\n",
    "\n",
    "### For 8 photons the following holds:\n",
    "\n",
    "The probability for all 8 indistinguishable photon ending up in the first two modes is: 0.240 %\n",
    "\n",
    "The probability for all 8 distinguishable photon ending up in the first two modes is: 0.002 %\n",
    "\n",
    "The probability for all 8 partially distinguishable photon ending up in the first two modes is: 0.294 %\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## References"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "<a id=\"1\">[1]</a>\n",
    "Agata M. Branczyk, Hong-Ou-Mandel Interference (2017) in arXiv:1711.00080 (http://arxiv.org/abs/1711.00080)\n",
    "\n",
    "<a id=\"2\">[2]</a>\n",
    "Benoit Seron, Leonardo Novo and Nicolas J. Cerf, Boson bunching is not maximized by indistinguishable particles (2022) in arXiv:2203.01306 (https://arxiv.org/abs/2203.01306)\n",
    "\n",
    "<a id=\"3\">[3]</a>\n",
    "V. S. Shchesnovich, Universality of Generalized Bunching and Efficient Assessment of Boson Sampling (2016) in arXiv:1509.01561 (https://arxiv.org/abs/1509.01561)\n",
    "\n",
    "<a id=\"4\">[4]</a>\n",
    "Stephen W. Drury, A counterexample to a question of Bapat and Sunder (2016) in Electronic Journal of Linear Algebra **Vol 31** 69-70 (https://journals.uwyo.edu/index.php/ela/article/view/1631/1631)\n"
   ]
  }
 ],
 "metadata": {
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}