atmosat.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Solar plane optimization\n",
    "The goal of this notebook is to determine if a solar plane can sustain flight, and if it can, figure out what the cheapest solar plane that can sustain flight is.\n",
    "\n",
    "The approach is to first make a model of a solar plane that can be fed different configuration parameters (wingspan, battery mass, cruising altitude, number of props, payload mass, etc). Different configutations are then fed to the model in an optimization stage to figure out which configuration leads to the optimal results.\n",
    "\n",
    "The plane is assumed to be a [flying wing](https://www.google.com/search?q=flying+wing&safe=off&source=lnms&tbm=isch) since this is generally the lowest drag configuration possible."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy import optimize\n",
    "\n",
    "from numba import jit  # The model runs very slowly without numba\n",
    "\n",
    "pd.set_option(\"max_columns\", 99)\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To do:\n",
    " - Sun following - if plane flies toward the sun, it can effectively get a slightly longer day\n",
    " - Wind speed vs. altitude, related to above\n",
    " - Add structural model to get mass of vehicle and prop\n",
    "\n",
    "Relevant work\n",
    " - [Noth 2008 - Design of Solar Powered Airplanes for Continuous Flight](http://www.sky-sailor.ethz.ch/docs/Thesis_Noth_2008.pdf)\n",
    " - [Leutenegger 2014 - Unmanned Solar Airplanes: Design and Algorithms for Efficient and Robust Autonomous Operation](https://www.research-collection.ethz.ch/bitstream/handle/20.500.11850/90524/eth-46751-02.pdf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Parameters\n",
    "These are the constants and bounds that we feed into the optimization."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SolidEnergy Hermes: http://www.solidenergysystems.com/hermes/\n",
    "battery_specific_energy = 415 # Wh/kg - design for a lower specific energy even though rated for 450 Wh/kg\n",
    "battery_cell_mass = 0.029 # kg\n",
    "battery_cell_nominal_voltage = 3.7 # V\n",
    "battery_high_water_mark = 0.05 # Can charge battery to 95% at winter solstice. Leaves room for lost capacity\n",
    "battery_low_water_mark = 0.08 # Don't deplete battery under 8% for reserve power\n",
    "\n",
    "# .036 for digital, .11 for AC: http://batteryuniversity.com/learn/article/how_to_measure_internal_resistance\n",
    "battery_cell_internal_resistance = 0.11 # Ohms\n",
    "\n",
    "motor_specific_power = 5000 # W/kg\n",
    "\n",
    "# Most MPPTs on the market have 95-99% efficiency\n",
    "mppt_efficiency = 0.98\n",
    "\n",
    "# Brushless motors typically have 95-98% efficiency\n",
    "motor_efficiency = 0.97\n",
    "\n",
    "# Other losses in wires and electronics\n",
    "processing_efficiency = 0.96\n",
    "\n",
    "prop_mass_per_diameter = 0.5 # kg / m / prop\n",
    "\n",
    "max_prop_ratio = 0.9 # Max fraction of span propellers take up\n",
    "\n",
    "solar_wing_coverage = 0.96 # Fraction of wing covered in solar cells\n",
    "solar_cell_mass_per_area = 0.2686 # 30g / (219mm * 510mm)\n",
    "\n",
    "avionics_power = 5 # W\n",
    "\n",
    "# Inmarsat SB-UAV: https://www.inmarsat.com/wp-content/uploads/2016/09/Inmarsat_Global_Government_SB-UAV-_June_2017_EN_lowres.pdf\n",
    "satcom_mass = 1.45 # kg\n",
    "satcom_power = 28.0 # W\n",
    "satcom_duty_cycle = 0.1 # 10% duty cycle\n",
    "\n",
    "# IAI MiniPOP: https://www.iai.co.il/p/minipop\n",
    "payload_mass = 8.0 # kg\n",
    "payload_power = 80.0 # W, this is a guess\n",
    "\n",
    "# POP300D: https://starkaerospace.com/products-services/sensors/pop300/\n",
    "# payload_mass = 16.3 # kg\n",
    "# payload_power = 120.0 # W\n",
    "\n",
    "\n",
    "# 3-5AM and 10AM-4PM usage\n",
    "payload_usage_1_start = 3.0\n",
    "payload_usage_1_end = 5.0\n",
    "payload_usage_2_start = 10.0\n",
    "payload_usage_2_end = 16.0\n",
    "\n",
    "max_max_engine_power = 50 * motor_specific_power # 50 kg motor max\n",
    "max_props = 30\n",
    "\n",
    "# The lower the min altitude, the more gravitational potential energy can be stored\n",
    "min_altitude = 1000\n",
    "max_altitude = 50000\n",
    "\n",
    "day_of_year = 355\n",
    "latitude = 40\n",
    "\n",
    "nt = 2000 # Time steps over 1 day\n",
    "\n",
    "likely_region = {\n",
    "    'starting_altitude': (min_altitude, max_altitude),\n",
    "    'battery_starting_energy_frac': (0.3, 0.7),\n",
    "    'battery_mass': (1, max_max_engine_power * max_props * 0.1 / battery_specific_energy),\n",
    "    'prop_diameter': (0.1, 10.0),\n",
    "    'wing_area': (1, 100),\n",
    "    'sweep_angle': (0, np.pi / 4),\n",
    "    'aspect_ratio': (3, 20),\n",
    "    'num_props': (4, 12), # At least two props for redundancy\n",
    "    'num_pusher_props': (0, 12),\n",
    "    # Engine power constants\n",
    "    # They define engine power throughout the day\n",
    "    'e1': (1, max_max_engine_power / 30),\n",
    "    'e2': (0, 400),\n",
    "    'e3': (0, 50),\n",
    "    'e4': (-6, 6),\n",
    "    'e5': (0, 100),\n",
    "    'e6': (0, 24),\n",
    "    # Airspeed constants\n",
    "    # They define airspeed throughout the day\n",
    "    'a1': (1, 20),\n",
    "    'a2': (0, 10),\n",
    "    'a3': (0, 10),\n",
    "    'a4': (-12, 12),\n",
    "    'a5': (0, 10),\n",
    "    'a6': (0, 24),\n",
    "}\n",
    "\n",
    "bounds = {\n",
    "    'starting_altitude': (min_altitude, max_altitude),\n",
    "    'battery_starting_energy_frac': (0.1, 0.8),\n",
    "    'battery_mass': (1, max_max_engine_power * max_props * 14 / battery_specific_energy),\n",
    "    'prop_diameter': (0.1, 10.0),\n",
    "    'wing_area': (0.01, 1000),\n",
    "    'sweep_angle': (0, np.pi / 2 * 0.99),\n",
    "    'aspect_ratio': (0.01, 50),\n",
    "    'num_props': (1, max_props),\n",
    "    'num_pusher_props': (0, max_props),\n",
    "    # Engine power constants\n",
    "    'e1': (1, max_max_engine_power / 2),\n",
    "    'e2': (0, max_max_engine_power / 2),\n",
    "    'e3': (0, 100),\n",
    "    'e4': (-12, 12),\n",
    "    'e5': (0, 100),\n",
    "    'e6': (0, 24),\n",
    "    # Airspeed constants\n",
    "    'a1': (1, 150),\n",
    "    'a2': (0, 150),\n",
    "    'a3': (0, 100),\n",
    "    'a4': (-12, 12),\n",
    "    'a5': (0, 100),\n",
    "    'a6': (0, 24)\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Constants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "R = 8314.46 # J/(kmol*K)\n",
    "gamma = 1.40 # air\n",
    "M = 28.895 # kg/kmol, air\n",
    "gravitational_constant = 6.674e-11\n",
    "earth_mass = 5.972e24 # kg\n",
    "earth_radius = 6375000 # m\n",
    "\n",
    "# Atmospheric data\n",
    "# From http://www.digitaldutch.com/atmoscalc/tableatmosphere.htm\n",
    "atm_data = pd.read_csv('st_atm_1976.csv').set_index('Altitude [m]')\n",
    "atmosphere = atm_data.reindex(np.arange(atm_data.index[0], atm_data.index[-1] + 1)).interpolate().values\n",
    "\n",
    "T_ground, P_ground, rho_ground, c_ground, u_ground = atmosphere[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Prepare the solar irradiance at different altitudes and times of day\n",
    "We pre-calculate the solar irradiance rather than calculating the solar irradiance itself\n",
    "every time it runs. We use the [Python airmass module](https://github.com/gusgordon/airmass).\n",
    "\n",
    "Different profiles can be generated with generate_irradiance_graph.ipynb."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "solar_graph = pd.read_csv(f'solar_graph_day_{day_of_year}_lat_{latitude}.csv', index_col='second').values\n",
    "\n",
    "@jit\n",
    "def solar_irradiance(altitude, second_of_day):\n",
    "    altitude = max(altitude, 0)\n",
    "    return solar_graph[int(round(second_of_day / 60)), int(round(altitude / 25))]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Wing, propeller thrust, etc. functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "@jit\n",
    "def swept_wing(Cl, wing_area, aspect_ratio, sweep_angle, Ma):\n",
    "    \"\"\"\n",
    "    Compute the coefficient of drag of the given wing.\n",
    "    This also supports the supersonic case, but I don't think that will be useful here :).\n",
    "    Induced drag is derived from \"Three-dimensional wings of minimum pressure drag\" by\n",
    "    Robert T. Jones. This material can be found at\n",
    "    https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19760011971.pdf page 661\n",
    "    The wing is treated as a yawed ellipse. See diagram in paper for variables.\n",
    "    \"\"\"\n",
    "\n",
    "    # Supersonic case, treat the wing as a yawed ellipse. Still need a, b for subsonic\n",
    "    psi = np.pi / 2 - sweep_angle\n",
    "    a_prime = 2 * np.sqrt(wing_area) / (np.pi * np.sqrt(aspect_ratio))\n",
    "    b_prime = np.sqrt(wing_area * aspect_ratio) / 2\n",
    "\n",
    "    # Solve for b given a', b', and angle of yaw\n",
    "    # http://bit.ly/2lY0T04, 4th solution\n",
    "    if b_prime**2 - 4 * a_prime**2 * np.sin(psi)**2 * np.cos(psi)**2 > 0:\n",
    "        within = b_prime ** 2 + np.sqrt(b_prime ** 2 * (b_prime**2 - 4 * a_prime**2 * np.sin(psi)**2 * np.cos(psi)**2))\n",
    "        b = np.sqrt((1 / np.sin(psi))**2 * within) / np.sqrt(2)\n",
    "    else:\n",
    "        return 1e50, 1, 1\n",
    "\n",
    "    a = a_prime * b_prime / b\n",
    "    wingspan = b_prime * 2 * np.sin(psi)\n",
    "\n",
    "    # Subsonic induced drag\n",
    "    # http://www.fzt.haw-hamburg.de/pers/Scholz/OPerA/OPerA_PUB_DLRK_12-09-10.pdf\n",
    "    if sweep_angle > np.pi / 6: # > 30 degrees, swept wing (Raymer)\n",
    "        e = 4.61 * (1 - 0.045 * aspect_ratio ** 0.68) * np.cos(sweep_angle) ** 0.15 - 3.1\n",
    "    else: # Unswept wing (Raymer)\n",
    "        e = 1.78 * (1 - 0.045 * aspect_ratio ** 0.68) - 0.64\n",
    "    if e < 0:\n",
    "        return 1e50, 1, 1\n",
    "\n",
    "    Cd_Cl2 = 1 / (np.pi * e * aspect_ratio)\n",
    "\n",
    "    # Karman-Tsien compressibility correction\n",
    "    beta = np.sqrt(np.abs(1 - Ma ** 2))\n",
    "    Cd_Cl2 = Cd_Cl2 / (beta + Ma ** 2 / (1 + beta) * Cd_Cl2 / 2)\n",
    "\n",
    "    # Supersonic pressure drag\n",
    "    if Ma > 1:\n",
    "        m = (b ** 2 - a ** 2) * np.sin(psi) * np.cos(psi) / b_prime ** 2\n",
    "\n",
    "        # There is a mistake in Jones' formula, a and b should be primed\n",
    "        Cd_Cl2 += 1 / 4 * np.sqrt(Ma ** 2 - 1 - (m + 1j * a_prime / b_prime) ** 2).real\n",
    "\n",
    "    Cd = Cd_Cl2 * Cl ** 2\n",
    "\n",
    "    wing_length = a_prime * 2\n",
    "\n",
    "    return Cd, wing_length, wingspan\n",
    "\n",
    "\n",
    "# Compute drag. The two sources of drag here are lift induced drag and skin friction drag.\n",
    "# Lift induced drag is computed from the output of swept_wing, skin friction drag is\n",
    "# the drag due to air rubbing against the wing.\n",
    "@jit\n",
    "def compute_drag(rho=0, v=0, c=0, L=0, dynamic_viscosity=0, T=0, p=0,\n",
    "                 wing_area=0, sweep_angle=0, aspect_ratio=0, prop_turb_length=0):\n",
    "    # Compute drag on a subsonic plane\n",
    "\n",
    "    Ma = v / c\n",
    "\n",
    "    ## Lift induced drag\n",
    "    Cl = 2 * L / (wing_area * rho * v ** 2) # Calculate required Cl\n",
    "    Cd, wing_length, wingspan = swept_wing(Cl, wing_area, aspect_ratio, sweep_angle, Ma)\n",
    "    induced_drag = 0.5 * Cd * wing_area * rho * v ** 2\n",
    "\n",
    "    ## Skin friction drag\n",
    "    # http://mail.tku.edu.tw/095980/drag.pdf\n",
    "    # Assume boundary layer is turbulent, which also leads to pressure drag being negligible\n",
    "    Re = v * wing_length * rho / dynamic_viscosity\n",
    "    Twall = T * (1 + 0.178 * Ma**2)\n",
    "    Tprime = T * (1 + 0.035 * Ma**2 + 0.45 * (Twall / T - 1))\n",
    "    Reprime = Re * (((T + 120) / (Tprime + 120)) * (T / Tprime) ** 1.5)\n",
    "\n",
    "    # Laminar: https://ipfs.io/ipfs/QmXoypizjW3WknFiJnKLwHCnL72vedxjQkDDP1mXWo6uco/wiki/Skin_friction.html\n",
    "    Cfprime_lam = 1.328 / np.sqrt(Reprime)\n",
    "    Cfprime_lam = Cfprime_lam * 1.04 # Roughness and gap scaling\n",
    "    Cf_lam = T / Tprime * Cfprime_lam\n",
    "\n",
    "    # Turbulent\n",
    "    Cfprime_turb = 0.455 / np.log10(Reprime) ** 2.58\n",
    "    Cfprime_turb = Cfprime_turb * 1.07 # Roughness and gap scaling\n",
    "    Cf_turb = T / Tprime * Cfprime_turb\n",
    "\n",
    "    # Fraction of laminar flow over the wing = non-prop span * frac of downwash (over non-prop span) that's laminar\n",
    "    turb_span_frac = prop_turb_length / wingspan\n",
    "    lam_span_frac = 1 - turb_span_frac\n",
    "    lam_turb_transition_loc = 5e5 * dynamic_viscosity / (rho * v)\n",
    "    laminar_downwash_frac = min(1, lam_turb_transition_loc / wing_length)\n",
    "    frac_laminar_total = lam_span_frac * laminar_downwash_frac\n",
    "\n",
    "    Cf = Cf_lam * frac_laminar_total + Cf_turb * (1 - frac_laminar_total)\n",
    "\n",
    "    S_total = wing_area * 2 * 1.07 # Scale slightly\n",
    "\n",
    "    parasitic_drag = 0.5 * Cf * S_total * rho * v ** 2\n",
    "\n",
    "    total_drag = induced_drag + parasitic_drag\n",
    "    return total_drag, wingspan\n",
    "\n",
    "\n",
    "# Compute the mass of the wing of a given area\n",
    "@jit\n",
    "def compute_wing_mass(wing_area):\n",
    "    # This part is sketchy\n",
    "    # Use Facebook Aquila: 43m wingspan, 400kg mass\n",
    "    # https://en.wikipedia.org/wiki/Facebook_Aquila\n",
    "    # Using example plane generated below with typical AR\n",
    "    # 10.12m wingspan, 6.53m wing area.\n",
    "    # Aquila wing area estimate: (43 / 10.12)^2 * 6.5 = 117.35 m^2\n",
    "    # this means 117.35 * 0.42 = 49 kg of solar cells\n",
    "    # ~40% total wing weight, - 49/400 = 28%\n",
    "    # Assume wing structure is 28% of mass\n",
    "    # Scale up by 20% for error\n",
    "    # wing mass = 400 * 0.28 * 1.2 = 134.4\n",
    "    # 134.4 * (this wing area / aquila wing area) ** (3 / 2)\n",
    "    # Assume a power law between ^1 and ^2 for relation\n",
    "    # of wing area and mass. 2.25 is a conservative guess.\n",
    "    # Lower is more conservative, because it means relatively more\n",
    "    # mass for a smaller wing.\n",
    "\n",
    "    aquila_wing_area = 117.35 # m^2\n",
    "    aquila_wing_mass = 134.4 # kg\n",
    "    return aquila_wing_mass * (wing_area / aquila_wing_area) ** (2.25 / 2)\n",
    "\n",
    "\n",
    "# Compute solar cell efficiency. Cell efficiency is dependent upon cell temperature,\n",
    "# the lower the temperature the more efficient we are.\n",
    "@jit\n",
    "def solar_cell_efficiency(T_cells):\n",
    "    # Alta Devices Anylight\n",
    "    # https://www.altadevices.com/wp-content/uploads/2019/10/HALE-HAPS-Solar-Products.pdf\n",
    "    # Assume there is no temperature dependence; GaAs cell temp dependence should be low\n",
    "    # The cell is the top layer of the wing, so Alta's figures include the cell coating\n",
    "\n",
    "    raw_eff = 0.23 # https://www.altadevices.com/wp-content/uploads/2019/10/HALE-HAPS-Solar-Products.pdf\n",
    "    pessimistic_scale = 0.95 # Assume cells are a little worse than advertised\n",
    "\n",
    "    total_eff = raw_eff * pessimistic_scale\n",
    "\n",
    "    return total_eff\n",
    "\n",
    "\n",
    "# Compute the thrust of a single propeller\n",
    "@jit\n",
    "def propeller_thrust(rho, airspeed, power, prop_diameter, pusher=False):\n",
    "    # Thrust of a propeller using actuator disk theory\n",
    "    # http://webserver.dmt.upm.es/~isidoro/bk3/c17/Propellers.pdf\n",
    "\n",
    "    # Scale down to account for friction, swirl losses, etc.\n",
    "    if pusher:\n",
    "        # Pusher props act in wing downwash\n",
    "        # https://en.wikipedia.org/wiki/Pusher_configuration#Propeller_efficiency_and_noise\n",
    "        other_losses = 0.79\n",
    "    else:\n",
    "        # Tractors act in undisturbed flow\n",
    "        other_losses = 0.85\n",
    "\n",
    "    if power == 0:\n",
    "        return 0, other_losses\n",
    "\n",
    "    airspeed = max(airspeed, 0.01)\n",
    "\n",
    "    A = np.pi * (prop_diameter / 2) ** 2 # Propeller frontal area\n",
    "\n",
    "    K = 2 * power / (rho * airspeed ** 3 * A)\n",
    "    inner = (9 * K ** 2 + 1.73205 * np.sqrt(27 * K ** 4 + 16 * K ** 3)) ** (1/3)\n",
    "    eta = (0.605707 * inner) / K - 2.20128 / inner\n",
    "\n",
    "    adj_eta = eta * other_losses\n",
    "    return power / airspeed * adj_eta, adj_eta\n",
    "\n",
    "\n",
    "# The sigmoid function is used to control the airspeed + thrust throughout the day\n",
    "@jit\n",
    "def sigmoid(x):\n",
    "    if x < -50:\n",
    "        return 0.0\n",
    "    elif x > 50:\n",
    "        return 1.0\n",
    "    else:\n",
    "        return 1 / (1 + np.exp(-x))\n",
    "\n",
    "\n",
    "@jit\n",
    "def compute_materials_cost(max_battery_energy_j, max_motor_power_w,\n",
    "                           solar_cell_efficiency, solar_wing_coverage,\n",
    "                           wing_area_m2, prop_diameter_m, num_props):\n",
    "    max_solar_irradiation = 1200.0 # W / m^2\n",
    "\n",
    "    battery_pack_cost = 1000.0 # $ / kWh, complete guess\n",
    "    solar_cell_cost = 5.00 # $ / peak W, complete guess\n",
    "\n",
    "    motor_cost = 100.0 # $ / kW\n",
    "    body_structural_cost = 300.0 # $ / m^2\n",
    "    prop_structural_cost = 200.0 # $ / m\n",
    "    fixed_costs = 10000.0 # $\n",
    "\n",
    "    peak_solar_power = wing_area_m2 * solar_wing_coverage * max_solar_irradiation * solar_cell_efficiency\n",
    "\n",
    "    return (\n",
    "        battery_pack_cost * max_battery_energy_j / (3600 * 1000) +\n",
    "        solar_cell_cost * peak_solar_power +\n",
    "        motor_cost / 1000 * max_motor_power_w +\n",
    "        body_structural_cost * wing_area_m2 +\n",
    "        prop_structural_cost * prop_diameter_m * num_props +\n",
    "        fixed_costs\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Helios Prototype drag test (HP01)\n",
    "https://www.nasa.gov/centers/armstrong/news/FactSheets/FS-068-DFRC.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "drag power: 16352.922471354654\n",
      "motor power: 16688.0\n",
      "error: -2.008%\n"
     ]
    }
   ],
   "source": [
    "altitude = 30480 # meters\n",
    "airspeed = 75.9968 # m/s\n",
    "engine_power = 1490 # W\n",
    "n_motors = 14\n",
    "motor_eff = 0.8\n",
    "\n",
    "T, P, rho, c, u = atmosphere[int(altitude)]\n",
    "\n",
    "computed_drag, a = compute_drag(rho=rho, v=airspeed, c=c, L=5880.549,\n",
    "                                dynamic_viscosity=u, T=T, p=P,\n",
    "                                wing_area=183.5764,\n",
    "                                sweep_angle=0, aspect_ratio=30.9, prop_turb_length=28.09 * 1.3)\n",
    "drag_power = computed_drag * airspeed\n",
    "engine_power = n_motors * engine_power * motor_eff\n",
    "print('drag power:', drag_power)\n",
    "print('motor power:', engine_power)\n",
    "print('error: {0:.3f}%'.format((drag_power - engine_power) / engine_power * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that the error is low, so our calculations seem to be accurate, at least for this case."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model function\n",
    "This loops through a day (in this case, the worst case scenario, the winter solstice) and checks if the plane can sustain flight. In order for a plane to sustain flight, it must meet the objectives defined below (mainly, ending altitude > starting altitude, and ending battery energy > starting battery energy).\n",
    "\n",
    "If a plane sustains flight, its loss function (function that the optimizer is trying to minimize) is related to the wing area. If a plane cannot sustain flight, its loss function is related to how close it was to being able to sustain flight."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "@jit\n",
    "def compute_flight(starting_altitude, battery_starting_energy_frac, battery_mass,\n",
    "                   prop_diameter, wing_area, sweep_angle, aspect_ratio, num_props,\n",
    "                   num_pusher_props, e1, e2, e3, e4, e5, e6, a1, a2, a3, a4, a5, a6):\n",
    "\n",
    "    # Convert # of props to int\n",
    "    num_props = np.rint(num_props)\n",
    "    e3 = np.rint(e3)\n",
    "    a3 = np.rint(a3)\n",
    "    num_pusher_props = min(np.rint(num_pusher_props), num_props)\n",
    "\n",
    "    # Get plane mass\n",
    "    motor_mass = (e1 + e2) / motor_specific_power * num_props\n",
    "    wing_mass = compute_wing_mass(wing_area)\n",
    "    prop_mass = prop_diameter * num_props * prop_mass_per_diameter\n",
    "    solar_cell_mass = solar_cell_mass_per_area * wing_area\n",
    "\n",
    "    mass = payload_mass + satcom_mass + battery_mass + motor_mass + wing_mass + prop_mass + solar_cell_mass # Total mass\n",
    "\n",
    "    # Get turbulence span over wing due to tractor props\n",
    "    prop_turb_length = (num_props - num_pusher_props) * prop_diameter * 1.3 # Scale for tip vortices\n",
    "\n",
    "    # We don't want to fully charge/discharge batteries, because that leads to a reduced lifetime of the\n",
    "    # battery. So we only charge and discharge the battery upto a limit.\n",
    "    max_battery_energy = battery_mass * battery_specific_energy * 3600 * (1 - battery_high_water_mark)\n",
    "    min_battery_energy = battery_mass * battery_specific_energy * 3600 * battery_low_water_mark\n",
    "    num_battery_cells = battery_mass / battery_cell_mass\n",
    "\n",
    "    # Power required to maintain battery temperature\n",
    "    # We insulate the battery with an insulator (aerogel etc.) and a radiation shield\n",
    "    battery_heating_power = 20 * (num_battery_cells / 400) ** (2 / 3) # Roughly - derived in craft_temperature.py\n",
    "\n",
    "    # Initialize variables\n",
    "    battery_starting_energy = (min_battery_energy + max_battery_energy) * battery_starting_energy_frac\n",
    "    battery_energy = battery_starting_energy\n",
    "    altitude = starting_altitude\n",
    "    lowest_altitude = np.Infinity\n",
    "    lowest_battery_energy = np.Infinity\n",
    "    wingspan = np.Infinity\n",
    "    prop_efficiency = np.nan\n",
    "    battery_eff = np.nan\n",
    "    time_store = np.empty(nt)\n",
    "    altitude_store = np.empty(nt)\n",
    "    engine_power_store = np.empty(nt)\n",
    "    airspeed_store = np.empty(nt)\n",
    "    battery_energy_store = np.empty(nt)\n",
    "    battery_SOC_store = np.empty(nt)\n",
    "    solar_power_store = np.empty(nt)\n",
    "    prev_kinetic_energy = 0.0\n",
    "    dt = 24 * 3600 * 1 / nt\n",
    "\n",
    "    # Loop over the day in nt steps of size dt seconds\n",
    "    for i in range(nt):\n",
    "        t_hour = (24 * i / nt + 7) % 24 # 7 AM launch time\n",
    "        t = t_hour * 3600\n",
    "\n",
    "        engine_power_t = e1 + e2 * (\n",
    "            sigmoid(e3 * (t_hour - e4)) - sigmoid(e5 * (t_hour - (e4 + e6)))\n",
    "        )\n",
    "        engine_power_t = max(engine_power_t, 0.0)\n",
    "        true_airspeed = a1 + a2 * (\n",
    "            sigmoid(a3 * (t_hour - a4)) - sigmoid(a5 * (t_hour - (a4 + a6)))\n",
    "        )\n",
    "        true_airspeed = max(true_airspeed, 1.0)\n",
    "\n",
    "        ### Determine change in altitude over dt\n",
    "\n",
    "        # Get lift required at altitude\n",
    "        T, P, rho, c, u = atmosphere[int(max(altitude, 0))]\n",
    "        g_a = gravitational_constant * earth_mass / (earth_radius + altitude) ** 2\n",
    "        g_a -= true_airspeed ** 2 / (earth_radius + altitude)\n",
    "        lift_t = mass * g_a\n",
    "\n",
    "        # Get drag force at starting altitude\n",
    "        drag_t, wingspan = compute_drag(rho=rho, v=true_airspeed, c=c, L=lift_t,\n",
    "                                        dynamic_viscosity=u, T=T, p=P,\n",
    "                                        wing_area=wing_area,\n",
    "                                        sweep_angle=sweep_angle, aspect_ratio=aspect_ratio,\n",
    "                                        prop_turb_length=prop_turb_length)\n",
    "\n",
    "        if drag_t > 1e40:\n",
    "            battery_energy  = -1e10\n",
    "            altitude  = -1e3\n",
    "            lowest_altitude = -1e3\n",
    "            lowest_battery_energy = -1e3\n",
    "            break\n",
    "\n",
    "        # Get thrust from propellers\n",
    "        prop_thrust_t_trac = 0.0\n",
    "        prop_efficiency_trac = 0.0\n",
    "        prop_thrust_t_push = 0.0\n",
    "        prop_efficiency_push = 0.0\n",
    "        prop_output_power = engine_power_t * motor_efficiency * processing_efficiency\n",
    "\n",
    "        # A propeller is a \"pusher\" if it's sitting behind the wing, and a normal\n",
    "        # prop if it's sitting in front of the wing.\n",
    "        # A normal prop has the drawback of making a turbulent boundary layer in its downwash\n",
    "        # over the wing, while a pusher prop has the drawback of slightly lower efficiency\n",
    "        # due to it acting in the downwash of the wing sitting in front of it.\n",
    "        if num_props - num_pusher_props > 0:\n",
    "            prop_thrust_t_trac, prop_efficiency_trac = propeller_thrust(rho, true_airspeed,\n",
    "                                                                        prop_output_power,\n",
    "                                                                        prop_diameter, pusher=False)\n",
    "        if num_pusher_props > 0:\n",
    "            prop_thrust_t_push, prop_efficiency_push = propeller_thrust(rho, true_airspeed,\n",
    "                                                                        prop_output_power,\n",
    "                                                                        prop_diameter, pusher=True)\n",
    "\n",
    "        prop_thrust_t = prop_thrust_t_trac * (num_props - num_pusher_props) + prop_thrust_t_push * num_pusher_props\n",
    "        prop_efficiency = (prop_efficiency_trac * (num_props - num_pusher_props)\n",
    "                           + prop_efficiency_push * num_pusher_props) / num_props\n",
    "\n",
    "        # Determine solar power at dt\n",
    "        # Assume T_cells = T_air + 1, from get_craft_temperature function\n",
    "        solar_power_t = (solar_irradiance(altitude, t) * wing_area\n",
    "                         * solar_wing_coverage * mppt_efficiency * solar_cell_efficiency(T + 1))\n",
    "\n",
    "        # Compute the change in energy in this time step\n",
    "        energy_delta_dt = solar_power_t * processing_efficiency * dt - engine_power_t * num_props * dt\n",
    "        energy_delta_dt -= battery_heating_power * dt\n",
    "\n",
    "        if (\n",
    "            ((t_hour > payload_usage_1_start) and (t_hour < payload_usage_1_end))\n",
    "            or ((t_hour > payload_usage_2_start) and (t_hour < payload_usage_2_end))\n",
    "        ):\n",
    "            payload_power_used = payload_power\n",
    "        else:\n",
    "            payload_power_used = 0.0\n",
    "\n",
    "        energy_delta_dt -= (avionics_power + satcom_power * satcom_duty_cycle + payload_power_used) * dt\n",
    "\n",
    "        # Account for extra power delta of acceleration/deceleration\n",
    "        kinetic_energy = 0.5 * mass * true_airspeed ** 2\n",
    "        energy_delta_dt -= (kinetic_energy - prev_kinetic_energy) / prop_efficiency\n",
    "        prev_kinetic_energy = kinetic_energy\n",
    "\n",
    "        # Account for battery self heating. If the battery is inefficient, it is converting energy\n",
    "        # into heat anyways through its internal resistance\n",
    "        cell_power = energy_delta_dt / dt / num_battery_cells\n",
    "        battery_eff = 1 - np.abs(cell_power) * battery_cell_internal_resistance / (battery_cell_nominal_voltage ** 2)\n",
    "        battery_eff = max(1e-6, battery_eff) # Efficiency lower bound is just above 0 to avoid /0\n",
    "        battery_self_heating_frac = 1 - battery_eff\n",
    "\n",
    "        # Add back unncessary heating, due to battery self-heating from inefficiency\n",
    "        energy_delta_dt += min(np.abs(energy_delta_dt) * battery_self_heating_frac, battery_heating_power * dt)\n",
    "\n",
    "        if energy_delta_dt > 0:\n",
    "            # Battery efficiency = 1 - Pcell * R / V^2, goes up with internal resistance\n",
    "            battery_energy = min(max_battery_energy, battery_energy + energy_delta_dt * battery_eff)\n",
    "        else:\n",
    "            battery_energy += energy_delta_dt * (1 / battery_eff)\n",
    "\n",
    "        # Compute the change in altitude of the airplane in the current time step\n",
    "        # https://en.wikipedia.org/wiki/Lift-to-drag_ratio\n",
    "        # Get the climb angle from the free body diagram of a climbing plane \n",
    "        # drag * cos(theta) + lift * sin(theta) = thrust * cos(theta), solve for theta, then get vertical speed\n",
    "        angle_of_climb = np.arctan((prop_thrust_t - drag_t) / lift_t)\n",
    "        vertical_airspeed = true_airspeed * np.sin(angle_of_climb) # Vertical true airspeed\n",
    "        altitude_delta = vertical_airspeed * dt\n",
    "\n",
    "        # Update altitude\n",
    "        altitude += altitude_delta\n",
    "        altitude = max(min(altitude, max_altitude), -1e50)\n",
    "\n",
    "        # Update our lowest altitude and battery energy\n",
    "        # One of our objectives should be that the lowest battery energy and altitude does not go past\n",
    "        # the lower bounds we have defined, otherwise the plane is invalid\n",
    "        lowest_altitude = min(lowest_altitude, altitude)\n",
    "        lowest_battery_energy = min(lowest_battery_energy, battery_energy)\n",
    "\n",
    "        time_store[i] = t / 3600\n",
    "        altitude_store[i] = altitude\n",
    "        engine_power_store[i] = engine_power_t * num_props\n",
    "        airspeed_store[i] = true_airspeed\n",
    "        battery_energy_store[i] = battery_energy\n",
    "        battery_SOC_store[i] = battery_energy / (max_battery_energy / (1 - battery_high_water_mark))\n",
    "        solar_power_store[i] = solar_power_t\n",
    "\n",
    "    objectives = np.array([\n",
    "            (battery_energy - battery_starting_energy) / battery_starting_energy, # Battery energy > battery starting energy]\n",
    "            (altitude - starting_altitude) / starting_altitude, # Altitude > starting altitude\n",
    "            (lowest_altitude - min_altitude) / min_altitude, # Lowest altitude > minimum altitude\n",
    "            (lowest_battery_energy - min_battery_energy) / min_battery_energy, # Lowest battery energy > minimum battery energy\n",
    "            (wingspan * max_prop_ratio - prop_diameter * num_props) / (prop_diameter * num_props) # Props diameter < wingspan * ratio\n",
    "        ])\n",
    "    energy_diff = battery_energy - battery_starting_energy + (altitude - starting_altitude) * mass * 9.8\n",
    "\n",
    "    materials_cost = compute_materials_cost(max_battery_energy / (1 - battery_high_water_mark),\n",
    "                                            (e1 + e2) * num_props,\n",
    "                                            solar_cell_efficiency(270),\n",
    "                                            solar_wing_coverage,\n",
    "                                            wing_area, prop_diameter, num_props)\n",
    "\n",
    "    # If all of the objectives are met, then the loss is given as a function of the winspan\n",
    "    # If the objectives are not met, then our loss is based on the objectives\n",
    "    if np.all(objectives > 0) and wingspan < 1000:\n",
    "        loss = -1 * np.power((1000 - wingspan), 2)\n",
    "    else:\n",
    "        loss = np.sqrt(np.array([np.abs(min(x, 0)) ** 2 for x in objectives]).sum())\n",
    "\n",
    "    result = {\n",
    "        'adl_battery_energy': objectives[0],\n",
    "        'adl_altitude': objectives[1],\n",
    "        'min_altitude_delta': objectives[2],\n",
    "        'min_battery_energy_delta': objectives[3],\n",
    "        'prop_wingspan_frac': objectives[4],\n",
    "        'energy_diff': energy_diff,\n",
    "        'wingspan': wingspan,\n",
    "        'prop_efficiency': prop_efficiency,\n",
    "        'materials_cost': materials_cost,\n",
    "        'cell_mass_fraction': battery_cell_mass * num_battery_cells / mass,\n",
    "        'loss': loss\n",
    "    }\n",
    "\n",
    "    detail = {\n",
    "        'time': time_store,\n",
    "        'altitude': altitude_store,\n",
    "        'engine_power': engine_power_store,\n",
    "        'airspeed': airspeed_store,\n",
    "        'battery_energy': battery_energy_store,\n",
    "        'battery_SOC': battery_SOC_store,\n",
    "        'solar_power': solar_power_store,\n",
    "    }\n",
    "\n",
    "    return result, detail\n",
    "\n",
    "\n",
    "# Used for applying compute_flight on a DataFrame\n",
    "def compute_flight_apply(x):\n",
    "    return compute_flight(**x.to_dict())\n",
    "\n",
    "\n",
    "# Used to map the optimization function into compute_flight\n",
    "def compute_flight_opt(a, *k):\n",
    "    return compute_flight(*a)[0]['loss']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Optimization\n",
    "For the optimization, we define an initial population in the `likely_region` bounds (this makes the optimization much faster than initializing in the general `bounds`). Then we use scipy's `differential_evolution` to evolve to a maxima. We repeat the optimization a number of times so it's more likely we find the global maxima."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "found plane, loss = -982947.4437567672, wingspan = 8.562940093135676\n",
      "found plane, loss = -984023.6050688532, wingspan = 8.020360557307965\n",
      "found plane, loss = -983774.6794763653, wingspan = 8.145837596894776\n",
      "found plane, loss = -982130.31636834, wingspan = 8.975118189084036\n",
      "found plane, loss = -981941.895109539, wingspan = 9.070186587597387\n",
      "found plane, loss = 2.6656411595680476\n",
      "found plane, loss = -982730.1432411518, wingspan = 8.672534809434545\n",
      "found plane, loss = 2.6656411118276027\n",
      "found plane, loss = -981941.7722761738, wingspan = 9.070248566441611\n",
      "found plane, loss = 2.665641110669588\n"
     ]
    }
   ],
   "source": [
    "opt_vars = ['starting_altitude', 'battery_starting_energy_frac', 'battery_mass', 'prop_diameter',\n",
    "            'wing_area', 'sweep_angle', 'aspect_ratio', 'num_props', 'num_pusher_props',\n",
    "            'e1', 'e2', 'e3', 'e4', 'e5', 'e6', 'a1', 'a2', 'a3', 'a4', 'a5', 'a6']\n",
    "opt_bounds = [bounds[b] for b in opt_vars]\n",
    "\n",
    "popsize = 50\n",
    "\n",
    "planes = pd.DataFrame(columns=opt_vars)\n",
    "\n",
    "while len(planes) < 10:\n",
    "    new_population = np.array([np.random.uniform(likely_region[v][0],\n",
    "                                                 likely_region[v][1], popsize) for v in opt_vars]).T\n",
    "\n",
    "    o = optimize.differential_evolution(compute_flight_opt, strategy='randtobest1bin', maxiter=5000,\n",
    "                                        tol=0.0000001, bounds=opt_bounds, init=new_population,\n",
    "                                        updating='deferred', workers=2)\n",
    "\n",
    "    params = dict(zip(opt_vars, o.x))\n",
    "    params.update(compute_flight(*o.x)[0])\n",
    "\n",
    "    if params['loss'] < 0:\n",
    "        print(f\"found plane, loss = {params['loss']}, wingspan = {params['wingspan']}\")\n",
    "    else:\n",
    "        print(f\"found plane, loss = {params['loss']}\")\n",
    "\n",
    "    planes = planes.append(params, ignore_index=True).sort_values('loss')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Results\n",
    "Each row is a plane — the lower the loss function, the better the plane."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>starting_altitude</th>\n",
       "      <th>battery_starting_energy_frac</th>\n",
       "      <th>battery_mass</th>\n",
       "      <th>prop_diameter</th>\n",
       "      <th>wing_area</th>\n",
       "      <th>sweep_angle</th>\n",
       "      <th>aspect_ratio</th>\n",
       "      <th>num_props</th>\n",
       "      <th>num_pusher_props</th>\n",
       "      <th>e1</th>\n",
       "      <th>e2</th>\n",
       "      <th>e3</th>\n",
       "      <th>e4</th>\n",
       "      <th>e5</th>\n",
       "      <th>e6</th>\n",
       "      <th>a1</th>\n",
       "      <th>a2</th>\n",
       "      <th>a3</th>\n",
       "      <th>a4</th>\n",
       "      <th>a5</th>\n",
       "      <th>a6</th>\n",
       "      <th>adl_altitude</th>\n",
       "      <th>adl_battery_energy</th>\n",
       "      <th>cell_mass_fraction</th>\n",
       "      <th>energy_diff</th>\n",
       "      <th>materials_cost</th>\n",
       "      <th>min_altitude_delta</th>\n",
       "      <th>min_battery_energy_delta</th>\n",
       "      <th>prop_efficiency</th>\n",
       "      <th>prop_wingspan_frac</th>\n",
       "      <th>wingspan</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>loss</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>-984023.605069</th>\n",
       "      <td>8340.950267</td>\n",
       "      <td>0.203636</td>\n",
       "      <td>8.440625</td>\n",
       "      <td>2.099477</td>\n",
       "      <td>9.468062</td>\n",
       "      <td>0.523413</td>\n",
       "      <td>9.056752</td>\n",
       "      <td>1.034535</td>\n",
       "      <td>9.654243</td>\n",
       "      <td>8.803818</td>\n",
       "      <td>292.944023</td>\n",
       "      <td>3.207843</td>\n",
       "      <td>0.041453</td>\n",
       "      <td>54.147832</td>\n",
       "      <td>16.387799</td>\n",
       "      <td>10.548946</td>\n",
       "      <td>7.542752</td>\n",
       "      <td>1.013355</td>\n",
       "      <td>7.206817</td>\n",
       "      <td>1.411952</td>\n",
       "      <td>11.274720</td>\n",
       "      <td>0.000238</td>\n",
       "      <td>0.000064</td>\n",
       "      <td>0.286510</td>\n",
       "      <td>7.445349e+02</td>\n",
       "      <td>2.870947e+04</td>\n",
       "      <td>0.006198</td>\n",
       "      <td>0.001569</td>\n",
       "      <td>0.768307</td>\n",
       "      <td>2.438153</td>\n",
       "      <td>8.020361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>-983774.679476</th>\n",
       "      <td>7140.299001</td>\n",
       "      <td>0.242248</td>\n",
       "      <td>8.442995</td>\n",
       "      <td>2.002509</td>\n",
       "      <td>9.591591</td>\n",
       "      <td>0.523355</td>\n",
       "      <td>9.221415</td>\n",
       "      <td>1.441492</td>\n",
       "      <td>16.575821</td>\n",
       "      <td>54.534445</td>\n",
       "      <td>279.900962</td>\n",
       "      <td>1.897744</td>\n",
       "      <td>2.746250</td>\n",
       "      <td>53.841843</td>\n",
       "      <td>12.945797</td>\n",
       "      <td>10.678800</td>\n",
       "      <td>8.362649</td>\n",
       "      <td>1.409987</td>\n",
       "      <td>8.441618</td>\n",
       "      <td>1.051592</td>\n",
       "      <td>10.374409</td>\n",
       "      <td>0.001047</td>\n",
       "      <td>0.000227</td>\n",
       "      <td>0.285525</td>\n",
       "      <td>2.882764e+03</td>\n",
       "      <td>2.888686e+04</td>\n",
       "      <td>0.699171</td>\n",
       "      <td>0.000813</td>\n",
       "      <td>0.757815</td>\n",
       "      <td>2.661034</td>\n",
       "      <td>8.145838</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>-982947.443757</th>\n",
       "      <td>9125.159803</td>\n",
       "      <td>0.163698</td>\n",
       "      <td>9.406302</td>\n",
       "      <td>1.926394</td>\n",
       "      <td>9.263414</td>\n",
       "      <td>0.523556</td>\n",
       "      <td>10.553392</td>\n",
       "      <td>1.413713</td>\n",
       "      <td>9.431942</td>\n",
       "      <td>113.536907</td>\n",
       "      <td>131.887015</td>\n",
       "      <td>59.919291</td>\n",
       "      <td>0.087287</td>\n",
       "      <td>62.819482</td>\n",
       "      <td>16.220418</td>\n",
       "      <td>13.867481</td>\n",
       "      <td>58.286788</td>\n",
       "      <td>61.672137</td>\n",
       "      <td>-11.241450</td>\n",
       "      <td>32.940051</td>\n",
       "      <td>8.765106</td>\n",
       "      <td>0.000105</td>\n",
       "      <td>0.000494</td>\n",
       "      <td>0.312701</td>\n",
       "      <td>1.452619e+03</td>\n",
       "      <td>2.875102e+04</td>\n",
       "      <td>4.573163</td>\n",
       "      <td>0.001745</td>\n",
       "      <td>0.766889</td>\n",
       "      <td>3.000555</td>\n",
       "      <td>8.562940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>-982730.143241</th>\n",
       "      <td>8872.238915</td>\n",
       "      <td>0.175079</td>\n",
       "      <td>9.394063</td>\n",
       "      <td>1.227067</td>\n",
       "      <td>9.600282</td>\n",
       "      <td>0.523578</td>\n",
       "      <td>10.445671</td>\n",
       "      <td>2.235513</td>\n",
       "      <td>18.443152</td>\n",
       "      <td>42.650233</td>\n",
       "      <td>91.051647</td>\n",
       "      <td>32.795039</td>\n",
       "      <td>-0.622877</td>\n",
       "      <td>15.980976</td>\n",
       "      <td>16.954699</td>\n",
       "      <td>13.163120</td>\n",
       "      <td>47.542790</td>\n",
       "      <td>35.925629</td>\n",
       "      <td>-10.981908</td>\n",
       "      <td>26.266709</td>\n",
       "      <td>2.385884</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>0.000073</td>\n",
       "      <td>0.305557</td>\n",
       "      <td>1.990788e+02</td>\n",
       "      <td>2.937872e+04</td>\n",
       "      <td>3.056988</td>\n",
       "      <td>0.000020</td>\n",
       "      <td>0.756254</td>\n",
       "      <td>2.180461</td>\n",
       "      <td>8.672535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>-982130.316368</th>\n",
       "      <td>9905.870870</td>\n",
       "      <td>0.136028</td>\n",
       "      <td>10.069388</td>\n",
       "      <td>1.695253</td>\n",
       "      <td>9.060768</td>\n",
       "      <td>0.523592</td>\n",
       "      <td>11.853605</td>\n",
       "      <td>1.407405</td>\n",
       "      <td>1.520354</td>\n",
       "      <td>198.008882</td>\n",
       "      <td>33.524704</td>\n",
       "      <td>59.659577</td>\n",
       "      <td>8.502559</td>\n",
       "      <td>1.390236</td>\n",
       "      <td>2.930155</td>\n",
       "      <td>16.058436</td>\n",
       "      <td>72.396663</td>\n",
       "      <td>13.734441</td>\n",
       "      <td>-2.845718</td>\n",
       "      <td>59.589063</td>\n",
       "      <td>0.905682</td>\n",
       "      <td>0.000025</td>\n",
       "      <td>0.000082</td>\n",
       "      <td>0.331434</td>\n",
       "      <td>2.440992e+02</td>\n",
       "      <td>2.866275e+04</td>\n",
       "      <td>8.887568</td>\n",
       "      <td>0.000004</td>\n",
       "      <td>0.772559</td>\n",
       "      <td>3.764837</td>\n",
       "      <td>8.975118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>-981941.895110</th>\n",
       "      <td>10287.976083</td>\n",
       "      <td>0.134318</td>\n",
       "      <td>10.404073</td>\n",
       "      <td>1.700234</td>\n",
       "      <td>9.165353</td>\n",
       "      <td>0.523597</td>\n",
       "      <td>11.967984</td>\n",
       "      <td>1.244618</td>\n",
       "      <td>17.369874</td>\n",
       "      <td>205.861993</td>\n",
       "      <td>0.095568</td>\n",
       "      <td>32.818047</td>\n",
       "      <td>-5.563158</td>\n",
       "      <td>75.577933</td>\n",
       "      <td>2.541226</td>\n",
       "      <td>16.434068</td>\n",
       "      <td>91.579600</td>\n",
       "      <td>54.532029</td>\n",
       "      <td>-7.735652</td>\n",
       "      <td>29.290668</td>\n",
       "      <td>6.645642</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>0.337363</td>\n",
       "      <td>1.651917e+01</td>\n",
       "      <td>2.896309e+04</td>\n",
       "      <td>9.287976</td>\n",
       "      <td>0.000070</td>\n",
       "      <td>0.772372</td>\n",
       "      <td>3.801203</td>\n",
       "      <td>9.070187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>-981941.772276</th>\n",
       "      <td>10287.544528</td>\n",
       "      <td>0.134314</td>\n",
       "      <td>10.408114</td>\n",
       "      <td>1.704179</td>\n",
       "      <td>9.168762</td>\n",
       "      <td>0.523599</td>\n",
       "      <td>11.963724</td>\n",
       "      <td>1.042137</td>\n",
       "      <td>6.373486</td>\n",
       "      <td>205.951441</td>\n",
       "      <td>0.503303</td>\n",
       "      <td>67.125560</td>\n",
       "      <td>-11.074156</td>\n",
       "      <td>34.749066</td>\n",
       "      <td>5.810523</td>\n",
       "      <td>16.427935</td>\n",
       "      <td>18.464062</td>\n",
       "      <td>16.857261</td>\n",
       "      <td>-5.990832</td>\n",
       "      <td>60.948458</td>\n",
       "      <td>2.720989</td>\n",
       "      <td>0.000007</td>\n",
       "      <td>0.000031</td>\n",
       "      <td>0.337382</td>\n",
       "      <td>8.765915e+01</td>\n",
       "      <td>2.897091e+04</td>\n",
       "      <td>9.287545</td>\n",
       "      <td>0.000011</td>\n",
       "      <td>0.772423</td>\n",
       "      <td>3.790120</td>\n",
       "      <td>9.070249</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2.665641</th>\n",
       "      <td>49999.977383</td>\n",
       "      <td>0.800000</td>\n",
       "      <td>253010.466588</td>\n",
       "      <td>0.150522</td>\n",
       "      <td>822.986076</td>\n",
       "      <td>1.042359</td>\n",
       "      <td>42.190009</td>\n",
       "      <td>1.971297</td>\n",
       "      <td>18.646581</td>\n",
       "      <td>55846.242584</td>\n",
       "      <td>90754.771881</td>\n",
       "      <td>75.709896</td>\n",
       "      <td>0.787387</td>\n",
       "      <td>68.098846</td>\n",
       "      <td>11.916420</td>\n",
       "      <td>67.706104</td>\n",
       "      <td>69.824429</td>\n",
       "      <td>51.066364</td>\n",
       "      <td>-8.843161</td>\n",
       "      <td>93.255031</td>\n",
       "      <td>14.891156</td>\n",
       "      <td>-1.020000</td>\n",
       "      <td>-1.032106</td>\n",
       "      <td>0.994139</td>\n",
       "      <td>-4.486700e+11</td>\n",
       "      <td>1.063214e+08</td>\n",
       "      <td>-2.000000</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.989587</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2.665641</th>\n",
       "      <td>49999.935526</td>\n",
       "      <td>0.799997</td>\n",
       "      <td>253011.541118</td>\n",
       "      <td>0.573439</td>\n",
       "      <td>428.328439</td>\n",
       "      <td>0.555223</td>\n",
       "      <td>41.600543</td>\n",
       "      <td>1.475114</td>\n",
       "      <td>3.589191</td>\n",
       "      <td>53186.675446</td>\n",
       "      <td>87999.546982</td>\n",
       "      <td>13.338456</td>\n",
       "      <td>-5.970357</td>\n",
       "      <td>98.429927</td>\n",
       "      <td>3.228363</td>\n",
       "      <td>132.544145</td>\n",
       "      <td>147.035839</td>\n",
       "      <td>62.520717</td>\n",
       "      <td>6.480120</td>\n",
       "      <td>25.382617</td>\n",
       "      <td>17.812599</td>\n",
       "      <td>-1.020000</td>\n",
       "      <td>-1.032106</td>\n",
       "      <td>0.997124</td>\n",
       "      <td>-4.482898e+11</td>\n",
       "      <td>1.056916e+08</td>\n",
       "      <td>-2.000000</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.569478</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2.665641</th>\n",
       "      <td>49999.914433</td>\n",
       "      <td>0.799999</td>\n",
       "      <td>253009.830972</td>\n",
       "      <td>0.130476</td>\n",
       "      <td>820.336505</td>\n",
       "      <td>1.126781</td>\n",
       "      <td>35.153265</td>\n",
       "      <td>5.899467</td>\n",
       "      <td>0.677959</td>\n",
       "      <td>27478.617911</td>\n",
       "      <td>60421.731512</td>\n",
       "      <td>27.346233</td>\n",
       "      <td>9.787806</td>\n",
       "      <td>29.949649</td>\n",
       "      <td>6.719711</td>\n",
       "      <td>108.282749</td>\n",
       "      <td>145.018866</td>\n",
       "      <td>60.450976</td>\n",
       "      <td>-1.857763</td>\n",
       "      <td>16.698225</td>\n",
       "      <td>19.663537</td>\n",
       "      <td>-1.020000</td>\n",
       "      <td>-1.032106</td>\n",
       "      <td>0.993975</td>\n",
       "      <td>-4.486896e+11</td>\n",
       "      <td>1.063405e+08</td>\n",
       "      <td>-2.000000</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.149636</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                starting_altitude  battery_starting_energy_frac  \\\n",
       "loss                                                              \n",
       "-984023.605069        8340.950267                      0.203636   \n",
       "-983774.679476        7140.299001                      0.242248   \n",
       "-982947.443757        9125.159803                      0.163698   \n",
       "-982730.143241        8872.238915                      0.175079   \n",
       "-982130.316368        9905.870870                      0.136028   \n",
       "-981941.895110       10287.976083                      0.134318   \n",
       "-981941.772276       10287.544528                      0.134314   \n",
       " 2.665641            49999.977383                      0.800000   \n",
       " 2.665641            49999.935526                      0.799997   \n",
       " 2.665641            49999.914433                      0.799999   \n",
       "\n",
       "                 battery_mass  prop_diameter   wing_area  sweep_angle  \\\n",
       "loss                                                                    \n",
       "-984023.605069       8.440625       2.099477    9.468062     0.523413   \n",
       "-983774.679476       8.442995       2.002509    9.591591     0.523355   \n",
       "-982947.443757       9.406302       1.926394    9.263414     0.523556   \n",
       "-982730.143241       9.394063       1.227067    9.600282     0.523578   \n",
       "-982130.316368      10.069388       1.695253    9.060768     0.523592   \n",
       "-981941.895110      10.404073       1.700234    9.165353     0.523597   \n",
       "-981941.772276      10.408114       1.704179    9.168762     0.523599   \n",
       " 2.665641       253010.466588       0.150522  822.986076     1.042359   \n",
       " 2.665641       253011.541118       0.573439  428.328439     0.555223   \n",
       " 2.665641       253009.830972       0.130476  820.336505     1.126781   \n",
       "\n",
       "                aspect_ratio  num_props  num_pusher_props            e1  \\\n",
       "loss                                                                      \n",
       "-984023.605069      9.056752   1.034535          9.654243      8.803818   \n",
       "-983774.679476      9.221415   1.441492         16.575821     54.534445   \n",
       "-982947.443757     10.553392   1.413713          9.431942    113.536907   \n",
       "-982730.143241     10.445671   2.235513         18.443152     42.650233   \n",
       "-982130.316368     11.853605   1.407405          1.520354    198.008882   \n",
       "-981941.895110     11.967984   1.244618         17.369874    205.861993   \n",
       "-981941.772276     11.963724   1.042137          6.373486    205.951441   \n",
       " 2.665641          42.190009   1.971297         18.646581  55846.242584   \n",
       " 2.665641          41.600543   1.475114          3.589191  53186.675446   \n",
       " 2.665641          35.153265   5.899467          0.677959  27478.617911   \n",
       "\n",
       "                          e2         e3         e4         e5         e6  \\\n",
       "loss                                                                       \n",
       "-984023.605069    292.944023   3.207843   0.041453  54.147832  16.387799   \n",
       "-983774.679476    279.900962   1.897744   2.746250  53.841843  12.945797   \n",
       "-982947.443757    131.887015  59.919291   0.087287  62.819482  16.220418   \n",
       "-982730.143241     91.051647  32.795039  -0.622877  15.980976  16.954699   \n",
       "-982130.316368     33.524704  59.659577   8.502559   1.390236   2.930155   \n",
       "-981941.895110      0.095568  32.818047  -5.563158  75.577933   2.541226   \n",
       "-981941.772276      0.503303  67.125560 -11.074156  34.749066   5.810523   \n",
       " 2.665641       90754.771881  75.709896   0.787387  68.098846  11.916420   \n",
       " 2.665641       87999.546982  13.338456  -5.970357  98.429927   3.228363   \n",
       " 2.665641       60421.731512  27.346233   9.787806  29.949649   6.719711   \n",
       "\n",
       "                        a1          a2         a3         a4         a5  \\\n",
       "loss                                                                      \n",
       "-984023.605069   10.548946    7.542752   1.013355   7.206817   1.411952   \n",
       "-983774.679476   10.678800    8.362649   1.409987   8.441618   1.051592   \n",
       "-982947.443757   13.867481   58.286788  61.672137 -11.241450  32.940051   \n",
       "-982730.143241   13.163120   47.542790  35.925629 -10.981908  26.266709   \n",
       "-982130.316368   16.058436   72.396663  13.734441  -2.845718  59.589063   \n",
       "-981941.895110   16.434068   91.579600  54.532029  -7.735652  29.290668   \n",
       "-981941.772276   16.427935   18.464062  16.857261  -5.990832  60.948458   \n",
       " 2.665641        67.706104   69.824429  51.066364  -8.843161  93.255031   \n",
       " 2.665641       132.544145  147.035839  62.520717   6.480120  25.382617   \n",
       " 2.665641       108.282749  145.018866  60.450976  -1.857763  16.698225   \n",
       "\n",
       "                       a6  adl_altitude  adl_battery_energy  \\\n",
       "loss                                                          \n",
       "-984023.605069  11.274720      0.000238            0.000064   \n",
       "-983774.679476  10.374409      0.001047            0.000227   \n",
       "-982947.443757   8.765106      0.000105            0.000494   \n",
       "-982730.143241   2.385884      0.000005            0.000073   \n",
       "-982130.316368   0.905682      0.000025            0.000082   \n",
       "-981941.895110   6.645642      0.000002            0.000005   \n",
       "-981941.772276   2.720989      0.000007            0.000031   \n",
       " 2.665641       14.891156     -1.020000           -1.032106   \n",
       " 2.665641       17.812599     -1.020000           -1.032106   \n",
       " 2.665641       19.663537     -1.020000           -1.032106   \n",
       "\n",
       "                cell_mass_fraction   energy_diff  materials_cost  \\\n",
       "loss                                                               \n",
       "-984023.605069            0.286510  7.445349e+02    2.870947e+04   \n",
       "-983774.679476            0.285525  2.882764e+03    2.888686e+04   \n",
       "-982947.443757            0.312701  1.452619e+03    2.875102e+04   \n",
       "-982730.143241            0.305557  1.990788e+02    2.937872e+04   \n",
       "-982130.316368            0.331434  2.440992e+02    2.866275e+04   \n",
       "-981941.895110            0.337363  1.651917e+01    2.896309e+04   \n",
       "-981941.772276            0.337382  8.765915e+01    2.897091e+04   \n",
       " 2.665641                 0.994139 -4.486700e+11    1.063214e+08   \n",
       " 2.665641                 0.997124 -4.482898e+11    1.056916e+08   \n",
       " 2.665641                 0.993975 -4.486896e+11    1.063405e+08   \n",
       "\n",
       "                min_altitude_delta  min_battery_energy_delta  prop_efficiency  \\\n",
       "loss                                                                            \n",
       "-984023.605069            0.006198                  0.001569         0.768307   \n",
       "-983774.679476            0.699171                  0.000813         0.757815   \n",
       "-982947.443757            4.573163                  0.001745         0.766889   \n",
       "-982730.143241            3.056988                  0.000020         0.756254   \n",
       "-982130.316368            8.887568                  0.000004         0.772559   \n",
       "-981941.895110            9.287976                  0.000070         0.772372   \n",
       "-981941.772276            9.287545                  0.000011         0.772423   \n",
       " 2.665641                -2.000000                 -1.000000              NaN   \n",
       " 2.665641                -2.000000                 -1.000000              NaN   \n",
       " 2.665641                -2.000000                 -1.000000              NaN   \n",
       "\n",
       "                prop_wingspan_frac  wingspan  \n",
       "loss                                          \n",
       "-984023.605069            2.438153  8.020361  \n",
       "-983774.679476            2.661034  8.145838  \n",
       "-982947.443757            3.000555  8.562940  \n",
       "-982730.143241            2.180461  8.672535  \n",
       "-982130.316368            3.764837  8.975118  \n",
       "-981941.895110            3.801203  9.070187  \n",
       "-981941.772276            3.790120  9.070249  \n",
       " 2.665641                 1.989587  1.000000  \n",
       " 2.665641                 0.569478  1.000000  \n",
       " 2.665641                 0.149636  1.000000  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "planes.set_index('loss')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Details\n",
    "\n",
    "Summary data and some plots for the top vehicle."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "starting_altitude                 8340.950267\n",
       "battery_starting_energy_frac         0.203636\n",
       "battery_mass                         8.440625\n",
       "prop_diameter                        2.099477\n",
       "wing_area                            9.468062\n",
       "sweep_angle                          0.523413\n",
       "aspect_ratio                         9.056752\n",
       "num_props                            1.034535\n",
       "num_pusher_props                     9.654243\n",
       "adl_battery_energy                   0.000064\n",
       "adl_altitude                         0.000238\n",
       "min_altitude_delta                   0.006198\n",
       "min_battery_energy_delta             0.001569\n",
       "prop_wingspan_frac                   2.438153\n",
       "energy_diff                        744.534873\n",
       "wingspan                             8.020361\n",
       "prop_efficiency                      0.768307\n",
       "materials_cost                   28709.472305\n",
       "cell_mass_fraction                   0.286510\n",
       "loss                           -984023.605069\n",
       "Name: Plane 0, dtype: float64"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA8YAAAOXCAYAAADmQ1iTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi40LCBodHRwOi8vbWF0cGxvdGxpYi5vcmcv7US4rQAAIABJREFUeJzs3XeYVdXVx/HvYmhSVBRCkC4iCkQRRsWuKIpYsCIoir4qGFsSKyQmGEuCMUAiKjYIEA2KnSioaFCjiEpTEAsDgg4iEJBiowzr/WOfMTc49Uw59878Ps9znnvOPuWum4d4Z92999rm7oiIiIiIiIhUVzWSDkBEREREREQkSUqMRUREREREpFpTYiwiIiIiIiLVmhJjERERERERqdaUGIuIiIiIiEi1psRYREREREREqrUKS4zNrKWZzTCzRWb2gZn9Imrfzcymm9ni6LVR1G5mdpeZ5ZjZ+2bWNeVZA6PrF5vZwJT2bma2ILrnLjOzivo8IiIi1ZGZ1TWzd8zsvej7/PdR+3gz+9TM5kdbl6i90O9zERGRdFWRPcbbgGvdvSPQHbjCzDoCQ4BX3L098Ep0DHAi0D7aBgFjICTSwDDgYOAgYFh+Mh1dc2nKfb0q8POIiIhUR5uBHu6+P9AF6GVm3aNz17t7l2ibH7UV+H0uIiKSziosMXb3le4+N9rfBHwINAf6ABOiyyYAp0X7fYCJHswCdjWzZsAJwHR3X+fuXwHTCV/KzYCd3X2WuzswMeVZIiIiUg6i7+Wvo8Na0eZF3FLY97mIiEjaqpQ5xmbWBjgAeBto6u4ro1NfAk2j/ebA5ym35UZtRbXnFtAuIiIi5cjMssxsPrCa8GP129Gp26Ph0qPMrE7UVtj3toiISNqqWdFvYGYNgCeBX7r7xtRpwO7uZlbUr87lFcMgwnAu6tev322fffap6LcUEZFqYs6cOf9x9yZJx1GR3D0P6GJmuwJPm1lnYCjhB+7awAPAjcAtJX2mvptFRKSixPlurtDE2MxqEZLiR9z9qah5lZk1c/eV0dCq1VH7CqBlyu0torYVwNE7tL8atbco4PofcfcHCF/aZGdn++zZs8vwqURERP7LzJYnHUNlcff1ZjYD6OXuf46aN5vZ34DrouPCvs93fJa+m0VEpELE+W6uyKrUBowFPnT3kSmnpgD5laUHAs+mtF8QVbPsDmyIhly/CBxvZo2iolvHAy9G5zaaWffovS5IeZaIiIiUAzNrEvUUY2Y7AT2Bj/LnDUffwacBC6NbCvs+FxERSVsV2WN8GHA+sCCalwTwa2A4MNnMLgaWA32jc1OB3kAO8C1wEYC7rzOzW4F3o+tucfd10f7lwHhgJ2BatImIiEj5aQZMMLMswg/qk939OTP7l5k1AQyYD1wWXV/g97mIiEg6q7DE2N3fIHxZFuTYAq534IpCnjUOGFdA+2ygcxnCFBERkSK4+/uEApo7tvco5PpCv89FRETSVaVUpRYRERERERFJV0qMRUREREREpFpTYiwiIiIiIiLVmhJjERERERERqdaUGIuIiIiIiEi1psRYREREREREqjUlxiIiIiIiIlKtKTEWERERERGRak2JsYiIiIiIiFRrSoxFRERERESkWlNiLCIiIiIiItWaEmMRERERERGp1pQYi4iIiIiISLWmxFhERERERESqNSXGIiIiIiIiUq0pMRYREREREZFqTYmxiIiIiIiIVGtKjEVERERERKRaU2IsIiIiIiIi1ZoSYxEREREREanWlBiLiIiIiIhItabEWERERERERKo1JcYiIiIiIiJSrVVYYmxm48xstZktTGl7zMzmR9syM5sftbcxs+9Szt2Xck83M1tgZjlmdpeZWdS+m5lNN7PF0WujivosIiIiIiIiUnVVZI/xeKBXaoO7n+PuXdy9C/Ak8FTK6SX559z9spT2McClQPtoy3/mEOAVd28PvBIdi4iIiIiIiJRKhSXG7v46sK6gc1Gvb19gUlHPMLNmwM7uPsvdHZgInBad7gNMiPYnpLSLiIiIiIiIlFhSc4yPAFa5++KUtrZmNs/MXjOzI6K25kBuyjW5URtAU3dfGe1/CTSt0IhFRERERESkSqqZ0Pv25397i1cCrdx9rZl1A54xs04lfZi7u5l5YefNbBAwCKBVq1YxQxYREREREZGqqNJ7jM2sJnAG8Fh+m7tvdve10f4cYAmwN7ACaJFye4uoDWBVNNQ6f8j16sLe090fcPdsd89u0qRJeX4cERGRKs3M6prZO2b2npl9YGa/j9rbmtnbUXHMx8ysdtReJzrOic63STJ+ERGRkkiix/g44CN3/2GItJk1Ada5e56Z7UkosrXU3deZ2UYz6w68DVwAjI5umwIMBIZHr89W5ocQERGpJjYDPdz9azOrBbxhZtOAa4BR7v5otJrExYSCmRcDX7n7XmbWD7gDOCep4Etr27ZtbNy4kU2bNvH111//8Lp161a2bt3Ktm3bfthSj92dUA6F2K877pdFtIiH7k/g/kyOPdPvb9++PUcffXSZ3l+qrwpLjM1sEnA00NjMcoFh7j4W6MePi24dCdxiZluB7cBl7p5fuOtyQoXrnYBp0QYhIZ5sZhcDywnFvERERKQcRcUvv44Oa0WbAz2Ac6P2CcDNhMS4T7QP8ARwt5mZl1fGVw62bdvGvHnzeO+991i4cCGffPIJX3zxBStXrmTNmjXllpyKSOUbP348AwcOTDoMyUAVlhi7e/9C2i8soO1JwvJNBV0/G+hcQPta4NiyRSkiIiLFMbMsYA6wF3APYcrTenffFl2SWhyzOfA5gLtvM7MNwO7Afyo16B188803PPXUU0yePJnXXnuNTZs2AVCvXj06dOhAy5YtOfjgg2nWrBmNGjWiYcOGNGjQgIYNG1K/fn3q1KlDzZo1qVmzJrVq1fqf/aysLGrUCLPT8nu74r7uuB9HWRN73R///kyOPdPv3759OyeeeKISY4ktqeJbIiIikiHcPQ/oYma7Ak8D+5T1mZVVGHPdunWMGDGC0aNHs2nTJlq3bs15551Hjx49yM7OpnXr1j8ktSKS2Q4//HAmT56cdBiSoZQYi4iISIm4+3ozmwEcAuxqZjWjXuPU4pgrgJZAblRwcxdgbQHPegB4ACA7O7vcxy67O4899hhXXXUVa9eu5ayzzuKqq67isMMOUyIsUkW1bduWr776im+++Yb69esnHY5kGH0ziIiISKHMrEnUU4yZ7QT0BD4EZgBnRZelFsHML45JdP5flT2/eOvWrQwePJj+/fvTtm1b5s+fz+TJkzniiCOUFItUYT/5yU8AWL260MVqRAqlbwcREREpSjNghpm9D7wLTHf354AbgWvMLIcwh3hsdP1YYPeo/RpgSGUGu2XLFvr06cODDz7I0KFDmTlzJvvtt19lhiAiCdl9990BWLv2R4NURIqlodQiIiJSKHd/HziggPalwEEFtH8PnF0Jof1IXl4eAwYMYNq0adx///0MGjQoiTBEJCENGzYEQrE9kdJSj7GIiIhUCcOHD+fxxx/nzjvvVFIsUg01aNAAgK+//rqYK0V+TImxiIiIZLw333yTYcOG0b9/f6699tqkwxGRBOT3GCsxljiUGIuIiEhG27JlC5dccgmtWrXivvvuK/M6wCKSmfIrUSsxljg0x1hEREQy2ujRo/noo4947rnn2HnnnZMOR0QSUqdOHQA2b96ccCSSidRjLCIiIhlrw4YN3HrrrfTu3ZuTTjop6XBEJEFKjKUslBiLiIhIxrr77rt/SI5FpHpTYixlocRYREREMtK3337LqFGjOOmkk+jatWvS4YhIwmrXrg2EugMipaXEWERERDLS5MmTWbt2Ldddd13SoYhIGqhRowY1a9ZUj7HEosRYREREMtL999/PPvvsw1FHHZV0KCKSJmrXrq3EWGJRYiwiIiIZZ8GCBcyaNYvBgwdreSYR+UGdOnWUGEssSoxFREQk40yaNImsrCwGDBiQdCgikkZq166tOcYSixJjERERySjuzmOPPcZxxx1H48aNkw5HRNJIzZo1ycvLSzoMyUBKjEVERCSjzJ07l6VLl9K3b9+kQxGRNFOzZk22bduWdBiSgZQYi4iISEZ56qmnyMrK4rTTTks6FBFJM0qMJS4lxiIiIpJRXnjhBQ499FB22223pEMRkTSjxFjiUmIsIiIiGWPVqlXMnTuXE044IelQRCQNKTGWuJQYi4iISMaYPn06gBJjESlQVlaWEmOJRYmxiIiIZIwXXniBxo0b07Vr16RDEZE0pB5jiavCEmMzG2dmq81sYUrbzWa2wszmR1vvlHNDzSzHzD42sxNS2ntFbTlmNiSlva2ZvR21P2ZmtSvqs4iIiEjy3J2XX36Znj17UqOGftsXkR/Tck0SV0V+q4wHehXQPsrdu0TbVAAz6wj0AzpF99xrZllmlgXcA5wIdAT6R9cC3BE9ay/gK+DiCvwsIiIikrCcnBxWrVrFUUcdlXQoIpKm1GMscVVYYuzurwPrSnh5H+BRd9/s7p8COcBB0Zbj7kvdfQvwKNDHzAzoATwR3T8B0JoNIiIiVdjMmTMBOOywwxKORETSlRJjiSuJcUhXmtn70VDrRlFbc+DzlGtyo7bC2ncH1rv7th3aRUREpIp688032XXXXenYsWPxF4tItaTEWOKq7MR4DNAO6AKsBEZUxpua2SAzm21ms9esWVMZbykiIiLl7M033+SQQw7R/GIRKZQSY4mrUr9Z3H2Vu+e5+3bgQcJQaYAVQMuUS1tEbYW1rwV2NbOaO7QX9r4PuHu2u2c3adKkfD6MiIiIVJp169axaNEiDj300KRDEZE0lpWVxdatW5MOQzJQpSbGZtYs5fB0IL9i9RSgn5nVMbO2QHvgHeBdoH1Ugbo2oUDXFHd3YAZwVnT/QODZyvgMIiIiUvlmzZoFaH6xiBQtKyuLkCqIlE7N4i+Jx8wmAUcDjc0sFxgGHG1mXQAHlgGDAdz9AzObDCwCtgFXuHte9JwrgReBLGCcu38QvcWNwKNmdhswDxhbUZ9FREQkk5nZIcAA4AigGfAd4cfp54GH3X1DguGVyJw5cwDIzs5OOBIRSWc1atRg+/btSYchGajCEmN3719Ac6HJq7vfDtxeQPtUYGoB7Uv571BsERERKYCZTQO+IIysuh1YDdQF9gaOAZ41s5HuPiW5KIs3b9482rdvT8OGDZMORUTSmBJjiUvVK0RERKq28939Ynef4u5fuPs2d//a3ee6+wh3PxqYWdjNZtbSzGaY2SIz+8DMfhG132xmK8xsfrT1TrlnqJnlmNnHZnZCeXyIefPmccABB5THo0SkClNiLHEpMRYREanaBpjZgSkFK3/E3f9TxP3bgGvdvSPQHbjCzPLXSxrl7l2ibSpAdK4f0AnoBdxrZlll+QDr1q1j2bJldO3atSyPEZFqoEaNGuTl5SUdhmQgJcYiIiJVWwvgr8BqM3vNzP5gZieb2W4ludndV7r73Gh/E/Ah0LyIW/oAj7r7Znf/FMihjFOf5s+fD6AeYxEplnqMJS4lxiIiIlWYu1/n7ocCPwWGAuuAi4CFZraoNM8yszbAAcDbUdOVZva+mY0zs0ZRW3Pg85Tbcik6kS7WvHnzACXGIlK8rKwsJcYSixJjERGR6mEnYGdgl2j7gv8muMUyswbAk8Av3X0jMAZoB3QBVgIjShOMmQ0ys9lmNnvNmjVFXjt37lxatGhBkyZNSvMWIlINqcdY4qqwqtQiIiKSPDN7gDDfdxMhEZ4JjHT3r0rxjFqEpPgRd38KwN1XpZx/EHguOlwBtEy5vUXU9j/c/QHgAYDs7OwiFx1dsGAB+++/f0nDFZFqTImxxKUeYxERkaqtFVAH+JKQoOYC60t6s5kZYbnFD919ZEp7s5TLTiesiwwwBehnZnXMrC3QHngnbvDbtm3j448/plOnTnEfISLViBJjiUs9xiIiIlWYu/eKkttOwKHAtUBnM1sHvOXuw4p5xGHA+cACM5sftf0a6G9mXQAHlgGDo/f7wMwmA4sIFa2vcPfYJWKXLl3Kli1b6NixY/EXi0i1p8RY4lJiLCIiUsW5uxOKba0HNkTbyYRq0UUmxu7+BmAFnJpaxD23A7fHDjjFokWhPti+++5bHo8TkSpOibHEpcRYRESkCjOzqwk9xYcCWwlzjGcC44AFCYZWIkqMRaQ0lBhLXEqMRUREqrY2wOPAr9x9ZcKxlNqiRYto2bIlDRs2TDoUEckASowlLiXGIiIiVdvv3P3roi4wswbFXZOURYsWaX6xiJSYEmOJS1WpRUREqrZnzWyEmR1pZvXzG81sTzO72MxeBHolGF+htm/fzkcffaTEWERKTImxxKUeYxERkSrM3Y81s96EqtGHmVkjQrXoj4HngYHu/mWSMRZm+fLlfPfdd5pfLCIlpsRY4lJiLCIiUsW5+1SKqCKdrj755BMAOnTokHAkIpIplBhLXBpKLSIiImkpJycHgPbt2ycciYhkCiXGEpcSYxEREUlLOTk51KtXj5/+9KdJhyIiGUKJscSlodQiIkX45htYsiRsOTnw6afwxRewciV8+SV06ADjx8MeeyQdqUjVk5OTw1577YWZJR2KiGSIGjVqkJeXl3QYkoGUGItItbd1a0h6P/gAPvrov0lwTk5IflM1agTNm0OzZrDXXvDMM3DeefDKK1BDY3AkjZnZ3939/OLa0klOTo4qUotIqajHWOJSYiwi1ca2bbB4cUiAFy0Krx98AJ98EpLjfM2bQ7t20Lt3eN1rr/Darh3suuv/PvPBB2HQIHjkETg/bdMLEQA6pR6YWRbQLaFYipWXl8fSpUs59dRTkw5FRDJIVlaWEmOJRYmxiFRJmzfDwoUwdy7Mmxde33sPvv8+nDeDtm2hUyc45ZTw2rEj7LMP1KtX8ve5+OKQHN9wA5x2GjRsWDGfRyQuMxsK/BrYycw25jcDW4AHEgusGJ9//jlbtmxR4S0RKRX1GEtcSoxFJOPl5YUe4Jkz4Z13QhK8cGHoIQbYeWfo2hUuvxz23x86dy59AlyYGjVg9Gjo3h1uuw3uuKPszxQpT+7+R+CPZvZHdx+adDwllV+Req+99ko4EhHJJEqMJS4lxiKScdavh1mz4K23QjL89tuwaVM4t/vu0K0bXHddSIa7dg09wxU5//fgg+HCC2HUqNCDvPfeFfdeImUwzcyO3LHR3V9PIpjiKDEWkThqRF/47q7CfVIqFZYYm9k44GRgtbt3jtruBE4hDN9aAlzk7uvNrA3wIfBxdPssd78suqcbMB7YCZgK/MLd3cx2Ax4D2gDLgL7u/lVFfR4RSc7atfDaazBjBrz6augNhpDs/uxnMGAAHHJI2Nq1C8OkK9sf/whPPgm/+hU8/3zlv79ICVyfsl8XOAiYA/RIJpyi5eTkULduXfZQyXcRKYX8xHj79u1kZWUlHI1kkorsMR4P3A1MTGmbDgx1921mdgcwFLgxOrfE3bsU8JwxwKXA24TEuBcwDRgCvOLuw81sSHR8YwH3i0iG2bABXn89JMIzZoS5we5h6PPhh8M554Qk+KCD0mdO709/CsOGhZ7q55+Hk05KOiKR/+Xup6Qem1lL4C8JhVOsxYsX065dux/+yBURKQklxhJXhSXG7v561BOc2vZSyuEs4KyinmFmzYCd3X1WdDwROI2QGPcBjo4unQC8ihJjkYzkDvPnw9SpMG1aGCK9fTvUqQOHHgq//z0cc0xIhGvXTjrawl11VSjE9atfwXHHhfhF0lgusG/SQRRm6dKltGvXLukwRCTDpCbGIqWR5Bzj/yMMhc7X1szmARuBm9z930Bzwhd3vtyoDaCpu6+M9r8EmlZwvCJSjtavh+nTQyI8bdp/1wvu1g2GDg2JZffuULdusnGWRu3a8Je/wIknwl//GipVi6QLMxsNeHRYA+gCzE0uosK5O8uWLaNHj7Qc5S0iaUyJscSVSGJsZr8BtgGPRE0rgVbuvjaaU/yMmXUq9AE7iOYce2HnzWwQMAigVatW8QMXkTL54gt45hl46qkwVzgvL6wLfPzxYc3gE04IQ5IzWa9ecOqpcOutYe6zpkdKGpmdsr8NmOTubyYVTFHWrVvH119/TZs2bZIORUQyjBJjiavSE2Mzu5BQlOtYd3cAd98MbI7255jZEmBvYAXQIuX2FlEbwCoza+buK6Mh16sLe093f4Borcbs7OxCE2gRKX9LlsDTT4fCVLNmhbYOHeD66+Hkk0NF55pVrD7+yJFhTeQhQ2DixOKvF6kM7j7BzGoD+xB6jj8u5pbELF++HECJsYiUmhJjiatSK1qYWS/gBuBUd/82pb2JmWVF+3sC7YGl0VDpjWbW3UK99QuAZ6PbpgADo/2BKe0ikrDPP4c//Qm6dIG99gpJ8JYtYZ3fRYvgo49CFefDDqt6STGEytjXXQd//3tYTkokHZhZb8KKEHcRimPmmNmJyUZVsGXLlgHQunXrZAMRkYyjxFjiqsjlmiYRimM1NrNcYBihCnUdYHq0rlj+skxHAreY2VZgO3CZu6+LHnU5/12uaVq0AQwHJpvZxcByoG9FfRYRKd769fDEE/DII2FpJfcwR3jkSDj9dKhuHT9Dh8KECXD11WGdZRXGlDQwEjjG3XMAzKwd8Dz//V5NG+oxFpG4lBhLXMUmxmZ2CDAAOAJoBnwHLCR8mT7s7hsKus/d+xfQPLaQa58Enizk3GygcwHta4Fji4tfRCrO5s1haaJHHoHnngu9wnvvDTffDOeeG3qLq6sGDeDOO8P/Dn/7G1xySdIRibApPymOLAU2JRVMUZYtW0bDhg1p1KhR0qGISIZRYixxFZkYm9k04AvCMOXbCfN46xLm/x4DPGtmI919SkUHKiLpY948GDs2JMTr10PTpvDzn4diU926QRgQIv36wb33ht7js84KhcZEEjTbzKYCkwlzjM8G3jWzMwDc/akkg0u1bNkyWrdujek/JiJSSkqMJa7ieozPd/f/7ND2NWF5h7nACDNrXCGRiUhaWbcuJMLjxoU1h+vUgTPPhAsugGOPrZpzhcvKDO66K/xYcPPNYSknkQTVBVYBR0XHawjTlE4hJMppkxgvX75cw6hFJJb8xDgvLy/hSCTTFPmn7I5JsZntnHqPu68rIHEWkSpi+3Z45ZXQO/z002GodNeucM890L8/aJRj8Q44AAYNgrvvhksvhU4lXohOpNw9tOPyTGZ2WDou2bRs2TKOOOKIpMMQkQykHmOJq0RVqc1ssJl9CbwPzIm22UXfJSKZ6rPPQg9n27ZhjeGXXoLBg8MQ6jlz4PLLlRSXxm23QcOG8ItfhKJkIgkZXcK2RK1fv54NGzaoIrWIxJIVVbtUYiylVdLBj9cBndU7LFJ15eWFBHjMmFBQyx169gzLLvXpA3XrJh1h5mrcGG69Fa66Ch59NPS2i1SWqIjmoUATM7sm5dTOQNrVS1dFahEpC/UYS1wlXcd4CfBtsVeJSMZZvRqGDw8VpHv3DksLDRkCS5fCiy/COecoKS4PP/85HHhg6DVeuzbpaKSaqQ00IPwY3jBl2wicVdzNZtbSzGaY2SIz+8DMfhG172Zm081scfTaKGo3M7vLzHLM7H0z61qaYPPXMFZiLCJxKDGWuEraYzwUmGlmbwOb8xvd/eoKiUpEKpQ7vPFG6B1+4gnYuhWOPhruuANOOw1q1046wqonKwseeigU4rrmmrDGsUhlcPfXgNfMbLy7L4/xiG3Ate4+18waAnPMbDpwIfCKuw83syHAEOBG4ESgfbQdDIyJXkskv8dYQ6lFJA4lxhJXSRPj+4F/AQsA/SsTyVAbN8LEiXDfffDBB7DLLmG+8ODBsO++SUdX9e23H9x4I9x+O5x3Xpi/LVKJxpvZj2a5u3uPom5y95XAymh/k5l9CDQH+gBHR5dNAF4lJMZ9gInu7sAsM9vVzJpFzynWsmXLqFevHo0ba9ELESk9JcYSV0kT41rufk3xl4lIOlq8GEaPhr/9Db7+OgzpHTs2rLNbr17S0VUvN90Ejz8efoxYuBDq1086IqlGrkvZrwucSegNLjEzawMcALwNNE1Jdr8Emkb7zYHPU27LjdpKlBh//vnntGrVSmsYi0gsSowlrpImxtPMbBDwT/53KPW6ColKRMrMHaZPh7/+FaZODcOj+/ULBaCys5OOrvqqWxcefBCOOgp+9zsYMSLpiKS6cPc5OzS9aWbvlPR+M2sAPAn80t03piau7u4F9UYX87xBwCCAVq1a/dCem5tLixYtSvMoEZEfKDGWuEpafKs/0TxjtFyTSFr75pswVLpTJzjhhLC80s03hyWYJkxQUpwOjjwSLrsM/vIXePfdpKOR6iIqlpW/NTazE4BdSnhvLUJS/Ii7PxU1rzKzZtH5ZsDqqH0F0DLl9hZR2/9w9wfcPdvds5s0afJDuxJjESkLJcYSV4l6jN29bUUHIiJls3w53HNP6I1cvz4UeZo4Efr2hTp1ko5OdjR8OEyZApdcArNnQ61aSUck1cAcwAEjDKH+FLi4uJssdA2PBT5095Epp6YAA4Hh0euzKe1XmtmjhKJbG0o6vzgvL4+VK1cqMRaR2JQYS1xFJsZmdri7v1HE+Z2BVu6+sNwjE5ESeeutMBz36afBDM44IywJdOih4VjS0y67wL33hirgd94Jv/510hFJVVeGH7kPA84HFpjZ/Kjt14SEeLKZXQwsB/pG56YCvYEcwlKPF5X0jVatWkVeXh7NmzePGaqIVHdKjCWu4nqMzzSzPwEvEH5pXkMo2LEXcAzQGri2QiMUkR/Zvj30Nt55J8ycCY0awfXXhwrTKVP1JM316QNnnw233AKnn67K4FKxouHQPweOjJpeBe53961F3Rf9QF7Yz2zHFnC9A1fEiTE3NxdAPcYiEpsSY4mryMTY3X9lZrsRKleeDTQDvgM+JHyZFtqbLCLl77vvwvDoESNCpek2beCuu+Cii6BBg6SjkzhGj4Z//QsuvBDefBNqlrQkokjpjQFqAfdGx+dHbZckFtEOlBiLSFkpMZa4iv0TLKo8/WC0iUgC1q4Nw25Hj4Y1a8L84UcfhTPPVCKV6Zo2hbvvhv79ww8eN96YdERShR3o7vunHP8NaT+CAAAgAElEQVTLzN5LLJoCKDEWkbLKT4zz8vISjkQyTUmrUotIApYuhSuvhJYtw9I+Bx4IM2aESsbnnKOkuKo455zwI8fvfgeLFiUdjVRheWbWLv/AzPYE0uovxxUrVlCnTh123333pEMRkQyVnxiHWR0iJac/q0XS0Pz58Mc/whNPQFYWDBgA114blmCSqscsjAh47TUYODAUVNOPHlIBrgdmmNlSwpzh1pSiMFZlyM3NpXnz5pgqB4pITPn//dBQaikt9RiLpJGZM+Gkk+CAA+CFF0JBrWXLYNw4JcVV3U9+Epbbmj07FFUTKW/u/grQHrgauAro4O4zko3qf2kNYxEpK/UYS1wlSozNrJ6Z/dbMHoyO25vZyRUbmkj14A4vvwzHHAOHHQZvvw233hrWJR4+HPbYI+kIpbL07RuqVN98MyzUInhSAdx9s7u/H22bk45nR0qMRaSsVHxL4ippj/HfgM3AIdHxCuC2ColIpJrYvh2efRa6d4eePeGTT2DkyNBDfNNNsOuuSUcoSbjnnrDG8YUXwtYiF9ERqVrcnRUrVmgNYxEpEw2llrhKmhi3c/c/AVsB3P1bCl/TUESKkJcHkybB/vvDaaeFKtP33RcKbf3qV1p2qbpr0iTMN54zB/70p6SjkarCgpZJx1GUtWvXsnnzZvUYi0iZaCi1xFXSxHiLme0EOEBU1TLthmCJpLO8PHjkEejYEc49Nxz//e+hp3jwYKhTJ+kIJV2cdVaoVP3738P77ycdjVQFHv5CnJp0HEXRUk0iUh40lFriKmliPAx4AWhpZo8ArwA3FHeTmY0zs9VmtjClbTczm25mi6PXRlG7mdldZpZjZu+bWdeUewZG1y82s4Ep7d3MbEF0z12mMpaShrZvD2sOd+4cqkvXrh2qTS9cGI5VfVgKcvfdsNtu4d/IZv0MKeVjrpkdmHQQhclPjDWUWkTKQkOpJa4SJcbuPh04A7gQmARku/urJbh1PNBrh7YhwCvu3p6QYA+J2k8kVMtsDwwCxkBIpAmJ+cHAQcCw/GQ6uubSlPt2fC+RxGzfDo8/DvvtB/37Q40aMHkyvPdeWLO2hmrCSxEaN4axY2HBgjDnXKQcHAy8ZWZLoh+gF5hZ2oxJWLFiBaAeYxEpG/UYS1xF9lWl9tpGVkavrcyslbvPLep+d3/dzNrs0NwHODranwC8CtwYtU+MhnvNMrNdzaxZdO10d18XxTQd6GVmrwI7u/usqH0icBowraiYRCra9u3wzDOhsvCCBbDPPmFO8dlnhzWJRUrqpJPgsstgxIiwf/TRSUckGe6EpAMoSm5uLllZWfz0pz9NOhQRyWCaYyxxFTeIc0T0WhfIBt4jFN3aD5jNf6tUl0ZTd89PsL8Emkb7zYHPU67LjdqKas8toF0kEe4wbRr85jcwfz7svXeYU3zOOUqIJb4//xleeQUGDgzzjXfZJemIJFO5+3IzOxxo7+5/M7MmQNqU+8vNzaVZs2Zk6T+YIlIG6jGWuIoczOnux7j7MYSe4q7unu3u3YADCEs2lUnUO1zhP+eY2SAzm21ms9esWVPRbyfV0FtvwVFHhV69jRth4kT44INQZEt/40lZ1K8PDz8MK1bAlVcmHY1kMjMbRhihNTRqqgU8nFxE/ys3N1fzi0WkzDTHWOIq6SzHDu6+IP/A3RcC+8Z8z1XREGmi19VR+wogdSmJFlFbUe0tCmj/EXd/IErqs5s0aRIzbJEfW7QoLLl06KGhuvQ998CHH8L556uolpSfgw6C3/42JMiTJycdjWSw04FTgW8A3P0LoGGiEaVYsWKF5heLSJlpKLXEVdLE+H0ze8jMjo62B4G4BTumAPmVpQcCz6a0XxBVp+4ObIiGXL8IHG9mjaKiW8cDL0bnNppZ96ga9QUpzxKpUJ9/Dv/3f/Czn8GMGXDbbZCTA5dfHqpOi5S33/wGDj44zDleUebxOlJNbUkdqWVm9ROO53/k5uYqMRaRMtNQaomrpInxRcAHwC+ibVHUViQzmwS8BXQws1wzuxgYDvQ0s8XAcdExhPUVlwI5wIPA5QBR0a1bgXej7Zb8QlzRNQ9F9yxBhbekgn31FVx3HbRvD//4B/zyl7BkSUhaGqTNTD2pimrWDOteb94MF10UiryJlNJkM7sf2NXMLgVeJnzfJi4vL49NmzYpMRaRMtNQaomrRIM93f17YFS0lZi79y/k1LEFXOvAFYU8ZxwwroD22UDn0sQkEsfWrXDffaHS9Pr1oRDSzTdDq1ZJRybVSfv2MHJk6DW++264+uqkI5JM4u5/NrOewEagA/C7aDnGxG3duhXQGsYiUnYaSi1xlSgxNrNPKaBIlrvvWe4RiaQRd5g6NfQSf/QRHHtsSEz22y/pyKS6GjQInnsObrghLN+kf4tSUtGordfd/fqkY9nRli1bAK1hLCJlp6HUEldJywNlp+zXBc4Gdiv/cETSx8KFcO218NJLYemlKVPg5JMhGqEjkggzGDsW9t8f+vWD2bOhXr2ko5IM0Qq438zaAHOA14F/u/v8JIMCJcYiUn40lFriKtEcY3dfm7KtcPe/ACdVcGwiiVizBn7+85B4vPMOjBoFCxbAKacoKZb08JOfhArVH30U5rmLlIS7D3P3HkAn4N/A9YQEOXH5Q6n32GOPhCMRkUynodQSV0mHUndNOaxB6EHWYjRSpeTlwf33h0JamzbBFVfAsGGw++5JRybyY8ceCzfeCMOHQ8+ecPbZSUck6c7MbgIOAxoA84DrCAly4rZs2UKTJk2oU6dO0qGISIbTUGqJq6TJ7YiU/W3Ap0Df8g9HJBmzZoVEeO5c6NEjFDbaN+5K3SKV5JZbwnJhl14KBx4IbdokHZGkuTMI3+HPA68Bb7n75mRDCrZu3cqee6psiYiUnYZSS1wlXa7pYnc/Jtp6uvsgYEtFBiZSGdasgUsugUMOgS+/hMceg5dfVlIsmaFWLZg0KRSJO/dc2LYt6Ygknbl7V8Iyie8APYEFZvZGslEFW7Zs0fxiESkXGkotcZU0MX6ihG0iGSEvD8aMgQ4dYMIEuP76MF+zb1/NI5bM0rYtPPAAvPVWWEJMpDBm1hk4DxgInAOsAP6VaFCRLVu2aKkmESkXGkotcRU5lNrM9iEU6djFzM5IObUzoTq1SMaZOzcseTNnDhxzTBg23bFj0lGJxHfOOaF6+h/+EOYeH3NM0hFJmhpOmFN8F/Cuu29NOJ4f5OXlqcdYRMqFhlJLXMXNMe4AnAzsCpyS0r4JuLSighKpCN98E4ppjRoVqvpOmhQSCvUQS1Vw113w5pswYAC89x40bpx0RJJu3P1kM6sN7A10MLOP0yk5VmIsIuVBQ6klriKHUrv7s+5+EXCyu1+Usl3t7jMrKUaRMnvpJfjZz2DECLj4Yli0KKwBq6RYqor69eHRR+E//4ELLgD9UC47MrOjgMXAPcC9wCdmdmQJ7htnZqvNbGFK281mtsLM5kdb75RzQ80sx8w+NrMTShqfEmMRKQ8aSi1xFZkYm9kN0e65ZnbXjlslxCdSJv/5D5x/PpxwQihU9OqrYT5mo0ZJRyZS/rp0gb/8BaZNgzvuSDoaSUMjgePd/Sh3PxI4ARhVgvvGA70KaB/l7l2ibSqAmXUE+hGmYfUC7jWzrJIEpznGIlIelBhLXMUNpf4wep1d0YGIlCd3ePhh+NWvYMMGuOmmsD5xXc2Mlyrussvg9dfDv/lDD4Wjjko6Ikkjtdz94/wDd//EzGoVd5O7v25mbUr4Hn2AR6NloD41sxzgIOCt4m5UYiwi5UFzjCWuIhNjd/9ntPutuz+ees7Mzq6wqETK4LPPQnGtF1+E7t3hwQehc+ekoxKpHGZhVMTcuWG6wPz50LRp0lFJmphtZg8BD0fH51G2H76vNLMLomdc6+5fAc2BWSnX5EZtRcrKyqJhw4ZlCEVEJNAcY4mrpMs1DS1hm0hi3ENC0LkzvPFGKEb0xhtKiqX6adgQnngC1q8P6xvn5SUdkaSJnwOLgKujbVHUFscYoB3QBVgJjCjtA8xskJnNNrPZ+X/IioiUlYZSS1zFLdd0ItAbaL7DnOKdgW0VGZhIaSxfDpdcAi+/DD16wNix0KZN0lGJJOdnP4N774X/+z+45Rb4/e+TjkiS5u6bzexu4BVgO/Cxu2+J+axV+ftm9iDwXHS4AmiZcmmLqK2gZzwAPACwyy67qGtHRMqFhlJLXMX9RPsFMAf4PnrN36YQinaIJMod7rsv9ArPmhX2X35ZSbEIwEUXwYUXwq23hsrsUr2Z2UnAEuCvwN1ATvQDeJxnNUs5PB3Ir1g9BehnZnXMrC3QHninuOfVqlXsVGcRkRLRUGqJq7g5xu8B75nZw+6uHmJJK8uWhaWX/vUvOPZYeOghJcQiO7rnHpg9G847D+bNA62IU62NAI5x9xwAM2sHPA9MK+omM5sEHA00NrNcYBhwtJl1ARxYBgwGcPcPzGwyYZj2NuAKdy92ML8Kb4lIedFQaomruKHUCwhfej8MS8g/Bbi771dxoYkULL+X+PrrQ6Gh++4Lxba0JrHIj9WrF+YbZ2eHYlwzZoSly6Ra2pSfFEeWApuKu8nd+xfQPLaI628Hbi9NYOoxFpHyoqHUEldxyzWdXClRiJTQF1+E4aEvvQTHHRd6iVu3TjoqkfTWoUMoTHfuuTB0KPz5z0lHJAmZbWZTgcmEH73PBt41szMA3P2pJIMTESkPGkotcRU3lHp5Qe1mdjjQH7iiIoISKcjjj8PgwfD992F46M9/rl5ikZLq3z9UaR8xAg4+GM7WgnvVUV1gFZC/uvUaYCfgFEKirMRYRDKehlJLXMX1GP/AzA4AziX8wvwp+gKVSrJ+PVx1FTz8MBx4YHjde++koxLJPKNGhXnGF10EHTtCp05JRySVyd0vSjoGEZGKpqHUEleRVanNbG8zG2ZmHwGjgc8Ac/dj3H10pUQo1dqMGbDffjBpEtx8M7z5ppJikbhq1w7zjRs0gDPOgA0bko5IRESkfGkotcRV3HJNHwE9gJPd/fAoGS62uqRIWX3/PVx7bViTuG5dmDkThg1T0SCRstpjjzAtYelSGDgQ9IO6iIhUJRpKLXEVlxifAawEZpjZg2Z2LKEidWxm1sHM5qdsG83sl2Z2s5mtSGnvnXLPUDPLMbOPzeyElPZeUVuOmQ0pS1ySPubPDxV0R46Eyy8PQz8POijpqESqjiOOCAW4nn0W/vjHpKMREREpPxpKLXEVmRi7+zPu3g/YB5gB/BL4iZmNMbPj47yhu3/s7l3cvQvQDfgWeDo6PSr/nLtPBTCzjkA/oBPQC7jXzLLMLAu4BzgR6Aj0j66VDJWXB3fcEZLgtWth6tRQZKt+/aQjE6l6rr46VKn+7W/hxReTjkYqg5k1NbOxZjYtOu5oZhcnHZeISHnSUGqJq7geYwDc/Rt3/4e7nwK0AOYBN5bD+x8LLCms+nWkD/Cou29290+BHOCgaMtx96XuvgV4NLpWMtBnn4Vh00OGwKmnwoIFcOKJSUclUnWZhSWcOncOFas//TTpiKQSjAdeBPaIjj8h/OAtIlJlaCi1xFWixDiVu3/l7g+4+7Hl8P79gEkpx1ea2ftmNs7MGkVtzYHPU67JjdoKa5cM88QTsP/+MHcujB8f5j82bpx0VCJVX/368PTT4B6KcX37bdIRSQVr7O6Tge0A7r4N1Q0RkSpGQ6klrlInxuXFzGoDpwKPR01jgHZAF8K85hHl+F6DzGy2mc1es2ZNeT1Wyuibb+DSS8N6qu3bh7nEAwdqbWKRytSuXVgCbf78sDa4Rp5Vad+Y2e6ENYsxs+6AapOLSJWiodQSV2KJMWFu8Fx3XwXg7qvcPc/dtwMPEoZKA6wAWqbc1yJqK6z9R6Ie7mx3z27SpEk5fwyJY9486NYNxo4Nw6fffBP22ivpqESqp5NOCsuhTZwIo7UQX1V2DTAFaGdmbwITgauSDUlEpHypx1jiqpnge/cnZRi1mTVz95XR4enAwmh/CvAPMxtJmBfVHniHUB27vZm1JSTE/YBzKyl2iWn7dvjrX0My3LgxvPxymFssIsn67W9Dr/E110DHjnDccUlHJOXN3eea2VFAB8J36MfuvjXhsEREypWZYWZKjKXUEkmMzaw+0BMYnNL8JzPrQhjitSz/nLt/YGaTgUXANuAKd8+LnnMloZBIFjDO3T+otA8hpfbll3DhhaEC7qmnht5izSUWSQ81aoQe40MOgb594Z13NIqjijoIaEP4/u9qZrj7xGRDEhEpX0qMJY5EEmN3/wbYfYe284u4/nbg9gLapwJTyz1AKXfTpoWkeONGuPdeuOwyzSUWSTcNG8KUKXDggdCnD7z1Fuy8c9JRSXkxs78TannM579Ft5wwpFpEpMqoUaOG5hhLqSU5lFqqgc2b4cYbw/Dpzp3hlVfCq4ikpz33DJXhjz8eBgyAZ54JvclSJWQDHV1/LYpIFVejRg31GEup6c8dqTBLlsChh4ak+Morw9BMJcUi6a9HDxg1Cv75T/jd75KORsrRQuCnSQchIlLRNJRa4lCPsVSIxx+HSy4JPU3PPBOGZYpI5rjySnjvPbj9dthvvzDvWDJeY2CRmb0DbM5vdPdTkwtJRKT8aSi1xKHEWMrV99+HqrZjxsDBB8Njj0Hr1klHJSKlZQb33AMffhjqA7RvDwcckHRUUkY3Jx2AiEhl0FBqiUOJsZSbxYtDr9L8+XDttfCHP0Dt2klHJSJx1akDTz0F2dlh1Mfs2fCTnyQdlcTl7q8lHYOISGXQUGqJQ3OMpVw8+ih07QqffRaq2v75z0qKRaqCpk3DdIg1a+DMM0NBPcksZvZG9LrJzDambJvMbGPS8YmIlDcNpZY4lBhLmXz3XVh6qX9/+NnPYN48OOWUpKMSkfLUrRuMHw9vvAGXXgr6WyOzuPvh0WtDd985ZWvo7lqQS0SqHA2lljg0lFpi++QTOPtseP99uOEGuO02qFUr6ahEpCKcc074//zvfgcdOsBvfpN0RFJaZrZbAc2b3H1rpQcjIlKBNJRa4lBiLLE89VQoyFO7Njz/PPTunXREIlLRbropJMc33RSKcalSdcaZC7QEvgIM2BX40sxWAZe6+5wkgxMRKS8aSi1xaCi1lMq2baF3+MwzYZ99YO5cJcUi1YUZPPQQHHYYDBwIb7+ddERSStOB3u7e2N13B04EngMuB+5NNDIRkXKkodQShxJjKbFVq6BnT7jzzjCv+N//hlatko5KRCpTnTrw9NOwxx5w6qmwfHnSEUkpdHf3F/MP3P0l4BB3nwXUSS4sEZHypaHUEocSYymRmTND1elZs2DChLBOcR39GSVSLTVpAs89FypUn3wybFRd40yx0sxuNLPW0XYDsMrMsgD9BSkiVYaGUkscSoylSO4wejQcdRTUrRsS4wsuSDoqEUnavvvCE0/Ahx9Cv35hmoWkvXOBFsAz0dYqassCNGNcRKoMDaWWOJQYS6G+/hrOOw+uvhpOPBHmzIH99086KhFJF8cdB/feC9OmwTXXJB2NFMfd/+PuV7n7AdF2pbuvcfct7p6TdHwiIuVFQ6klDlWllgLl5MBpp4XeoNtvhyFDoIZ+RhGRHQwaFCpVjxgRKlVfdVXSEUlhzGxv4DqgDSnf/+7eo5j7xgEnA6vdvXPUthvwWPSsZUBfd//KzAz4K9Ab+Ba40N3nlvdnEREpioZSSxxKjOVHXnoprFlaowa88EIouCUiUpg77oAlS+AXv4AWLeD005OOSArxOHAf8BCQV4r7xgN3AxNT2oYAr7j7cDMbEh3fSKh03T7aDgbGRK8iIpVGQ6klDvUByg/cYeTIMGy6ZUt4910lxSJSvKwseOQROPhgOPdceOutpCOSQmxz9zHu/o67z8nfirvJ3V8H1u3Q3AeYEO1PAE5LaZ/owSxgVzNrVl4fQESkJJQYSxxKjAWA774LRbWuvTb09sycCXvumXRUIpIp6tWDKVNCj/Epp4Th1ZJ2/mlml5tZMzPbLX+L+aym7r4y2v8SaBrtNwc+T7kuN2oTEak0mmMscSgxFnJz4cgj4eGH4ZZbYPJkaNAg6ahEJNM0aRKmX9SoAb16hbXPJa0MBK4HZgJzom12WR/qYSJfqSfzmdkgM5ttZrPXrFlT1jBERH6gOcYShxLjam7mTMjOho8+gmeegd/+VkW2RCS+du3CGsdffhnWOP7mm6Qjknzu3raALe7YoFX5Q6Sj19VR+wqgZcp1LaK2guJ5wN2z3T27SZMmMcMQEfkxDaWWOJQCVWMPPQRHHx16h2fNgj59ko5IRKqCgw6Cxx6DuXNDIT+tcZwsM7shZf/sHc79IeZjpxB6oIlen01pv8CC7sCGlCHXIiKVQkOpJQ4lxtVQXl6oHnvppSExfucd6NQp6ahEpCo55ZSwxvHzz8MVV4TifpKYfin7Q3c416u4m81sEvAW0MHMcs3sYmA40NPMFgPHRccAU4GlQA7wIHB5GWMXESk1DaWWOLRcUzWzcSP07w9Tp8Ivfwl33gk19a9ARCrA4MHw2Wfwhz9A69bw618nHVG1ZYXsF3T8I+7ev5BTxxZwrQNXlDw0EZHyp6HUEkdiPcZmtszMFpjZfDObHbXtZmbTzWxx9Nooajczu8vMcszsfTPrmvKcgdH1i81sYGHvJ7B8ORx2GLz4IowZA6NGKSkWkYp1220wYAD85jfwt78lHU215YXsF3QsIpLxNJRa4kg6LTrG3f+TcjwEeMXdh5vZkOj4RuBEoH20HQyMAQ6OlpkYBmQTvtznmNkUd/+qMj9EJnj77TCH+PvvYdo0rU8sIpXDDMaOhdWrw/SN3XeHU09NOqpqZ38z20joHd4p2ic6rptcWCIiFUNDqSWOdJtj3AeYEO1PAE5LaZ/owSxg16gK5gnAdHdfFyXD0ynBfKnqZvLkMJe4Xj146y0lxSJSuWrXhiefhG7doG9feP31pCOqXtw9y913dveG7l4z2s8/rpV0fCIi5U1DqSWOJBNjB14yszlmNihqa5pSvfJLoGm03xz4POXe3KitsHYhFLu57bZQFTY7O/Qa77tv0lGJSHXUoEEoxNW2bSjMNX9+0hGJiEhVpaHUEkeSifHh7t6VMEz6CjM7MvVkVMCjXMZAmNkgM5ttZrPXrFlTHo9Me5s3wwUXhHWJBwyAl18GLRMpIklq3Bheegl23hl69YIlS5KOSEREqiINpZY4EkuM3X1F9LoaeBo4CFgVDZEmel0dXb4CaJlye4uorbD2Hd/rAXfPdvfsJtUgO1y3Do4/Hh5+GG69FSZOhDp1ko5KRARatgzJ8bZt4b9TK7XCrYiIlDMNpZY4EkmMzay+mTXM3weOBxYCU4D8ytIDgWej/SnABVF16u7AhmjI9YvA8WbWKKpgfXzUVm0tXw6HHw6zZsGkSXDTTaH4jYhIuth337Bk3KpVoed4/fqkIxIRkapEQ6kljqSqUjcFnraQsdUE/uHuL5jZu8BkM7sYWA70ja6fCvQGcoBvgYsA3H2dmd0KvBtdd4u7r6u8j5Fe5s2D3r1D5emXXoKjjko6IhGRgh10EDz9NJx0UqhS/eKLsNNOSUclIiJVgYZSSxyJJMbuvhTYv4D2tcCxBbQ7cEUhzxoHjCvvGDPNSy/BmWdCo0ZhPnGnTklHJCJStJ49w5SPfv/P3n2HSVmebx//niwtVAvlRUBBBP1ZUQmxF9TYxRqxxBIVe4yxp9gSY4zGkqhEbKBRiDUYu7H3iB07toiiooKgKLBwvX/cs3FByrLszj2zc36O4zl25pnn2T03E7n3mrsNSYsE3nwztPAayWZmtoQ8lNrqo9S2a7J6GDEi9br06ZOGULsoNrNy8ZOfwMUXw7/+BQccALNn505kZmblzkOprT5yDaW2BlCzHdOpp6ael5tuSqu9mpmVkyOOgC+/hF/9Ctq2hcsu89oIZmZWf+4xtvpwYVymZs1Kf0xecQXsvz9cfrmHIJpZ+TrlFJg2Dc4+O+15/Oc/uzg2M7P68Rxjqw8XxmXo66/T8MM770z7FJ9xhv+ANLPyd9ZZ8NVXcMEF0L59+rfNzMxscTVr1ozq6urcMazMuDAuM198ATvsAE8/nYYbDh2aO5GZWcOQ4MIL04d/Z56Zeo5POCF3KjMzKzeeY2z14cK4jHz4IWy9Nbz1Ftx4I+y6a+5EZmYNq1kzGD48FccnnpiK48MPz53KzMzKiYdSW324MC4Tb74JP/5x6jG++27YfPPciczMGkdVFVx7LUyfntZSaNsW9tsvdyozMysXXnzL6sPbNZWB556DjTZKfyQ++KCLYjNr+lq0gBtugC22gAMPTHscm5mZ1YWHUlt9uDAucQ8+CJttBm3awGOPwbrr5k5kZlYcrVvDP/8J660He+0Fd92VO5GZmZUDD6W2+nBhXMJuvRW22QaWXx4efxz69cudyMysuNq1gzvugDXWgF12gfvuy53IzMxKnYdSW324MC5RV1wBu++eeogfeQS6d8+dyMwsj6WWgnvvhZVXhp12ggceyJ3IzMxKmYdSW324MC5B55wDhxySVqC+7z5YZpnciczM8lp2Wfj3v2GlldKWdQ8/nDuRmZmVKvcYW324MC4hEfCb38DJJ6f5dGPGpNVYzcwMOneG+++H3r1h++3h0UdzJzIzs1JUVVXF7Nmzc8ewMuPCuEREwC9/CWedBQcfnLYqadEidyozs9LSpUsqjnv2hO22gyeeyJ3IzMxKTfPmzV0Y22JzYVwC5syBww+HCy+En/8chg9P+3iamdn3/b//l+YZL7dcWqDw6adzJzIzs1LSvHlzqqurc8ewMuPCOLPqajjgALjsMjjllFQcS7lTmZmVtm7dUnHcpQv8+MfwzDO5E5mZWalwYWz14cI4o5kz01zia6+F3/8e/vAHF8VmZnXVvXva633ZZVNx/NxzuROZmVkpqKqqcmFsi82FcSbfftETGxUAACAASURBVAu77go33QTnnw+//nXuRGZm5adnz1Qcd+wIW24JL7yQO5GZmeXmOcZWHy6MM/j6a9hxR7jjDhg2DI49NnciM7PytcIKqThu397FcbFJek/Sy5JekDS2cG4ZSfdJeqvwdencOc2ssngotdWHC+Mimzo1LRbzwAMwciQcdljuRGZm5a9371Qct2kDgwZ5WHWRbR4R/SNiQOH5ycD9EdEXuL/w3MysaFwYW324MC6iL76ALbaAp56C0aNhv/1yJzIzazpWXBEefhg6dEj/1npBrmwGAyMLj0cCO2fMYmYVyHOMrT5cGBfJJ5/AZpvBSy/BLbfAHnvkTmRm1vT07p2K46WXTsOqn3oqd6ImL4B7JT0raWjhXNeImFh4/DHQNU80M6tUnmNs9eHCuAgmTIBNN4Xx4+H229P8YjMzaxwrrJCK486d02rVjz+eO1GTtlFErANsCxwpaZPaL0ZEkIrn75E0VNJYSWMnTZpUhKhmVik8lNrqo+iFsaSekh6U9KqkVyQdUzh/uqQPCwt4vCBpu1r3nCJpvKQ3JG1d6/w2hXPjJZXkHKZ334VNNoGPPoJ77oGttsqdyMys6evZMxXH3brB1lvDI4/kTtQ0RcSHha+fArcCA4FPJHUDKHz9dAH3Do+IARExoHPnzsWKbGYVwEOprT5y9BhXA8dFxKrAeqRPmFctvHZBYQGP/hFxJ0DhtSHAasA2wKWSqiRVAZeQPqVeFdir1vcpCW++mYriyZPh3/+GjTfOncjMrHJ07w4PPZSK5G23TY+t4UhqK6l9zWPgx8A44DZg/8Jl+wNj8iQ0s0pVM5Q6DVoxq5uiF8YRMTEinis8nga8BnRfyC2DgdERMSMi3gXGkz6RHgiMj4h3ImImMLpwbUkYNy4VxTNmpD/GBg7MncjMrPJ065b+De7VC7bbDu6/P3eiJqUr8JikF4H/AHdExN3AH4GtJL0FbFl4bmZWNM2bNwdgzpw5mZNYOck6x1hSL2Bt4OnCqaMkvSTpqlr7HnYHPqh124TCuQWdz+7ZZ9Oc4qqqNJRvrbVyJzIzq1xdu6atnFZaCXbYIU1rsSVX+GB6rcKxWkScVTj/eURsERF9I2LLiPgid1Yzqyw1hbGHU9viyFYYS2oH3Az8IiKmAsOAPkB/YCLw5wb8WUVb4OOJJ9Iemu3bpzlt//d/jfrjzMysDrp0SfvHr7wyDB4Md96ZO5GZmTWWqqoqwIWxLZ4shbGkFqSi+LqIuAUgIj6JiNkRMQe4nDRUGuBDoGet23sUzi3o/PcUa4GPBx9MK6B26ZKK4j59Gu1HmZnZYurUKRXHq60Gu+wCt92WO5GZmTUG9xhbfeRYlVrAlcBrEXF+rfPdal22C2kBD0iLeAyR1EpSb6AvaS7TM0BfSb0ltSQt0JXtz5w77kiLu6ywQiqKl18+VxIzM1uQZZZJiyGutRbsthvccEPuRGZm1tBqCmPvZWyLo3mGn7kh8FPgZUkvFM79irSqdH/SfofvAYcCRMQrkm4AXiWtaH1kRMwGkHQUcA9QBVwVEa8U8xepceONsPfesOaaae5ap045UpiZWV0svTTcdx9svz3stRd88w3sv/+i7zMzs/LgHmOrj6IXxhHxGKD5vLTAGV+FBT3Oms/5Oxd2XzGMHAk/+xmsv37qNe7YMWcaMzOri44d0weZgwfDAQek4viww3KnMjOzhuA5xlYfWVelLneXXpr+oBo0KP2B5aLYzKx8tG0Lt9+eeo4PPxzOP3/R95iZWenzUGqrDxfG9XTuuXDkkbDjjvCvf6U/sMzMrLy0bg233AJ77AHHHQe/+x1E5E5lZmZLwkOprT5yzDEuaxFw2mnpj6c994Rrr4UWLXKnMjOz+mrZEq6/PhXJp54KX38NZ58Nmt+kHzMzK3kujK0+XBgvhupqOOIIuPzyNK94+HAoTGEwM7My1rw5jBgBbdrAOefA9Olw4YXQzOOqzMzKjucYW324MK6j6dPT6qW33Qa//nXqMXZvgplZ09GsGQwblorjCy5I/+5fdpk/ADUzKzeeY2z14cK4Dr74Is0lfvJJuPjiNLfYzMyaHgn+/Gdo1y59APrNN6kn2VNmzMzKh4dSW324MF6E11+HnXaC//437Ve82265E5mZWWOS4MwzU8/xKaek4njUKGjVKncyMzOrCw+ltvrw7KmFuPNO+NGP4Msv4f77XRSbmVWSk0+Giy6CW2+FnXdOBbKZmZW+FoVhPrNmzcqcxMqJC+P5mD0b/vAH2GEH6NMHnnkGNtwwdyozMyu2n/88Lbh4zz2w3XYwbVruRGZmtiht2rQBYPr06ZmTWDlxYTyPCRNgyy3TAltDhsBjj8Hyy+dOZWZmuRx8MPz97/Doo6l9+OKL3InMzGxh2rZtC8DXX3+dOYmVExfGBXPmpO2X1lwz9RBfdRVcd12aY2ZmZpVt773h5pvhhRdgs83g449zJzIzswWpKYy/+uqrzEmsnFR8YRwBd98N660Hhx4Ka6wBzz8PBx7o7ZjMzOw7gwfDHXfA22/DxhvD++/nTmRmZvPTrl07wD3GtngqsjCOgPHj0z6V66wD224LEyfCtdfCQw9B3765E5qZWSnacku47z6YNCkVx2++mTuRmZnNy0OprT4qbrumV16BZZeFyZPT83XXhSuugJ/+FFq2zJvNzMxK3wYbpA9Rf/zjVBybmVlpcWFs9VFxhXHr1rDnnqmneNNNoV+/3InMzKzc9O+fFuPad1/49NPcaczMrLaWLVvSokULzzG2xVJxhXGfPjBsWO4UZmZW7lZeGf7zH2hWkZOSzMxKW9u2bd1jbIvFzbmZmVk9eZFGM7PS1K5dO6Z583lbDC6MzczMzMysSenZsyfvvfde7hhWRlwYm5mZmZlZk7Lyyivzxhtv5I5hZaTi5hibmZmZmVnTttZaazFixAjOPfdcll9+ebSEc1+W9H4rfS6MzczMzMysSTnwwAO57LLLOPHEE3NHsTLhwtjMzMzMzJqUjh07Mm7cOD744AOmT5++RN8rIhoolRXL6quvvtj3uDA2MzMzM7Mmp6qqil69euWOYWWi7BffkrSNpDckjZd0cu48ZmZmlc5ts5mZlZuyLowlVQGXANsCqwJ7SVo1byozM7PK5bbZzMzKUVkXxsBAYHxEvBMRM4HRwODMmczMzCqZ22YzMys75V4Ydwc+qPV8QuGcmZmZ5eG22czMyk5FLL4laSgwtPB0hqRxOfPY93QCPssdwr7H70vp8XtSmlbOHaAcuW0uef73pjT5fSk9fk9K02K3zeVeGH8I9Kz1vEfh3FwiYjgwHEDS2IgYUJx4Vhd+T0qT35fS4/ekNEkamztDiXHb3AT4PSlNfl9Kj9+T0lSftrnch1I/A/SV1FtSS2AIcFvmTGZmZpXMbbOZmZWdsu4xjohqSUcB9wBVwFUR8UrmWGZmZhXLbbOZmZWjsi6MASLiTuDOxbhleGNlsXrze1Ka/L6UHr8npcnvyzzcNjcJfk9Kk9+X0uP3pDQt9vuiiGiMIGZmZmZmZmZlodznGJuZmZmZmZktkYopjCVtI+kNSeMlnZw7jyWSlpJ0k6TXJb0maf3cmSqNpKskfVp7qxRJ5xbek5ck3SppqZwZK9EC3pf+kp6S9IKksZIG5sxYaST1lPSgpFclvSLpmHleP05SSOqUK2O5cdtcmtw25+e2uTS5bS49Ddk2V0RhLKkKuATYFlgV2EvSqnlTWcFFwN0RsQqwFvBa5jyVaASwzTzn7gNWj4g1gTeBU4odyub7vvwJOCMi+gOnFp5b8VQDx0XEqsB6wJE1bYmknsCPgf9mzFdW3DaXNLfN+Y3AbXMpGoHb5lLTYG1zRRTGwEBgfES8ExEzgdHA4MyZKp6kjsAmwJUAETEzIqbkTVV5IuIR4It5zt0bEdWFp0+R9iG1Iprf+wIE0KHwuCPwUVFDVbiImBgRzxUeTyMVC90LL18AnEh6j6xu3DaXILfNpcFtc2ly21x6GrJtLvtVqeuoO/BBrecTgB9lymLf6Q1MAq6WtBbwLHBMRHydN5bN42fAP3KHMAB+Adwj6TzSB5sbZM5TsST1AtYGnpY0GPgwIl6UlDVXmXHbXJrcNpcHt82lw21ziVjStrlSeoytNDUH1gGGRcTawNeA55iVEEm/Jg1RuS53FgPgcODYiOgJHEuhR8eKS1I74GbSH0PVwK9Iw+fMmgK3zSXObXPJcdtcAhqiba6UwvhDoGet5z0K5yyvCcCEiHi68PwmUmNsJUDSAcAOwD7hfd1Kxf7ALYXHN5KGoloRSWpBanivi4hbgD6kHrYXJb1Hal+ek/T/8qUsG26bS5Pb5hLmtrkkuW3OrKHa5kopjJ8B+krqLaklMAS4LXOmihcRHwMfSFq5cGoL4NWMkaxA0jakORk7RcT03Hnsfz4CNi08HgS8lTFLxVEai3Ul8FpEnA8QES9HRJeI6BURvUhFxTqFf99s4dw2lyC3zaXLbXPJctucUUO2zRUxxzgiqiUdBdwDVAFXRcQrmWNZcjRwXeGPoneAAzPnqTiSRgGbAZ0kTQBOI6102Qq4rzAv46mIOCxbyAq0gPflEOAiSc2Bb4Gh+RJWpA2BnwIvS3qhcO5XEXFnxkxly21zSXPbnJnb5tLktrkkNVjbLI/CMDMzMzMzs0pWKUOpzczMzMzMzObLhbGZmZmZmZlVNBfGZmZmZmZmVtFcGJuZmZmZmVlFc2FsZmZmZmZmFc2FsVmZkPTVPM8PkHRxkTPsIek1SQ8u4roRknYvVi4zM7Mc3DabNR0ujM0qXGHfvbo6CDgkIjZvrDxmZmaVzm2zWfG5MDZrAiT1kvSApJck3S9p+cL5uT4drvlkW9Jmkh6VdBvw6ny+316SXpY0TtI5hXOnAhsBV0o6d57rJeliSW9I+jfQpdZrp0p6pvC9hheu7SPpuVrX9K393MzMrNy5bTYrLy6MzcrHDyS9UHMAZ9Z67a/AyIhYE7gO+Esdvt86wDER0a/2SUnLAecAg4D+wA8l7RwRZwJjgX0i4oR5vtcuwMrAqsB+wAa1Xrs4In4YEasDPwB2iIi3gS8l9S9ccyBwdR0ym5mZlRK3zWZNhAtjs/LxTUT0rzmAU2u9tj5wfeHxtaRPjxflPxHx7nzO/xB4KCImRUQ1qTHfZBHfaxNgVETMjoiPgAdqvba5pKclvUxq0FcrnL8COFBSFbBnrfxmZmblwm2zWRPhwtisaaum8N+5pGZAy1qvfd3YP1xSa+BSYPeIWAO4HGhdePlmYFtgB+DZiPi8sfOYmZmVALfNZiXIhbFZ0/AEMKTweB/g0cLj94B1C493AlrU4Xv9B9hUUqfCJ8Z7AQ8v4p5HgD0lVUnqBtQsAFLT0H4mqR3wvzlVEfEtcA8wDA/VMjOzpsdts1kZWZwV78ysdB0NXC3pBGASaV4QpE+Bx0h6EbibOnwSHRETJZ0MPAgIuCMixizitltJQ7FeBf4LPFn4XlMkXQ6MAz4GnpnnvutIc6DuXeRvaGZmVl7cNpuVEUVE7gxmVqEkHQ90jIjf5s5iZmZmbputcrnH2MyykHQr0If0abaZmZll5rbZKpl7jM3MzMzMzKyiefEtMzMzMzMzq2gujM3MzMzMzKyiuTA2MzMzMzOziubC2MzMzMzMzCqaC2MzMzMzMzOraC6MzczMzMzMrKK5MDYzMzMzM7OK5sLYzMzMzMzMKpoLYzMzM8tO0lWSPpU0rg7XXiDphcLxpqQpxchoZmZNlyIidwYzMzOrcJI2Ab4CromI1RfjvqOBtSPiZ40WzszMmjz3GJuZmVl2EfEI8EXtc5L6SLpb0rOSHpW0ynxu3QsYVZSQZmbWZDXPHcDMzMxsAYYDh0XEW5J+BFwKDKp5UdIKQG/ggUz5zMysiXBhbGZmZiVHUjtgA+BGSTWnW81z2RDgpoiYXcxsZmbW9LgwNjMzs1LUDJgSEf0Xcs0Q4Mgi5TEzsybMc4zNzMys5ETEVOBdSXsAKFmr5vXCfOOlgSczRTQzsyakrApjSVWSnpd0e+F5b0lPSxov6R+SWubOaGZmZotP0ihSkbuypAmSDgL2AQ6S9CLwCjC41i1DgNHh7TXMzKwBlNV2TZJ+CQwAOkTEDpJuAG6JiNGS/ga8GBHD8qY0MzMzMzOzclI2PcaSegDbA1cUnou0MuVNhUtGAjvnSWdmZmZmZmblqmwKY+BC4ERgTuH5sqRFOaoLzycA3XMEMzMzMzMzs/JVFqtSS9oB+DQinpW0WT3uHwoMBWjbtu26q6yySgMnNDOzSvXss89+FhGdc+coZ506dYpevXrljmFmZk1EfdrmsiiMgQ2BnSRtB7QGOgAXAUtJal7oNe4BfDi/myNiODAcYMCAATF27NjipDYzsyZP0vu5M5S7Xr164bbZzMwaSn3a5rIYSh0Rp0REj4joRVqF8oGI2Ad4ENi9cNn+wJhMEc3MzMzMzKxMlUVhvBAnAb+UNJ405/jKzHnMzMzMzMyszJTLUOr/iYiHgIcKj98BBubMY2ZmZmZmZuWt3HuMzczMzMzMzJaIC2MzMzMzMzOraC6MzczMrE4kXSXpU0njFvD6PpJekvSypCckrVXsjGZmZvXhwtjMzMzqagSwzUJefxfYNCLWAH5HYatEMzOzUld2i2+ZmZlZHhHxiKReC3n9iVpPnwJ6NHYmMzOzhuDC2MzMzBrDQcBduUOYWRMwaxZ89hm0bAlt2kDr1iDV73vNng3Tp8OcOdCuHVRVNWxWK1sujM3MzKxBSdqcVBhvtJBrhgJDAZZffvkiJTOzsjFxIlx1FfzrX/DMM6mQrdGiBXTtCh06QKdOsPbasOOOsNlmcxe6EfDYY3DHHfD44/Dmm/Dpp3P/nA4doEcPWHFF6N07fe3XLx29ekFzl0uVwu+0mZmZNRhJawJXANtGxOcLui4ihlOYgzxgwIAoUjwzK3WffAK/+Q2MGJF6dwcOhJNOSsVrdXXq7Z0yJV03bRp89BEMHw4XXQTdu8OBB8KQIfDUU3DhhTBuXCpuf/jDVDz37Alt26Ye52nTYPJkeP99ePddePjhdK5G8+bQp08qkvv2/a5g7tcPlluu/r3WVpJcGJuZmVmDkLQ8cAvw04h4M3ceMysjEXDZZXDiifDtt3DYYXDMMbDSSou+d/p0uP32VEyfdRb8/vfp/JprwtVXw267Qfv2dcvw2Wfw1lupd7n28e9/wzfffHdt27apWJ63YO7XD5ZZpl7/E1heLozNzMysTiSNAjYDOkmaAJwGtACIiL8BpwLLApcq9aRUR8SAPGnNrGxMmQI/+xnceitsuSVcckkqMOuqTRv4yU/S8cEHMGZMKoo33njxenUl6Nw5HRtsMPdrc+bAhx9+v2B+/nm45ZbUu11j2WXn38u80kqpoLaSpIjKGr00YMCAGDt2bO4YZmbWREh61sXfknHbbFbB3n0Xtt8+9dL+8Y9w7LHQrMx2lJ01K/0e8xbNb76ZiunaevSYfy9z795p7rQ1iPq0ze4xNjMzMzOz4nv+edhmG5g5E+67Ly2eVY5atPiuwJ3X11/D+PHfL5hvvBG++OK766qqUnE8b8Hcr1+aO11uHxaUIRfGZmZmZmZWXC++mIZNt2+fFr1aZZXciRpH27aw1lrpmNfnn39/PvMbb8BDD6V50zVat55/L3PfvmlVbi8C1iBcGJuZmZmZWfG8+moqitu2hQcfTD2llWjZZdOx3npzn49Iq23P28v88stp/nR19XfXLrXU/Avmvn3rtuCY/Y8LYzMzMzMzK45PP01zips3hwceqNyieGGkNHy6e3fYfPO5X6uuhvfem7tgfusteOQR+Pvf5762W7fvF8z9+qW9mlu1KtqvUy5cGJuZmZmZWeP79lvYeee0B/Ejj9RtKyabW/Pm6X+3lVaC7bab+7Xp0+Htt+cumN98E/75T5g06bvrmjWDXr2+XzD365f2ea6qKuqvVCpcGJuZmZmZWeM77DB48sm08NQAL+bf4Nq0gTXWSMe8Jk/+rlCuPa/5scfgq6++u65Vq1R0z2+7qS5dmvR8ZhfGZmZmZmbWuEaOTMepp8Luu+dOU3mWXhoGDkxHbRHw8cffL5hffx1uvz1tRVWjQ4cFLwLWsWNxf59G4MLYzMzMzMwaz+uvwxFHpO2YTj01dxqrTUpzkbt1g003nfu12bPh/fe/v3L2k0/C6NGpqK7Rtev8i+Y+fdKq2mXAhbGZmZmZmTWOmTNhyJA0zPe66yp2/mpZqqpKC3WtuCJsvfXcr337LbzzzvdXzr7zTrjqqu+uk2D55effy7zCCmnOdIkonSRmZmZmZta0/PGPac/iMWNgueVyp7GG0ro1rLpqOuY1der3e5nfeguuvTa9VqNFi9SjPL+Vs7t1K/p8ZhfGZmZmZmbW8F5+GX7/e9h7b9hpp9xprFg6dIB1101HbRFpdex5C+Y334R77oEZM767tm3b+RfM/fql+dKNwIWxmZmZmZk1rOpqOPBAWGopuOii3GmsFEhpZesuXWCjjeZ+bc4c+OCD7xfMY8emVcznzPnu2k6d5l8wr7RSGrJfT2VRGEtqDTwCtCJlvikiTpM0AtgU+LJw6QER8UKelGZmZlYvX3656GvMrLxceCE8+yz84x+pkDFbmGbN0pzjFVaArbaa+7WZM9N85nmHZ997L4wYMfe1PXumIrkeyqIwBmYAgyLiK0ktgMck3VV47YSIuCljNjMzM1sS48enHqWf/7xJ75FpVjE++gjOOAO23x722CN3Git3LVvCKqukY15ffZXakHkXAauHsiiMIyKAmp2nWxSOWPAdZmZmVjaWWgp+8Yv0x8xf/5p6DsysfJ18curlu/BCf9hljatdO+jfPx211eP/d2XT8kiqkvQC8ClwX0Q8XXjpLEkvSbpAUquMEc3MzKw++vSB44+HSy9NcxJnz86dyMzq64kn0urDxx2X5nyalYmyKYwjYnZE9Ad6AAMlrQ6cAqwC/BBYBjhpfvdKGipprKSxkyZNKlpmMzMzq6Nzz4Uzz4RrroF9900L95hZeZkzJ02J6N4dfvWr3GnMFkvZFMY1ImIK8CCwTURMjGQGcDUwcAH3DI+IARExoHPnzsWMa2ZmZnX129/COefA6NFw+OFpaw8zKx/XXZcW3PrTn9IQV7MyUhZzjCV1BmZFxBRJPwC2As6R1C0iJkoSsDMwLmtQMzMzWzInnghTp8JZZ0HXrmkPVDMrfTNmpA+31lkHhgzJncZssZVFYQx0A0ZKqiL1ct8QEbdLeqBQNAt4ATgsZ0gzMzNrAL/7HXz66XfF8dFH505kZoty2WXw/vswfLgX0LOyVBaFcUS8BKw9n/ODMsQxMzOzxiTBsGEwaVJarbpfP9h669ypzGxBpk1LozsGDfr+HrRmZcIf55iZmVnpqaqCv/8d1lgD9twT3ngjdyIzW5Dzz08fZJ19trdnsrLlwtjMzMxKU9u2MGYMtGwJO+0EkyfnTmRm8/r8czjvPNhtNxg433VwzcqCC2MzMzMrXSusALfcAu++C3vvnbaDMbPSceGF8NVXcMYZuZOYLREXxmZmZlbaNtoI/vpXuPtu+MMfcqcxsxqTJ8Nf/gK77w6rrZY7jdkScWFsZmZmpW/oUNhnHzjtNHjggdxpzAxSUTx1KvzmN7mTmC0xF8ZmZmZW+iT4299g5ZVhr73go49yJzKrbF9+mYZRDx4Ma62VO43ZEnNhbGZmZuWhXTu46aY0n3GvvaC6Oncis8p18cUwZQr89re5k5g1CBfGZmZmVj5WXRUuuwweeQROPz13GrPKNG1a2qJp++1h3XVzpzFrEC6MzczMrLzsuy/87GdpIa6HH86dxqzyXHEFfPGF5xZbk+LC2MzMzMrPRRfBSiulItn7G5sVz6xZaW7xxhvDeuvlTmPWYFwYm5mZWflp1w6uvx4+/jitWB2RO5FZZbjpJvjvf+H443MnMWtQLozNzMysPA0YAGedlf5Qv+qq3GnMmr4IOO886NcPdtghdxqzBuXC2MzMzMrX8cfDoEHw85/DG2/kTmPWtD38MDz3HBx3HDRzGWFNi/8fbWZmZnUi6SpJn0oat4DXJekvksZLeknSOo0eqlkzuOYa+MEPYO+9YebMRv+RZhXrvPOgc2f46U9zJzFrcC6MzczMrK5GANss5PVtgb6FYygwrAiZoHt3uPLK1JPlPVXNGserr8Idd8BRR6UPosyaGBfGZmZmVicR8QjwxUIuGQxcE8lTwFKSuhUl3ODBcMghcO658OijRfmRZhXloougdWs44ojcScwahQtjMzMzayjdgQ9qPZ9QOFcc558PvXvDfvvB1KlF+7FmTd7kyfD3v8M++0CnTrnTmDUKF8ZmZmZWdJKGShoraeykSZMa5pu2a5fmG//3v3DssQ3zPc0MRoyA6dPhyCNzJzFrNC6MzczMrKF8CPSs9bxH4dz3RMTwiBgQEQM6d+7ccAk23BBOOilt3zRmTMN9X7NKNWcOXHJJ+m9r7bVzpzFrNC6MzczMrKHcBuxXWJ16PeDLiJhY9BSnnw79+6c5x59+WvQfb9ak3HsvvP12WnTLrAlzYWxmZmZ1ImkU8CSwsqQJkg6SdJikwwqX3Am8A4wHLgfyrNLTsiVcey18+SUMHQoRWWKYNQkXXwxdu8Kuu+ZOYtaomucOYGZmZsUjaX1gX2BjoBvwDTAOuAP4e0R8uaB7I2KvhX3viAigNCYhrr46/OEPcPzxcPXV8LOf5U5kVn7efhvuvDNtg9ayZe40Zo3KPcZmZmYVQtJdwMHAPaT9iLsBqwK/AVoDYyTtlC9hAzv2WNh0UzjmGHj33dxpzMrPsGFQVQWHHpo7iVmjK4vCWFJrSf+R9KKkVySdUTjfW9LTksZL+ockf5RlZma2YD+NiIMi4raI+CgiqiPiq4h4LiL+HBGbAU/kDtlgmjWDkSNBgv33h9mzcycy+jAQ9wAAIABJREFUKx/Tp8OVV6Yh1MstlzuNWaMri8IYmAEMioi1gP7ANoVFPc4BLoiIlYDJwEEZM5qZmZW0iPisIa4pKyusAH/9Kzz6aNrn2Mzq5sYbYcoUOCLPUgFmxVYWhXEkXxWetigcAQwCbiqcHwnsnCGemZlZWZA0TdLUBRyTJD0laYvcORvcfvvBLrvAb34DL72UO41Zebj8cujXDzbZJHcSs6Ioi8IYQFKVpBeAT4H7gLeBKRFRXbhkAtA9Vz4zM7NSFxHtI6LD/A7g/wGHAhdljtnwJLjsMlh6adh3X5gxI3cis9L26qvw+ONw8MHpvx+zClA2hXFEzI6I/kAPYCCwSl3vlTRU0lhJYydNmtRoGc3MzEqZpHYLeq3Qzr4IXFbESMXTuTNccQW8/DKccUbuNGal7coroUWLNDffrEKUTWFcIyKmAA8C6wNLSarZcqoH8OEC7hkeEQMiYkDnzp2LlNTMzKzkjJH0Z0mbSGpbc1LSioU9ie8BJmbM17h22CFt23TOOfDkk7nTmJWmGTPSonWDB0OXLrnTmBVNWRTGkjpLWqrw+AfAVsBrpAJ598Jl+wNj8iQ0MzMrfRGxBXA/acj0K5K+lPQ58HfSUOr9I+KmhX2PsnfBBdCjR+oJmz49dxqz0vPPf8Lnn6dh1GYVpPmiLykJ3YCRkqpIxfwNEXG7pFeB0ZJ+DzwPXJkzpJmZWamLiDuBO3PnyKZDB7j6athiCzj5ZPjLX3InMistl1+eVnPfaqvcScyKqiwK44h4CVh7PuffIc03NjMzM6ubQYPg6KPTNk4775yemxm88w7cfz+ceWbaB9ysgvj/8WZmZlZ5/vhH6NsXDjwQpk7NncasNFx5ZSqIDzwwdxKzonNhbGZmZpWnTZu0wNCECXDssbnTmOVXXZ2mGWy7bZqHb1ZhXBibmZlVGEnLzOdokTtX0a2/Ppx4Ilx1Fdx+e+40Znnddx9MnAgHHZQ7iVkWLozNzMwqz3PAJOBN4K3C4/ckPSdp3azJiu3002GNNeCQQ9JKvGaVauRIWGYZ2H773EnMsnBhbGZmVnnuA7aLiE4RsSywLXA7cARwadZkxdaqFVxzTSqKjzwydxqzPKZMSds07bUXtGyZO41ZFi6MzczMKs96EXFPzZOIuBdYPyKeAlrli5VJ//5w2mnwj3+kw6zS3HgjzJiR9vc2q1AujM3MzCrPREknSVqhcJwIfCKpCpiTO1wWJ50EAwfCEUekeZZmleSaa2CVVWDAgNxJzLJxYWxmZlZ59gZ6AP8sHMsXzlUBP8mYK5/mzdMcy+nTYehQiMidyKw43n4bHnss9RZLudOYZePC2MzMrMJExGcRcXRErF04joqISRExMyLG586XzSqrwNlnpxWqr746dxqz4rj22lQQ77tv7iRmWbkwNjMzqzCS+kkaLuleSQ/UHLlzlYSf/xw23RR+8Qt4//3cacwa15w5aRj1Flt472KreM1zBzAzM7OiuxH4G3AFMDtzltLSrFnqLV5zTTjwQPj3v9M5s6boscfg3XfhjDNyJzHLrqj/0itZQ9LWkjaRtGwxf76ZmZkBUB0RwyLiPxHxbM2RO1TJ6N0bzj8fHnwQLrkkdxqzxnPNNdCuHey6a+4kZtkVpcdYUi/gRGAb4F1gEtAa6CtpCulT679HeKULMzOzIviXpCOAW4EZNScj4ot8kUrMwQfDrbem1aq33hr69cudyKxhTZ8ON9wAu+8ObdvmTmOWXbGGUv8JGAYcFRFzbQMhqRuwD7A/MKJIeczMzCpZzWalJ9Q6F8CKGbKUJgmuuAJWXz2t1vvoo2nlarOmYswYmDYN9tsvdxKzklCUf+EjYoFbP0TEROC8YuQwMzMziIjeuTOUheWWS0Op994bzj0XTjkldyKzhnPttbD88mmxOTMr+hzjXSW1Lzw+WdINkvoXM4OZmVmlkjSo8HXX+R2585WkIUPSUNPTToOXXsqdxqxhfPYZ3Hsv7LWXF5czKyj2fwmnR8Q0SRsA2wHXkeYXm5mZWeOr6RracT7HDrlClTQJhg2DpZdOQ05nzsydyGzJ3XQTzJ6dCmMzA4q/XVPNlhA7AJdFxBhJpxc5g5mZWUWKiNMKXw/MnaWsdOoEl18OgwfDmWfC73+fO5HZkhk1ClZZJW1LZmZA8QvjiZIuIa1OPUBSS4rfa21mZlbRJLUCdgN6UetvgYg4M1emkrfTTmkRrrPPhh13hB/9KHcis/qZMCEtJnf66WlEhJkBxS9KfwI8DGwfEZOBTsDJRc5gZmZW6cYAg4Fq4Otahy3MRRdB9+6pQP7mm9xpzOrnhhsgwsOozeZRrH2MnwEeAe4C/hkRMwEi4iPgo2JkMDMzs//pERHb5A5Rdjp2hKuugq22gl/9Ci64IHcis8U3ahSsuy707Zs7iVlJKVaP8Yakongb4DFJ/5J0pCTvl2hmZlZ8T0haI3eIsrTllnDkkXDhhfDQQ7nTmC2et96CsWPTautmNpeiFMYRMTMi/h0Rx0fEQOAIYBZwnqRnJf2lGDnMzMwMgI2AZyW9IeklSS9L8l5EdXXOObDSSnDAATBtWu40ZnU3enT6uueeeXOYlaAsC19FxAcRMTwidgV+BNy8sOsl9ZT0oKRXJb0i6ZjC+dMlfSjphcKxXTHym5mZlbltgb7Aj/luq6Yd63KjpG0KBfV4Sd9bJ0TS8oU2+/lC0d302ua2bWHkSPjgAzjuuNxpzOomIg2j3nhj6NkzdxqzklPUVaklrQOcAqzA3KtgrrOIW6uB4yLiOUntSZ9y31d47YKIOK9RApuZmTUhkjpExFSgXt2ckqqAS4CtgAnAM5Jui4hXa132G+CGiBgmaVXgTtLq103LBhvA8cfDn/4EO+8M2zW9+t+amJdfhtdeg0svzZ3ErCQVe7umUaTC+GVgTl1vioiJwMTC42mSXgO6N0pCMzOzput6Uu/ws0AAtfdqCWBRa38MBMZHxDsAkkaTVreuXRgH0KHwuCNNeZHNM86AO+6Agw+GceNgmWVyJzJbsFGjoKoKdt89dxKzklTsodSfRcQtEfFWRLxdcyzON5DUC1gbeLpw6qjCUK2rJC3dwHnNzMyajIjYofC1d0SsWPhac9RlQczuwAe1nk/g+x9Unw7sK2kCqbf46AaIXppat4ZrroFJk+Coo3KnMVuwiDS/eMstoXPn3GnMSlKxC+MzJP1N0h6Sdqo56nqzpHak+ci/KAwFGwb0AfqTepT/vID7hkoaK2nspEmTGuDXMDMzK2+SlpY0UNImNUcDfeu9gBER0QPYDrhW0vf+3mgybfM668Bvf5t64268MXcas/l7+ml47z3vXWy2EMUujPchLba1M7BH4ajTeA5JLUhF8XURcQtARHwSEbMjYg5wOWmI1/cUFvoaEBEDOvtTMjMzq3CSDgYeAe4Bzih8Pb0Ot34I1F61p0fhXG0HATcARMSTQGug07zfqEm1zaecAgMGwOGHw2ef5U5j9n2jRkGrVmk+vJnNV7EL4/UiYu2I2Cciflo49lvUTZIEXAm8FhHn1zrfrdZluwDjGj6ymZlZk3MM8EPg/YjYnDRFaUod7nsG6Cupt6SWwBDgtnmu+S+wBYCk/yMVxmXcJVwHLVrA1VfD5Mnw+9/nTmM2t9mz4YYb0gJxHTvmTmNWsopdGD8taeV63Lch8FNg0DxbM/2p1t6LmwPHNmRYMzOzJurbiPgWQFKriHgdWGT7HBHVwFGkHubXSKtPvyLpzFpTo44DDpH0ImnRzQMiIhrltyglq68OBx2UVvx9e7GWTzFrXA89BB9/7GHUZotQ7FWp1wZekjQemEFaDTMWtV1TRDzG3Ctn1riz4SOamZk1eRMkLQX8E7hP0mTg/brcGBF3Mk/7GxGn1nr8KukD7cpz+ulw3XXw61+nhY7MSsGoUdCuHeywQ+4kZiWt2IWxJzaYmZllFhG7FB6eLulB0rZKd2eM1DQstxwcdxz87nfwy1/CwPkufWJWPDNnws03p7nFP/hB7jRmJa0oQ6kltQGovUXTvNs11VxjZmZmja+wKvWawDTStkurZ47UNJxwAnTpkr5WwAhyK3H33ANTpsCQIbmTmJW8Ys0xvl3SOZI2kNS65qSk5SXtL+lOwOM7zMzMikDS74CXgL+Stjr8M3Be1lBNRfv2aUj1I4/A7bfnTmOVbtQoWGYZ2Gqr3EnMSl6xCuMtgMdJq2C+JWmKpC+Am4BewCERcUORspiZmVW6nwB9ImLTiNi8cAzKHarJOPhg6NcPTjoJqqtzp7FK9fXXMGYM7L47tGyZO41ZyStKYRzJbRGxZ0T0jIilImKZiBgYEWdExLx7IJqZmVnjGQcslTtEk9WiBfzxj/Daa2kbJ7Mcbr8dpk/3atRmdVTsxbfMzMwsv7OB5yWNI+0SAUBE7LTgW2yx7LwzbLghnHpqKkzatcudyCrNqFHQrRtsvHHuJGZlwYWxmZlZ5RkJnAO8DMzJnKVpkuDcc2GDDeD881OBbFYsU6bAXXfBEUdAVVXuNGZlwYWxmZlZ5ZkeEX/JHaLJW3992G03+NOf4NBDoWvX3ImsUtxyS9qqycOozeqsWNs1dVjYUYwMZmZm9j+PSjpb0vqS1qk5codqks4+G2bMSCtVmxXL6NGw4orwwx/mTmJWNorVY/wKEICA5Uh7JgpoB3wE9CxSDjMzM4O1C1/Xq3UuAK9M3dD69oXDDoNhw+CYY2CVVXInsqbuk0/g/vvh5JPTkH4zq5NirUrdMyKWB+4AdimsSt0R2BnwJn9mZmZFVGuLps29XVMR/Pa30KYNnHJK7iRWCW68EebM8TBqs8VUrH2Ma2wYEbfVPImIfwEbFjmDmZlZRZPUVdKVku4qPF9V0kG5czVZXbqkPY3/+U947LHcaaypGzUKVl89HWZWZ8UujCdKOllSj8JxEvBJkTOYmZlVuhHAPaTpTQBvAr/IlqYSHHssLLccnHACROROY03V++/DE0+4t9isHopdGO9Nmk98F3Bn4bH/yzUzMyuuThFxA4WtmiKiGpidN1IT16YN/O538NRTcPPNudNYU/WPf6Sve+6ZN4dZGSpqYRwRn0XEkcAPI2LNiDgqIj4rZgYzMzPja0nLkhbcQtJ6wJd5I1WA/feH1VZLc41nzsydxpqiUaNg4EDo0yd3ErOyU9TCWNKPJL1MGrKFpLUk/bWYGczMzIxfArcBfSQ9DlwDHJ03UgWoqkp7Go8fD8OH505jTc3rr8MLL3gYtVk9FXso9UXADsDnABHxIrB5kTOYmZlVLEnNgNbApsAGwKHAahHxUtZglWLbbWHQIDjjDPjSnfTWgEaPTtsz/eQnuZOYlaViF8bNIuL9ec55TpOZmVmRRMQc4JKIqI6IVyJiXETMyp2rYkip1/izz9JXs4YQkYZRb7ppWuTNzBZbsQvjDyQNBEJSlaRfUBhWbWZmZkVzv6TdJCl3kIq07rqw995w/vkwYULuNNYUPP88vPmmh1GbLYFiF8aHk+Y1LU/apmm9wjkzMzMrnkOBG4EZkqZKmiZpau5QFeWss2DOHDj11NxJrCkYNQqaN4fddsudxKxsFXtV6k8jYkhEdCocQ7wqtZmZWXFFRPuIaBYRLSOiQ+F5h9y5KkqvXnD00TBiBLz8cu40Vs7mzEnbNG29NSy7bO40ZmWr2KtSryTpHkkvFp6vKemUYmYwMzMzKwm/+hV07AgnnZQ7iZWzJ56ADz6AIUNyJzEra8UeSn0FcAYwp/D8ZWDfImcwMzMzy2+ZZeDXv4a77oL778+dxsrVqFHQujUMHpw7iVlZK3Zh3DYinqh5EhEBLHIlTEk9JT0o6VVJr0g6pnB+GUn3SXqr8HXpRsxuZmZm1rCOOgpWWAFOOCENiTVbHNXVcOONsOOO0L597jRmZa3YhfHnknoDASBpZ+DjOtxXDRwXEauSFuw6UtKqwMnA/RHRF7i/8NzMzMwWQtKfJa2WO4eRevrOOiutKjxqVO40Vm4eeAAmTfJq1GYNoNiF8VHAlcAqkt4nFbKHLeqmiJgYEc8VHk8DXgO6A4OBkYXLRgI7N0ZoMzOzJuY1YLikpyUdJqlj7kAVba+9YO2107Dqb7/NncbKyahR0KEDbLtt7iRmZa/Yq1KPj4hBQDdgrYhYLyLeW5zvIakXsDbwNNA1IiYWXvoY6Npwac3MzJqmiLgiIjYE9gN6AS9Jul7S5nmTVahmzeDcc+H99+Hii3OnsXLx7bdwyy2wyy5p5IGZLZFir0q9tKTzgfuAewpDueo8L1hSO+Bm4BcRMdd+i4X5yrGA+4ZKGitp7KRJk5bgNzAzM2saJFUBqxSOz4AXgV9KGp01WKXaYovU63fWWfDFF7nTWDm46y6YOtXDqM0aSLGHUo8GpgH7kFajngr8oy43SmpBKoqvi4hbCqc/kdSt8Ho34NP53RsRwyNiQEQM6Ny58xL+CmZmZuVN0gXAG8B2wB8iYt2IOCcidiSNyrIczjkHvvwS/vCH3EmsHIweDZ06pQ9VzGyJFbsw7h4Rp0XEW4XjDGC5Rd0kSaS5ya9FxPm1XroN2L/weH9gTIMnNjMza3peIk1pOjQi/jPPawNzBDJgjTXggAPgr3+F997LncZK2Vdfwb/+BXvsAc2b505j1iQUuzC+X9LuNU8k7UoaVr0oGwI/BQZJeqFwbAf8EdhK0lvAloXnZmZmtnAvAitLWqfW0UdS84j4Mne4inbmmVBVlRbiMluQMWPgm288jNqsARX7I6b9gKMlzQJU+PlfSjqANE14mfndFBGPFa6fH48fMTMzWzyXAuuQeo4FrA68AnSUdHhE3JszXEXr0QOOPTYNp/7lL2HddXMnslJ0/fXQsydsuGHuJGZNRrF7jDsBLYA2wA8KjzsVDk/+NTMzK46PgLUL62+sS5pX/A6wFfCnrMkMTjwxzR094QSI+a4rapXss8/g3ntTb3GzYv8pb9Z0Ffu/ph8CrSJiNvAT4GxguYiYXThnZmZmja9fRLxS8yQiXgVWiYh3MmayGh07wqmnwoMPwt13505jpeamm6C62sOozRpYsQvj4cA3ktYETgI+BK4tcgYzM7NK96qkYZI2LRyXFs61AmYt7EZJ20h6Q9J4SScv4JqfSHpV0iuSrm+MX6DJO/RQWGml1Hs8230HVsv118P//R+stVbuJGZNSrEL4+rCfsODgYsj4iKgQ5EzmJmZVbr9gfHALwrHO8ABpKJ48wXdVNj7+BJgW2BVYC9Jq85zTV/gFGDDiFit8P1tcbVsCWefDePGwciRudNYqfjgA3j00dRbrAUtv2Nm9VHsxbe+lnQCaQ/jzSQ1I80zNjMzsyIoFLdXRMQ+wJ/nc8lXC7l9IDC+Zsi1pNGkD7tfrXXNIcAlETEZICI+bZDglWi33eBHP4Lf/haGDIE2bXInstxGj05fPYzarMEVu8d4T9Lql4dFxESgB3D+wm8xMzOzhlJY02MFSS3rcXt34INazycUztXWD+gn6XFJT0naZn7fSNJQSWMljZ00aVI9olQACc47Dz76CC64IHcaKwWjRsHAgWmYvZk1qKL1GBc+oR4ZEVvVnIuI/wJXFyuDmX0nIvi2+lumz5rO9FnTmTF7BrNmz2LWnFnMmj2LmbNn/u9x7a8zZ89k1uxZVM+pZk7MIQjmxJy5joi5zy3sGgBJNFMzhOr1uHmz5rSsakmrqla0rGq5wKNV81a0adGGDq060L5le6qaVWV+F8yyeQd4XNJtwNc1JyOiIT6sbg70BTYjfQD+iKQ1ImJK7YsiYjhp7REGDBjgpZcXZKONYOed4Zxz4JBDoEuX3Iksl9dfh+ef94ckZo2kaIVxRMyWVCWpQ0RMLdbPNWuK5sQcvvjmCz7+6mM++eoTJn87mSnfTmHyN+nrlG+nMGVGej51xtT/Fb+1j2+qv8n9a2T3g+Y/SEVyq/a0b9mepVovRZe2XejStgtd23ala7uudG3blR4derDi0ivSsXXH3JHNGsrbhaMZ0H4x7vsQ6FnreY/CudomAE9HxCzgXUlv8v/Zu/M4G+v+j+Ovz2wYM8PYd4MIZW1SpFBalLQXpaTSpu1W3aU71e3u195d3WmjpCRtEpUoRSqRIaLsJDsxdrN/f39ch8aWYWbOdc6c9/PxuB7nOte5zjnvcWq+8znXd/EK5RlHHzfCPf44HH88/Oc/8OKLfqcRv4wc6fUiuOIKv5OIlEjBHmO8FZhjZl+y7zfU/YKcQyRkOedIz0hnWfoylqcvZ/mW5SxLX8aqbatYt2OdVwzvXE9OXs5Bnx9t0ZQvXX7vllQqieQyycTHxntbTPxf+4GtTGwZSkWXIjY6ltioWGKjY4mLjtu7n/82LjqO2OhYYqJiiLKoAzbD9r1vdshzLDBxiHNu71Xlo9nPycshKzfroFtmbuZf+zmZ7MzeyfbM7WzL3Mb2rO3eftY2tmduJz0jnZlrZ7Jh5wa2ZR74/V2FMhWon1yf+sn1Ob7y8bSs1pIW1VpQO6n23p9FJBw45/4NYGbxzrldR/DUGUBDM6uHVxB3B67c75xPgB7Am2ZWCa9rtZaBKozGjb2rxa++CnfcAQ0b+p1Igs05bzbqTp2genW/04iUSMEujD8LbCICbNi5gXkb5u3dft34K79u+JWtmVv3Oa9imYrULleb6gnVaV61OdUSqu3dqpatSoUyFfYWwglxCeFXpIVg3N3Zu9mwcwPrd65n1bZVLEtftndLW5PGh79+iMPr/ZlcOpm2tdvSsW5HOqR0oHX11sREBfvXq0jBmVlb4A0gAahjZi2Am5xzt/7d85xzOWZ2GzABiAaGOud+NbOBQJpzbmzgsbPM7DcgF7jXObepOH+eiPDwwzB8OPTv761jK5Fl5kxYsgTuP+gKaSJSBMxbPSlypKamurS0NL9jSATKzMnk53U/M23VtL3biq0r9j5esUxFjq9yPMdVPo5jKhxDveR61Ctfj3rJ9UgqpVXNQs2OrB3MXT+X2etm8/O6n/nuj+9Y8OcCAJJKJXFuw3O5qPFFdDmmC4mljqSnqoQbM5vpnEv1O8eRMLPpwKXAWOdcq8Cxec654/3Io7a5gAYO9ArkqVOhbVu/00gw9esHgwbB+vWQnOx3GpGQdzRtc1AKYzMb6ZzrYWY/Awe8oXOudbGHCFDjK8HinGPuhrl8tfQrvlr2FVNWTNk7rrdOuTqcXOtk2tRoQ4tqLTi+yvFULVs1/K70yj7W7VjHlBVTmLBkAp8u+pSNuzZSKroUFza+kBta38Dp9U4nyoK9GIAUt3AtjJ1zJ5nZz/kK4znOuRZ+5FHbXEA7dnjdqBs08NayVZsRGXJzoU4dOPFE+OQTv9OIhIWjaZuD1dfv3sDtpUF6PxFf5OTl8O3v3zJq/ig+WfAJa3esBaBxpcbc0PoGOqZ05ORaJ1MjsYbPSaU4VEuoxuXHXc7lx11Obl4uU1dO5YNfP2DE3BG8/+v7pJRP4Y42d9DnhD4kxCX4HVci20ozawc4M4sF7gTm+5xJDichAf79b7jpJhgzxputWkq+KVO8Jbu0drFIsVJXapFCcs7xw8ofeGv2W4xeMJpNuzcRHxtPl2O60LVRVzrX70ytpFp+xxQfZeRkMHr+aF5Je4Xv/viO5NLJ3N7mdvq17aeZrkuAML1iXAl4AeiMN8r/S+BOv8YCq20+Ajk50Ly5dxVx3jyIjfU7kRS3G2/0Jt7asAHi4/1OIxIWQrYr9d43MzsReBFoApTCa4wznXNBG0CpxleKytrta3l7ztsMnT2URZsWkRCXQLdju3Fpk0s5+5iziY9V4yUHmrpyKk/+8CRjF46lUnwlHunwCDeecCOx0frjNlyFY2EcatQ2H6FPP4Vu3eDll+GWW/xOI8UpKwuqVYPzzvMmXxORAgmHwngG0BN4D2gDXAvUdc49GKwManylsGatncVz057jvXnvkZOXw6l1TuW6VtdxWdPLKBtX1u94EiZmrpnJPV/dw+TfJ9O4UmNe6/oap9U9ze9YchTCsTA2s8pAHyCFfMOqnHPX+ZFHbfMRcg46doQFC7yZihM1wV+JtedLkM8/h3PP9TuNSNg4mrY52LPARDnnFgIxzrls59wQ4LwgZxA5Ys45xi0eR6e3OnHC4BP4ZMEn9D2xLwtvW8iU3lO4tuW1KorliJxQ4wS+ueYbxnQfQ2ZOJh2GdeDmz25ma8bWwz9ZpPDGAOWAicDn+TYJB2bw9NNe19qnn/Y7jRSn4cOhUiU480y/k4iUeMFeaHOnmcUBc8zsMWAt3jqIIiHJOceXS7/kockP8dPqn6idVJtnznyGG1rfoLGhUmhmRrdju3FGvTN4aNJDPD/9ecYtHse7l7xL+zrt/Y4nJVu8c+4+v0NIIbRpA5dfDs8+CzffDDU0qWOJs2ULjB0LffpoLLlIEAT7ivG1gfe8DcgFGqKZqiVETVo+ifZvtuecEeewfsd6Xj//dZbesZS7292toliKVNm4sjx79rNMvW4qcdFxdBjWgUenPEpuXq7f0aTk+szM1C8z3D32GGRnwyOP+J1EisNHH0FmJlx9td9JRCKCZqUW2c+SzUu458t7GLNwDLWSavHgqQ/Su1Vv4qLj/I4mEWBb5jZu/uxmRs4byZn1z+T9S98nuUyy37Hkb4TpGOPtQFkgK7AZ4II5GWZ+apsL4a674MUXYe5caNrU7zRSlDp2hLVrvbHkWrNa5IiEwxhjkZC1NWMr9355L01fasrXy7/msdMfY/Hti7kp9SYVxRI0SaWSGHHxCIacP4TJv0/mpNdPYuGfC/2OJSWMcy7RORflnCvtnEsK3PelKJZCevBBb33j+9QzvkRZsQK+/da7WqyiWCQoVBhLxHPOMXLuSBoNasSzPz5Lz+Y9WXTbIvqf2p/SMaX9jicRyMy4ofUNfNPrG7ZkbOGk109iwpIJfseSEsQ8Pc1sQOB+bTNr43cuOQqVKsEDD8Bnn8HkyX6nkaIZALihAAAgAElEQVQyYoR327OnvzlEIogvhbGZlfLjfUX29/uW3zn33XO58uMrqVuuLj/1+YmhFwylemJ1v6OJ0L5Oe2b0mUFK+RS6juzKO7+843ckKTleBtoCVwbu7wBe8i+OFModd0CtWnDvvZCX53caKSznvNmoTz0VUlL8TiMSMYJaGJtZGzObCywO3G9hZi8W4HlDzWyDmc3Ld+wRM1ttZrMDmyYRkQLLycvh2anPctzLx/Hdiu944ZwX+PH6H0mtEVbDBCUC1C1flym9p3Ba3dO4evTV/PfH//odSUqGk5xzfYEMAOdcOqAxI+GqTBl49FFIS4MPPvA7jRTWzJneuGJdLRYJqmBfMf4f0BXYBOCcmwN0KsDzhgHnHOT4c865loFtXJGllBJt4Z8LafdGO+756h5Or3c6v/X9jTtOuoPoKK0cJqEpqVQS464cx6VNL+XuL+/m/on3E2kTJ0qRyzazaMABmFllQJcaw1nPntC8udetOjPT7zRSGMOHQ1wcXHaZ30lEIkqwC+Mo59yK/Y4ddj0S59wUYHPxRJJI4ZzjlRmv0Oq1VixNX8p7l7zH2O5jqVOujt/RRA6rVEwp3rvkPW4+4Wae/OFJbht3G3lOdYwctf8Bo4EqZvZ/wPfAY/5GkkKJjoann4bly+GVV/xOI0crOxtGjoTzz4dkrUggEkwxQX6/lYHJPVzgm+rbgUWFeL3bzOwaIA24O9AVTOQA63as47ox1/HFki84u8HZDL1gKDUSa/gdS+SIREdF8/J5L5NYKpGnpz5NTl4Or3R9hSjTPIpyZJxzI8xsJnAG3lJNFzrn5vscSwrrrLPgzDPhP/+Ba6+F8uX9TiRH6quvYONGrV0s4oNg/zV1C9APqAOsB04OHDsarwANgJbAWuDZQ51oZjeaWZqZpW3cuPEo307C1ej5ozn+5eOZ9PskXuzyIl9c9YWKYglbZsaTnZ/kgfYPMHjWYPqM7aMrx3JUnHMLnHMvOecGqSguQZ58EtLTveJYws/w4VCxInTp4ncSkYgT1CvGzrkNQPcieq31e/bNbAjw2d+cOxgYDJCamqqBeRFie+Z27hx/J2/OfpPW1VvzzkXv0KRyE79jiRSamfHo6Y8SExXDwCkDyXW5vNHtDY2TFxFo1Qr69IHnnoNLLoF27fxOJAW1bRt88glcd503xlhEgiqohXGggD2gMHXO3XgUr1XdObc2cPciYN7fnS+R5Yc/fuDq0VezYusKHmj/AA93fJi4aDUyUnKYGf/u9G9iomJ4aPJD5OTlMOzCYcREBXuEjIiEnGeegS+/hGuugTlzoGxZvxNJQYwaBRkZmo1axCfB/gtqYr790ngF7crDPcnMRgIdgUpmtgp4GOhoZi3xCu3fgZuKOqyEn6zcLP49+d888cMT1C1XlynXTuGUOqf4HUuk2AzoMIDY6Fj6f92fXJfL8IuGqziWwzKz24F3NDdHCZWYCMOGQadOcN99MGiQ34mkIN58Exo2hJNP9juJSEQKdlfq9/PfN7PheDNhHu55PQ5y+I2iyiUlw/yN8+k5uiez1s6id8vePH/O8ySVSvI7lkixu7/9/URbNP+c+E9y83IZcfEIYqNj/Y4loa0qMMPMZgFDgQlOa4CVLB06wF13eV2qL7wQOnf2O5H8ncWL4bvv4LHHwMzvNCIRye+pTOvhNc4iRy3P5THop0G0HtyaFVtWMOryUQy9YKiKYoko955yL8+e9Swf/vYh3Ud1Jys3y+9IEsKccw8CDfG+ZL4WWGxmj5lZA1+DSdH6v/+Dxo2hd2/YutXvNPJ3hg2DqCiv+7uI+CKohbGZpZvZ5sC2BfgK6B/MDFKyrN62mnPeOYfbv7idTimdmHvLXC5ucrHfsUR80a9tP1445wU+nv8xl394uYpj+VuBK8TrAlsOkAx8ZGZP+RpMik6ZMvDWW7B2Ldx5p99p5FByc73P6eyzoWZNv9OIRKygFcZmZkALoHJgS3bO1XfOfRCsDFKyvD/vfZq90owfVv7AK+e9wudXfk71xOp+xxLx1R0n3cGgLoMYs3AMl3xwCZk5mX5HkhBkZncG1jF+CvgBaOacuwU4AbjE13BStNq0gf79vcJrzBi/08jBfPUVrF7tzUYtIr4JWmEc+GZ6nHMuN7BpLJMclfTd6Vz18VV0H9WdRhUbMfum2dycejOmMTkiAPRt05dXznuFzxZ9xkXvX0RGTobfkST0JAMXO+fOds596JzLBnDO5QFd/Y0mRW7AAGjZEm68ETZu9DuN7O/NN6FCBTj/fL+TiES0YI8xnm1mrYL8nlKCTFgygeavNueDXz9gYMeBfH/d9zSs2NDvWCIh5+bUmxncdTBfLPmCC9+7kN3Zu/2OJCHCzKKB7s65FQd73Dk3P8iRpLjFxcHbb8OWLXDLLaBrE6Fj82Zv7eKePaFUKb/TiES0oBTGZrZn9utWeLNgLjSzWWb2c2BGTJG/tWnXJq4ZfQ3njDiHxLhEfrz+RwZ0GKBlaUT+Rp8T+vBGtzf4cumXdBnRhS0ZW/yOJCHAOZcLLDSzOn5nkSBq1gwGDvTWyh050u80sse770JWljdBmoj4KlhVxU9Aa6BbkN5PSgjnHB/8+gG3f3E76RnpDDhtAP869V+UitG3qiIFcV2r6ygTU4Zen/Ti1DdP5YurvqBWUi2/Y4n/koFfzewnYOeeg845tdMl2T33eOOM+/b1lnPSRE/+GzoUWrXyurqLiK+CVRgbgHNuaZDeT0qA5enLuXP8nXy66FNSa6QysdtEmldt7ncskbDTo1kPqpStwkXvX0TbN9oy/qrxHFflOL9jib8G+B1AfBAd7U3C1aIF3HADjBunNXP9NHs2/Pwz/O9/ficREYI3xriymfU71BakDBImdmXv4uFJD9P05aZ8vfxrnjnzGX68/kcVxSKFcEb9M5jSewo5eTm0f7M9E5dN9DuS+Mg59y3wOxAb2J8BaGhTJGjYEJ56CsaPhyFD/E4T2d580xv/feWVficREYJXGEcDCUDiITYRnHN89NtHNHmpCQOnDOTCxhey8LaF3N3ubo0lFikCLau15Mfrf6RmYk3Ofudsnpn6DFogIDKZWR/gI+C1wKGawCf+JZKguvVWOOMM6NcPli3zO01kysyEESPgggugYkW/04gIwetKvdY5NzBI7yVhaNLySfT/uj/TV0+nWZVmTO41mQ4pHfyOJVLipJRPYdoN0+g9pjf3fnUvM9fOZMj5Q0iIS/A7mgRXX6ANMB3AObfYzKr4G0mCJirKG9varJk36dOkSd4xCZ5PPoFNm7R2sUgICdZvQQ1gkYNKW5PGWcPP4vS3T2f19tUMOX8Is26apaJYpBglxCXwwaUf8PgZj/P+vPdp9Vorpq+a7ncsCa5M51zWnjuB1SPUfSCS1KkDL7wAU6Z4txJcr70GKSlw1ll+JxGRgGAVxmcE6X0kDDjn+HrZ15w1/CxOHHIis9bO4tmznmXx7Yu5ofUN6jYtEgRmxv3t72fytZPJys3ilKGnMPDbgeTk5fgdTYLjWzN7AChjZmcCHwKfFuSJZnZOYNnFJWZ2/9+cd4mZOTNLLaLMUtR69YJu3aB/f5iv5auDZtEi7yp9nz66Ui8SQoLyf6NzbnMw3kdCW1ZuFu/Pe58Th5xI5+GdmbthLk+c8QTL7lxGv7b9KB1T2u+IIhHntLqn8cvNv9CjWQ8envwwJww+gR/++MHvWFL87gc2AnOBm4Bxzrl/He5JZhYNvAR0AZoCPcys6UHOSwTuJNBVW0KUmXflMiEBrrkGcvTFWFAMHgwxMepGLRJi9DWVFLtl6cvoP7E/tZ+rTfdR3dmWuY3BXQez/M7l3Nf+PpJKJfkdUSSilStdjuEXDWf0FaNJ351O+zfbc92Y61i9bbXf0aT43O6cG+Kcu8w5d6lzboiZ3VmA57UBljjnlgW6Yr8HXHCQ8/4DPAlkFGFmKQ7VqsGrr0JaGjz+uN9pSr6MDG826gsu8P7tRSRkqDCWYrF592Zen/U6Z7x9Bg3+14Cnpj5F21pt+eKqL5jfdz59TuijK8QiIebCxhfyW9/f+Ge7fzL8l+Ec8+Ix3PPlPfy560+/o0nR63WQY9cW4Hk1gZX57q8KHNvLzFoDtZ1znx91OgmuSy+FHj1g4ECYpVW7itWoUbB5M9x8s99JRGQ/GswpRWbdjnWMXzKej377iAlLJ5CTl0PDCg15pMMjXN/6emol1fI7oogcRkJcAk+e+SQ3pd7Ev7/9N89Ne45X017l2pbXcsdJd9CoYiO/I0ohmFkP4EqgnpmNzfdQIlDoYU9mFgX8lwIU2WZ2I3AjQJ06dQr71lJYgwbB5Mlel+qZM6FUKb8TlUyvvQYNGsDpp/udRET2o8JYjlpGTgYzVs9g/JLxfLHkC35e9zMAdcrV4R8n/4Pux3enVbVWmGlScpFwUz+5Pm9d+Bb3nXIfT099miGzhvDSjJfockwXerXoRbdju1EmtozfMeXITQXWApWAZ/Md3w78UoDnrwZq57tfK3Bsj0TgeGBy4Hd/NWCsmXVzzqXlfyHn3GBgMEBqaqpmxPZbhQrw+utw3nnw0EPw5JN+Jyp5fvsNvvvO+7fVpFsiIceci6y2KDU11aWlpR3+RNmHc451O9aRtiaN7//4nh9W/sCMNTPIys0i2qJpV7sdXY7pQpeGXWhRtYWKYZESZv2O9bya9ipDZg1h9fbVJMYlclGTizi/0fl0rt+Z8qXL+x3RN2Y20zkXVjMvm9mTzrn7DnfsIM+LARbhrTaxGpgBXOmc+/UQ508G7tm/KN6f2uYQcuONXoH8/ffQrp3faUqWu+6Cl1+GVaugipYNFylOR9M2qzCWA2zJ2MLSzUuZ/+d85qybw+z1s5mzbg4bd20EIDYqlhNqnED72u05pc4pdEzpGNF/FItEkty8XKasmMI7v7zDxws+ZkvGlr1fjnVK6UTb2m05qeZJJJdJ9jtq0IRpYTzLOdd6v2O/OOeaF+C55wLPA9HAUOfc/5nZQCDNOTd2v3Mno8I4vGzfDs2bQ3Q0zJkDZcv6nahk2LULatXy1i1+7z2/04iUeCqMCyDSG9+cvBw27NzA2u1rWbN9DWu2r2H19tUsTV/K0s1LWbJ5CZt2b9p7fqnoUhxX5ThaVG1Bi6otaFW9FSfWOFFdKEWEnLwcpq+azrjF4xi/dDyz180mz+UBcGzFY2letTlNKzfluMrH0aRyE+qVr0fZuJL3R3Y4FcZmdgtwK9AAWJLvoUTgB+dcTz9yRXrbHHK+/RY6doS+fb2xx1J4Q4Z4V+OnTIFTT/U7jUiJp8K4AMK58c3Ny2V3zm52Ze9iZ9ZOdmXv8vazvf1tmdtI351OekY66bvT2bx7s7cfuL9h5wbW71y/9w/XPaIsijrl6nBMhWNokNxg722jio04ttKxxERpKLqIHN72zO3MWDODH1f+yE9rfuK3jb+xdPNSHH+1MxXLVKROuTrUKVeH2km1qRRfae9WMb4ileIrUaFMBRLiEkiISyAuOs7Hn6hgwqwwLgckA4/jrWW8x3bnXKEn3zpa4dw2l1j/+Ac8/zx89RV07ux3mvDmnHcVPjbWm9hMw81Eip0K4wKo06SO6/dmP5xz5Lk8HIHbYriffz87L9vbco/8dk8xnJFT8OUgS8eUpkKZCiSXTia5TDLJpZOpHF+Zmkk1qZ5QnRqJNfZuVcpWITY6thj/1UUkUu3O3s3CTQuZv3E+K7au4I+tf+zdVm5byZaMLX/7/JioGBLiEigbW9a7jStL6ZjSxEXHERsVS1x0nLcfHXvAsZioGKIs6pCbYX/7eJRFYWYY3h+x++8DGMatbW4Nm8J4f2ZWBdi7dp5z7g8/cqgwDkG7d0OrVrBzJ8ybB+XK+Z0ofH3zDZxxhrd+8bXX+p1GJCKU2MLYzIYCXYENzrnjA8cqAO8DKcDvwOXOufTDvlYNc9x05Bny/xFlZoe8f7DHDCM2OpbYqNgjv42KpUxsGeJj4ykbW5b42HhvPy7ffuB4QlyCVwyXSdYawSISFrJzs9m8ezObdm/iz11/8ueuP9m8ezM7s3ayI2sHO7MDt1k72ZG9gx1ZO8jMySQ7L5us3CyycrPIzvX29z+WnZe990vL/bf8V7EL7RHCrjA2s/PxllWqAWwA6gLznXPH+ZFHhXGI+uknbwKunj1h2DC/04SvCy+EH36AlSuhtP4+EwmGoymMw6WP7DBgEPB2vmP3A187554ws/sD9/92Nk2AFtVaMOmfkw5b4OYvajXDsohI8YiNjqVqQlWqJlQN6vvu6dVzqMI5/wbsPXfPfv7XqP5I9aBmLyKPAicDE51zrcysE+DL+GIJYW3aQP/+8OijcNFFcMEFficKP8uXw9ix8MADKopFQlxYFMbOuSlmlrLf4QuAjoH9t4DJFKAwjomKiajZUkVE5EB7u0UbRBPtdxw/ZDvnNplZlJlFOecmmdnzfoeSEDRgAHz2mTdxVLt2ULmy34nCy0sveTN833KL30lE5DDCeXXxqs65tYH9dUBwLzeIiIiEry1mlgBMAUaY2QvATp8zSSiKi4O334YtW7ziLgyG4IWMHTu8NaEvvRRq1vQ7jYgcRjgXxns5r3/bIX9Tm9mNZpZmZmkbN24MYjIREZGQdAGwC/gHMB5YCpzvayIJXc2awcCBMGoUjBzpd5rw8c47sHUr3HGH30lEpADCuTBeb2bVAQK3Gw51onNusHMu1TmXWlldgEREJMI553Y65/KccznA58CLzrlNh3ueRLB77oG2bb21jVev9jtN6MvLgxdegNRUOPlkv9OISAGEc2E8FugV2O8FjPExi4iISMgzs5PNbLKZfWxmrcxsHjAP78vmc/zOJyEsOhreegsyM+GGG9Sl+nA+/xwWLIB+/bRusUiYCIvC2MxGAj8Cx5rZKjO7HngCONPMFgOdA/dFRETk0AYBjwEjgW+AG5xz1YDTgMf9DCZhoGFDeOopGD8ehgzxO01oe+opqFsXLrvM7yQiUkDhMit1j0M8dEZQg4iIiIS3GOfclwBmNtA5Nw3AObdASxNKgdx6K3zyiXcltHNnqF/f70Sh58cf4fvvva7UMWHxp7aIECZXjEVERKRI5OXb373fY+obK4cXFQVDh3pdq3v39sbSyr6efhqSk+G66/xOIiJHQIWxiIhI5GhhZtvMbDvQPLC/534zv8NJmKhTx7saOmWKdyt/WbTIu6Lety8kJPidRkSOgApjERGRCOGci3bOJTnnEp1zMYH9Pfdj/c4nYaRXLzj/fOjfH+bP9ztN6Hj2WW/t59tu8zuJiBwhFcYiIiIicmTMYPBg76roNddATo7fify3fr03c/e110LVqn6nEZEjpMJYRERERI5ctWrwyiuQlgaPa1JzXnwRsrLg7rv9TiIiR0GFsYiIiIgcncsugx49YOBA+Plnv9P4Z8sWGDQILr7YW9ZKRMKOCmMREREROXqDBkHlyl6X6sxMv9P448UXYetWePBBv5OIyFFSYSwiIiIiR69CBXj9dZg3Dx56yO80wbdtGzz3HHTrBi1b+p1GRI6SCmMRERERKZxzz4U+fbw1fKdO9TtNcL30EqSnw4ABficRkUJQYSwiIiIihffss1C3rteleudOv9MEx44d8N//el8MpKb6nUZECkGFsYiIiIgUXmIiDBsGS5fCfff5nSY4Xn0V/vxTV4tFSgAVxiIiIiJSNDp0gLvu8roXT5zod5ritWuX13X8zDPh5JP9TiMihaTCWERERESKzmOPwbHHQu/e3kzNJdWLL8KGDfDww34nEZEioMJYRERERIpOmTLw9tuwdi3ceaffaYrHli3w5JPe2OJTTvE7jYgUARXGIiIiIlK02rSB/v3hrbdgzBi/0xS9Z57xZqL+v//zO4mIFBEVxiIiIiJS9AYM8Nb1vfFG2LjR7zRFZ/16eP55uOIKrVssUoKoMBYRERGRohcX53WpTk+HW24B5/xOVDQeewwyMmDgQL+TiEgRUmEsIiIiIsWjWTOvgBw1CkaO9DtN4a1Y4S3R1Ls3NGrkdxoRKUIqjEVERESk+Nx7r7ecUd++sGaN32kK51//AjN46CG/k4hIEVNhLCIiIiLFJzra61KdmQnXXx++XaqnTYMRI+Cee6B2bb/TiEgRU2EsIiIiIsWrYUN46ikYPx6GDPE7zZFzDu66C6pVg/vv9zuNiBQDFcYiIiIiUvxuvRXOOAP69YNly/xOc2RGjoTp0+HxxyEhwe80IlIMVBiLiIhIgZnZOWa20MyWmNkBl87MrJ+Z/WZmv5jZ12ZW14+cEoKiomDoUK9rde/ekJfnd6KC2bUL7rsPWreGa67xO42IFJOwL4zN7Hczm2tms80sze88IiIiJZWZRQMvAV2ApkAPM2u632k/A6nOuebAR8BTwU0pIa1OHXjhBZgyxbsNB888A6tWwXPPecW9iJRIJeX/7k7OuZbOuVS/g4iIiJRgbYAlzrllzrks4D3ggvwnOOcmOed2Be5OA2oFOaOEul694PzzoX9/mD/f7zR/b+1aePJJuOwyOO00v9OISDEqKYWxiIiIFL+awMp891cFjh3K9cAXxZpIwo8ZDB7sjdW95hrIyfE70aE9+ihkZcETT/idRESKWUkojB3wpZnNNLMb/Q4jIiIiYGY9gVTg6UM8fqOZpZlZ2saNG4MbTvxXrRq88gqkpXkTWoWiZcu8Ar5PH6hf3+80IlLMSkJh3N451xpvvFNfMzugn4saXxERkSKxGsi/gGutwLF9mFln4F9AN+dc5sFeyDk32DmX6pxLrVy5crGElRB32WXQowcMHAg//+x3mgM98gjExMCDD/qdRESCIOwLY+fc6sDtBmA03vin/c9R4ysiIlJ4M4CGZlbPzOKA7sDY/CeYWSvgNbyieIMPGSWcDBoElSt7XaozD/odij9+/RXeeQduvx1q1PA7jYgEQVgXxmZW1swS9+wDZwHz/E0lIiJSMjnncoDbgAnAfOAD59yvZjbQzLoFTnsaSAA+DKwYMfYQLycCFSrA66/DvHnw8MN+p/nLvfdCYqK3TJOIRIQYvwMUUlVgtJmB97O865wb728kERGRkss5Nw4Yt9+xh/Ltdw56KAlv554LN9wATz8N3bpBu3b+5hk3Dr74wlumqWJFf7OISNCEdWHsnFsGtPA7h4iIiIgUwn//CxMnel2q58yBsmX9yZGVBf36QcOGXjdqEYkYYd2VWkRERERKgMREePNNWLrU3+7LL70ECxfCc89BXJx/OUQk6FQYi4iIiIj/OnaEu+7yitOJE4P//qtXezNRn322171bRCKKCmMRERERCQ2PPQbHHgu9e8PmzcF7X+fg1lu9rtSDBoE3f42IRBAVxiIiIiISGsqUgeHDYf16b7xxXl5w3vfDD2HsWPjPf+CYY4LzniISUlQYi4iIiEjoOPFEbzKuzz+HJ54o/vfbuNGbaCs11evKLSIRSYWxiIiIiISWvn2he3cYMAC+/rr43icvD669FrZuhTfegJiwXrBFRApBhbGIiIiIhBYzGDLEG2/cowf88UfxvM/zz3vrFj/7LDRvXjzvISJhQYWxiIiIiISehAQYNcqbEOvcc2HLlqJ9/UmTvKWhLrzQm3hLRCKaCmMRERERCU1NmsDo0bBoEVx8sVckF4VFi+CSS6BRIxg2TLNQi4gKYxEREREJYZ06wZtveld4r7oKsrML93rr1kHXrhAdDZ9+CuXKFU1OEQlrmmFARERERELbVVd5s0f/4x+QkwPvvQelSh3562zYAKefDmvWwJdfQv36RZ9VRMKSrhiLiIiISOi76y548UX45BM480xvreMjsWABtGsHK1Z4S0G1a1c8OUUkLKkwFhEREZHwcNttMHIkpKV56w5/++3hn+McjBgBJ58M27fDxInQoUPxZxWRsKLCWERERETCR/fuMHUqxMVBx45eN+t58w48zzmYMgXOPht69oTGjeGnn6Bt26BHFpHQpzHGIiIiIhJeWraEuXPhsce8NYjffReaNoXWrSEpyZtga9o0byxxpUreesW33eZNuCUichAqjEVEREQk/MTHw6OPehNyvfWW10V6yhSvu3TVqnDKKd76x5df7p0rIvI3VBiLiIiISPiqWBH69fM2EZGjpDHGIiIiIiIiEtFUGIuIiIiIiEhEU2EsIiIiIiIiEU2FsYiIiIiIiEQ0FcYiIiIiIiIS0VQYi4iIiIiISEQL+8LYzM4xs4VmtsTM7vc7j4iIiIiIiISXsC6MzSwaeAnoAjQFephZU39TiYiIiIiISDgJ68IYaAMscc4tc85lAe8BF/icSURERERERMJIuBfGNYGV+e6vChwTERERERERKZAYvwMEg5ndCNwYuJtpZvP8zCMHqAT86XcIOYA+l9CjzyQ0Het3gHA3c+bMHWa20O8csg/9vglN+lxCjz6T0HTEbXO4F8argdr57tcKHNuHc24wMBjAzNKcc6nBiScFoc8kNOlzCT36TEKTmaX5naEEWKj/tkOLft+EJn0uoUefSWg6mrY53LtSzwAamlk9M4sDugNjfc4kIiIiIiIiYSSsrxg753LM7DZgAhANDHXO/epzLBEREREREQkjYV0YAzjnxgHjjuApg4srixw1fSahSZ9L6NFnEpr0uRSe/g1Djz6T0KTPJfToMwlNR/y5mHOuOIKIiIiIiIiIhIVwH2MsIiIiIiIiUigRUxib2TlmttDMlpjZ/X7nEY+ZlTezj8xsgZnNN7O2fmeKNGY21Mw25F/GzMyeDnwmv5jZaDMr72fGSHSIz6WlmU0zs9lmlmZmbfzMGGnMrLaZTTKz38zsVzO7c7/H7zYzZ2aV/MoYbtQ2hya1zf5T2xya1DaHnqJsmyOiMDazaOAloAvQFOhhZk39TSUBLwDjnXONgRbAfJ/zRKJhwDn7HfsKON451xxYBPQPdig56OfyFPBv51xL4KHAfQmeHOBu5yYQ8QAAACAASURBVFxT4GSg7562xMxqA2cBf/iYL6yobQ5papv9Nwy1zaFoGGqbQ02Rtc0RURgDbYAlzrllzrks4D3gAp8zRTwzKwecBrwB4JzLcs5t8TdV5HHOTQE273fsS+dcTuDuNLw1wiWIDva5AA5ICuyXA9YENVSEc86tdc7NCuxvxysWagYefg74J95nJAWjtjkEqW0ODWqbQ5Pa5tBTlG1z2M9KXUA1gZX57q8CTvIpi/ylHrAReNPMWgAzgTudczv9jSX7uQ543+8QAsBdwAQzewbvi812PueJWGaWArQCppvZBcBq59wcM/M1V5hR2xya1DaHB7XNoUNtc4gobNscKVeMJTTFAK2BV5xzrYCdgMaYhRAz+xdeF5URfmcRAG4B/uGcqw38g8AVHQkuM0sARuH9MZQDPIDXfU6kJFDbHOLUNocctc0hoCja5kgpjFcDtfPdrxU4Jv5aBaxyzk0P3P8IrzGWEGBm1wJdgauc1nULFb2AjwP7H+J1RZUgMrNYvIZ3hHPuY6AB3hW2OWb2O177MsvMqvmXMmyobQ5NaptDmNrmkKS22WdF1TZHSmE8A2hoZvXMLA7oDoz1OVPEc86tA1aa2bGBQ2cAv/kYSQLM7By8MRndnHO7/M4je60BOgT2TwcW+5gl4pjXF+sNYL5z7r8Azrm5zrkqzrkU51wKXlHROvD7Tf6e2uYQpLY5dKltDllqm31UlG1zRIwxds7lmNltwAQgGhjqnPvV51jiuR0YEfijaBnQ2+c8EcfMRgIdgUpmtgp4GG+my1LAV4FxGdOcczf7FjICHeJz6QO8YGYxQAZwo38JI9IpwNXAXDObHTj2gHNunI+Zwpba5pCmttlnaptDk9rmkFRkbbOpF4aIiIiIiIhEskjpSi0iIiIiIiJyUCqMRUREREREJKKpMBYREREREZGIpsJYREREREREIpoKYxEREREREYloKoxFwoSZ7djv/rVmNijIGS4zs/lmNukw5w0zs0uDlUtERMQPaptFSg4VxiIRLrDuXkFdD/RxznUqrjwiIiKRTm2zSPCpMBYpAcwsxcy+MbNfzOxrM6sTOL7Pt8N7vtk2s45m9p2ZjQV+O8jr9TCzuWY2z8yeDBx7CGgPvGFmT+93vpnZIDNbaGYTgSr5HnvIzGYEXmtw4NwGZjYr3zkN898XEREJd2qbRcKLCmOR8FHGzGbv2YCB+R57EXjLOdccGAH8rwCv1xq40znXKP9BM6sBPAmcDrQETjSzC51zA4E04Crn3L37vdZFwLFAU+AaoF2+xwY55050zh0PlAG6OueWAlvNrGXgnN7AmwXILCIiEkrUNouUECqMRcLHbudcyz0b8FC+x9oC7wb2h+N9e3w4Pznnlh/k+InAZOfcRudcDl5jftphXus0YKRzLtc5twb4Jt9jncxsupnNxWvQjwscfx3obWbRwBX58ouIiIQLtc0iJYQKY5GSLYfA/+dmFgXE5XtsZ3G/uZmVBl4GLnXONQOGAKUDD48CugBdgZnOuU3FnUdERCQEqG0WCUEqjEVKhqlA98D+VcB3gf3fgRMC+92A2AK81k9ABzOrFPjGuAfw7WGeMwW4wsyizaw6sGcCkD0N7Z9mlgDsHVPlnMsAJgCvoK5aIiJS8qhtFgkjRzLjnYiErtuBN83sXmAj3rgg8L4FHmNmc4DxFOCbaOfcWjO7H5gEGPC5c27MYZ42Gq8r1m/AH8CPgdfaYmZDgHnAOmDGfs8bgTcG6svD/oQiIiLhRW2zSBgx55zfGUQkQpnZPUA559wAv7OIiIiI2maJXLpiLCK+MLPRQAO8b7NFRETEZ2qbJZLpirGIiIiIiIhENE2+JSIiIiIiIhFNhbGIiIiIiIhENBXGIiIiIiIiEtFUGIuIiIiIiEhEU2EsIiIiIiIiEU2FsYiIiIiIiEQ0FcYiIiIiIiIS0VQYi4iIiIiISERTYSwiIiIiIiIRTYWxiIiIiIiIRDQVxiIiIiIiIhLRVBiLiIiIiIhIRFNhLCIiIiIiIhFNhbGIiIiIiIhENBXGIiIiIiIiEtFUGIuIiIiIiEhEU2EsIiIiIiIiEU2FsYiIiIiIiEQ0FcYiIiIiIiIS0WL8DhBslSpVcikpKX7HEBGREmLmzJl/Oucq+50jnKltFhGRonQ0bXPEFcYpKSmkpaX5HUNEREoIM1vhd4Zwp7ZZRESK0tG0zepKLSIiIiIiIhFNhbGIiIiIiIhENBXGIiIiIiIiEtFUGIuIiIiIiEhEU2EsIiIiIiIiEU2FsYiIiIiIiEQ0FcYiIiIiIiIS0VQYi4iIiIiISESL8TuAiEgoycnIYeeGnUTFRhEbH0tc2TiiYo78O0TnHDkZOWTvysblOUollSKmlH7liohIMXAOsrdBdCmILu13GpGwpL/SRCTibVu1jVmvz2Lh2IWs+3ndvg8axFeKp1RSKeIrxVOtVTUadW1Eg7MaEB0bvfc0l+dYPmk5i8ctZuUPK9m0cBMZWzL2eanoUtEk1UoiuX4y5euVJ7l+MhUbVaRio4pUaFCBmNL6lSwiIgW0ex0seQ1WfwqbZwEOMChVAaJKQZ3L4bh/QelKficVCQvmnPM7Q1Clpqa6tLQ0v2OISAjYvmY73/zrG+YMnwMOap9Sm3qn1yOpdhJ5OXlk78omY0sGO9fvJGt7FtvXbmftzLVkbsskvnI8LXq14LjLj2PtzLVMf2E6fy74k+hS0dQ6uRaVm1YmqXYScWXjsCgjc3smGekZbP1jK1uWbyF9eTq7Nu76K4xBuTrl9hbK+bdydcsRFa2RL6HKzGY651L9zhHO1DaLHIHsbTBnACweBC4P4ipAtTOgbD3I3QU7lsGacd650fFQ51I49i6o0Mrf3CJBdDRtsy5PiEjEcc4x4+UZTLxvInnZebS5vQ0n33ky5VPKH/a5uVm5LJmwhDnD5jD9+en8+MyPAFQ/oToXDb+IJhc3ITY+tkA5MrdlsmnxJjYv3symRZv2br8M/4XMbZl7z4uKjaJCgwreleVGFfYpmhOqJWBmR/cPISIi4WXdN5B2K2xbCNU6w3EPQtUOB56Xlwt//gCLX4M/PoLlb0OtC6Dt2xCbFPzcImFAV4xFJKLsTt/NmGvHsHDsQo455xi6vNiFCsdUOKrX2rF+B4s+XUSV46tQ86SaRVagOufYtXGXVygv9orlzYs2772fm5m799y4hDgqNKxw0CvNpctrnFkw6Ipx4altFjkM52DeozD3IShbF054wSt0CyJrC3x3Caz/xnvuqaN19VhKvKNpm1UYi0jESF+WzohzR5C+LJ0znz6Tk+44Keyutro8x9aVWw+4yrxp0Sa2LN+Cy/vrd3p85XivSG6475XmCsdUILZMwa5qy+GpMC48tc0ifyMvF2beDotf8a4Snzrq6K76rv4cvr8McNDpK6jSvsijioQKdaUWETmENTPXMKLLCFyu45qvr6HuqXX9jnRULMooX7c85euWp37n+vs8lpuVS/qy9AOuNC/9cimzh83e59w945n3v9pcPqX8Uc3CLSIixWD3WvjxGlg3EZr8E1o+DnaUv6NrngddZsNXp8DE0+DM76Fyu6LNKxLGVBiLSIm39ue1DD9zOKXLlabnhJ5UbFTR70jFIjoumkqNK1Gp8YEzkGbtyDpwPPPCTcwbOW+f2bOjYqL2zpa9/3jmxBqJYXeFXUQkbG1dABNOhJwdcMKLcOxthX/NpEZw9gyYkAqTzoau8yG+VuFfV6QEUGEsIiXahl83MPzM4ZRKLEWvSb0KNMFWSRSXEEf1VtWp3qr6Psedc+zetHvvFeb845mXfb2MnN05e8+NjY894ArznvvxFeOD/SOJiJRcu9bA1x29orjjOKjRpeheOyEFOn4BE06CMXXh0q0Qm1B0ry8SplQYi0iJtWP9Dt49712i46K55ptrIrYo/jtmRnyleOIrxVO7be19HnN5ju1rth8wlnnd7HUsGL2AvJy8veeWqVDmr2K5UQUqNvyrcI4rGxfsH0tEJHztXuddKc5YDx3HQ42zi/49Kp4IrZ+DWXfBz/dAm1eL/j1EwowKYxEpkXIycnj/wvfZuWEnvaf0pkKDo5t5OpJZlJFUK4mkWknUO73ePo/lZuey5fct+xTMmxdtZvmk5cx5e84+5ybWTDzoeObkeslEx0UH80cSEQltLg9+vBp2r4EOnxZPUbxH4zth9aew5DWvUG5wffG9l0gYUGEsIiXSZzd/xqppq7h81OXUSK3hd5wSJzo22rsq3LAinLfvY9m7stm85K+xzHvGNS/4eAG7/ty19zyLNsqnlD/oUlNJtZKwKI1nFpEI89ON3kRbbV6Dml2L//1OedfrUj39Bkhqosm4JKKpMBaREmf2W7OZ89YcTnvoNJpc3MTvOBEnNj6Wqs2rUrV51QMe2735r/HM+ScCWzFlBdk7s/eeF1M6hgrHVNh3ErBA9+z4yvGaBExESp6lb3hb7YuhQZ/gvGfpKtB5CoypA1N7QrclRz/rtUiYU2EsIiXKnwv+ZNyt40jpmEKHhzr4HUf2U6ZCGWqdVItaJ+07C6pzjh1rdxyw1NTG3zay8NOF5GX/NZ65VLlSB50ArGLDipRKKhXsH0lEpPC2LoDpfSCpMbQdDsH88q9sbWjaH357HH4fCfWuCt57i4QQFcYiUmLkZuXyUfePiI2P5eIRFxMVrW+9w4WZkVgjkcQaiaR0TNnnsbycPLb+sfWAScD++P4P5r47F9xf5yZUSzj4eOYGycSUUpMnIiEoLxd+uAJw0P4DiPFhlv/m/4Flb0BaX6hzCUSXDn4GEZ/prwQRKTG+f+J71s9ZzxWfXEFijUS/40gR2bO2cnL9ZI4555h9HsvJyGHz0s0HTAK26LNF7Fy/c+95FmWUq1uOig33W5+5YUXK1S2nL1FExD+/PAhbfoG270D5Zv5kiIqGpvfDrH6w4n2o38ufHCI+UmEsIiXC+rnrmfLoFJpd2YzGFzT2O44ESUzpGKocV4Uqx1U54LGMrRn7jGPes//L27+QuS1z73nRcdEkN0g+aPfshGoJGs8sIsVny1yY/7RXEKf08DdLozu8LDPvgnpXa6yxRBwVxiIS9vJy8hh73VhKly/NOS+c43ccCRGly5WmRmqNA2Yld86xc8POAwrmTYs2sWT8EnIzc/eeG5cQd+BY5sBWury6GopIIWRv88YVR5eBjl/4X4hGRXvF8Zz+MP8ZaPpPf/OIBFnEFcYZ6Rl+RxCRIjbthWmsSVvDpR9cSnwlH8ZmSVgxMxKqJpBQNYG6p9bd57G83Dy2rdy2dwKwPV2zV/+0ml8/+BWX99eA5vjK8VRsVDHY8YuNmQ0FugIbnHPHB45VAN4HUoDfgcudc+nmXUZ/ATgX2AVc65ybFXhOL+DBwMs+6px7K5g/h0jYmDMANk2H9h9CfE2/03ga94M5D8CSwdDk3uBOAibis4grjDcv28x3j31H+/7t1T1OpATYvnY73z7yLY26NqLppU39jiNhLio6ivIp5SmfUp4GZzbY57GczBzSl6Xvc4V506JNPiUtFsOAQcDb+Y7dD3ztnHvCzO4P3L8P6AI0DGwnAa8AJwUK6YeBVLxp0Waa2VjnXHrQfgqRcLBpBiz6H9Q8H+pc6neav0THQcsnYPZ9sOoTqH2R34lEgibiCuP4CvF8869v2LRoE91e70ZUjMZPiISzr+//mtysXM5+7mx92SXFKqZUDJWbVKZyk8r7HO9tvX1KVLScc1PMLGW/wxcAHQP7bwGT8QrjC4C3nXMOmGZm5c2seuDcr5xzmwHM7CvgHGBkMccXCS/zHvVuUwf5m+NgGt3mFcY/9YFaF+qqsUSMiKsKy9crT4eHOzDnrTmMunIUudm5h3+SiISklT+uZM7bc2h7d1sqHFPB7zgiJVFV59zawP46oGpgvyawMt95qwLHDnVcRPbYMAVWj4Wm90HZOn6nOVBMvNelOnMTpP/sdxqRoIm4whig4yMdOevZs/jtw9/48LIPyc1ScSwSblye44vbvyCxRiKnPnCq33FESrzA1WF32BMLyMxuNLM0M0vbuHFjUb2sSGhzDqb29PaPvdPfLH+n6f3e7U83+5tDJIgisjAGaNuvLV1e7MLCMQsZc92YfSZUEZHQN/fduayduZbOT3UmLiHO7zgiJdX6QBdpArcbAsdXA7XznVcrcOxQxw/gnBvsnEt1zqVWrlz5YKeIlDxrx8Ould4V2TLV/U5zaKUrQ7XOsHkGbPnV7zQiQRGxhTFAm9vacPr/nc7cEXOZ0G8C3pfhIhLqcjJzmDRgEtVaVaNZj2Z+xxEpycYCvQL7vYAx+Y5fY56Tga2BLtcTgLPMLNnMkoGzAsdEBP66AtvsEV9jFMgJL3i3y4f5GkMkWCJu8q39te/fnp0bdjL9hekkVEug/f3t/Y4kIocxc/BMtvy+hatevQqL0qQgIkXBzEbiTZ5VycxW4c0u/QTwgZldD6wALg+cPg5vqaYleMs19QZwzm02s/8AMwLnDdwzEZdIxPvjI9j1BzToA7GJfqc5vHJNodpZ3prGxz0IceX8TiRSrCK+MDYzzv7v2excv5Ov+39NpcaVaHxhY79jicghZG7PZMp/ppDSKYUGZzU47PkiUjDOuR6HeOiMg5zrgL6HeJ2hwNAijCYS/vKPLW79X3+zHIl6V8O6L2H1p1Cvp99pRIpVRHel3sOijG5Du1HjxBp83PNj1s9d73ckETmEac9NY9fGXZzx+BlanklERMLDmi8gLxMa9oXYBL/TFFzKlVC6Gsy8A7J3+J1GpFipMA6ILRNL90+6U7pcad7r9h47N+70O5KI7Gf35t1MfWYqjS9qTK2TavkdR0RE5PCcg2/P8/ZbPelvliNlUdDsYchKh18G+J1GpFipMM4nsUYiV3xyBTvW7eCjyz8iLyfP70giks+056eRtT2LTgM7+R1FRESkYNYG5p+rejrElPU3y9FoeDPUugB+H+4V+SIllArj/dQ8sSZdX+vK75N/Z9LDk/yOIyIBGVsymP7CdJpc0oQqx1fxO46IiEjBLAzM7nzKSH9zFEbN8yFzEyx70+8kIsVGhfFBtLimBa1uaMX3j33P4nGL/Y4jIsD0/00nc1smpz14mt9RRERECmb9ZG/t4ka3Q+kw/lI35Wrvds6/dNVYSiwVxofQ5X9dqNayGqOvHs3WP7b6HUckomVuy2Ta89M49v/Zu+84q6pz/+OfZwplYIYuvUkVEJBmu6IIWLBAYjSaaywxmquJmp+Jid6oMT3RJMaoSdQYo+aKGo2CiqKgEitFRRh6770MMAxMe35/7D3JaJA656xz5nzfr9d+rXP22XPWFwtn1tlrPev8HrTq3yp0HBERkQNzh4/+X/S4xw1hsxyp7DpR4bA962H77NBpRBJCA+PPkVs/lwv/fiEVZRX8/aK/U1FaETqSSMaa9sA09mzbw9DbdbdYRETSxNtfgG0zocvXIb9r6DRHrs9tUTv7zqAxRBJFA+P9aNq1KaMfHc2aqWuYdOuk0HFEMlLprlLe/837dBvVjTaD2oSOIyIicmAl62D1uOjxoAfCZqkp9VtBoz6w+nko1WxKqX00MD6AXhf0YtB1g/jgtx+w5LUloeOIZJyPHvmIki0lnHLbKaGjiIiIHJy5d0ftiCnRNOTa4rhfR+3C+8LmEEkADYwPwhm/PoMWvVrw/GXPU7xR+xuLJEtleSUf3PMBHU7pQPsT24eOIyIicmDb58CCe6CgJxxVy5YAtT4D8trDiqdCJxGpcRoYH4Tc+rlcMPYC9mzfw7ivjcNVjU8kKeY+N5eiFUWc9N2TQkcRERE5ODO/H7V9fxo2RyKYQcdLoGgObHo/dBqRGpXSA2MzO8vMFpjZYjO7ZR+vdzCzN83sYzObZWajEpWlZd+WjLxrJIteXsT0B6YnqhsRibk77939Hs26N6P7ud1DxxERETmwzdNg7cvR4w4XhM2SKD3jStur/xE2h0gNS9mBsZllAw8AZwO9gEvMrNdnLrsNeMbdjwMuBv6QyExDrh9C17O78tp3X2Nj4cZEdiWS8Vb8cwXrPlzHCTedgGVZ6DgiIiIH9trxUXvCX4PGSKj6raIp4vN+DeVaYii1R8oOjIEhwGJ3X+rupcBTwOjPXONAQfy4EbA2kYHMjDF/HUO9xvV47pLnKCspS2R3Ihnt/V+/T17zPPpd1i90FBERkQOrmlpsOXD05WGzJFqnS6N20Z/C5hCpQak8MG4LrKr2fHV8rro7gUvNbDUwAbg+0aEaHNWAMX8dw8bCjUz6vrZwEkmETfM2sfClhQz+1mBy6+eGjiMiIrJ/leXwelwP4/TXwmZJhs6XRe0qTaeW2iOVB8YH4xLgr+7eDhgFPGFm//FnMrNrzGyGmc3YtGnTEXfa9ayuDLlhCNPum8aS17WFk0hNm3rvVHLq5TD4usGho4iIiBzYxin/ftxyWLgcyZJdF3rfBpvfg+JVB75eJA2k8sB4DVB9f5Z28bnqrgKeAXD394F6QPPPvpG7P+Tug9x9UIsWLWok3IhfjqB5z+aMu3IcJdtKauQ9RQT2bN/DrCdm0ecrfWjQokHoOCIiIvvnDm+MiB6PXhE2SzK1Oz9qFz0QNodIDUnlgfF0oJuZdTazOkTFtcZ/5pqVwHAAMzuGaGB85LeED0Ju/Vy+8MQXKN5QzCvfeiUZXYpkhJmPzaRsdxlDvjkkdBQREZEDm/3Dfz9u0CFcjmRrNhga94O5v4KKvaHTiByxlB0Yu3s58C1gIjCPqPr0HDP7sZnFX1HxHeBqM/sEGAtc4UncZLjNoDYMvX0os5+cTeHThcnqVqTW8kpn+gPTaXdiO1oPaB06joiIyP65Q+FPosfnLwubJYSOF0ft5vfC5hCpASk7MAZw9wnu3t3du7j7z+Jzd7j7+PjxXHc/2d37uXt/d096tYNT/vcU2g5py8vXvsyONTuS3b1IrbLk9SVsXbSVId/S3WIREUkDy56I2vZfhIadgkYJovt1YNkw/3ehk4gcsZQeGKeDrJwsvvDEFyjfU874q8aTxBvWIrXO9Pun06BlA3p96bNblouIiKSgJQ9FbZ/bw+YIJbcA8rvD+tegojR0GpEjooFxDWjWvRln/PoMlkxcwow/zggdRyQtbVu2jYUvL2TgNQPJrpMdOo6IiMj+bXwbNr0b3S1u0j90mnD6/Qwq9nx6rbVIGtLAuIYMunYQXc7swmvffY0tC7eEjiOSdmb8cQaWZQz8xsDQUURERPavsgwmDY0eD/x92CyhtRkFjfrAvLui/ZxF0pQGxjXEzBj9l9Hk1s/l+a8+T2V5ZehIImmjrKSMjx/5mGO+eAwFbQtCxxEREdm/WfHU6YZHQ17bsFlCy64L3a4Fr4SVz4ZOI3LYNDCuQflt8jnnj+ewZtoa3v7526HjiKSNuc/OpWRrCYOuHRQ6ioiIyP6VbIC5d0WD4vMWh06TGrpeAxhsfCt0EpHDpoFxDet9UW+O/e9jmfLjKaydsTZ0HJG08NHDH9G0a1M6ndYpdBQREZH9W/8a4DDgd2AWOk1qyMqBNmfD4gehbFfoNCKHRQPjBBh1/yjyW+fz/Fefp6ykLHQckZS2ecFmVr69kuO+fhymXzBERCSVuUPhT6FOE2h7Tug0qaXtuVG7dkLYHCKHSQPjBKjXuB6j/zqazfM388YP3ggdRySlffTnj8jKyaL/FRlc0VNERNLDutdg50LodCmYfo3+lM6XR+38e8LmEDlM+j86QY4efjSDvzmYD373AcunLA8dRyQlVZRW8Mljn9Dj/B40bNkwdBwREZH9W/ynqO39g7A5UlFOHrQbDUWzVZ1a0pIGxgk04lcjaNqlKeOuGMfenXtDxxFJOfPHzWf3pt0c9/XjQkcRERHZv4q9sOal6M5o/Zah06SmjpdAeTGsfCZ0EpFDpoFxAtVpUIcxj42haGURr33ntdBxRFLOx3/+mEYdGtHljC6ho4iIiOzfwgfAy6MiU7JvrUZE7ZqXw+YQOQwaGCdY+5Pac9LNJ/HRwx+x6JVFoeOIpIzty7ez5PUl9P9af7Ky9VeRiIikuA2TIbcRdLgodJLUVbcZtP8SrHgSSreHTiNySPTbaBKc9qPTOKrPUYy/ajwlW0tCxxFJCR898hEAx12padQiIpLiildE1ZY7fUVbNB1Im1FRu/GfYXOIHKKkDIzNrL6Z9UhGX6kop24OYx4fw+5Nu3nl+ldCxxEJrrKikpmPzqTrWV1p1KFR6DgiIiL7N/fuqO14Sdgc6aDTJZCVC6vHhU4ickgSPjA2s/OAmcCr8fP+ZjY+0f2mmtbHtWboHUOZ/eRs5j47N3QckaCWTlrKzjU7Oe5rulsskg7M7P+Z2RwzKzSzsWZWz8w6m9lUM1tsZk+bWZ342rrx88Xx653Cphc5Qu6w6tlomvBRp4ROk/qy60GjPrD0L6pOLWklGXeM7wSGANsB3H0m0DkJ/aacU249hTaD2/DS/7zErg27QscRCWbW47Oo16Qe3c/rHjqKiByAmbUFbgAGuXsfIBu4GPgVcI+7dwW2AVfFP3IVsC0+f098nUj62rUE9myAPj8MnSR9dLo0atdPCptD5BAkY2Bc5u5FnznnSeg35WTlZDHmsTGU7irlpWtewj0j/zFIhtu7Yy/znp9Hn4v7kFM3J3QcETk4OUB9M8sB8oB1wOnAs/HrjwFj4sej4+fErw8306JMSWPz4mnUrYaHzZFOOn45alc8FTaHyCFIxsB4jpl9Bcg2s25mdh/wXhL6TUktjmnB8J8PZ8H4BXzy+Ceh44gk3dxn51JeUk6/y/qFjiIiB8Hd1wC/BlYSDYiLgA+B7e5eNU9yNdA2ftwWWBX/bHl8fbNkZhapMe6w6nmo2xwKjgmdJn3ktYWWw2Hr9NBJRA5aMgbG1wO9gb3AWGAH8O0k9JuyTvj2CXQc2pFXb3iVolWfvZkuUrt9S3RoFgAAIABJREFU8tgnNOvejLbHtz3wxSJySMzsRDN7wMxmmdkmM1tpZhPM7JtmdliV7sysCdFd4M5AG6ABcFYNZL3GzGaY2YxNmzYd6duJJMbGKbB3E3T/lqpRH6qWp0HRXCiaFzqJyEFJ+MDY3Xe7+w/cfbC7D4of70l0v6nMsozRj46msqKS8V8bj1dqSrVkhm3LtrHinyvod3k/NLNSpGaZ2SvA14GJRAPX1kAv4DagHjDOzM4/jLceASxz903uXgb8AzgZaBxPrQZoB6yJH68B2seZcoBGwJbPvqm7PxT/XjCoRYsWhxFLJAk2Tonabt8MmyMdVW3btPqFsDlEDlLCF/iZ2Yv855riImAG8GCmDpKbHN2EM397Ji994yVm/GkGg68bHDqSSMLNemIWGPS9tG/oKCK10VfdffNnzu0CPoqP35hZ88N435XACWaWB5QAw4k+w98EvgQ8BVwOVO3NMj5+/n78+huuohqSjtxh0Z+gcV+odzj/62S4pgMgvzts/iB0EpGDkoyp1EuJPpgfjo8dwE6ge/w8Yw24egBdz+rK6ze/ztbFW0PHEUkod+eTxz+h87DO2rtYJDEuNbPB1e7i/od9DJwPyN2nEhXR+giYTfS7w0PA94GbzGwx0RriR+IfeQRoFp+/CbjlUPsUSQk7F8Ge9dCkf+gk6avZ8bBmPJRuC51E5ICSMTA+yd2/4u4vxselwGB3/yYwIAn9pywz47w/n0d2nWxeuOIFKisqQ0cSSZhV761i25Jt9L1Md4tFEqQdcC+w0cymmNnPzexcM2t6pG/s7j90957u3sfdv+rue919qbsPcfeu7n6hu++Nr90TP+8av770SPsXCWLhA1Hb6/thc6SzqkreG94Mm0PkICRjYNzQzDpUPYkfN4yfliah/5RW0LaAs+8/m1XvruL9374fOo5Iwnzy2CfkNsil1wW9QkcRqZXc/bvufhLQCrgV2ApcCRSa2dyg4UTS0eb3ICsXCnqGTpK+On4ZLAvm3hU6icgBJWNg/B3gHTN708zeAt4GvmtmDfj3PocZ7divHMsxFxzDm7e9ycbCjaHjiNS4spIy5jwzh14X9KJOwzqh44jUdvWBAqKiV42AtcDUoIlE0k3FXtj+CfS8KRrYyeHJrgdNBsCOeeCaGSmpLaH/p5tZFrAd6Ea0RdONQA93f9ndi939d4nsP12YGef88RzqNqrL85c9T0VZRehIIjVqwfgF7C3aq2nUIglkZg+Z2bvA08CJwHvAhXHl5yvDphNJMwvvh8oyaKriqEes27VQtiNasy2SwhI6MHb3SuCBeC3SJ/GRkVWoD6RBiwac99B5rP94Pf/86T9DxxGpUbOemEVBuwI6D+scOopIbdYBqAusJ9oyaTXRl9Micqg2x8vbWp8ZNkdt0GxI1M7X/TBJbcmYGzLZzC4wbVp6QD3H9KTfZf14+2dvs3bG2tBxRGrE7i27WTJxCX0u6YNl6a8BkURx97OAwcCv41PfAaab2Wtm9qNwyUTSUFEhtBsNuQ0PfK3sX6PekNMAts8KnURkv5IxMP4G8Hdgr5ntMLOdZrYjCf2mpbPuPYv81vk8f9nzlJWUhY4jcsTmPTePyvJK+lzSJ3QUkVrPI4XABOAV4F2gC9FSJhE5GNsLYccCaDIwdJLawQy6fD0qZrZrWeg0Ip8r4QNjd8939yx3r+PuBfHzgkT3m67qNa7H+Y+cz+Z5m3nzdpW2l/RX+FQhzbo3o1X/VqGjiNRqZnaDmT1lZiuBKcC5wHzgi8ARb9kkkjHWT47atqPC5qhN2p4btdq2SVJYUsrsmVkTMxtiZkOrjmT0m666nNGFQdcO4v3fvs+Kt1eEjiNy2Hau3cnyt5ZH06i1mkIk0ToRzdA63t27xPsN/zGu76FysCIHq/BHUKcJNNUd4xrT8nTIyYetH4ZOIvK5Ej4wNrOvA/8EJgI/its7E91vuht510iaHN2EcVeMo3RXxm/3LGlqzjNzwNE0apHkuMPdn3P3dZ93gZlpwaTI/pTvhtJt0LBr6CS1i2VB0wGw6A9RtW+RFJSMO8Y3EhUDWeHuw4DjUJXMA6rTsA5j/jqGbcu28drNr4WOI3JYCscW0uq4VjTv0Tx0FJFMMM7MfhPPzGpQddLMjjazq8xsInBWwHwiqW97YdT2vjVsjtroqFOjtmhO2BwinyMZA+M9VVs0mVldd58P9EhCv2mvw3914KTvnsSHf/qQxRMXh44jcki2Ld3Gmmlr6HOx7haLJIO7DwcmExW9nGNmRWa2Bfgb0Aq43N2fDZlRJOWtGBu1TfqFzVEbdbo0ahfcGzaHyOfISUIfq82sMfAC8LqZbQO0cPYgDfvxMBa9vIjxV43n2tnXUr9J/dCRRA5K4VPRt+4aGIskj7tPIKpILSKHY9O7UdugU9AYtVJ+V8gtgKK5oZOI7FMyqlJ/wd23u/udwO3AI8CYRPdbW+TUy2HM42Mo3lDMqze8GjqOyEErHFtI+5Pb06hDo9BRREREDqyyHLZOhx43RmtipWaZQZerYdsnWmcsKSlZVamzzawNsAyYSTSlSw5Sm4FtOOW2U5j1t1nM+8e80HFEDmhj4UY2Fm5U0S0REUkfK+OVBo16h81RmzUdCJV7//3PWiSFJKMq9fXABuB14OX4eCnR/dY2p/zvKbQe2JqX/uclijcWh44jsl+FTxViWUavL/UKHUVEROTgbJ8VtVVrYaXmtTk7aqumrIukkGRVpe7h7r3d/dj46JuEfmuV7Nxsxjw2hr079vLS/7yEu4eOJLJP7k7h2EI6D+9Mw5baGUYk2czsiYM5JyKfsX12dLc4R/VcEqZOY2hxCmz7OHQSkf+QjIHxKqAoCf3Uekf1PorTf3o685+fz+z/mx06jsg+rZ2+lm1Lt2katUg4n5oHambZwMBAWUTSQ3kxrH0JGh8bOknt13QAbH4P9mwMnUTkUxI2MDazm8zsJmAp8JaZ3Vp1Lj4vh+GE/3cCHf6rAxO+NYFd63eFjiPyH2aPnU12nWyO+cIxoaOIZJT4c3Yn0NfMdsTHTmAjMC5wPJHUtmV61DYdFDZHJqjaz3jdxLA5RD4jkXeM8+NjJdH64jrVzuUnsN9aLSs7i/P/cj5lu8t46863QscR+ZTKikrmPD2Hrmd3pV7jeqHjiGQUd/+Fu+cDd7t7QXzku3szd781dD6RlLYmLn/T8eKwOTJB2/Mgux4sfjB0EpFPSdg+xu7+o0S9d6Zr1q0Zg64dxPQHpnP8jcfT4pgWoSOJALDy7ZXsWrdL06hFwnrFzIZ+9qS7/zNEGJG0sOGNqK3fJmyOTJCVAw2Phm0zwSu1NZakjGRUpX7dzBpXe97EzDR34ggNvW0odRrUYfKtk0NHEfmX2WNnk9sglx7n9QgdRSST3VztuB14EbgzZCCRlOaVsHMhdL0m2mtXEq/nTdG67p1LQicR+ZdkfEXTwt23Vz1x923AUQfzg2Z2lpktMLPFZnbL51xzkZnNNbM5ZvZkDWVOeQ1aNODkW05mwbgFrHh7Reg4IlSUVjDv2Xn0HN2T3Lzc0HFEMpa7n1ftGAn0AbaFziWSsopXRoO0pqpRlzRNjova1c+HzSFSTTIGxhVm1qHqiZl1BA6411BcRfMB4GygF3CJmfX6zDXdgFuBk929N/Dtmgye6k648QTy2+bz+ndf1/ZNEtzSSUsp2VpC74t7H/hiEUmm1YCq4Yl8nhXxfZVG+vxKmqp/1irAJSkkYWuMq/kB8I6ZTQEMOAW45iB+bgiw2N2XApjZU8BoYG61a64GHojvQuPuGVX3PTcvl9N/ejrjrhzH3Gfn0vtC/YUu4RSOLaRek3p0PbNr6CgiGc3M7uPfX0BnAf2Bj8IlEklx22ZGre4YJ092Xej0VVj+hNYZS8pI+H+F7v4qMAB4GngKGOjuB/P1UFuiPZCrrI7PVdcd6G5m75rZB2Z2Vk1kTid9v9qXo449ism3TqaitCJ0HMlQZbvLmP/CfI654Biy62SHjiOS6WYAH8bH+8D33f3SsJFEUljRvH9XSpbkaTogaqsqgosElpSvZ9x9s7u/BAxy9801+NY5QDfgNOAS4OHqhb6qmNk1ZjbDzGZs2rSpBrsPLys7i5F3jWTbkm3M+NOM0HEkQy18eSGlu0rpc7GqUYuE5u6PAWOBj4FZwPSwiURSWGU57FwAjXod+FqpWV2uitptH4fNIRJL9ryF8w/h2jVA+2rP28XnqlsNjHf3MndfBiwkGih/irs/5O6D3H1Qixa1b2ujLmd2ofPwzkz58RT2FO0JHUcy0Jyn5tCwVUM6ndYpdBSRjGdmo4AlwO+B+4HFZnZ22FQiKWrrh1BZBgUaGCddbj4U9Ij+HYikgGQPjA+lBv50oJuZdTazOsDFwPjPXPMC0d1izKw50dTqpTWQM62YGSPvHknJlhLe/dW7oeNIhtlTtIeFLy+k10W9yMrWGiGRFPBbYJi7n+bupwLDgHsCZxJJTVX7F7cZFTZHpmrcF9a8CKUqnC/hJey3WDP7VdxeWO30QVc1cPdy4FvARGAe8Iy7zzGzH5tZ1Z3nicAWM5sLvAnc7O5bauQPkGZaH9eavpf25YN7PmDH6h2h40gGmf/CfCr2VnDsJceGjiIikZ3uvrja86XAzlBhRFLa9lnQoBPUax46SWZqfWbUblV9QAkvkbd3RpmZEW2nBIC7Vx7KG7j7BHfv7u5d3P1n8bk73H18/Njd/SZ37+Xux7r7UzX6J0gzw346DK903rz9zdBRJIMUji2kcafGtD3+s7XxRCSQGWY2wcyuMLPLgReB6Wb2RTP7YuhwIinDK2H9JGg2OHSSzNU2vtc17zdhc4iQ2IHxq8A2oK+Z7TCzndUO3dJMgMYdGzPkhiHMfGwmG2ZtCB1HMkDxpmKWTlpK74t7E30PJiIpoB6wATiVaLnRJqA+cB5wbrhYIilm1zLYuxlanRE6Seaq1wJyG0Nxxq2ElBSUsH2M3f1m4GYzG+fuoxPVj3zaKf97Ch8/8jGTvj+J/37lv0PHkVpu7rNz8QrXNGqRFOLuV4bOIJIWdiyIWlWkDqvHjTDnJ1C+G3LyQqeRDJaMfYxHm1lLMzs3PmpfWegUUr9JfYbeNpTFry5m6SR9+yaJVTi2kBa9WnDUsUeFjiIiMTNrYWb/a2YPmdlfqo7QuURSzvZPoja/e9gcma5Jv2ha++oXQieRDJfwgXFcfGsacCFwETDNzL6U6H4z2eBvDqZxp8a8fvPreKWHjiO1VNGqIla+vVLTqEVSzzigETAJeLnacdjMrLGZPWtm881snpmdaGZNzex1M1sUt03ia83Mfm9mi81slpkNOOI/kUgiLH4IcvKhbrPQSTJbi1Oidv3ksDkk4yVjb5XbgMHufrm7XwYMAW5PQr8ZK6duDqf/7HTWz1zP7Cdnh44jtdScZ+YAaBq1SOrJc/fvu/sz7v5c1XGE73kv8Kq79wT6Ee0WcQsw2d27AZPj5wBnA93i4xrgj0fYt0jNqyyDkrXQagToy92w6jWPqlOvnQDlJaHTSAZLxsA4y903Vnu+JUn9ZrQ+F/eh9cDWvPGDNyjfUx46jtRChWMLaTOoDU27Ng0dRUQ+7SUzq7FNWc2sETAUeATA3UvdfTswGngsvuwxYEz8eDTweLxzxAdAYzNrXVN5RGrElhlQWQodLjzwtZJ4rc+EPeth5d9DJ5EMlowB6qtmNjHeNuIKoulcE5LQb0azLGPk3SMpWlnE1Pumho4jtcyWRVtY9+E6+lzSJ3QUEflPNxINjkuq7QpxJLtBdCaqbP2omX1sZn82swZAS3dfF1+zHmgZP24LrKr286vjcyKpY+kjkF0P2pwVOokAdL8BLAfWTQydRDJYMopv3Qw8CPSNj4fc/fuJ7leg87DOdBvVjbd/9ja7t+wOHUdqkcKxhWDQ+6LeoaOIyGe4e767Z7l7fXcviJ8XHMFb5gADgD+6+3FAMf+eNl3VpwOHVNTCzK4xsxlmNmPTpk1HEE/kMOyYD82OhzpNQicRgKzsqDr4irFRIS6RAJIypdnd/+HuNwFvuPvzyehTIiN+NYLSnaW8/fO3Q0eRWsLdKRxbSMdTOlLQ7kh+1xaRmmRmPeN2wL6OI3jr1cBqd6+afvQs0UB5Q9UU6bitWja1Bmhf7efbxec+xd0fcvdB7j6oRQttWCFJVFkBm96F/K6hk0h1HS4EHIqXh04iGSrZa31/nOT+Mt5RfY6i/5X9mX7/dLYt2xY6jtQCG2ZtYPP8zZpGLZJ6borb3+zj+PXhvqm7rwdWmVmP+NRwYC4wHrg8Pnc5UTVs4vOXxdWpTwCKqk25Fglv87tRm9chbA75tFYjo3bbJ2FzSMbKSXJ/KvsXwGk/Oo3ZT87mjR+8wQVPXhA6jqS5wrGFWLbR60u9QkcRkWrc/Zq4HZaAt78e+D8zqwMsBa4k+nL9GTO7ClhBtCUjRHVERgGLgd3xtSKpY+eiqO381bA55NMax1+4L7gX2n8hbBbJSMkeGH8jyf0JUNC2gBNvOpG3f/Y2J950Im0GtQkdSdKUu1P4VCFdRnYhr3le6DgikiTuPhMYtI+Xhu/jWge+mfBQIodr1fNg2ZDX/sDXSvLkNID8brBrSegkkqESPpXazPLM7HYze9jdp5lZNzM7N9H9yqed/L2TyWuRx+s3v070O4vIoVv9wWqKVhRpGrWIiKSvosJoEJaV7PtDckBdvg67V8P22aGTSAZKxhrjR4G9wInx8zXAT5PQr1RTt6Aup/7wVJa/tZxFExaFjiNpqnBsIdl1s+k5pmfoKCKyD/G6Xt0GE/k8lWXRwKv79aGTyL60PD1qtW2TBJCMgXEXd78LKANw991orXEQA68ZSNNuTZn0vUlUlqsUvhyayvJK5jwzh+7ndKduQd3QcURkH+JpzBNC5xBJWcUrwSugYZfQSWRfmg2C+m1g26zQSSQDJWNgXGpm9Yn3NzSzLkR3kCXJsnOzGf6L4Wyau4mZj80MHUfSzPK3llO8oVjTqEVS30dmNjh0CJGUtCnevlJbNaWuxn1hnb7fk+RLxsD4TuBVoL2Z/R8wGfh+EvqVfTjmi8fQ7sR2vHXHW5QWl4aOI2mk8KlC6uTXods53UJHEZH9Ox5438yWmNksM5ttZrr9IgKw+oWobdw3bA75fPldYe8W2F4YOolkmIQPjN39NeCLwBXAWGCQu7+Z6H5l38yMkXePZOfanXxwzweh40iaKN9bzrzn5tFzTE9y6+eGjiMi+3cm0AU4HTgPODduRWTHQmg5DOo0Cp1EPk+Xq6N264dhc0jGSUZV6snuvsXdX3b3l9x9s5lNTnS/8vk6nNyBnl/oybu/epfijcWh40gaWDJxCXu276HPxZpGLZLq3H0F0B44PX68m+TMEBNJbRV7YedCaH5S6CSyP416Rdtpzf9N6CSSYRL2QWlm9cysKdDczJqYWdP46AS0TVS/cnCG/2I4ZSVlTPnxlNBRJA0Uji2kfrP6HD3y6NBRROQAzOyHREuWbo1P5QJ/C5dIJEXsWBAV3mqkL3lTWlYONBkAu9eAthiVJErkN8jfAD4EesZt1TEOuD+B/cpBaN6jOQO/MZAPH/yQLQu3hI4jKay0uJQF4xfQ60u9yM7NDh1HRA7sC8D5QDGAu68F8oMmEkkFRfGa1cYaGKe8zpdC6VbYsz50EskgCRsYu/u97t4Z+K67H+3uneOjn7trYJwCTr3jVHLq5TD5Vs1sl8+3YPwCynaXqRq1SPoojbdtqtoNokHgPCKpYXshWA7kdw+dRA6k8bFRu3pc2BySUZJRfOs+M+tjZheZ2WVVR6L7lQNr2LIhJ33vJOb9Yx6r3lsVOo6kqMKxheS3zafjKR1DRxGRg/OMmT0INDazq4FJwMOBM4mEVzQHCnpAdp3QSeRAmgyI2g1vBY0hmSUZxbd+CNwXH8OAu4imeEkKOPGmE2nYuiGv3/w6rnUc8hklW0tY/Opi+lzcB8uy0HFE5CC4+6+BZ4HngB7AHe5+X9hUIimgqFDri9NFnUbQZhTsmBs6iWSQZFSp/BIwHFjv7lcC/QDVyE8RdRrUYdiPh7HqvVXMf2F+6DiSYuY+N5fKskqO/cqxoaOIyEEys6uA5e5+s7t/191fD51JJLjyYti1VOuL00mjPlHBtMry0EkkQyRjYFzi7pVAuZkVABuJtpGQFNH/iv606NWCSd+fREVZReg4kkIKnyykWfdmtDquVegoInLwOgAPmtlSM/u7mV1vZv1DhxIJqii+89iod9gccvAa9YbKUq0zlqRJxsB4hpk1Jlrf9CHwEfB+EvqVg5SVk8WIX41g66KtfPTwR6HjSIrYsWYHy6csp89X+mCmadQi6cLdf+jupwO9gbeBm4k+f0Uy184lUavCW+mjzaio3aq/viQ5klF86zp33+7ufwJGApfHU6olhXQ7pxsdT+3IW3e+xd6de0PHkRQw5+k54HDsJZpGLZJOzOw2M3sFeA3oCnwXaBc2lUhgxcujtoEKSaaNes2hUS9Y+XToJJIhklF86197Abn7cnefVf2cpAYzY+TdI9m9aTfv3vVu6DiSAgrHFtJ6YGuadW8WOoqIHJovAs2IqlH/Axjn7uvCRhIJbOXfoU5TyG0YOokcinqto7XhpdtDJ5EMkLCBsZnVM7OmQHMza2JmTeOjE9A2Uf3K4Ws7uC19Lu7D+795n51rd4aOIwFtWbSFtTPWau9ikTTk7gOAEcA0oplas83snbCpRAIrXg65BaFTyKHqcUPUFs0Lm0MyQiLvGH+DaE1Tz7itOsYB9yewXzkCp//sdCrLK3nzh2+GjiIBFY4tBIM+X9bAWCTdmFkf4L+By4EvA2uAN4KGEgmpvBhKt0LXq0MnkUNVVSxthwbGkngJGxi7+73u3hn4rrsf7e6d46Ofu2tgnKKaHN2EId8awsy/zGTjnI2h40gA7s7sJ2fTcWhHCtrp23WRNPRLoAD4PXCMuw9z9zsCZxIJZ9fSqM3T+uK006ATZNWBxQ+HTiIZIBlVqdebWT78qyDIP8xsQBL6lcN0yg9OoU5+HSZ9f1LoKBLA+o/Xs2XBFu1dLJKm3P1c4B5gB9DDzHIDRxIJa0VcvCm/W9gccuiysqOCabuWhE4iGSAZA+Pb3X2nmf0X0ZqnR4A/JqFfOUx5zfI45X9PYdHLi1j25rLQcSTJZo+dTVZuFsdccEzoKCJyGMzsVGAR8ADwB2ChmQ0Nm0okoD3xDLhmg8PmkMPT9RuwdxPs3RI6idRyyRgYV8TtOcBD7v4yUCcJ/coRGHL9EAraFzDpe5PwSg8dR5LEK505T82h65ldyWuWFzqOiBye3wJnuPup7j4UOJPoDrJIZtq9EpoOArPQSeRwNOoVtUsfDZtDar1kDIzXmNmDRAVAJphZ3ST1K0cgt34up//sdNbOWMucZ+aEjiNJsvKdlexYvYM+X1HRLZE0luvuC6qeuPtCQNOpJXMVr4QGHUKnkMPV6oyo3aVZjJJYyRigXgRMBM509+1AU+DmJPQrR6jvf/elZb+WTL51MuV7y0PHkSSY/eRscvNy6XF+j9BRROTwzTCzP5vZafHxMDAjdCiRIPZuiSoaN+wSOokcrqxsaHY87Fhw4GtFjkDCB8buvtvd/+Hui+Ln69z9tUT3K0fOsoyRd49k+/LtTP/D9NBxJMEqSiuY+/e59BjdgzoNtNpBJI1dC8wFboiPufE5kcyz4a2obaSZUGmtoAdsfi90CqnlNKVZ9qvLyC50OaML//zJPynZVhI6jiTQkteXULK1hD6X6JcHkXTm7nuB+4EfAT8EHojPiWSevXHhrdYjw+aQI1PQAypKYMOU0EmkFtPAWA5oxF0j2LN9D+/84p3QUSSBZv9tNvWb1qfrmV1DRxGRI2Bm5wBLgHuJBsiLzezssKlEAtm1DCwL6rYInUSORNdvRO3H3wmbQ2q1hA+Mzex6M2uS6H4kcVr1a0W/y/ox9fdT2b5ie+g4kgB7d+xl/gvz6X1xb7LrZIeOIyJH5jfAMHc/zd1PBYahqtSSqRb9EfLaQ1ZO6CRyJOo2g2ZDoGgeuHZLkcRIxh3jlsB0M3vGzM4yU638dDTsJ8MAePP2NwMnkUSY+9xcyveU0++r/UJHEZEjt9PdF1d7vhTYGSqMSDDuULEbmhwXOonUhM6XRf8+S9aFTiK1VDKKb90GdAMeAa4AFpnZz81M5QHTSKP2jTjh2ycw62+zWPex/kKqbWY9MYumXZvS9vi2oaOIyJGbYWYTzOwKM7sceJHoC+ovmtkXQ4cTSZqyIvBKaHFK6CRSE/K7R+3ORWFzSK2VlDXG7u7A+vgoB5oAz5rZXcnoX2rGf936X9RvWp9J35uEaxpLrVG0qojlby3n2EuPRRM6RGqFesAG4FTgNGATUB84Dzg3XCyRJNu1JGq1h3HtkN8talc8GTaH1FoJX3BhZjcClwGbgT8DN7t7mZllAYuA7yU6g9SMeo3qMfT2oUz89kSWvLZERZpqidn/Nxsc+l7aN3QUEakB7n5l6AwiKWH7nKjVVk21Q9UXHLpjLAmSjDvGTYAvuvuZ7v53dy8DcPdKDvDNdbwmeYGZLTazW/Zz3QVm5mY2qGajy2cNvnYwTY5uwqTvTaKyojJ0HDlC7s6sJ2bR/qT2NO3SNHQcEUlRZpZtZh+b2Uvx885mNjX+fH7azOrE5+vGzxfHr3cKmVsyXNEcyKoD+foiv1awLOj0Vdi5+MDXihyGhA6MzSwbuNjdV+zrdXefd4CffQA4G+gFXGJmvfZxXT5wIzC1RkLLfmXXyeb0n5/OhlkbmPW3WaHjyBFa//F6Ns3dRN+v6m6xiOzXjUD1z+xfAfe4e1dgG3BVfP4qYFt8/p74OpEwiuZE+9+qInXtkd8Vdq+C8pLQSaQWSujA2N0rgAVmdjiLO4YAi919qbuXAk8Bo/dx3U+IPnj3HH5SORS9L+qIVrVnAAAgAElEQVRNm8FtePO2NykrKQsdR47AJ098QnadbHpf1Dt0FBGpAWaWZWYX1fB7tgPOIVoORby7xOnAs/EljwFj4sej4+fErw/XbhQSTNFcaKTPt1qlap3xFt0Pk5qXrKnUc8xsspmNrzoO4ufaAquqPV8dn/sXMxsAtHf3l2surhyImTHy7pHsWL2DqffqL6Z0VVleSeHYQrqd0436TeuHjiMiNSBeplTTtTt+F79n1fqZZsB2dy+Pn1f/fP7XZ3f8elF8vUhyVVbA7pXQUJug1CpNB0btuolhc0itlIy5Jbcn4k3j4l2/JdoC6kDXXgNcA9ChgyoT1oROp3ai+3ndeecX7zDg6wPIa54XOpIcoiWvL6F4Q7GmUYvUPpPM7LvA00Bx1Ul333qob2Rm5wIb3f1DMzut5iLqs1kSbM968ArIaxc6idSkgu6Q1wGKl4dOIrVQMvYxngIsB3Ljx9OBjw7iR9cA7as9bxefq5IP9AHeMrPlwAnA+H0V4HL3h9x9kLsPatGixWH9OeQ/jfjlCEp3lfLmHW+GjiKHYdYTs6jXpB7dRnULHUVEataXgW8C/wQ+jI8Zh/leJwPnx5+zTxFNob4XaGxmVV+uV/98/tdnd/x6I2DLvt5Yn82SUDviJfF57fd/naSfgp6w4inQ1qFSwxI+MDazq4nWGT0Yn2oLvHAQPzod6BZXvqwDXAz8awq2uxe5e3N37+TunYAPgPPd/XA//OUQtejVgiHXD2HGH2ew7I1loePIIdi7cy/zX5hP7y/3JqeuipKI1Cbu3nkfx9GH+V63unu7+HP2YuANd/9v4E3gS/FllwPj4sfj4+fEr7/h2vheQlg+Nmrzu4fNITUvL165UbVPtUgNScYa428SfeO8A8DdFwFHHeiH4rVJ3wImElXCfMbd55jZj83s/ATmlUMw/OfDada9GeOuHMeeItU/SxfznptHeUk5/b7aL3QUEUkAM+tjZheZ2WVVRw138X3gJjNbTLSG+JH4/CNAs/j8TcDnbrUoklA75keFtwo0K6rWOfqKqN2pgbHUrGTcKtrr7qVVRSnjqVUH9e2xu08AJnzm3B2fc+1pRxZTDkduXi5jHhvDX07+CxNvmsjoR/ZVOFxSzcy/zqRp16a0O1Frr0RqGzP7IXAa0VaHE4i2PXwHePxI3tfd3wLeih8vJdo94rPX7AEuPJJ+RI5YeTFsfg861/T3QZISqgqq6Y6x1LBk3DGeYmb/C9Q3s5HA34EXk9CvJEm7E9px8i0nM/MvM1nw4oLQceQAti7ZyoopK+h/ZX+0i4pIrfQlYDiw3t2vBPoRrfUVyQxF8frigh5hc0hi1G8N2fWgqDB0EqllkjEwvgXYBMwGvgFMcPcfJKFfSaJT7ziVln1b8uLVL7J7y+7QcWQ/Zv51JpZl9LtM06hFaqmSeNumcjMrADby6WKWIrXb7tVR2/rMsDkkMSwL6reDJY8c+FqRQ5CMgfH17v6wu1/o7l9y94fN7MYk9CtJlFM3hzGPj6FkawkTrptw4B+QICorKvnksU/ockYXCtoVhI4jIokxw8waAw8TVaT+CHg/bCSRJKoaGKside3V4iSoLIUK1beRmpOMgfHl+zh3RRL6lSRr1a8Vp915GnOemUPh05rekoqWTV7GjlU76H9l/9BRRCRB3P06d9/u7n8CRgKXx1OqRTLD7pWQVQfqNg+dRBKl1RlRu0u7okjNSVjxLTO7BPgK0NnMxld7KR/Ymqh+JayTv3cyC8YvYMJ1E+g4tCP5rfNDR5JqZj46k/pN69NjtNZdidQ2ZjZgf6+5+0fJzCMSzML7Ib9rNOVWaqeCeBuupY/CcXeFzSK1RiKrUr8HrAOaA7+pdn4nMCuB/UpAWTlZjHlsDA8e9yAvXv0il7x4iQo8pYiSbSXMe34eA64eoL2LRWqn3+znNQdOT1YQkWAqK6CiBPI6hE4iidR0UNSWrA+bQ2qVhP127O4rgBVm9k93n1L9NTP7FdEeiFILNe/RnBG/HMGrN77Kx3/5mAFXfe5NDEmiwrGFVOyt4LivHRc6iogkgLsPC51BJLjSeFJim1Fhc0himcFRp0Hx0tBJpBZJxhyTkfs4d3YS+pWAhnxrCJ2GdWLityeyffn20HGEaBp1y34taX1c69BRRCSBzCzXzG4ws2fj41tmlhs6l0hS7IoHSvX1WVfrNTwadmovY6k5CRsYm9m1ZjYb6Glms6ody9BU6lrPsozRj44Gg3FXjsMrPXSkjLZh9gbWzliru8UimeGPwEDgD/ExMD4nUvsteyxqG/cNm0MSL78L7FkPxStDJ5FaIpF3jJ8EzgPGxW3VMdDdL01gv5IiGndszFm/O4vlby1n6n1TQ8fJaDMfnUlWbhbHfuXY0FFEJPEGu/vl7v5GfFwJDA4dSiQpynZGRbeqijNJ7dW4X9Sunxw2h9QaCRsYu3uRuy9390vi9cYlRMU/GpqZKiJkiP5X9qfbOd2YfMtkNi/YHDpORqoorWDW32bRc3RP8prnhY4jIolXYWZdqp6Y2dFARcA8IsmzdxM0GRg6hSRD6zOiL0GKtWWT1IyErzE2s/PMbBGwDJgCLAdeSXS/khrMjPMePo/cvFxeuOwFKssrQ0fKOPPHzWf3pt3au1gkc9wMvGlmb5nZFOAN4DuBM4kkx65l0ED3XzJCVi7Ubwcrng6dRGqJZBTf+ilwArDQ3TsDw4EPktCvpIj81vmM+sMo1kxbw7t3vRs6Tsb58MEPadShEV3O7HLgi0Uk7bn7ZKAbcANwPdDD3d8Mm0okCSr2wK7F0Kh36CSSLHWbRv/OXTde5MglY2Bc5u5bgCwzy4o/nAcloV9JIX2+3IfeX+7NW3e+xfpPtOdcsmxdvJVlk5cx4JoBZGUn4393EQnFzAabWSsAd98L9Ad+AtxtZk2DhhNJhh0LogFSo16hk0iydP1G9O+8ZG3oJFILJOM35e1m1hD4J/B/ZnYvUJyEfiXFjHpgFHnN8njhshco31seOk5G+PChD8nKyVI1apHM8CBQCmBmQ4FfAo8DRcBDAXOJJEfR3KjVHePM0fDoqN2l/YzlyCVjYDwa2A38P+BVYAlRdWrJMHnN8jjv4fPYMGsDU340JXScWq98bzkzH51Jj/N7kN86P3QcEUm8bHffGj/+MvCQuz/n7rcDXQPmEkmOojlg2ZDfLXQSSRYNjKUGJXxg7O7F7l7p7uXAy8B98dRqyUDdz+1O/6/1591fvcvqD1aHjlOrzfvHPHZv3s3Ab6g6p0iGyDaznPjxcKKiW1Vy9nG9SO1SNDcaFGfXDZ1EkiUvLrS26E9hc0itkLCBsZmdEFfE/IeZHWdmhUAhsMHMzkpUv5L6zrrnLAraF/DC5S9QtrssdJxa68MHP6TJ0U04esTRoaOISHKMBaaY2TiiLRLfBjCzrkTTqUVqt6I5Wl+cabLrRHeNy7aHTiK1QCLvGN8P/Jzog/oN4Ovu3goYCvwigf1KiqtbUJfRj45my8ItTLplUug4tdLm+ZtZMWUFA64egGVZ6DgikgTu/jOibZn+CvyXu3v8UhZRdWqR2qtirypSZ6r2F0TbdKkytRyhRA6Mc9z9NXf/O7De3T8AcPf5CexT0kTnYZ0ZcsMQpt03jWVvaGP2mlZVdEt7F4tkFnf/wN2fd/fiaucWuvtHIXOJJNzOeMuegmNCJ5Fka3g0VJaqMrUcsUQOjKt/bVPymdccyXgjfjGCZt2bMe7Kcewp2hM6Tq1RtruMTx77hJ5f6EnDlg1DxxEREUm84hVR26BT0BgSQFUBro3vhM0haS+RA+N+ZrbDzHYCfePHVc+PTWC/kiZy83IZ89gYdqzewcSbJoaOU2vMfnI2JVtLGPzNwaGjiIiIJMfuVVHboH3YHJJ8jeNhxfrXw+aQtJewgbG7Z7t7gbvnu3tO/LjqeW6i+pX00u6Edpx8y8nM/MtMFry4IHSctOfuTP39VFr2a0nHoR1DxxEREUmOVc9Fbb3WYXNI8tVvHU2h11RqOULJ2MdYZL9OveNUWvZtyYtXv8juLbtDx0lrK6asYOPsjRx/w/GYqeiWiIhkiN2rIacBZGWHTiIhNOoN6yaCa7WmHD4NjCW4nLo5jHl8DCVbS5hw3YTQcdLa1N9PpX6z+vS5pE/oKCIiIsnhHk2l7vL10EkklHpHAR7tZS1ymDQwlpTQql8rTrvzNOY8M4fCpwtDx0lL25dvZ8G4BQy8ZiC59bVaQUREMkRZEZTvgjytL85YXb4WtTsXhs0haU0DY0kZJ3/vZNoe35YJ101g57qdoeOknWkPTAODQdcOCh1FREQkeYqXR22DDkFjSEANu0TtrqVhc0ha08BYUkZWThZjHhtDWUkZL179Iq51IgettLiUj//8Mb0u6EWj9o1CxxEREUme7XOitqBX2BwSTp3G0XTqxQ+FTiJpTANjSSnNezRnxC9HsOjlRXz8l49Dx0kbs/42iz3b9zDkhiGho4iIiCTX9lmQlQsF3UMnkZCOOi2aSl2yLnQSSVMaGEvKGfKtIXQa1omJ357I9uXbQ8dJeV7pTL13Kq0HtKb9SVpfJSIiGWb77OhucZbqa2S0bv8TtZveCZtD0pYGxpJyLMsY/ehoMBh35Ti8UlOq92fRhEVsnreZE246QVs0iYhI5tk+Cxr3DZ1CQis4JmpXvxg2h6QtDYwlJTXu2JizfncWy99aztT7poaOk9LevetdCtoX0Pui3qGjiIiIJNferVCyBppoYJzx6reCgh5QsjZ0EklTGhhLyup/ZX+6ndONybdMZvOCzaHjpKTVH6xm5dsrOfGmE8nOzQ4dR0REJLm2z45a3TEWgCYDYdfi0CkkTWlgLCnLzDjv4fPIzcvlhcteoLK8MnSklPPe3e9Rr3E9Bnx9QOgoIlJLmVl7M3vTzOaa2RwzuzE+39TMXjezRXHbJD5vZvZ7M1tsZrPMTH9BSeIUxRWpG2nWlAD5XWD3KqgoDZ1E0pAGxpLS8lvnM+oPo1gzbQ3v3vVu6DgpZcuiLcx7fh6DrhtEnYZ1QscRkdqrHPiOu/cCTgC+aWa9gFuAye7eDZgcPwc4G+gWH9cAf0x+ZMkYxcsgux7UbxM6iaSChl3AK2HdK6GTSBrSwFhSXp8v96H3l3vz1p1vsf6T9aHjpIz3f/s+2bnZHH/98aGjiEgt5u7r3P2j+PFOYB7QFhgNPBZf9hgwJn48GnjcIx8Ajc2sdZJjS6bYtQwadAQVnxSAViOidtussDkkLWlgLGlh1AOjyGuWxwuXvUD53vLQcYIr3ljMzEdn0u/yfjRs1TB0HBHJEGbWCTgOmAq0dPeqDUPXAy3jx22BVdV+bHV8TqRmVZbDquegfrvQSSRV5LWF+m21zlgOiwbGkhbymuVx3sPnsWHWBqb8aEroOMFNu38aFaUVnPidE0NHEZEMYWYNgeeAb7v7juqvubsDh7S3npldY2YzzGzGpk2bajCpZIySNVHb9LiwOSS15HeBZY+Da7tPOTQaGEva6H5ud/p/rT/v/updVn+wOnScYPYU7WHq76fSc0xPmvdoHjqOiGQAM8slGhT/n7v/Iz69oWqKdNxujM+vAdpX+/F28blPcfeH3H2Quw9q0aJF4sJL7bVrWdS2PjNsDkktTeIvSkrW7f86kc/QwFjSyln3nEVB+wJeuPwFynaXhY4TxLT7prG3aC9DbxsaOoqIZAAzM+ARYJ67/7baS+OBy+PHlwPjqp2/LK5OfQJQVG3KtUjNWTE2aht0DptDUkubUVGr6dRyiDQwlrRSt6Auox8dzZaFW5h0y6TQcZJu7869vP/b9+l+bndaD1AtGxFJipOBrwKnm9nM+BgF/BIYaWaLgBHxc4AJwFJgMfAwcF2AzJIJiuOl7A01MJZq8rtG7U4NjOXQ5IQOIHKoOg/rzJAbhjDt99PoOaYnnU/PnA/E6Q9MZ8+2PQy9XXeLRSQ53P0d4PNK/g7fx/UOfDOhoUQAipdC+y+B6T6PVJPXASwH1oyHLl8LnUbSiP4mkbQ04hcjaNa9GeOuHMeeoj2h4yRFaXEp7//mfbqe1ZW2Q1TgVUREMphXRmuMGx4dOomkmqwcqN9aWzbJIdPAWNJSbl4uYx4bw47VO5h408TQcZJixp9msHvzbobeobvFIiKS4XYsgMpSaHRM6CSSio6+AnavgIrS0EkkjaT0wNjMzjKzBWa22Mxu2cfrN5nZXDObZWaTzaxjiJwSRrsT2nHyLScz8y8zWfDigtBxEqpsdxnv3f0enYd3pv2J7Q/8AyIiIrXZ1hlR23Rw2BySmvK7RbMKipeFTiJpJGUHxmaWDTwAnA30Ai4xs16fuexjYJC79wWeBe5KbkoJ7dQ7TqVl35a8ePWL7N6yO3SchJl2/zSKNxRz6g9PDR1FREQkvC0zIDsPCnqGTiKpqGFcgGvz1LA5JK2k7MAYGAIsdvel7l4KPAWMrn6Bu7/p7lWjoQ+I9kqUDJJTN4cxj4+hZGsJE66bEDpOQuwp2sM7v3yHrmd1peMpmhQhIiLClmnQdABkZYdOIqmoaor92pfC5pC0ksoD47bAqmrPV8fnPs9VwCsJTSQpqVW/Vpx252nMeWYOhU8Xho5T49779Xvs2baH0392eugoIiIi4ZWsh63ToOWw0EkkVdVpDM1OgF3LQyeRNJLKA+ODZmaXAoOAuz/n9WvMbIaZzdi0aVNyw0lSnPy9k2l7fFsmXDeBnet2ho5TY4o3FvPBPR/Q68Je2rdYREQEYM2L0frRDheGTiKprOkA2DodKjJj9xI5cqk8MF4D/P/27jy+qvrM4/jnCYEAYoIYBNkEkUWkECygKEVBRgGVaJVBdKrF1nbqYFvrtKUbjra2tTq1nWqdkYJUi1pXxIpSdnlhQVABEUQQqATZkX3L8swf54IRAgRI8js35/t+vfI69557zznf8CN58pztlr7LULPUvM8xs77AT4CB7r6vrBW5+2Pu3tXduzZs2LBSwkpYGZkZXPPnayjcU8grt71C9DGa6W/mL2dStKeI3j/XXnEREREA1k+DjCzI6Rg6icRZznnRdP20sDkkbcS5MZ4LtDGzVmZWC7gBGF/6DWbWBfg/oqZ4Q4CMEiO57XLp++u+LHt1Ge+Ofjd0nJO27eNtzHt0Hp2/2pncdrmh44iIiMTDukmQ1QDMQieROGvxr9F0e/X+5BKpOLFtjN29CBgGTASWAM+6+/tmdq+ZDUy97QGgHvCcmc03s/FHWJ0kRPdh3WnZuyUTvzuRrau2ho5zUqb+ZCoYXHr3paGjiIiIxEPRHti3CZpfHzqJxF3tXKjVALZ/EDqJpInYNsYA7j7B3du6e2t3vy81b4S7j0897uvujdw9L/U18OhrlOrOMoz8x/PB4OWhL+Ml6XlKdcGcAhb+ZSE97upBTouc0HFERETiYceyaNqwZ9gckh6y28Pq50OnkDQR68ZY5ETUP6s+/X7Xj1XTVzHnD+n3+XXuzsTvTqRe43r0HK7CLyIictCK0dG0vq4vlnKo1xr2bYbdh92mSOQwaoylWsobmkebK9swZfgUNi3dFDrOcVn0zCIKZhfQ55d9yDo1K3QcERGR+Ni6MJpmtw+bQ9JD66HRdNvisDkkLagxlmrJzLh65NXUrFuTcTePo6SoJHSkcincU8jkH06mcZfG5N2SFzqOiIhIfLhHN1JqeROY/oSVcjhwZ+plj4bNIWlBv1Wk2jr1zFMZ8McBrHlrDbN+Myt0nHJ588E32b56O1c8dAWWobttioiIHLTzI9jzCeReFDqJpIvaZ0SnU+/6Z+gkkgbUGEu11nFwR84bfB7T/2s66xasCx3nqHas3cGsX8+iw/UdaHlJy9BxRERE4mX91Gja+LKwOSS9NMuH7YuhpDh0Eok5NcZS7Q14ZAB1T6/LuJvHUbSvKHScI5p530yK9xdz2a9U8EVERA6zeW708Tuntg2dRNJJTkco3gubZ4dOIjGnxliqvbqn1+XqkVezfuF6ZtwzI3ScMn268lPefuxtunytCw3OaRA6joiISLx4CXz0JzitM5guNZLj0DB16v2aV8PmkNhTYyyJ0PaqtuTdmses+2dRMLsgdJzDzLhnBhk1Muj1s16ho4iIiMTP1kXRtH6nsDkk/WS3g5wOsH5a6CQSc2qMJTH6PdSP7ObZjLtlHIW7C0PHOWjj4o0sfHIh3YZ1I7tpdug4IiIi8bPyiWja5lthc0h6yu4QnUq9d2PoJBJjaowlMbKys8h/PJ/NH25m8vDJoeMcNOkHk6hVrxY9f9gzdBQREZF42jwnmtY7J2wOSU9n3xJNt8wLm0NiTY2xJEqr3q3o/u3uvPWHt1g5dWXoOCx7bRnLXl1GrxG9qJtbN3QcERGR+CneC5++C21uh4waodNIOmr4pWi6YWbYHBJraowlcfr+qi+ntz2dl4e+zN5te4PlKC4sZuKdE2nQpgEX3HFBsBwiIiKxtm4yFO2CpleHTiLpqlYO1Dsblv9v6CQSY2qMJXFq1q3JNX++hu0F25n4vYnBcsx9ZC6bl27mit9eQY1a2gMuIiJSptUvQs0caNQndBJJZy0Gwf5PYd2U0EkkptQYSyI1u7AZFw+/mPmj57P0laVVvv3ta7Yz/e7ptL68NW2ubFPl2xcREUkLhTujxrjp1VCjVug0ks7O/UE0XfzrsDkkttQYS2JdMuISGnVqxCu3vcKuDbuqbLvuzoTbJ1C8v5gBjwzA9HmMIiIiZdswAwq3fXbzJJETldUAmn85OjW/cHvoNBJDaowlsTKzMrn2yWvZt20fL9z4AiXFJVWy3cXPLWbp+KVceu+lNDinQZVsU0REJC0t/hXUqA25F4dOItXBWUOi6T//GjaHxJIaY0m0Rp0aMeCRAaycspLpd0+v9O3t2riL1+54jTO/eCY97uxR6dsTERFJW/u2wMZZ0LAnZNYJnUaqg+bXQZ0zoWB86CQSQ2qMJfG63NqFvFvzmHnfTD589cNK246XOONuGcfebXvJH51PRqZ+/ERERI5o5RPRNE/XhEoFMYMzesMnf4NdH4dOIzGjv8xFgAEPD6BxXmNe+spLbFm+pVK2Mft3s1n+2nIu/+/LadSpUaVsQ0REpFoo3gvvfh8aXgynnR86jVQn7e6Ipu/dEzaHxI4aYxGgZp2aDHp+EJZhjO0/lt2bdlfo+ldOXcnkH06m/TXt6XZ7twpdt4iISLXzwUPgRdBmWHSUT6Si5F4IOR2hYByUFIZOIzGixlgkpUHrBgx5ZQjbC7bz9MCnKdxTMb8sN3+4mWeve5bT255O/ph83YVaRETkaHauggU/hgZd4azBodNIddT5Pti/BZY8EDqJxIgaY5FSmvdozrV/uZaC2QU8N+g5ivYWndT6dqzdwVNXPkVGzQyG/G0ItXNqV1BSERGRaqh4P8xK3Tm40y90tFgqR5MBULcZLPoF7N0UOo3EhBpjkUN0uK4DVz56JcteXcYz+c9QuPvEjhzvXL+TJ/o8wc51OxkyfgintTqtgpOKiIhUIyVFMOfrsHk2XPQUNLkidCKprjIyoftIKN4DC34UOo3EhBpjkTJ0/WZXBo4ayEeTPmLMJWPYtnrbcS2/cfFGRvUYxbaPt3HjhBtpdmGzSkoqIhI/ZtbPzJaa2XIzGx46j6SB4v3wj5th1ZPQ/i5oOSR0IqnumvSDM/vBR3+CT14LnUZiQI2xyBF0ubULg18azKYPNjGy60iWv778mMu4OwueWMCoHqMo3F3IzVNv5qwvnVUFaUVE4sHMagCPAP2BDsAQM+sQNpXEljtseAOm9IZ/Pg1th8H5D4ZOJUnR86/R5xq/kQ+bZodOI4Flhg4gEmft89vz9be+zrNffpax/cdy7pfPpeePetKka5PPvc9LnFXTV/HGL95g1bRVNL+4Odc9dR05LXICJRcRCaY7sNzdVwCY2TNAPrD4iEsU74Gtiw6Zeci1pWVea3rovDLec9hy5VhPeZapkPWWtVzMv+8TyZeRCZb6k7NoJ+zfBjuXw6Y5sPZ12DwHshrCRWOh5Y1lrE+kktTMhj5TYOpl8PeLoNPPof33ILPO4e8tKYw++3jfJijcHn2kGCWQkQU1aqemWWA1wDKAjNT0RPkJLnaCy53MNqsJNcYix9Dw3IZ8c/43efOBN5l1/yyWvLiE084+jSZdm1Aruxa71u9izZw17Nqwi1ManUL/h/vT7VvdsAzdMEREEqkpsLrU8wLggqMusW0xTPhCZWaSWDKo/wXo+jCcPRQy64YOJEmUcy5cPhve/DdY+FNY8iDk9oC6TaOddvu2RDtydq4ALw6dViqRGmORcsjMyqTXT3txwbcvYOFfFrJi0grWvruW/Tv2c8oZp3D2v5zNOf3PocN1HcisrR8rEZFjMbNvAN8AaHNWLvR8tNSrhxy1KPMIyAm8p6yjIYctVxHLlDHvRPJV2vdUnm1V4Hq9OPpMYi+JjtDVzI7uCNzgi1Dz1DLWIVLFTmkBfWdEZzB8/BxseRs+fQcyT4Ga9aF+Z2gxCOqdA7XPiP4P16gTHREu3gcle1PTfdH/cy8BSqCk+CTvrH6Cy4bYZuwMOu4l9Be8yHHIys6i2+3d6HZ7t9BRRETiag3QvNTzZql5n+PujwGPAXTt2tVpcX3VpBMRKYsZNOkffUki6eZbIiIiUpHmAm3MrJWZ1QJuAMYHziQiInJUOmIsIiIiFcbdi8xsGDARqAGMdvf3A8cSERE5KjXGIiIiUqHcfQIwIXQOERGR8tKp1CIiIiIiIpJoaoxFREREREQk0dQYi4iIiIiISKKpMRYREREREZFEU2MsIiIiIiIiiabGWERERERERBJNjbGIiIiIiIgkmrl76AxVysx2AEtD55DPyQU2hQ4hh9G4xI/GJJ7aufupoUOkM9XmWNLvm3jSuMSPxiSejrs2Z1ZWkhhb6u5dQ4eQz5jZPI1J/Ghc4kdjEk9mNi90hmpAtXFGr0gAAAarSURBVDlm9PsmnjQu8aMxiacTqc06lVpEREREREQSTY2xiIiIiIiIJFoSG+PHQgeQw2hM4knjEj8ak3jSuJw8/RvGj8YknjQu8aMxiafjHpfE3XxLREREREREpLQkHjEWEREREREROSgxjbGZ9TOzpWa23MyGh84jETOrb2bPm9kHZrbEzHqEzpQ0ZjbazDaY2aJS8x5IjclCM3vJzOqHzJhERxiXPDObbWbzzWyemXUPmTFpzKy5mU0zs8Vm9r6ZfeeQ1+8yMzez3FAZ041qczypNoen2hxPqs3xU5G1ORGNsZnVAB4B+gMdgCFm1iFsKkn5PfC6u7cHOgNLAudJojFAv0PmTQI6unsn4EPgR1UdSsocl98A97h7HjAi9VyqThFwl7t3AC4E/uNALTGz5sDlwMcB86UV1eZYU20ObwyqzXE0BtXmuKmw2pyIxhjoDix39xXuvh94BsgPnCnxzCwH6AWMAnD3/e6+NWyq5HH3N4Ath8z7u7sXpZ7OBppVebCEK2tcAAeyU49zgE+qNFTCuftad38n9XgHUbPQNPXyQ8APiMZIyke1OYZUm+NBtTmeVJvjpyJrc2alJIyfpsDqUs8LgAsCZZHPtAI2Ao+bWWfgbeA77r4rbCw5xK3AX0OHEAC+C0w0sweJdmxeFDhPYplZS6ALMMfM8oE17r7AzILmSjOqzfGk2pweVJvjQ7U5Jk62NifliLHEUyZwPvCou3cBdgG6xixGzOwnRKeojA2dRQD4FnCnuzcH7iR1REeqlpnVA14g+mOoCPgx0elzItWBanPMqTbHjmpzDFREbU5KY7wGaF7qebPUPAmrAChw9zmp588TFWOJATP7KnAVcJPrc93i4hbgxdTj54hORZUqZGY1iQrvWHd/EWhNdIRtgZmtIqov75hZ43Ap04ZqczypNseYanMsqTYHVlG1OSmN8VygjZm1MrNawA3A+MCZEs/d1wGrzaxdatZlwOKAkSTFzPoRXZMx0N13h84jB30CXJJ63AdYFjBL4lh0LtYoYIm7/xbA3d9z9zPcvaW7tyRqKs5P/X6To1NtjiHV5vhSbY4t1eaAKrI2J+IaY3cvMrNhwESgBjDa3d8PHEsidwBjU38UrQCGBs6TOGb2NHApkGtmBcDdRHe6zAImpa7LmO3u/x4sZAIdYVxuA35vZpnAXuAb4RIm0sXAV4D3zGx+at6P3X1CwExpS7U51lSbA1NtjifV5liqsNpsOgtDREREREREkiwpp1KLiIiIiIiIlEmNsYiIiIiIiCSaGmMRERERERFJNDXGIiIiIiIikmhqjEVERERERCTR1BiLpAkz23nI86+a2cNVnGGQmS0xs2nHeN8YM7u+qnKJiIiEoNosUn2oMRZJuNTn7pXX14Db3L13ZeURERFJOtVmkaqnxlikGjCzlmY21cwWmtkUM2uRmv+5vcMH9myb2aVmNtPMxgOLy1jfEDN7z8wWmdn9qXkjgJ7AKDN74JD3m5k9bGZLzWwycEap10aY2dzUuh5Lvbe1mb1T6j1tSj8XERFJd6rNIulFjbFI+qhjZvMPfAH3lnrtD8Cf3b0TMBb4n3Ks73zgO+7etvRMM2sC3A/0AfKAbmZ2jbvfC8wDbnL37x+yrmuBdkAH4GbgolKvPezu3dy9I1AHuMrdPwK2mVle6j1DgcfLkVlERCROVJtFqgk1xiLpY4+75x34AkaUeq0H8FTq8ZNEe4+P5S13X1nG/G7AdHff6O5FRMW81zHW1Qt42t2L3f0TYGqp13qb2Rwze4+ooJ+Xmv8nYKiZ1QAGl8ovIiKSLlSbRaoJNcYi1VsRqZ9zM8sAapV6bVdlb9zMagN/BK539y8AI4HaqZdfAPoDVwFvu/vmys4jIiISA6rNIjGkxlikengTuCH1+CZgZurxKuCLqccDgZrlWNdbwCVmlpvaYzwEmHGMZd4ABptZDTM7EzhwA5ADhXaTmdUDDl5T5e57gYnAo+hULRERqX5Um0XSyPHc8U5E4usO4HEz+z6wkei6IIj2Ar9sZguA1ynHnmh3X2tmw4FpgAGvuvvLx1jsJaJTsRYDHwP/SK1rq5mNBBYB64C5hyw3lugaqL8f8zsUERFJL6rNImnE3D10BhFJKDP7TyDH3X8WOouIiIioNkty6YixiARhZi8BrYn2ZouIiEhgqs2SZDpiLCIiIiIiIommm2+JiIiIiIhIoqkxFhERERERkURTYywiIiIiIiKJpsZYREREREREEk2NsYiIiIiIiCSaGmMRERERERFJtP8HOC74B8/dnaEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1152x1152 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plane_no = 0\n",
    "result, detail = compute_flight(**planes.iloc[plane_no].loc[opt_vars])\n",
    "detail_df = pd.DataFrame.from_dict(dict(detail)).set_index('time').sort_index()\n",
    "\n",
    "# Summary stats\n",
    "skip_vars = ['a1', 'a2', 'a3', 'a4', 'a5', 'a6', 'e1', 'e2', 'e3', 'e4', 'e5', 'e6']\n",
    "\n",
    "display(pd.concat([\n",
    "    pd.Series(dict(planes.iloc[plane_no].loc[opt_vars])),\n",
    "    pd.Series(dict(result))\n",
    "]).rename('Plane {}'.format(plane_no)).drop(skip_vars))\n",
    "\n",
    "# Plots of day\n",
    "f, axarr = plt.subplots(3, 2, figsize=(16,16))\n",
    "y_labels = {\n",
    "        'altitude': 'Altitude (m)',\n",
    "        'engine_power': 'Engine power output (W)',\n",
    "        'airspeed': 'True airspeed (m/s)',\n",
    "        'battery_energy': 'Battery energy remaining (J)',\n",
    "        'battery_SOC': 'Battery state-of-charge',\n",
    "        'solar_power': 'Solar power input (W)',\n",
    "}\n",
    "colors = ['blue', 'black', 'green', 'red', 'purple', 'orange']\n",
    "for i, col in enumerate(detail_df):\n",
    "    ax = axarr.flatten()[i]\n",
    "    detail_df[col].plot(ax=ax, color=colors[i])\n",
    "    ax.set_ylabel(y_labels[col])\n",
    "    ax.set_xlabel('Hour of day')\n",
    "    ax.set_xlim(0, 24)\n",
    "    ax.get_yaxis().get_major_formatter().set_useOffset(False)\n",
    "    ax.set_xticks([0, 6, 12, 18, 24])\n",
    "    if col == 'altitude':\n",
    "        ax.set_ylim(0, 20000)\n",
    "    elif col == 'engine_power':\n",
    "        ax.set_ylim(0, 350)\n",
    "    elif col == 'airspeed':\n",
    "        ax.set_ylim(0, 40)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}