time-series-forecast.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "Untitled0.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zIkxCA9e86xX"
      },
      "source": [
        "# Title : Implementation of regression techniques on time-series data to generate future predictions.\r\n",
        "\r\n",
        "- A Cloud platform continuously records the resource utilization of instances per minute and stores this information in the resource utilization log. For example, Instance A, has its CPU, Memory, Disk and Network Bandwidth utilization with a timestamp stored in a log file. This log file then contains the historical data of resource utilization from the time that Instance A was created and till the time it was destroyed.\r\n",
        "\r\n",
        "### Problem statement:\r\n",
        "- How can we use Machine Learning constructs, to implement an intelligent system, that allow the user to see a predicted forecast of resources. Input to the system is resource log files. The system is expected to perform required data cleaning and transformation. Then the system is expected to perform required analysis on the processed data and generate results in terms of future predictions. The output of the system should be a predicted resources log.\r\n",
        "\r\n",
        "### Functional Requirements:\r\n",
        "1. System should convert given data into required data model. This stage generates 'input-data.\r\n",
        "2. System should implement relevant and required data cleaning and data\r\n",
        "transformations techniques in the input data. This stage generates\r\n",
        "'pre-processed-data'.\r\n",
        "3. System should (if required) perform training on pre-processed-data and\r\n",
        "generate machine learning model. This stage will generate a 'model'.\r\n",
        "4. System should process pre-processed-data and generate prediction\r\n",
        "results. This stage generates 'predicted-transactions'.\r\n",
        "\r\n",
        "### Expected Results:\r\n",
        "Given the data-set below, following are the expected results from the system,\r\n",
        "1. Predictions per Instance (required)\r\n",
        "2. Predictions per Group (required)\r\n",
        "3. Predictions per Resource (CPU/Memory/Network/Storage) per group or per\r\n",
        "instance (optional)\r\n",
        "\r\n",
        "### Expected Deliverables:\r\n",
        "1. Link to the code base as a GitHub/GitLab repo. (required)\r\n",
        "2. Assignment documentation which should include the following, (required)\r\n",
        "    \r\n",
        "    a. Problem Statement\r\n",
        "\r\n",
        "    b. Brief description (your understanding about the problem and its\r\n",
        "    brief explanation)\r\n",
        "\r\n",
        "    c. Proposed solution (this section must include your solution to\r\n",
        "    the problem, relevant algorithm, techniques, frameworks used)\r\n",
        "\r\n",
        "    d. Block diagram (a diagram explaining, data flow, processing and\r\n",
        "    overall architecture of your solution)\r\n",
        "\r\n",
        "    e. Sample Test Results\r\n",
        "\r\n",
        "    f. Future scope (possible improvements, performance tweaks, ability\r\n",
        "    to process other time-series data etc)\r\n",
        "    \r\n",
        "3. Document containing references, bibliography items. (optional)\r\n",
        "\r\n",
        "\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RAnHK6ddyrfA",
        "outputId": "ebbe2f4a-be68-4713-be27-45033b8a712c"
      },
      "source": [
        "!wget --header=\"Host: nsk-ecos.enlightcloud.com:8080\" --header=\"User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/87.0.4280.101 Safari/537.36\" --header=\"Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.9\" --header=\"Accept-Language: en-US,en;q=0.9\" \"http://nsk-ecos.enlightcloud.com:8080/v1/AUTH_e4fa818cb8d34f06bb6274ac8d5b30f9/assignment-files/group82_resource_utilization.zip\" -c -O 'group82_resource_utilization.zip'"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2020-12-24 09:04:34--  http://nsk-ecos.enlightcloud.com:8080/v1/AUTH_e4fa818cb8d34f06bb6274ac8d5b30f9/assignment-files/group82_resource_utilization.zip\n",
            "Resolving nsk-ecos.enlightcloud.com (nsk-ecos.enlightcloud.com)... 115.124.98.240\n",
            "Connecting to nsk-ecos.enlightcloud.com (nsk-ecos.enlightcloud.com)|115.124.98.240|:8080... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 215449235 (205M) [application/zip]\n",
            "Saving to: ‘group82_resource_utilization.zip’\n",
            "\n",
            "group82_resource_ut 100%[===================>] 205.47M  3.46MB/s    in 56s     \n",
            "\n",
            "2020-12-24 09:05:31 (3.70 MB/s) - ‘group82_resource_utilization.zip’ saved [215449235/215449235]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OOKSITMHytdv"
      },
      "source": [
        "!mkdir /content/group82_resource_utilization\r\n",
        "!unzip /content/group82_resource_utilization.zip -d /content/group82_resource_utilization"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1B57ynbEy47n"
      },
      "source": [
        "# import essential libraries\r\n",
        "import numpy as np\r\n",
        "import pandas as pd\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "import seaborn as sns\r\n",
        "from tqdm import tqdm\r\n",
        "import re\r\n",
        "import os\r\n",
        "import warnings\r\n",
        "warnings.filterwarnings('ignore')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BdviJG_pkPGf"
      },
      "source": [
        "# 1. Exploratory data analysis (EDA) - Understanding the data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k4qSIL1Zy-mh",
        "outputId": "5b63d0ff-c764-4594-d241-14b6ab1c6264"
      },
      "source": [
        "# 1. Total no. of unique folders i.e., instances\r\n",
        "dir_lst = os.listdir('/content/group82_resource_utilization')\r\n",
        "print('Total no. of unique folders i.e., instances : ', len(dir_lst))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Total no. of unique folders i.e., instances :  3570\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mXg0IDLxzBeX",
        "outputId": "16c900ce-5e9e-4dc1-964f-9557f0723184"
      },
      "source": [
        "# 2. Total no. of unique groups\r\n",
        "dir_lst = os.listdir('group82_resource_utilization')\r\n",
        "grp_lst = []\r\n",
        "for i in dir_lst:\r\n",
        "    x = i.split('_')\r\n",
        "    y = '_'.join([x[0], x[1]])\r\n",
        "    grp_lst.append([y, x[2]])\r\n",
        "# pd.DataFrame(grp_lst, columns = ['group', 'instance'])\r\n",
        "grp_lst = np.array(grp_lst)\r\n",
        "print('Total no. of unique GROUPS : ', len(np.unique(grp_lst[:, 0])))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Total no. of unique GROUPS :  82\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PWYzpnAgNS8l",
        "outputId": "a0e8dbfc-fec2-4c20-d947-e9b76650a8ba"
      },
      "source": [
        "# 3. How many datapoints each instance has?\r\n",
        "cnt_per_instance_lst = []\r\n",
        "for i in tqdm(dir_lst):\r\n",
        "    folder = 'group82_resource_utilization/'+i\r\n",
        "    if os.listdir(folder):\r\n",
        "        f = open(folder + '/' + os.listdir(folder)[0], 'r')\r\n",
        "        cnt = 0\r\n",
        "        for dp in f:\r\n",
        "            if len(dp)>0 and not re.findall('\\x00', dp):\r\n",
        "                cnt+=1\r\n",
        "        cnt_per_instance_lst.append(cnt)\r\n",
        "        f.close()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 3570/3570 [00:34<00:00, 104.77it/s]\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Gyk6wZSI1T0h",
        "outputId": "2376b756-a2a5-4989-cdde-db45ee903dab"
      },
      "source": [
        "len(cnt_per_instance_lst)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3570"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 361
        },
        "id": "sdQGTRkw1YUA",
        "outputId": "11ea231a-a89d-43d8-d1ff-03d935884c76"
      },
      "source": [
        "sns.set_style('darkgrid')\r\n",
        "plt.figure(figsize = (16, 5.5))\r\n",
        "plt.plot(cnt_per_instance_lst)\r\n",
        "plt.xlabel('instance')\r\n",
        "plt.ylabel('count')\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 450
        },
        "id": "EdU3S4lo1ah_",
        "outputId": "0857978d-34ce-44fa-f6b9-67a489fe5fb1"
      },
      "source": [
        "# PDF\r\n",
        "plt.figure(figsize = (16, 5.5))\r\n",
        "plt.subplot(1,2,1)\r\n",
        "plt.title('Probability density function of count')\r\n",
        "sns.distplot(cnt_per_instance_lst, hist=False)\r\n",
        "\r\n",
        "# CDF\r\n",
        "plt.subplot(1,2,2)\r\n",
        "plt.title('Cumulative distruibution function of count')\r\n",
        "sns.distplot(cnt_per_instance_lst, hist=False, kde_kws= {'cumulative': True})\r\n",
        "\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `kdeplot` (an axes-level function for kernel density plots).\n",
            "  warnings.warn(msg, FutureWarning)\n",
            "/usr/local/lib/python3.6/dist-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `kdeplot` (an axes-level function for kernel density plots).\n",
            "  warnings.warn(msg, FutureWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x396 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lF77tals1sD5",
        "outputId": "b657d742-fcbf-478d-a258-51810b18581e"
      },
      "source": [
        "# understanding extreme counts using percentile method\r\n",
        "for i in np.arange(90, 100.5, 0.5):\r\n",
        "    print('percentile:', i, '| count:', int(np.percentile(cnt_per_instance_lst, i)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "percentile: 90.0 | count: 46728\n",
            "percentile: 90.5 | count: 47065\n",
            "percentile: 91.0 | count: 48085\n",
            "percentile: 91.5 | count: 48836\n",
            "percentile: 92.0 | count: 49081\n",
            "percentile: 92.5 | count: 49262\n",
            "percentile: 93.0 | count: 49497\n",
            "percentile: 93.5 | count: 49795\n",
            "percentile: 94.0 | count: 49942\n",
            "percentile: 94.5 | count: 50179\n",
            "percentile: 95.0 | count: 50353\n",
            "percentile: 95.5 | count: 50543\n",
            "percentile: 96.0 | count: 50936\n",
            "percentile: 96.5 | count: 51493\n",
            "percentile: 97.0 | count: 52232\n",
            "percentile: 97.5 | count: 54364\n",
            "percentile: 98.0 | count: 91545\n",
            "percentile: 98.5 | count: 101488\n",
            "percentile: 99.0 | count: 102456\n",
            "percentile: 99.5 | count: 104337\n",
            "percentile: 100.0 | count: 894794\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kffb3aaSk0RN",
        "outputId": "28ce96f6-649e-4e23-ac9a-fce596883f93"
      },
      "source": [
        "# understanding extreme counts using percentile method\r\n",
        "for i in np.arange(0, 42, 2):\r\n",
        "    print('percentile:', i, '| count:', int(np.percentile(cnt_per_instance_lst, i)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "percentile: 0 | count: 1\n",
            "percentile: 2 | count: 76\n",
            "percentile: 4 | count: 434\n",
            "percentile: 6 | count: 724\n",
            "percentile: 8 | count: 1169\n",
            "percentile: 10 | count: 1633\n",
            "percentile: 12 | count: 1913\n",
            "percentile: 14 | count: 1942\n",
            "percentile: 16 | count: 1955\n",
            "percentile: 18 | count: 1975\n",
            "percentile: 20 | count: 1998\n",
            "percentile: 22 | count: 2045\n",
            "percentile: 24 | count: 2046\n",
            "percentile: 26 | count: 2051\n",
            "percentile: 28 | count: 2052\n",
            "percentile: 30 | count: 2054\n",
            "percentile: 32 | count: 2055\n",
            "percentile: 34 | count: 2056\n",
            "percentile: 36 | count: 2058\n",
            "percentile: 38 | count: 2100\n",
            "percentile: 40 | count: 2149\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pher5alek3gF"
      },
      "source": [
        "# Observations:\r\n",
        "    1. Almost 97.5% instances has less than 55000 datapoints in it\r\n",
        "    2. 40% instances has less than 2200 data points in it\r\n",
        "    \r\n",
        "# Conclusion:\r\n",
        "    1. Some instances do not have enough datapoints, hence few instance based models may perform poorly."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 361
        },
        "id": "MZQ8L419k39n",
        "outputId": "6e3e8bc9-d9e2-45a5-e1f0-df646557b402"
      },
      "source": [
        "# 4. How many instances each group has?\r\n",
        "val, cnt = np.unique(grp_lst[:, 0], return_counts=True)\r\n",
        "sns.set_style('darkgrid')\r\n",
        "plt.figure(figsize = (16, 5.5))\r\n",
        "plt.plot(cnt)\r\n",
        "plt.xlabel('group_no.')\r\n",
        "plt.ylabel('count')\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hPf6zNIkk4At",
        "outputId": "fcef0ad1-3c50-49ab-b4b4-266852b4eefa"
      },
      "source": [
        "print(sorted(cnt))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 4, 4, 4, 4, 4, 4, 6, 6, 6, 6, 6, 8, 8, 10, 12, 12, 12, 12, 14, 14, 14, 16, 18, 18, 18, 20, 22, 24, 24, 24, 28, 28, 34, 34, 36, 40, 42, 42, 42, 48, 50, 54, 56, 58, 58, 66, 70, 72, 86, 88, 88, 88, 90, 100, 102, 104, 132, 154, 156, 162, 192, 192, 196, 196, 198]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5FmsaM2nk4FB"
      },
      "source": [
        "# Observations:\r\n",
        "    1. each group has atleast 2 instances in it.\r\n",
        "    2. few groups has more than 190 instance in it\r\n",
        "\r\n",
        "# Conclusion\r\n",
        "    1. Primary conclusion is ML models may work better on groups than single instance based models \r\n",
        "    2. But we need to take care of data variability i.e., group has data from multiple instances, each instance may have different range of values in it. This can cause a problem to ML model."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Oj_EMFQYk4MW",
        "outputId": "9ab3b6db-0d54-4cd2-94fe-94c229145723"
      },
      "source": [
        "# 5. How many datapoints each group has?\r\n",
        "\r\n",
        "# creating path list for each group\r\n",
        "path_lst = []\r\n",
        "for grp in val:\r\n",
        "    grp_arr = grp_lst[np.where(grp_lst[:,0] == grp)]\r\n",
        "    path_arr = grp_arr[:, 0].astype(object) + '_' + grp_arr[:, 1].astype(object)\r\n",
        "    path_lst.append(path_arr)\r\n",
        "\r\n",
        "# counting datapoints for each group\r\n",
        "cnt_per_grp_lst = []\r\n",
        "for h in tqdm(path_lst):\r\n",
        "    cnt = 0\r\n",
        "    for i in h:\r\n",
        "        folder = 'group82_resource_utilization/'+i\r\n",
        "        if os.listdir(folder):\r\n",
        "            f = open(folder + '/' + os.listdir(folder)[0], 'r')\r\n",
        "            for i in f:\r\n",
        "                if len(dp)>0 and not re.findall('\\x00', dp):\r\n",
        "                    cnt+=1\r\n",
        "            f.close()\r\n",
        "    cnt_per_grp_lst.append(cnt)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 82/82 [00:33<00:00,  2.44it/s]\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 462
        },
        "id": "Hf9Sf73-k4TM",
        "outputId": "889d38ac-3c2c-42fd-be56-09d56e6f9dde"
      },
      "source": [
        "plt.figure(figsize = (16, 5.5))\r\n",
        "plt.subplot(1,2,1)\r\n",
        "plt.title('Probability density function of count')\r\n",
        "sns.distplot(cnt_per_grp_lst, hist=False)\r\n",
        "\r\n",
        "plt.subplot(1,2,2)\r\n",
        "plt.title('Cumulative distruibution function of count')\r\n",
        "sns.distplot(cnt_per_grp_lst, hist=False, kde_kws= {'cumulative': True})\r\n",
        "\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `kdeplot` (an axes-level function for kernel density plots).\n",
            "  warnings.warn(msg, FutureWarning)\n",
            "/usr/local/lib/python3.6/dist-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `kdeplot` (an axes-level function for kernel density plots).\n",
            "  warnings.warn(msg, FutureWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA6YAAAFnCAYAAAC1qVTTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd1hT9/4H8HdICBvCHoKIiuLEKIgLB+49u2tvHW291eq92lptf3qvdtqhvb33Vm9ra2vnta0tDure1i1qFbcs2SusBAhJfn8guYCAqISTk7xfz+NjIDnnvIEk3/PJ+Q6JwWAwgIiIiIiIiEggNkIHICIiIiIiIuvGwpSIiIiIiIgExcKUiIiIiIiIBMXClIiIiIiIiATFwpSIiIiIiIgExcKUiIiIiIiIBMXClJpdx44dkZyc/EDbxsTE4Pfff6/3vtOnT2PkyJH1PnbdunV4/fXXH+iY93L79m107NgRlZWVJtk/AGzZsgUzZ85stv1999136NevH5RKJQoKCpptv/diyr9DY3bv3o1BgwZBqVQiISGhxY9PRGSO/vnPf+Lll19+4O3Hjh2LEydONGOi+rVUe15X3bb3Yc5f6qNUKpGamtps+6t269YtTJw4EUqlEhs3bmz2/TckPT0dSqUSOp2uxY4JALm5uXjqqaegVCrx7rvvtuixqWXJhA5A5iEmJga5ubmQSqVwcHDAwIEDsWzZMjg5OQkdzSgiIgI7d+6s9745c+YYb9++fRtDhw7FpUuXIJOJ4yk+YcIETJgwwfh1x44dsWvXLgQHB9/3vrRaLd59911s2rQJYWFhzRmzlhMnTuCVV17BoUOHjN+r+XdoSatWrcKyZcswbNgwQY5fn4f5GxKRuG3duhUbNmxAYmIinJycEBYWhjlz5iAiIkLoaA1asmQJfH198de//tX4ve3bt7d4jqa2I9OnT8eECRPwyCOPPPCx6ra9D6O+PPHx8c2y77rWr1+PqKgoxMbGmmT/1WJiYvDmm2+iX79+AICAgACT/UyN+e9//wt3d3ecPXsWEomkxY9fn/peL/TweMWUjNatW4f4+Hj88ssvuHjxItauXXvXY0x51ZCaR15eHsrLy9G+fXuho7SY9PR0hIaGCh2DiAgbNmzA22+/jTlz5uDo0aPYv38/nnzySezdu1foaFbFks9XrK3NS09PR7t27cymKCXTYWFKd/H19UV0dDSuX78OoOrKz7fffosRI0ZgxIgRAIBNmzZh+PDh6N27N+bMmYOsrKxa+zh48CCGDh2KqKgorFq1Cnq9HgCQkpKCZ555BlFRUYiKisKiRYtQVFRUa9s//vgDY8aMQWRkJJYuXYry8nIAVVfoBg4cWG/mmt2Vnn76aQBAZGQklEolTp48id69e+Pq1avGx+fl5SE8PBz5+fl37Uun02HVqlWIiorC0KFDcfDgwVr3FxcX47XXXsOAAQMQHR2NNWvWGLu1bN68GU888QRWrVqFyMhIxMTE1Np+8+bNGDp0KJRKJWJiYrBly5Za2wHAU089BQDGbjpxcXEYN24c9u3bZ9yPVqtFVFTUXd1WExMTMWrUKOPP/8wzz9TbFXn69On48ccfm5RZpVJh6dKlGDBgACIjI/Hiiy9CrVbjueeeQ3Z2NpRKJZRKJbKysu7qNrZ3716MHTsWERERmD59Om7evGm8LyYmBp9//jnGjx+PXr164S9/+Yvxb12XXq/HJ598giFDhqBv375YvHgxiouLUVFRYexWNHHixAavmF6/fh0zZsxA79690a9fP6xbtw4AUFFRgbfeegsDBgzAgAED8NZbb6GiouKuv0m1mt28lixZghUrVuD555+HUqnEI488gpSUlAb/hkRk+YqLi/Hxxx9j+fLlGDFiBBwdHWFra4uYmBi8+uqrAKreO9asWWPcpm7bFhMTg/Xr12P8+PHo0aMHXnvtNeTm5mL27NlQKpV49tlnUVhYWO+21ds3NCRm/vz56N+/P3r16oWnnnrK2M7/97//xdatW/H5559DqVQar1pW7ysrKwvdu3eHSqUy7ishIQFRUVHQarUAgJ9++gmjR49GZGQkZs2ahbS0tAZ/T7/++iuGDBmCqKiouz4Er9mOlJeX4+WXX0ZUVBQiIiIwdepU5ObmYs2aNTh9+jRWrlwJpVKJlStXArj7fKWp7V9NDZ2/1G3fau67sTzVbUZxcTEWL16MPn36YMiQIfjkk0+M+75XO1zTM888gxMnThiPlZiYWOtnqu/n6tixI77//nuMGDECERERWLFiBQwGg/H+TZs2YfTo0VAqlRgzZgwuXbqEV155Benp6ZgzZw6USiU+++yzu36fWVlZmDNnDnr37o3hw4dj06ZNtf6OCxYswOLFi6FUKjF27Fj88ccfDTwjgLNnz2Lq1Kno1asXpk6dirNnzwKoer38+uuvxudmfc/tsrIyvPvuuxgyZAh69eqFJ554AmVlZQAaPw+p23W75muz+rX1xRdfoG/fvhgwYAB+/vlnAA2/XujhsTClu2RkZODQoUPo1KmT8Xt79uzBpk2bEBcXh2PHjuHDDz/ERx99hCNHjqBVq1ZYuHBhrX3s3r0bP//8M3755Rfs27fP+GI2GAx44YUXcPjwYfz222/IzMzEP//5z1rbVr/Yd+/ejcTERHzyySf3lf+bb74BAJw6dQrx8fHo3bs3xowZYywCAWDbtm3o27cvPDw87tp+06ZN2L9/P3799Vf8/PPP2LFjR637lyxZAplMhl27duHXX3/F0aNHazUIFy5cQEhICI4fP47Zs2fj9ddfh8FggFqtxptvvonPPvsM8fHx+OGHH2r9jqt9++23AIDY2FjEx8djzJgxmDhxYq38Bw8ehI+PDzp37lxr25CQEGzbts348zd17ElDmQFg8eLF0Gg02L59O37//Xc8++yzcHR0xGeffQYfHx/Ex8cjPj4evr6+tfaZmJiIRYsW4bXXXsOxY8cwcOBAzJkzx1j4AcBvv/2G9evXY+/evbh69So2b95cb77Nmzfjl19+wcaNG7Fnzx6o1WqsXLkScrnc2K0oNjYWe/bsuWvbkpISzJgxA9HR0Th8+DB27dqFvn37AgDWrl2L8+fPIzY2Flu2bMEff/xxX8+3uLg4zJs3D6dOnULr1q2NDVp9f0Misnzx8fEoLy/H8OHDH2o/u3btwoYNG7Bz507s378fzz33HBYuXIjjx49Dr9fj66+/fqD9Dhw4EDt37sSxY8fQuXNnY6H12GOPYfz48Zg1axbi4+ONH95V8/X1RY8ePbBr1y7j97Zu3YqRI0fC1tYWe/bswX/+8x/861//wrFjx9CrVy8sWrSo3gw3btzAihUr8N577+Hw4cNQqVTIzMys97G//PILSkpKcODAAZw4cQIrVqyAvb09/vrXvyIiIgLLly9HfHw8li9fbtym5vnKg2jo/KUxjeWp9sYbb6C4uBh79uzB119/jdjY2Fr7bqwdrmnjxo21jhUSEtKkn+vAgQP46aefsGXLFvz22284fPgwgKp2+J///CdWrVqFs2fPYu3atVAoFHj//fcREBBg7E333HPP3bXPhQsXws/PD4cPH8bHH3+M1atX49ixY8b79+3bh7Fjx+L06dOIiYnBG2+8UW82lUqFF154AdOnT8eJEycwY8YMvPDCCygoKMC7775b67lZ3a24plWrVuHSpUv44YcfcPLkSbzyyiuwsbFp0nlIY3Jzc1FcXIxDhw7hrbfewsqVK1FYWHjP1ws9OLMrTJcuXYq+ffti3LhxD72v48ePY+LEicZ/3bp1q/fElarMnTsXERERePLJJxEZGVnrE6Dnn38eCoUC9vb22Lp1K6ZOnYouXbpALpdj4cKFOHfuHG7fvm18/HPPPQeFQoGAgAA888wzxmIpODgY/fv3h1wuh4eHB2bMmIFTp07VyvHUU0/B398fCoUCf/7zn5tljMvkyZOxfft245t8bGxsg+NKfvvtN/zpT38yZnjhhReM9+Xm5uLgwYN47bXX4OjoCE9PTzz77LO1MgYEBODRRx+FVCrF5MmTkZOTg9zcXACAjY0Nrl+/jrKyMvj4+DS5K86ECRNw8OBBlJSUAKiasKG5xsU0ljk7OxuHDh3CihUr4ObmBltbW/Tu3btJ+4yLi8OgQYPQv39/2NraYtasWSgrK6s1PmX69Onw9fWFQqHAkCFDcPny5Xr3tXXrVjz77LMICgqCk5MTFi5ciLi4uCZ11Tpw4AC8vLwwc+ZM2NnZwdnZGeHh4cb9zp07F56envDw8MDcuXNrfQBwL8OGDUP37t0hk8kwYcKEBvMTkXVQqVRwd3d/6PkNnn76aXh5ecHX1xcRERHo3r07OnfuDDs7OwwfPvyBJ3mbNm0anJ2dIZfL8dJLL+HKlSsoLi5u0rbjx483tuUGgwFxcXEYP348AOCHH37A888/j3bt2kEmk2HOnDm4fPlyvVdNd+zYgcGDByMyMhJyuRwLFiyAjU39p6MymQwqlQrJycmQSqXo2rUrnJ2dG81Z83zlQTR0/vIwdDod4uLisGjRIjg7OyMwMBAzZsyo1d40du7QHJ577jm4uroiICAAUVFRuHLlCoCqK92zZ89G9+7dIZFIEBwcjFatWt1zfxkZGTh79ixefvll2NnZoVOnTnjkkUdqjXvt1asXBg0aBKlUiokTJxqPWdeBAwcQHByMSZMmQSaTYdy4cWjbti32799/zxx6vR4///wzXn/9dfj6+kIqlaJnz56Qy+VNOg9pjEwmw9y5c2Fra4tBgwbB0dERiYmJTdqWHozZzQwzZcoUPP3008YuLw+jT58+xheISqXCiBEj0L9//4fer6X697//Xe8nUQDg7+9vvJ2dnY0uXboYv3ZycoJCoUBWVhYCAwPvenyrVq2QnZ0NoKqwe+utt3D69GmUlpbCYDDA1dW1wWMFBAQYt30Y4eHhsLe3x4kTJ+Dt7Y2UlBQMHTq03sdmZ2fflaFaeno6KisrMWDAAOP39Hp9rcd7eXkZbzs4OAAA1Go1vL29sWbNGnzxxRd4/fXX0bNnT7z66qto167dPfP7+vqiZ8+e2LlzJ4YPH45Dhw4166yFDWUuLCyEm5sb3Nzc7nuf2dnZtX53NjY28Pf3r9Xt29vbu9ZxG/pbZ2dn12ooW7VqhcrKSuTl5d11pbaujIwMtG7dukkZ7/f5VvP3Zm9vD7Va3eRticjyKBQKFBQUoLKy8qGK05rvLXZ2ds3yXqPT6bBmzRrs2LED+fn5xmKwoKAALi4u99x+xIgReOONN5CdnY2kpCTY2NgYJ3NKT0/H22+/jVWrVhkfbzAYkJWVdVeRk52dDT8/P+PXjo6OUCgU9R5z4sSJyMzMxMKFC1FUVIQJEybgr3/9K2xtbRvMWbM9fhANnb88jIKCAmi12rvam5rtYUPtcHOp296WlpYCaLyNbEx2djbc3NxqfVAQEBCAixcvGr+u+7wtLy+v97VRty2u3lfdYWL1KSgoQHl5OYKCgurNeK/zkMYoFIpaWR0cHNjOm5jZXTGNjIy86yQ4JSUFs2bNwpQpU/Dkk0/W6h/eVDt37kR0dLTxxU73p+aAcx8fn1qfgqrVaqhUqloFQkZGhvF2eno6fHx8AACrV6+GRCLB1q1bcfbsWbz//vt3dVVpaNsHyVrT5MmTsWXLFmzZsgUjR46EnZ1dvY/z9vaulaHmbT8/P8jlchw/fhynT5/G6dOncfbs2SZf1Y2OjsaGDRtw5MgRtG3bFsuWLWvyz1Wdf8eOHejRo8c9C7Jqjo6OAGAcbwEAOTk5TdrWz88PhYWFd40DBhr+PVfz8fFBenq68WuDwYCMjIwm5667r5rPufT0dMhkMnh6et5zW39//wan66+bMSMjw/h8c3BweKDfGRFZL6VSCblc3mjvrLrvLQ9zVazuvnQ6Xb1zJwBVPUT27t2LDRs24MyZM8Z5C6rb4Hu9p7u5uaF///6Ii4vDtm3bMGbMGOM2/v7+WLFihbFdPH36NC5cuICePXvetR8fH59aXXc1Gk2tsas12draYt68eYiLi8MPP/yAAwcO4Ndff200Z82f40Hav4bOQR7m7+bu7g5bW9u72psHaQ/r4+DgAI1G80DZ/P39jfMj3A8fHx8UFhYae3IBD/4z1W2L72df7u7usLOzq7edv9d5SN3f2/2085yIyTTMrjCtz7Jly7Bs2TJs3rwZr776KlasWHHf+9i+fXuzdA8mYNy4cdi8eTMuX76MiooKrF69Gt27dzdeLQWAzz//HIWFhcjIyMDGjRuNY+xKS0vh6OgIFxcXZGVlYf369Xft/7vvvkNmZiZUKhXWrVt33+PzPDw8YGNjc9eb1IQJE7Bnzx5s2bIFkyZNanD70aNH4+uvv0ZmZiYKCwvx6aefGu/z8fFB//798e6776KkpAR6vR4pKSk4efLkPXPl5uYax0fK5XI4Ojo22H3Jy8vrrvzDhg1DQkICNm7c2Gj+ujw8PODr64vY2FjodDr89NNPTV5XzcfHBwMHDsSKFStQWFgIrVZr7Hrt6ekJlUrVYDew0aNH4+DBgzh27Bi0Wi2++OILyOVyKJXKJmevNm7cOHz11VdITU1FaWkp1qxZg9GjRzfpisTgwYORk5ODL7/8EhUVFSgpKcH58+cBVK3Rt3btWuTn5yM/Px///ve/jV3TwsLCcP36dVy+fBnl5eV3jYW+l/r+hkRk2VxcXDB//nysXLkSe/bsgUajgVarxcGDB/Hee+8BADp16oSDBw9CpVIhJycHX3311QMfLyQkBOXl5Thw4AC0Wi3Wrl3b4Pi50tJSyOVyuLu7Q6PRYPXq1bXu9/T0rDUkpz7jx49HbGwsdu7caXyvBIDHH38cn376qXEypeLiYvz222/17mPkyJE4cOAATp8+jYqKCnz88cfGSYDqOn78OK5evQqdTgdnZ2fIZDJju9mU99gHaf8aOn/p1KkTTp06hfT0dBQXF+M///lPre0ayyOVSjFq1CisWbMGJSUlSEtLw4YNG5ptSE6nTp2we/duaDQaJCcn46effmryttOmTcMXX3yBixcvwmAwIDk52fhBcGM/k7+/P5RKJVavXo3y8nJcuXIFP/300wP9TIMGDUJSUhK2bt2KyspKxMXF4caNGxg8ePA9t7WxscHUqVPxzjvvICsrCzqdDvHx8aioqLjneUhYWBi2bdsGnU6HQ4cO3TW0rDFNeb3Q/TP7wrS0tBTx8fFYsGABJk6ciOXLlxs/0di1axfGjRt3179Zs2bV2kd2djauXbtWq/slPbh+/fphwYIFeOmllzBgwACkpqbWmmEQAIYOHYopU6Zg0qRJGDx4MKZNmwYAmDdvHhISEhAREYHnn3/eOMtvTePGjcPMmTMxbNgwtG7dGn/+85/vK5+DgwPmzJmDJ554AhERETh37hyAqjfRzp07QyKRNLqW3KOPPooBAwZg4sSJmDx58l0Z33vvPWi1WuPMwfPnz2/Sp2x6vR5ffvkloqOj0bt3b5w6dQp///vf633svHnzsGTJEkRERBgncLC3tzfOMni/E2u88cYb+PzzzxEVFYUbN27cV3H43nvvQSaTYfTo0ejXr5/xJKpdu3YYO3Yshg0bhoiIiLu6xrRt2xbvv/8+3njjDfTp0wf79+/HunXrIJfL7ys7AEydOhUTJkzA008/jaFDh0Iulzf5arOzszO++OIL7N+/H/3798fIkSONC8a/+OKL6Nq1q3Etuy5duuDFF18EUHXCN3fuXDz77LMYMWIEevXqdV+Z6/sbEpHlmzlzJpYsWYJPPvkEffv2xeDBg/Htt98aZw2fOHEiwsLCEBMTg5kzZz7U5GguLi7429/+hv/7v//DwIED4eDgUKubbE2TJk1CQEAAoqOjMXbsWPTo0aPW/dOmTcONGzcQERFhfB+sKyYmBklJSfDy8qq1Tvbw4cMxe/ZsLFy4ED179sS4ceNqrXFdU2hoKJYvX46XX34Z0dHRcHV1bTBzbm4u5s+fj169emHMmDHo3bs3Jk6cCKBqdtqdO3ciMjISb775ZoO/o/tt/xo6f+nfvz/GjBmDCRMmYMqUKRgyZEit7e6VZ9myZXBwcMCwYcPw5JNPYty4cZg6dWqjWZrqT3/6E2xtbdGvXz+8+uqrtT40uJfRo0djzpw5WLRoEXr27Im5c+caZ31+/vnnsXbtWkRERODzzz+/a9vVq1cjLS0N0dHRmDdvHl566aUGh4Q1xt3dHevWrcOGDRsQFRWF9evXY926dfVOUFmfV199FR06dMC0adPQu3dvfPDBB9Dr9fc8D3n99dexf/9+REREYOvWrfe1FnpTXi90/ySG+qb8Etjt27cxZ84cbNu2DSUlJRg1ahSOHDnywPv76quvcOPGjQZnAyPrsXTpUvj4+Ih2QeR//etfSEpKwgcffCB0FCIiIiKiZmP2V0yrZy+r7hJiMBganNWrIdu3b8fYsWNNEY9E5Pbt29i9e7fx00+xUalU+Pnnn/HYY48JHYWIiIiIqFmZXWG6cOFCPP7440hMTMTAgQPx448/4v333zf2Wx87dux9Lfly+/ZtZGRkNHmJC7JMH330kXHNqfpmbjN3mzZtwuDBgxEdHY3IyEih4xARERERNSuz7MpLRERERERE1sPsrpgSERERERGRdWFhSkRERERERIK69yKALUiv10OnM/+exVKpRBQ56xJjbjFmBsSZW4yZAXHmFmNmQHy5bW2lQkewCGybTUuMucWYGRBnbjFmBsSZW4yZAfHlbqxtNqvCVKczQKVSCx3jnhQKR1HkrEuMucWYGRBnbjFmBsSZW4yZAfHl9vZ2ETqCRWDbbFpizC3GzIA4c4sxMyDO3GLMDIgvd2NtM7vyEhERERERkaBYmBIREREREZGgWJgSERERERGRoFiYEhERERERkaBYmBIREREREZGgWJgSERERERGRoFiYEhERERERkaBYmBIREREREZGgWJgSERFZgKVLl6Jv374YN25cvfcbDAa8+eabGD58OMaPH49Lly61cEIiIqKGsTAlIiKyAFOmTMH69esbvP/QoUNISkrCrl278MYbb+Dvf/97y4UjIiK6BxamREREFiAyMhJubm4N3r93715MmjQJEokEPXr0QFFREbKzs1swIRERUcNkQgcg85JZVIbDt/JxMaMI3QNcMayDN9wcbIWORUREDykrKwt+fn7Gr/38/JCVlQUfHx8BUxERWYdKvQFlWh3UFTqotTpotDqUafWo0OlRqTOgQqeHVqeH1njbcOfrO7f1eugNVcMydHrAAAP0BsDWVgpNmRYGAHqDAXr9nf9R/VgDDAZUbQsDAMBQ9d+dr6oeV1Pd+wFgarg/+oV4mPJXxMKU/ufIrTy8uiUBFToD3OxliEvIxpoDt/DOuE6IbucpdDwiIhKAVCqBQuEodIx7kkptRJGzLjHmFmNmQJy5xZgZEGfupmQur9Qjp7gMBWotVBotCtVaFGq0KFBXoFBTfbvq/+KySqgrKqHR6lBaoUNFpf7h8tlIYCMBbCSSqn82NW5L6ty+81jJne9J79yWSAAJJAAASdV/d76q+XXtO6rv17XA35SFKQEA9l3LwWvbr6CDtxNWjglDsLsDrmWX4q3d17B4SwJWTeiMgSxOiYhEy9fXF5mZmcavMzMz4evre8/tdDoDVCq1KaM1C4XCURQ56xJjbjFmBsSZW4yZAXHmltrLcTklHxlFZcgoKkdOSQXySsuRW1px53YFCssqG9zeSS6Fm70Mrva2cLWXobXCHg5yKRxtpXCwlcJRbgOH6tu2UjjIpbCX2UAutYGtVAJbqc2df1W35TW+J7ORQGojqfe4Lfm7bo7jeHu7NHgfC1NCRlEZlv92FZ19XfDx1K5wtqt6WnT0dca/p3XH3J8uYMnWBHwzvSfaejoJnJaIiB5ETEwMvvnmG4wdOxbnz5+Hi4sLu/ESkdUwGAwo0GiRmKdGUr4at1VlSC+s+pdRVHZX0SmzkcDLSQ5vZzlauzugZ6AbvJ3t4OUkh8LR1liEujnI4Gong0zKqXseFgtTwof7bkIC4O1xYcaitJqLvQwfTemKRzecxtu7ruPTx8NhI6n/ExsiIhLOwoULcfLkSRQUFGDgwIF46aWXUFlZdaL1xBNPYNCgQTh48CCGDx8OBwcHvP322wInJiIyjdzSClzLLkFinhqJ+Wok3SlGaxafcqkE/q72CHCzRxd/F7Tzc4W7rQ0C3Ozh72oHhYMtJDznbVEsTK3coZt5OHgzDy9Fh8DP1b7ex3g4yrFgUFus3HkNv1zIwNTwgBZOSURE97J69epG75dIJPjb3/7WQmmIiEzPYDAgq7gcV7NLcCWrBFfu/J9bWmF8jMLBFiEeDojp4IU2Ho4I8XREiIcjfFzsal1sEWP3Y0vDwtSKGQwGfHIkEW08HPBEr1aNPnZcF1/EXc7Gvw8nYWSYz11XVomIiIiITKm8Uo+EzGKcTyvE+fQiXMooRoFGCwCwkQDBHo6IbK1AmK8zOvo4o62nI9wd5QKnpqZidWHFTqeqcDNXjeUjO8D2Hv3iJRIJ5kWH4Nlv4/HLhQxMjwxqoZREREREZI1UGi3OpxUZC9HLWcXQ6qoWMWnj4YD+bT3QydcZYb4uCPV2goOtVODE9DBYmFqx/55Nh8LBFiPCmjb5RRc/F/RurcC3Z9LwqLIV7GQc5E1EREREzaNMq8P5tCKcSC7AyRQVrmaXAKiaiKiTrwseV7ZCeCs3hAe4QuFoK3Baam4sTK1UWqEGh27mYUZU0H0VmM9GBeHFH//AtkuZHGtKRERERA9MbzDganYJTiarcCK5AOfTClGhM0BmI0G3AFe80C8YvYIU6OTrDHteDbV4LEyt1ObzGbCR4L6Ly4ggBbr4ueC7M2mY0t2fs5URERERUZNptDqcSCrA4Vt5OHIrH/nqqjGi7bwcMa1HAHq3docy0A2Ochai1oaFqRUyGAzYfTUHfdp4wMfF7r62lUgkmBLujzd2XsOF9CKEt3IzUUoiIiIisgSZRWU4cisfh2/l4XSKChU6A5zkUvRt44Hodh7o3VoBL+f7Oycly8PC1AolZBYjo6gcz/cLfqDth3Xwxof7biL2j0wWpkRERERUi8FgQEJmMQ7ezMz5jogAACAASURBVMPhm3m4nlMKAAhU2GNqeACi23lA2coNsntMvknWhYWpFdp9NRcyGwkGtfN6oO0d5VIMD/PGrivZWBTTDk5yPo2IiIiIrJneYMCljGLsvZaLAzdzkaYqg40ECA9wxfyBIYhu64lgDwcOA6MGsaKwMgaDAXuv5SAq2B0u9g/+55/Y1Q+xf2Ri95UcTOru34wJiYiIiEgMdHoDLqQXYd/1XOy7loPskgrIbCQY0N4Ls6JaI7qdJxQOnD2XmoaFqZW5lFmMzOJyvND/wbrxVuvq74I2Hg7YcSWbhSkRERGRlajUG3DudiH2XsvB/ht5yCutgFwqQd82Hpgb7YWB7TwR6OsKlUotdFQSGRamVubQzTxIJcDAdp4PtR+JRIJhHbzxxYkU5JZWwMtJ3kwJiYiIiMicVOr0OJNaiL3Xc3Dgeh4KNFrYyWwwoK0HYkK90L+tB4d20UPjM8jKHE8qQLcAV7jaP3y3iqEdvbH+eAr2X8/FIz24pikRERGRpdDq9DiZrMLeazk4dDMPhWWVcLSVYkBbDwzt4IW+IR5w4Nqi1IxYmFqRvNIKXMkqeeDZeOtq5+mIEA9H7L2Ww8KUiIiISOTKK/U4nlSAfderitGSch2c5FIMbOeJoR28EBXsDnsWo2QiLEytyO83c2EA0LeNe7PsTyKRYGgHL3xxIgV5pRXwZHdeIiIiIlEp0+pwLKkAe6/l4PDNfKi1OrjayzCkvRdiOnihd2t3yGVc1oVMj4WpFTlyIw9u9jKE+bo02z5rduedxqumRERERGavvmLUzV6GEWHeGNrBCxFBCq4xSi2OhamVMBgMOHojF5Gt3SG1ab71o9p5OiJIYY/Dt/JYmBIRERGZqfJKPY4l5mNPPcXosI7e6BWkgKwZzxGJ7hcLUytxM0+NrOJyPNdX0az7lUgkGNDWEz+fT4dGq+MgeCIiIiIz0Wgx2sEbvYLceGWUzAYLUytxNrUQABDZunnGl9Y0oK0Hvj+bhpPJKgxq/3DL0BARERHRg6uawCgfu6/m4MitfJRWVBWjw8O8MZzFKJkxFqZW4lxaIfxc7eHvatfs+1YGusFJLsXhW3ksTImIiIhaWHUxuudaLg7fzDMWo8M6emMYx4ySSLAwtQIGgwHn0grRO8QDEknzjx2wldqgbxt3HL2VD73BABsTHIOIiIiI/kerq1raZdfVHGMx6movw7AO3hja0QuRLEZJZFiYWoH0ojLklFQgMtjDZMcY0NYTe67l4kpWCTr7Nd+sv0RERERURX/nYsPOyznYey0HhWWVcLWXYWgHLwzr6M1ilESNhakVOJ9WBADoFdz840ur9Q2p2vfxpAIWpkRERETNxGAw4HpOKXZeycbOKznIKi6HvcwGg9p7YlQnH0QFu8OWxShZABamVuBcWiGc7aTo4OOMoiKNSY7h4ShHRx9nHE8uwMw+rU1yDCIiIiJrkVaowXfnMvDruTQk5qkhtZGgbxt3zIsOwcB2nnCUcyUEsiwsTK3AudtFCA9wg42J16aKCnbHt2duo7SiEk5yPrWIiIiI7keZVod913Ox9WImTt9ZUUHZyhVLhrXH0FBvKBxtBU5IZDqsHiycSqNFYr4aozv7mPxYfdu4Y+OpVJxJLcTAdpydl4iIiOheDAYDEjKLseViFnZeyUZphQ6t3Ozx5/5t8FifYDjBIHREohbBwtTCXcosBgB0D3A1+bG6B7jCXmaDE0kFLEyJiIiIGqHSaLH9Uha2XMzErTw17GQ2GNbBC+O7+kEZ6AYbiQQKhQNUKrXQUYlaBAtTC5eQWQwJgDBfZ5MfSy6zQa8gBY4nF5j8WERERERilJBZjB/PpWP31RyUV+rR1d8FS4eHYkRHbzjb8dScrBef/RYuIbMYbTwdW2zMZ1Qbdxzdn4+MojL4u9q3yDGJiIiIzFlFpR57ruXgx3PpuJhRDAdbG4zr4otp4QFo7+0kdDwis8DC1IJVj1noG2K69UvrimytAACcTlFhfFe/FjsuERERkbkp1Gjx8/kM/Dc+DflqLYLdHfDykHYY28WXV0eJ6uArwoJlFZcjX61FlxZcV7SdpyMUDrY4k8rClIiIiKxTRlEZvjuThtg/MqDR6tE/xANP9GqF3q0VkEhMu0oCkViZtDD98ssv8eOPP0IikaBDhw545513YGdnZ8pDUg3VEx91bsHCVCKRoFeQG86kFsJgMPDNl4iIiKzGjdxSfHkiBXuu5gASCUaFeePpiCB21yVqApMVpllZWdi4cSPi4uJgb2+PBQsWYPv27ZgyZYqpDkl1JGQWw1YqQahXy74Z9gpSYO+1XKQVliFQ4dCixyYiIiJqaYl5anx2LBl7rubAwVaKx3sG4vGeAfDjfBtETWbSK6Y6nQ5lZWWQyWQoKyuDj4/p19Kk/0nILEaotzPkMpsWPW6vIDcAwJlUFQtTIiIislhJ+WqsP5aMXVeqCtJno4LwVK9AuDnYCh2NSHRMVpj6+vpi5syZGDJkCOzs7NC/f38MGDDAVIejOvQGAy5nlWB0p5b/MCDEwxEejrY4k1qIid38W/z4RERERKaUWVSGdUeT8NvlbMilNpgeGYTpEYFQOLIgJXpQJitMCwsLsXfvXuzduxcuLi5YsGABYmNjMXHixAa3kUolUCgcTRWp2UilNmafMzmvFKUVOvQM8TRmbcncfdp64kxKAdzcHB5qnKkYftf1EWNuMWYGxJlbjJkB8eYmImoupRWV2HgyFd+eSYPBYMDjPVvhT72D4OEoFzoakeiZrDD9/fffERgYCA+PqqVKRowYgfj4+EYLU53OAJVKbapIzUahcDT7nKdu5AIAgpxtjVlbMnd3P2fEXczExaR8BLk/eHdeMfyu6yPG3GLMDIgztxgzA+LL7e3dchO/EZFl0+kN2HoxE2uPJiFfrcXIMG/MjQ7hmu1EzchkhWlAQADOnz8PjUYDe3t7HDt2DF27djXV4aiOazklkNpI0NZTmFngegXeWc80VfVQhSkRERGRkM6kqvDBvpu4kVuK7gGu+HBSF3T1dxU6FpHFMVlhGh4ejpEjR2Ly5MmQyWTo1KkTHnvsMVMdjuq4ml2Ctp6OLT7xUbVgDwd4OslxJlWFyd05zpSIiIjERaXW4qNDt7D9UhYCXO3wzrhOGNrBi0vhEZmISWflnT9/PubPn2/KQ1ADrmWXIipYIdjxJRIJegVyPVMiIiISF4PBgK2XsvDxwVsoqdBhRlQQZka1hr2tVOhoRBbNpIUpCSOvtAK5pRXo4OMsaI5erRXYdTUHyQUatPHghClERERk3pLy1Hhnz3WcvV2I8ABXLB0einYtvB48kbViYWqBruWUAAA6ClyYRgRVXbE9m6piYUpERERmS6c34NvTt7Hu9yTYy6R4bXgoJnbzgw17fBG1GBamFuhadikAINRb2E/4ghT28HaW43RqIaaEBwiahYiIiKg+aYUarPjtKuLTijAk1AuvDm0PTycu/0LU0liYWqCr2SXwd7WDq72wizxLJBL0ClLgZHIBx5kSERGRWakeS/rhvpuQSIC/j+qIMZ19eL5CJBAWphboWnYJOngL2423mjLQDTsuZyOlQINgduclIiIiM1CgrsDbu6/jwI089Ax0w99Hd+SapEQCY2FqYcq0OqSqNBjW0VvoKACAnq3cAADn0gpZmBIREZHgzqSq8Pr2Kygq02LBoLZ4slcrjiUlMgPCLHJJJpOcr4HeALQ3kxnkgj0c4O5gi/jbhUJHISIiIium1xuw4UQKXvzxApzkUnz5pBJPRwSyKCUyE7xiamFu5FZNfGQuU5tLJBIoA91YmBIREZFgVBotFm1JwKHruRjR0RuvjQiFk5ynwUTmhK9IC3MjtxRyqQRB7g5CRzFSBrph3/VcZBaVwY/jN4iIiKgFXcsuwSuxl5BTWoElw9pjSnd/TnBEZIbYldfC3MwtRRsPR8hszOcNV3lnnGl8Gq+aEhGZyqFDhzBy5EgMHz4cn3766V33p6enY/r06Zg0aRLGjx+PgwcPCpCSqGXtvpqDWd+fg1ZvwPezozA1PIBFKZGZYmFqYW7mlppNN95q7b2d4CSXsjsvEZGJ6HQ6rFy5EuvXr8f27duxbds23Lhxo9Zj1q5di9GjR+PXX3/FmjVrsGLFCoHSEpme3mDAJ0cS8dq2ywj1dsbGp5QID1QIHYuIGsGuvBakqEyL7JIKs5n4qJrURoIerdxw7naR0FGIiCzShQsXEBwcjKCgIADA2LFjsXfvXrRv3974GIlEgpKSEgBAcXExfHx8BMlKZGplWh3+9ttV7Luei4nd/LA4pj3kMl6LITJ3LEwtyM1cNQDzmfioJmWgG44m5iNfXQEPR7nQcYiILEpWVhb8/PyMX/v6+uLChQu1HjNv3jzMmjUL33zzDTQaDTZs2NDSMYlMLl9dgUW/XsKljGL85c5SMOy6SyQOLEwtyP9m5DW/9UKVgdXrmRYhJtRL4DRERNZn+/btmDx5MmbOnIn4+HgsXrwY27Ztg41N41eSpFIJFArza1fqkkptRJGzLjHmNtfMN7JL8NwP55FbUo5/PaHEiM6+te4319yNEWNmQJy5xZgZEG/u+rAwtSA3c0vhbCeFr4ud0FHu0snXGXYyG8TfLmRhSkTUzHx9fZGZmWn8OisrC76+tU/Kf/rpJ6xfvx4AoFQqUV5ejoKCAnh6eja6b53OAJVK3fyhm5lC4SiKnHWJMbc5Zv4jvQh/+eUiZDYSrHukO7r4u9yV0Rxz34sYMwPizC3GzID4cnt7uzR4HzvcW5BbuaVo6+lkll1WbKU26BbgygmQiIhMoFu3bkhKSkJqaioqKiqwfft2xMTE1HqMv78/jh07BgC4efMmysvL4eHhIURcomZ1PCkfL/54Aa72Mnz+RA908XcVOhIRPQBeMbUgifkaDGrf+CffQlK2csX6YykoKa+Esx2fekREzUUmk2H58uWYPXs2dDodpk6ditDQUPzjH/9A165dMXToUCxZsgT/93//hy+//BISiQTvvvuuWX6QSXQ/dl/NwfK4KwjxdMTHU7vBy4nzWBCJFasDC6FSa6HSaBHiYb59zJWBbjAAOJ9WhP5t+Sk9EVFzGjRoEAYNGlTrewsWLDDebt++PX744YeWjkVkMr9cyMA7u68jvJUrVk/qChd7ntYSiRm78lqIxPyqvuVtPM23MO3m7wqpjQTxaezOS0RERA/ux3PpeHv3dfRp445/Tu3GopTIAvBVbCGqC1NzvmJqbytFZ18XjjMlIiKiB/bD2TR8uP8mott64N3xnblGKZGF4CvZQiTlqWEvs4Gfq/nNyFuTMtANCZnFKNPqhI5CREREIvPt6dv4cP9NDG7viVUTWJQSWRK+mi1EYr4awR6OsDHziSx6BrqhUm/AxYxioaMQERGRiHx/Ng0fHbyFoR288M64TrCV8jSWyJLwFW0hkvLUaOPhIHSMewpv5QoJwO68RERE1GS/XMjA6v03MSTUC2+O7QQZi1Iii8NXtQVQV+iQWVyOEDOe+Kias50Mod5OOMsJkIiIiKgJfruchXd2X0e/EHe8NTYMMhvz7h1GRA+GhakFSC4w/4mPalIGuuGP9CJodXqhoxAREZEZ2389Fyt+u4qeQW5YNb4zu+8SWTC+ui1AYp75LxVTU89AN5RX6nElq0ToKERERGSmTqeo8Pr2y+js54IPJ3WBva1U6EhEZEIsTC1AUr4aUgkQpDD/MaYA0CPQDQDHmRIREVH9rmWX4OXYSwhUOGDN5K5wknOFQyJLx8LUAiTmqRGocBBN9xYPRznaeDggnuNMiYiIqI6MojIs2HwRTnIpPp7SFW4OtkJHIqIWII5KhhqVnK8RxcRHNSkD3XAurRA6vUHoKERERGQmVBot5v/8B8or9fjH1G7wc7UXOhIRtRAWpiJXqdMjRaVBG5FMfFRNGeiGknIdbuSWCh2FiIiIzEBFpR6vxF5CemEZPpjUGe29nISOREQtiIWpyN1WlUGnN4jvimkrjjMlIiKiKgaDAW/vvoZzaUX4++gw9AxUCB2JiFoYC1ORS8y/MyOvyK6Y+rnaw9/VDuc4zpSIiMjqfXkyFdsTsvFCv2AM7+gtdBwiEgALU5FLEmlhClR1542/XQiDgeNMiYiIrNW+67n45EgSRoZ5Y1af1kLHISKBsDAVucQ8NXxd7OAoF9/aXspWbshXa5FcoBE6ChEREQngSlYxlsddQTd/Vywb2RESiUToSEQkEBamIpeUr0aICK+WAlVXTAGOMyUiIrJGKrUWr8QmQOFgiw8mdYadjKelRNaM7wAipjcYkJSvRhuRTXxUrbW7AzwcbVmYEhERWZlKvQGvbb+MfHUF3pvQGR6OcqEjEZHAWJiKWHZxOTRaPUI8HISO8kAkEgl63hlnSkRERNZj7ZEknEpR4dVhoejs5yJ0HCIyAyxMRcw4I69Ir5gCVd15M4vLkVFUJnQUIiIiagH7ruVg46lUTA33x4SufkLHISIzwcJUxJLyqyYNEuOMvNV6cD1TIiIiq5GUr8aKHdfQzd8FCwe3EzoOEZkRFqYilpKvhoudDO4OtkJHeWDtvJzgYidjYUpERGThyiv1eG3bZdhKJXhnfGfIOdkREdUgEzoAPbiUAg2CPRxEPbW61EaC8FauLEyJiIgs3EcHbuJ6TinWTO4CXxc7oeMQkZnhR1UillygQWt3cU58VFPPQDckF2iQV1ohdBQiIiIygX3XcvDT+Qw81SsQA9p6Ch2HiMwQC1ORKtPqkFVcbhGFafV6pufSeNWUiIjI0qQVavDGrmvo4ueCudFthI5DRGaKhalIpaqqJj5q7S7eiY+qhfk4w15mw+68REREFqZSb8Cy7VcAAG+NC4OtlKeeRFQ/vjuIVEpBdWEq/iumMqkNugdwnCkREZGl+epkCv7IKMbSYaFo5Sb+cxYiMh0WpiJVXZgGKSzjTb5HoBuu55SiuKxS6ChERETUDC5nFeOzYykYGeaNEWE+QschIjPHwlSkkgs08HGWw1EuFTpKs+gZ6AYDgPPpvGpKREQkdmVaHf4WdxWejrZYPLS90HGISARYmIpUSr5lzMhbrYufC2Q2EnbnJSIisgCfHElCYr4ay0d2hKu9eNdbJ6KWw8JUpFIK1Aj2EP/ER9XsbaXo4ufCwpSIiEjkTqeo8P3ZNDzSIwBRbdyFjkNEIsHCVIRUGi0Kyyot6oopULVsTEJWCdQVOqGjEBER0QPQaHV4Y9c1tHZ3wPyBIULHISIRYWEqQpY0I29NEa0V0OkNiOd6pkRERKK07mgS0gvL8PqIUNjbWsY8GETUMliYilBKgRqAZaxhWlN4gCvkUglOJhcIHYWIiIju06XMYvxwNg1TuvujZ6BC6DhEJDIsTEUopUADqY0EAa52QkdpVva2UnRv5YZTKSqhoxAREdF9qNTp8ebOa/B0kuMlduElogdg0sK0qKgI8+fPx6hRozB69GjEx8eb8nBWI6VAg1Zu9pBJLe9zhd6tFbieU4p8dYXQUYiIiKiJNp66jRu5pXh1aCic7WRCxyEiETJpZfPWW28hOjoaO3bsQGxsLNq1a2fKw1mNZAtbKqam3sFVs/ed5lVTIiIiUUjOV2P98WQM6+CFQe09hY5DRCJlssK0uLgYp06dwrRp0wAAcrkcrq6upjqc1dAbDEhVaRBsYeNLq4X5OMPFToaTySxMiYiIzJ3BYMAH+25CLrXBopj2QschIhEzWV+L27dvw8PDA0uXLsWVK1fQpUsXvP7663B0bLigkkolUCjMv+CSSm0Ey5mu0qC8Uo+wVm73nUHI3PejT1tPnL6tgpubg2gy1yXG3GLMDIgztxgzA+LNTUSms/96Lo4nF2DRkHbwcpILHYeIRMxkhWllZSUSEhKwbNkyhIeH480338Snn36Kv/zlLw1uo9MZoFKpTRWp2SgUjoLl/OPOjLXe9tL7ziBk7vvRw98Fuy9n4VJyPrq28RRF5rrE8ruuSYyZAXHmFmNmQHy5vb1dhI5AZNE0Wh1WH7iFUG8nTOsRIHQcIhI5k3Xl9fPzg5+fH8LDwwEAo0aNQkJCgqkOZzUsdQ3TmnoHV00xz2VjiIiIzNeGEynIKi7H4pj2kNlIhI5DRCJnssLU29sbfn5+uHXrFgDg2LFjnPyoGaQUaOBga2PR3WWC3R3g4yzHSU6AREREZJaS89X45vRtjOnsgx6BbkLHISILYNL5vJctW4aXX34ZWq0WQUFBeOedd0x5OKuQUqBGa3dHSCSW+8mkRCJBZLA7jtzMg15vEDoOERER1WAwGPDh/qoJj14a2FboOERkIUxamHbq1AmbN2825SGsTnK+Bp39LH/cVO/WCmy/lIXLmUVo5WgrdBwiIiK64/CNXBxLKsBfB7e16B5cRNSyTLqOKTWviko9MorKLHp8abXeravGmR69mSdwEiIiIqpWqTfg3R1XEaiwxyOc8IiImhELUxFJKyyD3gAEe1h+YerlbIcQT0ccvcHClIiIyFxsu5iJ69kleCk6BLZSnkYSUfPhO4qIpBRULdPQ2t061hHs18YDp5Lzoa7QCR2FiIjI6qkrdFj3ezJ6tlZgSKiX0HGIyMKwMBUR41IxCsu/YgoA/ULcodUZcIqz8xIREQnum9OpyCutwJJRYRY9CSMRCYOFqYgkF2jg4WgLF3uTzlllNnq0coOTXIrfE/OFjkJEZPYOHTqEkSNHYvjw4fj000/rfUxcXBzGjBmDsWPHYtGiRS2ckMQsp6QcX5+6jWEdvKEMUggdh4gskHVUOBYiJV9tFRMfVZPLbNCvnSeOJubDYDDw01kiogbodDqsXLkSGzZsgK+vL6ZNm4aYmBi0b9/e+JikpCR8+umn+P777+Hm5oa8PI7hp6b7z+/JqNQbMDe6jdBRiMhC8YqpiCQXaKyqMAWAQR28kVVcjlt5aqGjEBGZrQsXLiA4OBhBQUGQy+UYO3Ys9u7dW+sxmzZtwlNPPQU3NzcAgKenpxBRSYRSCzTYdjETU8P9EWglw4mIqOWxMBWJkvJK5Ku1VjPxUbWBod4AwO68RESNyMrKgp+fn/FrX19fZGVl1XpMUlISEhMT8fjjj+PRRx/FoUOHWjomidRnx5Ihk9rg2ajWQkchIgvGrrwiYZz4yMqumPq72SPU2wlHE/MxPTJI6DhERKKl0+mQnJyMr7/+GpmZmXj66aexdetWuLq6NrqdVCqBQmH+H4pKpTaiyFmXuee+nl2CHVeyMbt/CNq3qhpbau6ZGyLG3GLMDIgztxgzA+LNXR8WpiJRXZhawxqmdfUL8cA3p2+jpLwSznZ8yhIR1eXr64vMzEzj11lZWfD19b3rMeHh4bC1tUVQUBDatGmDpKQkdO/evdF963QGqFTmP5xCoXAURc66zD33hzuvwNFWike7+xlzmnvmhogxtxgzA+LMLcbMgPhye3u7NHgfu/KKREqBGhIAgW7WV5j2D/GATm/AyeQCoaMQEZmlbt26ISkpCampqaioqMD27dsRExNT6zHDhg3DyZMnAQD5+flISkpCUBB7olDDrmaXYO+1XDzRsxUUDrZCxyEiC8fLTyKRUqCBv5s95DLr+yyhW4ArnO2k+D2xADEdvIWOQ0RkdmQyGZYvX47Zs2dDp9Nh6tSpCA0NxT/+8Q907doVQ4cORXR0NI4ePYoxY8ZAKpVi8eLFcHd3Fzo6mbH/HE2Ci50MT/YKFDoKEVkBFqYikZxvfTPyVpPZSNAn2J3LxhARNWLQoEEYNGhQre8tWLDAeFsikWDp0qVYunRpS0cjEbqUUYTDt/Lx4oA2VrN+OhEJq0mX3+bNm4cDBw5Ar9ebOg/Vw2AwIKVAg9ZWPEV7vxAP5JZW4FpOqdBRiIhMju0uCe3z4ylws5fhUWWA0FGIyEo0qTB98sknsXXrVowYMQIffPABbt26ZepcVENeaQXUWp1VTnxUrW+IByQADt/kgvBEZPnY7pKQrueU4PCtfDzesxWc5LxaSkQto0nvNv369UO/fv1QXFyMbdu2YcaMGfD398cjjzyCCRMmwNaWA+JNKbl6Rl4rW8O0Ji8nOboFuOLAjTzM7hssdBwiIpNiu0tC+upkatVMvLxaSkQtqMkz6RQUFGDz5s348ccf0alTJzzzzDNISEjAzJkzTZmP8L/CtLUVXzEFgCGhXriaXYK0Qo3QUYiITI7tLgkhtUCD3VdzMK2HP1zt+QEIEbWcJl0xnTt3LhITEzFx4kSsW7cOPj4+AIAxY8ZgypQpJg1IQEq+BnYyG/i62AkdRVCD23viHwdv4cD1PDwVwRkCichysd0loXx1KhUyGwme4Ey8RNTCmlSYPvroo3fN9FdRUQG5XI7NmzebJBj9T3KBGkEKB9hY+Wy0gQoHhHo7Yf/1XBamRGTR2O6SELKKy7H9UhYmdfODl5Nc6DhEZGWa1JX3o48+uut7jz32WLOHofqlFFjvUjF1DQn1woX0IuSWVggdhYjIZNjukhC+PX0bBoMB0yODhI5CRFao0SumOTk5yMrKQllZGRISEmAwGAAAJSUl0Gg4zq8lVOr0SCssQ0yol9BRzMKQUC98+nsyDt3IxZRwTspARJaF7S4JRaXR4pcLGRjVyQcBbvZCxyEiK9RoYXrkyBFs3rwZmZmZeOedd4zfd3JywsKFC00ejoC0wjLo9AarXiqmpnaejmjt7oD91/NYmBKRxWG7S0L55UIGyir1eJpXS4lIII0WppMnT8bkyZOxc+dOjBw5sqUyUQ0p1TPyWvFSMTVJJBIMbu+Fb8/cRlGZljMGEpFFYbtLQqio1OO/8enoE+yO9l5OQschIivVaGEaGxuLiRMnIi0tDRs2bLjr/hkzZpgsGFX53xqmvGJabUioJzaeSsWRW/kY09lX6DhERM2G7S4JYffVHOSVVuBvozoIHYWIrFijhWn1eBa1Wt0iYehuKQVquNnL4ObAK4PVOvu5wMdZjv3Xc1mY+CKVVQAAIABJREFUEpFFYbtLLc1gMOC7M7cR4umIPsHuQschIivWaGH6+OOPAwDmzZvXImHoblUz8rIbb002d7rzxl7MhEarg4OtVOhIRETNgu0utbQzqYW4llOK14eHQmLly9IRkbCatFzMe++9h5KSEmi1WvzpT39Cnz59EBsba+psBCA5X8OJj+oR08EL5ZV6HLyRJ3QUIqJmx3aXWsp3Z27D3cEWozr5CB2FiKxckwrTo0ePwtnZGQcOHECrVq2we/dufP7556bOZvVKKyqRW1rBNUzroQx0g6+LHXZczhY6ChFRs2O7Sy0hOV+Nw7fyMTXcH/bsfUREAmtSYarT6QAABw4cwKhRo+Di4mLSUFQllRMfNchGIsHIMG8cT8pHgbpC6DhERM2K7S61hB/OpsFWKsG0Hlx+jYiE16TCdPDgwRg1ahQuXbqEvn37Ij8/H3Z2dqbOZvWS8+8sFePBMab1GdXJBzoDsPtqrtBRiIiaFdtdMrWS8kpsT8jCyDAfeDrJhY5DRNT45EfVXn75ZcyePRsuLi6QSqVwcHDAJ598YupsVi+lQAMJgEA3e6GjmKVQb2e083LEjsvZeFTJT3uJyHKw3SVTi0vIgkarxyO8WkpEZqJJhSkA3Lp1C2lpacbuRQAwadIkk4SiKskFavi52nHcRyNGd/LFvw4n4rZKg0AFuzwTkeVgu0umYjAY8NO5DHT2c0FnP3YTJyLz0KTC9JVXXkFqairCwsIglVYVSRKJhA2kiaUUaBDMpWIaNTLMG/86nIgdl7Mxu2+w0HGIiJoF210ypbO3C5GYr8bykR2EjkJEZNSkwvTixYuIi4vj+lYtyGAwIKVAg7GdXYWOYtb8XO2hDHTDjsvZmNWnNZ+jRGQR2O6SKf14Lh1u9jIM7+gtdBQiIqMmTX4UGhqKnJwcU2ehGvJKK1BaoeMapk0wqpMPkgs0uJJdInQUIqJmwXaXTCWnpBwHrudifFc/DhUiIrPSpCumBf/f3p2HR1nd7QO/Z802SSaTZRJIyAIJYQkBZEdBw6aETRa1Lm2pVuUVwddqq/aVKq3UutTS2ira/kTEIiIVFJBFokBBiKwBshCWLCSZySSZLJNlJjPz/P4A0yIRAszMmcncn+vyksw8ydx5Lpgz3+c553vMZmRnZ2PQoEFQqVQdj7/99ttuC+bvSi9uFcM9TK9uQmoUXt15GlsLqtFPz7UyROT7OO6Su2zIM8AhAXMy40RHISK6RJcK08cff9zdOeh7yjoKU64xvZrwIBVuTtFha0E1Ft6SDJWiSxMBiIi8Fsddcge7w4lPj1dhdFIEGwYSkdfp0if4ESNGoGfPnrDb7RgxYgQyMjLQv39/d2fza6V1rVArZIgN4751XTFjYCzqWtqx50yt6ChERDeM4y65w+4ztTBZbJjLLWKIyAt1qTD9+OOPsWjRIixZsgQAYDQa8dhjj7k1mL8rM7cgISIIcja+6JLRyTrEaNT49LhBdBQiohvGcZfc4dM8A/ShARibrBMdhYjoMl0qTD/88EOsWbMGGo0GAJCUlIS6ujq3BvN3peZWTuO9Bkq5DDMzYnGgxIzKhjbRcYiIbgjHXXK1yoY2HCg1Y8ZAPRRyXvQmIu/TpcJUrVZDrVZ3fG23290WiC6sAaloaEMiGx9dkxkDYwEAG0/wrikR+TaOu+Rqn18cG6dfHCuJiLxNlwrT4cOH4+2330ZbWxv27t2LxYsXIysry93Z/FZloxUOp8SOvNcoNiwQY5J1+PyEAXanJDoOEdF147hLruRwSvjshAEjkyIQFxYoOg4RUae6VJg+9dRT0Ol0SEtLw9q1azF+/Hg88cQT7s7mt0rrWgAAiTpO5b1Wdw6Khcliw96znPJGRL6L4y650v5SM6otNszK4N1SIvJeXdouRi6XY+LEiZg4cSJ0Oi6Yd7cy7mF63camRCIqRI0Nx6swvk+k6DhERNeF4y650sbjBkQEqTCuN8dFIvJeVyxMJUnCm2++idWrV0OSLkyNlMvluP/++7Fw4UKPBPRHZeZWhAcqoQ1SXf1guoRSLsOMgXqszC2HobENsZyyREQ+hOMuuVptsw27z9TiR0N7cp9vIvJqV3yHWrlyJQ4fPoxPPvkEubm5yM3Nxbp163DkyBGsXLnSQxH9T6m5hR15b8CMjFhI0oUrxEREvoTjLrnalnwjHE4JM9n0iIi83BUL040bN+L1119HQkJCx2MJCQl49dVXsWHDBreH81dl5lb00nEa7/XqGR6EsSk6/CuvCla7U3QcIqIu47hLriRJEjYcN2BwzzAkRfKCNxF5tysWpna7vdO1LTqdjq3r3aTZZofJYuNWMTfoR0N7oq6lHdsKq0VHISLqMo675ErHKhpRZm7t2E6NiMibXbEwVal+eI3jlZ6j61d+sfERC9MbM7yXFn2iQvDR4YqOdVpERN6O4y650uZ8I4JUckxIixYdhYjoqq7Y/KiwsBBDhw697HFJkmCz2dwWyp/9pyMvp9zcCJlMhh8N7Ynfbj+Fg+X1GN4rQnQkIqKr4rhLrtLW7sCXp0zISo1CsFohOg4R0VVdsTAtKCjwVA66qLSuFTIA8Vp2k71RU/rF4M0957DmUAULUyLyCRx3yVX2nK2DxerA1P560VGIiLqEfcO9zLm6FsSFBSBQxaubNypAKceczDj8+2xdx51oIiIif7Al34gYjRo3JWhFRyEi6hK3F6YOhwOzZs3CI4884u6X6hZK6lrYOc+F5gzuAaVCho8OV4iOQkRE5BG1zTZ8c64Od/TXQyGXiY5DRNQlbi9MV61ahd69e7v7ZboFh1NCaV0LknUhoqN0G1EhakxOj8HnJwxobGsXHYeIiMjtthVWwyEB2ZzGS0Q+xK2FqcFgwNdff425c+e682W6jarGNtgcEpIj2ZHXle4d2hNtdic+OVolOgoREZHbbT5pRP/YUCRzBhYR+RC3FqbLli3D008/DbmcS1m74lxtCwAgSceBxJXSYjS4OUWHfx46jxabQ3QcIiIityk2WXDK1Izs/jGioxARXZMrduW9EV999RV0Oh0GDhyIAwcOdOl7FAoZtFrvL8oUCrlbchpajQCAwSlRCA9y/X517srtTq7KvGhCKu569wC2FtfgZ2OTXZDsyvz5XHuaL+b2xcyA7+Ym8idb8quhkMswuS8LUyLyLW4rTA8fPoycnBzs3r0bVqsVFosFTz31FF577bUf/B6HQ0J9fYu7IrmMVhvslpz55+uhC1ZBsraj3ur69ZDuyu1OrsqcHBaAYQnheHfPOWT3jUaA0r138f35XHuaL+b2xcyA7+WOjg4VHYHIo+xOCV8UVOPmZB20wa6/wE1E5E5u+3T+i1/8Art370ZOTg7++Mc/YtSoUVcsSulCR94Urgdxm/kje6Gm2YZNJw2ioxAREbnct2Vm1DbbMHUAmx4Rke/h4k8vIUkSztW2cH2pGw3vpUVGXChW5ZbD7nCKjkNERORSm08aER6oxM3JOtFRiIiumUcK05EjR2LFihWeeCmfVdNsQ7PNwQ56biSTyTB/ZC9UNlqxtbBadBwiIiKXsVjt+Pp0LSb1jYbazctViIjcge9cXoIdeT3j5hQdUqNDsPJAORxOSXQcIiKX2b17N6ZMmYJJkybhnXfe+cHjtm3bhr59++L48eMeTEfulnOqBla7E9mcxktEPoqFqZf4rjDlHVP3kslkeGhUL5SaW7G1gHdNiah7cDgcWLp0Kf7+979j8+bN2LRpE06fPn3ZcRaLBatWrUJmZqaAlOROWwqM6BURhAGxbPpFRL6JhamXOFfXAk2AAlEhatFRur3bUqOQHqPBO/tK0M61pkTUDeTl5SExMREJCQlQq9XIzs7Gzp07Lztu+fLl+PnPf46AgAABKcldjE1WHC5vwO39YiCTyUTHISK6LixMvURJXQuSdcEcUDxAJpNhwc1JqGy0YuNxduglIt9nNBoRGxvb8bVer4fRaLzkmJMnT8JgMODWW2/1cDpytx1FJkgApqRz71Ii8l1u28eUrs252haMZRc9jxmdFIHBPcPwj/1lmDZAj0CVQnQkIiK3cTqdePnll/H73//+mr9XoZBBq/X+ZSYKhdwncn6fK3LvLK5BRs8wDEqOdFGqK/Pnc+1pvpgZ8M3cvpgZ8N3cnWFh6gUaWttR19LO9aUe9N1d00fW5mHd0Uo8MDxBdCQiouum1+thMPxnBojRaIRe/58mOM3NzTh16hR+/OMfAwBMJhMWLFiAt956CxkZGVf82Q6HhPr6FvcEdyGtNtgncn7fjeYurWvBicpG/O+tKR77/f31XIvgi5kB38zti5kB38sdHf3D6+A5ldcLlNSxI68IQ+O1GJUUgfdzy2Gx2kXHISK6bhkZGSgpKUF5eTlsNhs2b96MrKysjudDQ0Nx4MAB5OTkICcnB4MHD+5SUUreb3uhCTIAk/pGi45CRHRDWJh6AXbkFWfB2CQ0tNmx5lCF6ChERNdNqVRiyZIleOihhzB16lTccccdSE1NxfLlyzttgkTdgyRJ2FpYjZsSwhGtYUMrIvJtnMrrBc7VtSBAKUdcWKDoKH6nf2woslKj8MHBctyZGceuyETks8aPH4/x48df8tjixYs7PfaDDz7wRCRys6JqC8rMrbh/WLzoKEREN4x3TL1ASV0LekUEQSFnR14RFt6SDJtDwoq9JaKjEBERddnWAhOUchmyUqNERyEiumEsTL1ASe2FrWJIjISIIMwb3AOfnTDgdE2z6DhERERX5ZQk7CiqxuikCIQHqUTHISK6YSxMBWttd6Cy0Yokri8V6sFRvRCiVuIvu8+KjkJERHRVR843oNpiw+39uHcpEXUPLEwFK73YkZd3TMXSBqnws1G9sO+cGQdKzKLjEBERXdH2QhOCVHLc0tsze5cSEbkbC1PBztRcKEx7R4UITkJ3De6BHmEBWL77LBxOSXQcIiKiTrU7nNh5yoRxvSMRpFKIjkNE5BIsTAUrNjVDrZAhISJIdBS/p1bK8dgtySg2NePzE4arfwMREZEAB0rNaGizcxovEXUrLEwFO1PTjOTIECjZkdcrTOobjcweYfjbv0vQ1GYXHYeIiOgyWwuqER6oxMjECNFRiIhchoWpYKdrmtEnmtN4vYVMJsPTWX1Q39qOd74pFR2HiIjoEq3tDuw+U4sJadFQKfgxjoi6D76jCVTf0o6aZhv6cH2pV+mr1+DOQXFYd6QCZ7h9DBEReZE9Z2rR2u7E5PRo0VGIiFyKhalA3+2ZmcrC1OssGJuEkAAlXvvqDCSJjZCIiMg7bC2oRoxGjSHx4aKjEBG5FAtTgYovFqa9OZXX62iDVXhkTBIOltXjq+Ia0XGIiIjQ0NqOb0rMmNQ3BnIZe1MQUffCwlSgM6ZmRASpEBmsEh2FOjE7Mw6p0SF44+uzaGt3iI5DRER+Lqe4BnanhNv7cRovEXU/LEwFKq5pRu/oEMh41dMrKeUXGiEZmqx495sy0XGIiMjPbS+sRq+IIPSN0YiOQkTkcixMBXFKEs7WNLPxkZcbEh+OmQNj8eGh8zhtYiMkIiISw2Sx4lB5A25Pj+EFbSLqlliYClJR34Y2u5ONj3zA4+OSERagxLIdp+BkIyQiIhJgR5EJEoBJ7MZLRN0UC1NB2PjId4QHqfDErSk4XtWET/OqRMchIiI/tLWgGv30GiTpgkVHISJyCxamgpwxNUMGoHckBxhfcEe/GAzvpcWbe86hxmIVHYeIiPxImbkVBUYLJqfHiI5CROQ2LEwFKa5pRkJEEAJVCtFRqAtkMhmemZgKm92JP359VnQcIiLyI9sKqyEDMKkvp/ESUffFwlSQM2x85HN6RQThZ6N6YUeRCXvP1YmOQ0REfkCSJGwvrMaQ+HDoQwNExyEichsWpgK0tjtQbm5lYeqDHhiWgCRdEF75sph7mxIRkdudMjWjpK4VU9j0iIi6ORamApytbYEEoA8bH/kctVKOZyelorLRinf2lYqOQ0RE3dy2gmoo5DJkpbEwJaLujYWpAGcu7ofJO6a+aWi8FrMyLuxteryyUXQcIiLqppyShO1FJoxOioA2SCU6DhGRW7EwFaC4phmBSjl6agNFR6HrtHh8CqI1AVi6rYhTeomIyC2OVTTC2GTFFHbjJSI/wMJUgNM1zegdFQK5TCY6Cl0nTYASz09OQ0ldK1ZwSi8REbnBtsJqBCjlGNc7UnQUIiK3Y2HqYZIk4bSJHXm7g5FJEZg9KA4fHjyPYxUNouMQEVE3Ync48WWRCeN7RyJYza3liKj7Y2HqYdUWG+pb25EWoxEdhVxg0fhkxIYFYOm2U5zSS0RELnOgrB4NbXZM5jReIvITLEw9rNBoAQCk61mYdgchaiX+b3IaysyteGtvieg4RETUTWwrqEZYoBJjkiNERyEi8ggWph5WVN0EuQxI5VYx3caIxAjMyYzDmkMVnNJLREQ3rK3dgV2na3FbahRUCn5UIyL/wHc7Dys0WpCoC0aQiutFupNF41IQFxaAF7cWoZVTeomI6AbsOVuHlnYHpqRz71Ii8h8sTD2sqNqCdK4v7XaC1Qosub0vzte3Yfmus6LjEBGRD9teWI2oEDWGxmtFRyEi8hgWph5U22xDtcXG9aXd1E0JWtw/LB7rj1VhZ0G16DhEROSDmtrs2HuuDpP6RkMh57ZyROQ/WJh6UFH1hcZHfXnHtNt6dGwS+sZo8OyG46ixWEXHISIiH/NVcQ3aHRKm9GM3XiLyLyxMPYiFafenVsrx26npaG134MWtp+CUJNGRiIjIh2wtrEa8NhD9ObuKiPwMC1MPKjRakKANhCZAKToKuVFyZDCevT0d+0vNWHukUnQcIiLyETUWKw6V12NKegxkMk7jJSL/wsLUgwqMTegbEyo6BnnAj4Yn4JYUHf6y+ywKjU2i4xARkQ/YcaoGTgmYks5pvETkf1iYekhtsw1VjVYMjGNh6g9kMhmWTOmLiCAVnt1UAIvVLjoSERF5uS/yjUiLDkFyZLDoKEREHsfC1EPyDRfumg2IZWHqL7TBKiyb1g9VDW14aXsxJK43JSKiH3CutgUFRguyB+hFRyEiEoKFqYecNDRBLgP6spmBX8nsGY4FNyfjy1MmrD9WJToOERF5qS35RihkwGRO4yUiP8XC1ENOGprQOyoEQSqF6CjkYQ8Mj8eY5Aj88eszXG9KRESXcUoSviioxsikCESFqEXHISISgoWpB0iShAJDE/pzGq9fkstkePH2dEQEqfCrzwvQ0NouOhIREXmRw+UNMDZZkd2f03iJyH+xMPWAioY2NLTZub7Uj2mDVfjDjP4wWaz4v82FcDi53pSIiC7YnG9EiFqBcb0jRUchIhKGhakHnKxi4yMCBsaF4emsPthfasbbe0tExyEiIi/QanMg51QNJqRFIZDLfYjIj7Ew9YCThiYEKOVIiQoRHYUEu3NQHGZlxGJlbjlyTplExyEiIsF2FBjR0u7AVE7jJSI/x8LUA45XNaKfXgOlXCY6CnmBp7P6YGBcKF7YWoTTNc2i4xARkUAbj1UiNjQAQ+LDRUchIhLKbYVpVVUVHnjgAUydOhXZ2dl4//333fVSXq2t3YFCowWDenDAoQvUSjn+ML0/gtVK/OLTE6hrsYmOREREAtRYrPj36RpM7R8DuYwXr4nIv7mtMFUoFHjmmWewZcsWrF27Fv/85z9x+vRpd72c1yowWmB3SsjsGSY6CnmRmNAAvD5rAGpb2vHUhnxY7U7RkYiIyMO25FfDKQF3cBovEZH7CtOYmBgMGDAAAKDRaJCSkgKj0eiul/NaxyoaAACD4liY0qUGxIbixTv64nhVI367rQiSxE69RET+QpIkfHbCgJt6aZGkCxYdh4hIOKUnXuT8+fMoKChAZmbmFY9TKGTQar3/zVmhkHc5Z76pGSlRIUjygqm815LbW/hiZqDrueeMSER1qx1//LIYfXuE4/Hb+nggXee6+7n2Jr6YGfDd3ETeKK+yEaXmVjx6a2/RUYiIvILbC9Pm5mYsWrQIzz33HDQazRWPdTgk1Ne3uDvSDdNqg7uUU5IkHC41Y1zvSK/4vbqa25v4Ymbg2nLfMygWRZUN+HPOaYSr5JgxMNbN6TrnD+faW/hiZsD3ckdH+9cWXbt378ZLL70Ep9OJefPm4eGHH77k+ffeew/r1q2DQqGATqfDsmXL0LNnT0FpaeNxA4JVCtwxIBbtrew1QETk1q687e3tWLRoEaZPn47Jkye786W8UmldKxra7FxfSlckk8nw68lpGJmoxbLtp7D7TK3oSETkYxwOB5YuXYq///3v2Lx5MzZt2nRZX4d+/fph/fr1+PzzzzFlyhS8+uqrgtJSs82OL0+ZMCk9GiEBHpm8RkTk9dxWmEqShF//+tdISUnB/Pnz3fUyXi2vshEA2JGXrkqlkOMPM/ojLUaD5zYVdKxNJiLqiry8PCQmJiIhIQFqtRrZ2dnYuXPnJceMGjUKQUFBAIDBgwfDYDCIiEoAdhSa0NruxExBM2SIiLyR2wrTQ4cOYePGjdi/fz9mzpyJmTNnYteuXe56Oa90pKIB4YFKJOqCREchHxCiVmL57IHQhwbgyQ0ncYZ7nBJRFxmNRsTG/qfI0ev1V2w4+Mknn2DcuHGeiEad+OyEAcmRwRgY51/TzYmIrsRt80eGDRuGoqIid/14n3C4vB5DE7Tcm4y6LCJYjb/MycCDa45i4SfHseLuTPSK4IUNInKdjRs34sSJE1i9enWXju+OjQlFKq624HhVE569vS8iIkJ8Jvd/88XMgG/m9sXMgG/m9sXMgO/m7gwXNrhJZUMbKhutuPemeNFRyMf0CA/Em3MzsODjPCz4+BhW3J2JeC2LUyL6YXq9/pKpuUajEXr95Xtj7tu3D2+//TZWr14NtVrdpZ/d3RoTirZ63zko5TLclhyB+voWn8n933wxM+CbuX0xM+CbuX0xM+B7ua/UmNCtzY/82eHz9QCAmxK0gpOQL+odFYK/zsuA1e7Ego/zUNnQJjoSEXmxjIwMlJSUoLy8HDabDZs3b0ZWVtYlx+Tn52PJkiV46623EBkZKSipf7PZnfgivxrjekciIrhrFwaIiPwFC1M3OVR+YX1pSlT3uLVOnpcarcFf5w5Cs82BBR8fQ1Uji1Mi6pxSqcSSJUvw0EMPYerUqbjjjjuQmpqK5cuXdzRBeuWVV9DS0oLFixdj5syZePTRRwWn9j85xTUwt7bjzkFsekRE9H2cyusmXF9KrtBXr8GbczPw2Cd5+PlHx/Dm3Awk6Xixg4guN378eIwfP/6SxxYvXtzx55UrV3o4EX3fuqOV6BURhBGJEaKjEBF5Hd4xdYPv1pcOS+A2MXTj+seG4u27MtHucOLhj46h0NgkOhIREV2jomoL8iobMSczjhetiYg6wcLUDQ6WX1hfOjSe60vJNfrGaPDO3ZkIUMrx6Md5OHKe+5wSEfmST45WIkApx7QBlzelIiIiFqZucaDEjMgQNXpzfSm5UKIuGO/ek4moEDUeX38cO0+ZREciIqIuaGqzY2tBNW7vF4OwQJXoOEREXomFqYs5nBIOlJoxKlELGafqkIvFhgXi3XsykRatwTOfF+C9A2WQJEl0LCIiuoLPTxrQZndiXmYP0VGIiLwWC1MXK6y2oKHNjlFJOtFRqJuKCFbjrbsGYUp6NP727xK8uO0UbHan6FhERNQJpyRh/bEqZMSFoa9eIzoOEZHXYmHqYvtL6gAAIxK5vpTcJ0Apx2+npuPhMYnYfNKI/1mXh+omq+hYRET0Pd+W1qPM3Ip5Q+JERyEi8mosTF3sQIkZ6TEa6LhxNrmZTCbDz0cnYtm0fjhlsuD+Dw53XBghIiLvsPZIBSKCVJiQGi06ChGRV2Nh6kIWqx15VU0YlcT9ychzJvWNxqr7hkIXosKi9Sfw1t4SOJxcd0pEJNq52hbsOVuHOZlxUCv5kYuI6Er4LulCuaVmOJwSC1PyuKTIYKy8dwimDdDj/+0vw88/OoaSuhbRsYiI/NqHB88jQCnHXUPY9IiI6GpYmLrQrjO1CA9UIrNnuOgo5IcCVQosub0vfjs1HaXmFty36hBW5ZbDzrunREQeZ7JYsaXAiOkD9Ijg8h4ioqtiYeoidqeEvWfrMDZFB6Wc28SQOLf3i8Hanw7DmGQd/rLnHB5acxQFxibRsYiI/MpHhyvhcEq4b1i86ChERD6BhamLHKtoQEObHeN6R4qOQoSoEDVemdEfL2Wno7KhDT9ZfQQvbC2CycLOvURE7max2vGvvEpkpUYhXhskOg4RkU9Qig7QXew+UwuVQsb1peQ1ZDIZJqfHYEyyDu8dKMOawxXYWWTCA8Pjcc/QnggLVImOSETULW04boDF6sD9wxNERyEi8hm8Y+oCkiRh95laDEvQIkTNWp+8iyZAicfHpeDjnw7D2BQd3v2mDNPfycXyXWd5B5WIyMXaHU6sOXQeNyWEY0BsqOg4REQ+g1WUC5yuacb5+jbcz3Uk5MXitUF4eXp/FJsseD+3HP88dB5rj1RgcnoMpg/QIyuc082IiG7U1oJqVFtseG5ymugoREQ+hYWpC2wvNEEhA25LjRIdheiqUqM1+F12Pzw6NgmrD57H1oJqbD5pRMKOYkztF4OstCgk64Ihk7GJFxHRtWh3OPH3/WVIj9FgDJf2EBFdExamN0iSJGwvMmF4rwjo2A6efEi8NgjPTEzF4vEp+Kq4BluLTFixrxQr9pUiLiwAY5J1GJ2kQ0aPUP7dJiLqgs9OGFDZ0IZfzh7Ii3tERNeIhekNyjc0obKhDQ+O6iU6CtF1CVIpMLW/HveOSUZRuRl7z9Vh39k6bMk3Yv2xKgBAbGgA+seGIjU6BPHaIPQMD0RPbSC0QSrI+eGLiAht7Q78Y38ZBvUI491SIqLrwML0Bm0vMkEpl+G2PpzGS75PHxqA2YPiMHtQHGx2J04YGlFgsCDf0IRJZKT2AAAX3ElEQVR8YxNyimsuOV4hA0IDVdAGKREaoEKAUgalXA6lQgaVQg6lXAa57EKHYBkAmQyQ4cIfZMB/PXbhie++DlAqEBqgQGigCqEBCkSHBEAfFoDY0AAEqhQePy9ERFfzr7wqmCw2/HZqOu+WEhFdBxamN8DhlLCjyITRSREIDeSppO5FrZRjaLwWQ+O1HY+1tTtQ2diGivo2VDS0oa7FhoZWOxra2tHYZke7w4k2ux3tDgl2pxPtDglOSYIkARIASBKkC/+7+H8JAC57zGp3otnm6DRXRJAKKVHB6BMVgt5RIRjWOwpxQUoo5fwgSERitNgcWHmgHMN7aXFTgvbq30BERJdhNXUD9peYYbLY8NRtetFRiDwiUKVASmQIUiJD3P5aDqcEi9WOJqsdJosNhqY2GBqtqGhow9maZnx2woDWdiewoxjBKgUG9QjDkPhwjEzUol9sKKcYE5HHrD1SAXNrOxaMTRIdhYjIZ7EwvQEbTxgQEaTCLb0jRUch6nYUchnCg1QID1IhXhsEIPyS552ShMqGNpQ22bD3lAlHzjfgrb0leGsvEBmixi0pOozrHYmRiRFQK7llMxG5R2NbOz749jxuTtEho0eY6DhERD6Lhel1qmuxYfeZWtw9pAdUCn7oJfI0uUyGeG0QBiZFYmzChaK1vrUd35TUYffpOuwoMmHDcQPCApWY1Dcad/SLwaAeYVz7RUQu9e43ZWi22fE/NyeJjkJE5NNYmF6nLfnVcDglzMyIFR2FiC7SBqlwRz897uinR7vDidyyenyRb8Smkxc6DCdoAzE7swemDdBDG6QSHZeIfFxJbQvWHa3EzIxYpEZrRMchIvJpLEyvg9Mp4dO8KmTEhXpkrR0RXTuVQo6xyTqMTdah2WbHV8U12HjcgOW7zuLtvSWY1Dcadw3pgX76UNFRicgHSZKEN3adQaBSjke5tpSI6IaxML0Ou4pNKDO34ndT00VHIaIuCFErMW1ALKYNiMVpUzM+OVaJL/KrsemkEcN6afHj4fEYlRjBab5E1GVfn67FvnNmPDE+Bbpgteg4REQ+j4sjr8PKfaWI0agxIY17lxL5mj7RIXhmYio2PzISi8Ylo7SuBYvWn8B9HxzGFwVG2B1O0RGJyMs12+x4Lec0UqNDcPeQHqLjEBF1CyxMr9FpUzP2na3FvME9oGTTIyKfpQlQ4oHhCdj40AgsmZIGu1PCki1FuPMf32Lt4Qq0tXe+jyoR0Yq9pTBZbHh2Yio/CxARuQjfTa/R6kPnEaiS485BcaKjEJELqBRyTB8Yi49+chPeuHMAYsMC8NpXZzDz77lYlVuOZptddEQi8iLHKhrw0eEKzM6M4/YwREQuxDWm16Dc3Iqt+UbcPyoR4ezoSdStyGUy3JwSiZtTInH4fD3e21+Ov+w5h/e/LcfdQ3rg7iE9+e+eyM+1tTuwdNspxIYF4PFxyaLjEBF1KyxMr8E/DpRBqZDj4VuSAa5DI+q2hsZrMXSuFicNTVh5oAzvflOGDw9WYE5mHO4dFo+oEDY6IfJHb+45hzJzK96aNwghan6EIiJyJb6rdlGZuRVf5Btxz9CeiAkNRH19i+hIRORmA2JD8erMAThd04yVB8rw4aHz+PhoJWYMjMWPh8cjNixQdEQi8pDdZ2qx9kgl7h7SA8N6aUXHISLqdrjGtIv+9u9zUCnkeGB4gugoRORhfaJC8Lvsfvhk/nDcnh6DT/OqMOsf32Lp1iKU1vEiFVF3Z2yyYunWIvSN0WDRuBTRcYiIuiXeMe2Cg2X12HmqBo+MSeQUPiI/lhARhP+bkoaHRvfC6oPnseG4AZvzjZiQFo35IxOQGq0RHZGIXMxmd+LZz/NhczjxUnY61Epe0ycicgcWpldhd0p4/asziAsLwP3D4kXHISIvEBsWiKey+mD+yF7456EKrD9WiR1FJtySosPPRvXCwDh26iTqDiRJwis7T+N4VRNent4Pibpg0ZGIiLotXva7io+PVOB0TTMWj09BoEohOg4ReZHIEDUeH5eMz34+Ag+PSUReZSPm//Mo/mddHg6W1UOSJNERiegGrDlcgY0nDPjZyARMSIsWHYeIqFvjHdMrKKlrwd/+XYJbUnTISo0SHYeIvFRYoAo/H52Ie2/qiX8dq8Lqg+exYF0eBsaF4u4hPTEhLQoqBa8DEvmSHUUm/Onrs7i1TyQeHpMkOg4RUbfHT0o/wO6UsHRrEQKUcjw3KRUymUx0JCLyciFqJR4YnoCND43ALyf0QWObHc9vKcT0d3Pxzr4S1FisoiMSURccKDXjN18UYlCPMPx2ajoUcn4GICJyN94x/QFv/bsEx6ua8Lup6YjSBIiOQ0Q+JFClwLzBPTAnMw77S8z4+Egl3v2mDP/vQDkmpkXhgTHJSNMG8IIXkRf6tsyMX2w4iSRdMF6fNYDLeIiIPISFaSdyTpmw6ttyzMmMw5R+MaLjEJGPkstkGJOsw5hkHcrNrVh3tBKfnTBgW6EJCdpATB8Yi2kD9IjmxS8ir7D3bB1+9Xk+4rWB+OvcDIQHqURHIiLyGyxMv+ekoQkvbC1CRlwonry1t+g4RNRNJEQE4cnbeuN/bk7CN+cbsSa3DH/7dwlW7C3BmGQdpg+MxdhkHbeiIBJk00kDfrftFFKjNVg+ZyAigrk9HBGRJ7Ew/S/naluweP1xRASp8IcZ/fkBkYhcLlClwJ1DeuK25AiUmVvx2QkDNp80Ys/ZfIQGKHFbaiSmpMfgpgQt17UReYDdKeGve85h9cHzGN5Li1dn9keImh+PiIg8je+8F502NePx9cehVMjx13mDOLWOiNyuV0QQFt6SjEfHJuHbMjO2FVRj56kafHbCiMgQNSamRWFiWjQyeoSxSCVyA5PFit98UYRvy+oxb3AP/O+tKeygTUQkCAtTAIfP1+OpDfkIUsnx5zkZiNcGiY5ERH5EKZdhdJIOo5N0aGt3YO+5OmwtqMa/8qqw9kgldMEq3NI7Erf2icTwXhEI4GwOohsiSRJyimvw+x3FsNqdeH5KGmYMjBUdi4jIr/l1YSpJEv55qAJ/2XMO8eGBeHNuBmLDAkXHIiI/FqhSYEJaNCakRcNitWPfuTp8fboWXxaZsPG4AUEq+cUiNgKjkiL4nkV0jSoaWvF6zhnsOVuH9BgNfpudjiRdsOhYRER+z28L08qGNrz8ZTG+KTHj1j6R+M3tfaEJ8NvTQUReSBOgxOT0GExOj4HN7sTB8nrsOl2LPWdrkVNcAwBI1gVj5MUidWh8OIK4tQVRp8wtNqzMLce6o5VQyGRYPD4F9wztCSWnyRMReQW/q8QsVjvWHK7AqtxyyGTA01m9MW9wD+4nSEReTa2Ud2w984zUB2dqW3CgxIz9JWZ8mleFjw5XQCGXIT1Gg8E9wzEkPgyZPcOh5XYX5OfKza1Ye6QCG44bYLM7MX2gHg+PSYI+lL0kiIi8id8UpjUWK9Yfq8K6o5VoaLMjKzUK/3trCqfBEZHPkclk6BMVgj5RIbhvWDza2h04UtGAw+UNOFrRgI+PVuDDQ+cBAMmRweiv1yBdH4r0GA3SYjQIVvOuKnVv9S3t+Pp0DXYUmZBbVg+FXIY7+sXgJ8MTkBTJabtERN7IrYXp7t278dJLL8HpdGLevHl4+OGH3flylzE2WZFbasaXp0w4UGKGUwLGpujw0OhEDIgN9WgWIiJ3CVQpOponAYDV7kS+oQlHKxqQV9mI/aX12JxfDQCQAUjSBaNPdAiSdcFI1AUhOTIYCdogBHIasE+72phrs9nwy1/+EidPnoRWq8Ubb7yB+Ph4QWldy+6UcLKyAbsLq7HnTC0Ol9fDIQE9wgPxyJhEzMqIRRS77RMReTW3FaYOhwNLly7Fe++9B71ej7lz5yIrKwt9+vRx10sCAM7UNGPtkQocLKtHeX0bACAuLAD3DUvArIxYJESw4y4RdW8BSjmGxIdjSHx4x2MmixWFRgsKjRYUGJuQb2jCl0UmSBefl+HCe2WP8EDEhgUiNjQAsWEBiA0LRIwmABHBKoSFSZ2+HonXlTF33bp1CAsLw44dO7B582a89tpr+NOf/iQw9fVpa3fgXF0Lzta04GxtC4qqm3C8sgkt7Q4AQGJEEH4yIgFZqdFIiwnhUh0iIh/htsI0Ly8PiYmJSEhIAABkZ2dj586dbi9MN500YnuhCUPjwzF3cA/clKBFanQI5ByYiMiPRWsCEK0JwC29Izsea2t3oLy+FSV1rSipa0FpXQuqGi/MNKlptsH5vTpUIZchPFCJiGAVIoLV0KgVCAlQIkSlQEiAAsGqi1+rFQhUKaBWyKBSyKFWyKHq5M8qhQxymQwKmQwKuQxy2YXX4D6S164rY25OTg4WLlwIAJgyZQqWLl0KSZI8WrhJkgSHBLQ7nLDZnWh3Smh3OGG1O9Fsc6DZar/wf5sdFqsD5tZ21FisqLbYUGOxwWSxoqHN3vHzlHIZkiODMbV/DMakRSNVG8glOkREPspthanRaERs7H/2BNPr9cjLy3PXy3VYPD4Fi8Yl8wopEdFVBKoUSI3WIDVac9lzdocT1RYbqhrbYLLYYG5tR6tTQlVdC+pb21HX0o7yFhuarQ402xxosdnhcNEN1UfGJOKh0Ymu+WF+oitjrtFoRFxcHABAqVQiNDQUZrMZOp3O7fne3VeK978th83uxLX8NZHLgMgQNaI1AegZHojBPcMQExqARF0wUnTBiNcGQnnxQoZWG4z6+hb3/AJEROR2XtX8SKVSIDraN9Z++krO7/PF3L6YGfDN3L6YGfDN3L6QOQ5ApugQJJwrxubnZg7EczMHuijRD/OFf1ed8cXcvpgZ8M3cvpgZ8M3cvpgZ8N3c3+e2+VJ6vR4Gg6Hja6PRCL1e766XIyIi8ltdGXP1ej2qqqoAAHa7HU1NTYiIiPBoTiIioh/itsI0IyMDJSUlKC8vh81mw+bNm5GVleWulyMiIvJbXRlzs7Ky8OmnnwIAtm3bhlGjRnHZCxEReQ2ZJElua7O4a9cuLFu2DA6HA3PmzMGCBQvc9VJERER+rbMxd/ny5Rg4cCAmTJgAq9WKp59+GgUFBQgPD8cbb7zR0SyJiIhINLcWpkRERERERERXw578REREREREJBQLUyIiIiIiIhKKhWkX1NfXY/78+Zg8eTLmz5+PhoaGTo/r168fZs6ciZkzZ+LRRx/1cMoLdu/ejSlTpmDSpEl45513LnveZrPhiSeewKRJkzBv3jycP39eQMrLXS33v/71L4waNarj/K5bt05Ayks9++yzGD16NKZNm9bp85Ik4Xe/+x0mTZqE6dOn4+TJkx5OeLmrZT5w4ABuuummjvP85ptvejhh56qqqvDAAw9g6tSpyM7Oxvvvv3/ZMd52vruS2RvPt9Vqxdy5czFjxgxkZ2fjz3/+82XHeOv7CPkXjs3ux7HZMzg2ew7HZi8n0VX94Q9/kFasWCFJkiStWLFCeuWVVzo9bvDgwZ6MdRm73S5NmDBBKisrk6xWqzR9+nSpuLj4kmNWr14tPf/885IkSdKmTZukxYsXi4h6ia7kXr9+vfTiiy8KSti53Nxc6cSJE1J2dnanz3/99dfSgw8+KDmdTunIkSPS3LlzPZzwclfLvH//funhhx/2cKqrMxqN0okTJyRJkqSmpiZp8uTJl/0d8bbz3ZXM3ni+nU6nZLFYJEmSJJvNJs2dO1c6cuTIJcd44/sI+R+Oze7FsdlzODZ7Dsdm78Y7pl2wc+dOzJo1CwAwa9YsfPnll4ITdS4vLw+JiYlISEiAWq1GdnY2du7ceckxOTk5uPPOOwEAU6ZMwTfffANJcP+rruT2RsOHD0d4ePgPPv/d3xuZTIbBgwejsbER1dXVHkx4uatl9lYxMTEYMGAAAECj0SAlJQVGo/GSY7ztfHclszeSyWQICQkBcGGvS7vdftmWIt74PkL+h2Oze3Fs9hyOzZ7Dsdm7sTDtgtraWsTExAAAoqOjUVtb2+lxVqsVs2fPxl133SVkgDQajYiNje34Wq/XX/aPzWg0Ii4uDgCgVCoRGhoKs9ns0Zzf15XcALB9+3ZMnz4dixYt6tgk3pt9//eKjY31iTe/o0ePYsaMGXjooYdQXFwsOs5lzp8/j4KCAmRmZl7yuDef7x/KDHjn+XY4HJg5cybGjBmDMWPGdHquve19hPwPx2b34tjsXbxxrPhvHJvdzx/GZqXoAN7ipz/9KWpqai57/Iknnrjka5lM9oMbkn/11VfQ6/UoLy/HT37yE6SlpaFXr15uyetvbrvtNkybNg1qtRofffQRfvWrX2HVqlWiY3U7AwYMQE5ODkJCQrBr1y489thj2L59u+hYHZqbm7Fo0SI899xz0Gg0ouN0yZUye+v5VigU2LhxIxobG/HYY4/h1KlTSEtLEx2L/BDHZu/GsdkzvHWs+A7HZs/wh7GZd0wvWrlyJTZt2nTZfxMnTkRkZGTHtIPq6mrodLpOf4ZerwcAJCQkYMSIEcjPz/dY/u9e32AwdHxtNBo7Mv33Md9d0bTb7WhqakJERIRHc35fV3JHRERArVYDAObNmyd88XxXfP/3MhgMl/1e3kaj0XRMFRk/fjzsdjvq6uoEp7qgvb0dixYtwvTp0zF58uTLnvfG8321zN58vgEgLCwMI0eOxJ49ey553BvfR6h74tgsDsdm7+HNYwXHZs/rzmMzC9MuyMrKwoYNGwAAGzZswIQJEy47pqGhATabDQBQV1eHw4cPo0+fPh7NmZGRgZKSEpSXl8Nms2Hz5s3Iysq65JisrCx8+umnAIBt27Zh1KhRP3iV2VO6kvu/1yPk5OSgd+/eno55zb77eyNJEo4ePYrQ0NCOaWfeymQydaxHyMvLg9Pp9Io3NUmS8Otf/xopKSmYP39+p8d42/nuSmZvPN91dXVobGwEALS1tWHfvn1ISUm55BhvfB8h/8Ox2b04NnsPbxwrAI7NnuQvY7NM8rVVsQKYzWY88cQTqKqqQo8ePfCnP/0JWq0Wx48fx0cffYSXXnoJhw8fxm9+8xvIZDJIkoQf//jHmDdvnsez7tq1C8uWLYPD4cCcOXOwYMECLF++HAMHDsSECRNgtVrx9NNPo6CgAOHh4XjjjTeQkJDg8ZzXmvv1119HTk4OFAoFwsPD8cILLwgfAJ988knk5ubCbDYjMjISjz/+OOx2OwDgRz/6ESRJwtKlS7Fnzx4EBQVh2bJlyMjI8OrMq1evxpo1a6BQKBAYGIhnnnkGQ4cOFZoZAA4ePIj77rsPaWlpkMsvXE978sknUVlZCcA7z3dXMnvj+S4sLMQzzzwDh8MBSZJw++23Y+HChT7xPkL+hWOz+Nwcmz2T2RvHCoBjsyf5y9jMwpSIiIiIiIiE4lReIiIiIiIiEoqFKREREREREQnFwpSIiIiIiIiEYmFKREREREREQrEwJSIir/Pss89i9OjRmDZtWpeO37JlC6ZOnYrs7Gz84he/cHM6IiIi/+PusVl5owGJiIhcbfbs2bj//vvxq1/96qrHlpSU4J133sGaNWsQHh6O2tpaDyQkIiLyL+4em3nHlIiIvM7w4cMRHh5+yWNlZWV48MEHMXv2bNx77704c+YMAODjjz/Gfffd13F8ZGSkx/MSERF1d+4em3nHlIiIfMLzzz+PF198EUlJSTh27BhefPFFrFq1CiUlJQCAe+65B06nEwsXLsS4cePEhiUiIvIDrhybWZgSEZHXa25uxpEjR7B48eKOx2w2GwDA4XCgtLQUH3zwAQwGA+6//358/vnnCAsLExWXiIio23P12MzClIiIvJ4kSQgLC8PGjRsve06v1yMzMxMqlQoJCQlISkpCSUkJBg0aJCApERGRf3D12Mw1pkRE5PU0Gg3i4+PxxRdfALgwGBYWFgIAJk6ciNzcXABAXV0dSkpKkJCQICwrERGRP3D12CyTJElyb2QiIqJr8+STTyI3NxdmsxmRkZF4/PHHMWrUKLzwwgswmUyw2+2YOnUqFi5cCEmS8PLLL2PPnj1QKBR49NFHkZ2dLfpXICIi6lbcPTazMCUiIiIiIiKhOJWXiIiIiIiIhGJhSkREREREREKxMCUiIiIiIiKhWJgSERERERGRUCxMiYiIiIiISCgWpkRERERERCQUC1MiIiIiIiISioUpERERERERCfX/Ab+USalZgWULAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 1152x396 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "d1OV1yArk4Z7",
        "outputId": "1c129f38-1101-4a68-c944-c69bc0449c24"
      },
      "source": [
        "print(sorted(cnt_per_grp_lst))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[96, 2086, 3276, 3521, 4137, 4201, 4215, 5926, 6230, 6489, 6603, 7869, 8157, 14983, 15578, 22620, 28594, 29048, 38455, 49561, 54133, 54566, 54629, 62237, 68216, 78071, 108520, 109369, 110414, 111605, 112186, 122256, 123722, 128401, 140572, 143763, 147070, 175531, 179100, 203911, 216458, 237746, 243160, 248659, 256606, 273597, 355399, 370438, 385390, 399156, 425664, 464338, 539647, 543858, 549484, 562079, 581659, 612181, 640842, 641313, 731868, 737822, 784832, 803943, 872184, 872985, 916758, 934800, 1074255, 1081396, 1161086, 1248156, 1263131, 1472984, 1503879, 1744741, 1792696, 1920159, 2012449, 2155376, 2166176, 2522333]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zq5uUvGMk4fr"
      },
      "source": [
        "# Observations:\r\n",
        "    1. Almost 13 groups has less than 10k datapoints in it.\r\n",
        "    2. Almost 14 roups has more than 10 lakh datapoints in it.\r\n",
        "    \r\n",
        "# Conclusion:\r\n",
        "    1. We may have to remove groups with very small no. of data points in it."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3eduXQu32owJ"
      },
      "source": [
        "# 2. TASK-I : predictions per instance\r\n",
        "### 2.1. Pre-processing and data cleaning -"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vom57muzk4k5"
      },
      "source": [
        "# 1. create dataframe for each instance\r\n",
        "def create_df_instance(i):\r\n",
        "    \"\"\"\r\n",
        "    in this function 'i' takes folder name as a string, which contain mem.log file\r\n",
        "    and return pandas dataframe as output\r\n",
        "    \"\"\"\r\n",
        "    splt = i.split('_')\r\n",
        "    grp = splt[0] + '-' + splt[1]\r\n",
        "    instance = splt[2]\r\n",
        "    folder = 'group82_resource_utilization/'+i\r\n",
        "    f = open(folder + '/' + os.listdir(folder)[0], 'r')\r\n",
        "    \r\n",
        "    data_lst = []\r\n",
        "    for dp in f:                            # dp : datapoint\r\n",
        "        if len(dp)>0 and not re.findall('\\x00', dp):\r\n",
        "            dp_splt = dp.split('\"')\r\n",
        "            timestamp = [dp_splt[1]]\r\n",
        "            features = dp_splt[3].split(':')[:-1]\r\n",
        "            storage_space = re.findall('[0-9]*\\.?[0-9]+',dp_splt[3].split(':')[-1])    # \r\n",
        "            final_datapt = ([grp] + [instance] + timestamp + features + storage_space )\r\n",
        "            data_lst.append(final_datapt)\r\n",
        "    f.close()\r\n",
        "    instance_df = pd.DataFrame(data_lst, columns = ['group', 'instance', 'timestamp', 'Memory_Allocated', 'Memory_Used', 'CPU_Allocated', 'CPU_Used', 'Network_bandwidth_utilization', 'Storage_space_utilization'])\r\n",
        "    instance_df = instance_df.astype({'Memory_Allocated':int, 'Memory_Used':int,'CPU_Allocated': int, 'Memory_Used':int,\r\n",
        "                                      'CPU_Allocated':int, 'CPU_Used':float, 'Network_bandwidth_utilization': float, 'Storage_space_utilization': float})\r\n",
        "    \r\n",
        "    return instance_df"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "jjdN9Xg5k4qg",
        "outputId": "b232c1bd-25a9-4dab-e05f-bc1b4f62122f"
      },
      "source": [
        "# 2. random test - generate random idx to test above function\r\n",
        "idx = np.random.randint(0, len(dir_lst)) #1251\r\n",
        "print(idx)\r\n",
        "instance_df = create_df_instance(dir_lst[idx])\r\n",
        "instance_df.head(3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "3307\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "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>group</th>\n",
              "      <th>instance</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>Memory_Allocated</th>\n",
              "      <th>Memory_Used</th>\n",
              "      <th>CPU_Allocated</th>\n",
              "      <th>CPU_Used</th>\n",
              "      <th>Network_bandwidth_utilization</th>\n",
              "      <th>Storage_space_utilization</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>group-1</td>\n",
              "      <td>e8dc56b5-d902-4b5f-a8d8-fd3f0c1a62a4</td>\n",
              "      <td>Sat Jan 04 21:31:04 IST 2020</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4096.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>group-1</td>\n",
              "      <td>e8dc56b5-d902-4b5f-a8d8-fd3f0c1a62a4</td>\n",
              "      <td>Sat Jan 04 21:37:25 IST 2020</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4096.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>group-1</td>\n",
              "      <td>e8dc56b5-d902-4b5f-a8d8-fd3f0c1a62a4</td>\n",
              "      <td>Sat Jan 04 21:43:45 IST 2020</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4096.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "     group  ... Storage_space_utilization\n",
              "0  group-1  ...                    4096.0\n",
              "1  group-1  ...                    4096.0\n",
              "2  group-1  ...                    4096.0\n",
              "\n",
              "[3 rows x 9 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c1ZF_e19217R"
      },
      "source": [
        "### 2.2. EDA after data cleaning :"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 287
        },
        "id": "nuemSXkzlMi7",
        "outputId": "9e36070e-849f-41e5-8ba0-5e8442c7af01"
      },
      "source": [
        "# 3. Basic info about instance_df\r\n",
        "instance_df = create_df_instance(dir_lst[cnt_per_instance_lst.index(49468)])  # let's take any one instance with good no. of datapointss\r\n",
        "instance_df.describe()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>Memory_Allocated</th>\n",
              "      <th>Memory_Used</th>\n",
              "      <th>CPU_Allocated</th>\n",
              "      <th>CPU_Used</th>\n",
              "      <th>Network_bandwidth_utilization</th>\n",
              "      <th>Storage_space_utilization</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>49468.000000</td>\n",
              "      <td>49468.000000</td>\n",
              "      <td>49468.0</td>\n",
              "      <td>49468.000000</td>\n",
              "      <td>49468.0</td>\n",
              "      <td>49468.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>4811.814668</td>\n",
              "      <td>3868.292128</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.191130</td>\n",
              "      <td>0.0</td>\n",
              "      <td>150.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>4473.460860</td>\n",
              "      <td>320.254088</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.050532</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>4096.000000</td>\n",
              "      <td>1291.000000</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>150.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>4096.000000</td>\n",
              "      <td>3931.000000</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.170000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>150.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>4096.000000</td>\n",
              "      <td>3967.000000</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.190000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>150.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>4096.000000</td>\n",
              "      <td>3979.000000</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.210000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>150.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>32768.000000</td>\n",
              "      <td>3979.000000</td>\n",
              "      <td>4.0</td>\n",
              "      <td>1.450000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>150.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "       Memory_Allocated  ...  Storage_space_utilization\n",
              "count      49468.000000  ...                    49468.0\n",
              "mean        4811.814668  ...                      150.0\n",
              "std         4473.460860  ...                        0.0\n",
              "min         4096.000000  ...                      150.0\n",
              "25%         4096.000000  ...                      150.0\n",
              "50%         4096.000000  ...                      150.0\n",
              "75%         4096.000000  ...                      150.0\n",
              "max        32768.000000  ...                      150.0\n",
              "\n",
              "[8 rows x 6 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 209
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 714
        },
        "id": "TDmwHLhplMmK",
        "outputId": "953f0cb9-0e0a-4d18-a146-e5f61983f5df"
      },
      "source": [
        "# 4. plotting target numerical features\r\n",
        "target = instance_df.columns[[4, 6, 7, 8]]\r\n",
        "plt.figure(figsize = (16, 12))\r\n",
        "for col in range(len(target)):\r\n",
        "    plt.subplot(2, 2, col+1)\r\n",
        "    plt.plot(instance_df[target[col]])\r\n",
        "    plt.xlabel(target[col])\r\n",
        "    plt.ylabel('Usage')\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x864 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6wWY0Hf0lMp-"
      },
      "source": [
        "# Observations:\r\n",
        "    1. For most of the instances 'Network_bandwith_utilization' and 'Storage_space_utilization' are constants\r\n",
        "    \r\n",
        "# Conclusion:\r\n",
        "    1. For regression tasks, we need to have variance in target_feature. Hence, for this exercise I'll consider 'Memory_Used' and 'CPU_Used' as target feature.\r\n",
        "    2. Let's neglect 'Network_bandwith_utilization' and 'Storage_space_utilization' target features."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        },
        "id": "kl49wb2KlMuD",
        "outputId": "936d1564-cf4d-4f7e-bd7e-e1b9f208b2f7"
      },
      "source": [
        "# 5. Understand target feature from pdfs\r\n",
        "plt.figure(figsize = (16, 5.5))\r\n",
        "plt.subplot(1,2,1)\r\n",
        "plt.title('PDF (Memory_Used)')\r\n",
        "sns.distplot(instance_df['Memory_Used'], hist=False)\r\n",
        "\r\n",
        "plt.subplot(1,2,2)\r\n",
        "plt.title('PDF (CPU_Used)')\r\n",
        "sns.distplot(instance_df['CPU_Used'], hist=False)\r\n",
        "\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x396 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gjXUwVCnlM1Q"
      },
      "source": [
        "# 3. Feature transformations - Transforming data into model ready format\r\n",
        "\r\n",
        "### 3.1. Instance based model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5sLjnhbkoaN1",
        "outputId": "748972c2-5e03-4b5f-fea1-1b9eed1f7df6"
      },
      "source": [
        "# 1. making data model ready : basic featurization : lag features\r\n",
        "window = 15\r\n",
        "lstm_x = []\r\n",
        "dense_x = instance_df[['Memory_Allocated', 'CPU_Allocated']].values[window:]\r\n",
        "y = []\r\n",
        "data = instance_df[['Memory_Used', 'CPU_Used']].values\r\n",
        "for i in range(len(data) - window):\r\n",
        "    if i+window+1 <= len(data):\r\n",
        "        lstm_x.append(data[i:i+window])\r\n",
        "        y.append(data[i+window])\r\n",
        "lstm_x = np.array(lstm_x)\r\n",
        "y = np.array(y)\r\n",
        "lstm_x.shape, dense_x.shape, y.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((49453, 15, 2), (49453, 2), (49453, 2))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 212
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FtJFeyxnv2HH",
        "outputId": "15c0c6c9-69dd-43ca-cefc-fc7aaf9f11d6"
      },
      "source": [
        "# 2. Train test split : time based spliting\r\n",
        "train_set = 0.70\r\n",
        "lstm_x_train = lstm_x[:int(len(lstm_x) * train_set)]\r\n",
        "dense_x_train = dense_x[:int(len(dense_x) * train_set)]\r\n",
        "y_train = y[:int(len(dense_x) * train_set)]\r\n",
        "\r\n",
        "lstm_x_cv = lstm_x[int(len(lstm_x) * train_set):]\r\n",
        "dense_x_cv = dense_x[int(len(lstm_x) * train_set):]\r\n",
        "y_cv = y[int(len(lstm_x) * train_set):]\r\n",
        "\r\n",
        "lstm_x_train.shape, dense_x_train.shape, y_train.shape, lstm_x_cv.shape, dense_x_cv.shape, y_cv.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((34617, 15, 2),\n",
              " (34617, 2),\n",
              " (34617, 2),\n",
              " (14836, 15, 2),\n",
              " (14836, 2),\n",
              " (14836, 2))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 213
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oYVrFkbbyzuC"
      },
      "source": [
        "# 4. Modelling - Deep learning based model - LSTM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zaWI8Wc2lgaQ"
      },
      "source": [
        "# 1. import essential libraries\r\n",
        "import tensorflow as tf\r\n",
        "from tensorflow.keras.layers import Input, LSTM, GRU, Dense, RepeatVector, TimeDistributed, Concatenate, Bidirectional\r\n",
        "from tensorflow.keras import Model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fwm95GOalgfg",
        "outputId": "b6c920dc-0772-4344-b056-5905e2acc237"
      },
      "source": [
        "# 2. building a model\r\n",
        "tf.keras.backend.clear_session()\r\n",
        "\r\n",
        "dense_ins = Input(name = 'constants', shape = (2,))\r\n",
        "\r\n",
        "lstm_ins = Input(name = 'Sequence', shape = (lstm_x.shape[1:]))\r\n",
        "lstm_out = LSTM(name = 'LSTM', units = 128, kernel_initializer = 'glorot_normal')(lstm_ins)\r\n",
        "\r\n",
        "concat = Concatenate(axis = -1)([dense_ins, lstm_out])\r\n",
        "\r\n",
        "dense_1 = Dense(units = 30, activation = 'relu')(concat)\r\n",
        "out = Dense(units = 2)(dense_1)\r\n",
        "model = Model(inputs = [dense_ins, lstm_ins], outputs = out)\r\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "Sequence (InputLayer)           [(None, 15, 2)]      0                                            \n",
            "__________________________________________________________________________________________________\n",
            "constants (InputLayer)          [(None, 2)]          0                                            \n",
            "__________________________________________________________________________________________________\n",
            "LSTM (LSTM)                     (None, 128)          67072       Sequence[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 130)          0           constants[0][0]                  \n",
            "                                                                 LSTM[0][0]                       \n",
            "__________________________________________________________________________________________________\n",
            "dense (Dense)                   (None, 30)           3930        concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "dense_1 (Dense)                 (None, 2)            62          dense[0][0]                      \n",
            "==================================================================================================\n",
            "Total params: 71,064\n",
            "Trainable params: 71,064\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Mj1wCEwS0Rzo"
      },
      "source": [
        "# 3. tensorflow-keras callbacks\r\n",
        "# 3.1 To save best models\r\n",
        "!rm -r models\r\n",
        "if not os.path.isdir('/content/models'):\r\n",
        "  os.mkdir('/content/models')\r\n",
        "save_model = tf.keras.callbacks.ModelCheckpoint(filepath = '/content/models/weights-{epoch:02d}-{val_root_mean_squared_error:.2f}.hdf5',\r\n",
        "                                                monitor='val_root_mean_squared_error', verbose=1,\r\n",
        "                                                save_best_only=True, save_weights_only=True)\r\n",
        "\r\n",
        "# 3.2 Reduce learning rate if model is not learning\r\n",
        "reduce_lr = tf.keras.callbacks.ReduceLROnPlateau(monitor = 'val_root_mean_squared_error', factor = np.sqrt(0.1), patience = 4, verbose = 1, min_lr = 10**-6)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uY_1QgQ7lglw",
        "outputId": "e3760459-b6dd-4036-8177-46f89092cd45"
      },
      "source": [
        "# 4. training and evaluating model on cross validation\r\n",
        "rmse = tf.keras.metrics.RootMeanSquaredError(name=\"root_mean_squared_error\", dtype=None)\r\n",
        "opt = tf.keras.optimizers.Adam(learning_rate = 0.001)\r\n",
        "\r\n",
        "model.compile(loss = 'mean_squared_error', optimizer = opt, metrics = [rmse])\r\n",
        "hist = model.fit([dense_x_train, lstm_x_train], y_train, batch_size = 128, epochs = 40,\r\n",
        "                 validation_data = ([dense_x_cv, lstm_x_cv], y_cv),\r\n",
        "                 callbacks = [save_model, reduce_lr]) "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/40\n",
            "271/271 [==============================] - 3s 6ms/step - loss: 4737778.6250 - root_mean_squared_error: 2175.8926 - val_loss: 3216818.0000 - val_root_mean_squared_error: 1793.5490\n",
            "\n",
            "Epoch 00001: val_root_mean_squared_error improved from inf to 1793.54895, saving model to /content/models/weights-01-1793.55.hdf5\n",
            "Epoch 2/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 4422735.6057 - root_mean_squared_error: 2102.9951 - val_loss: 3293271.0000 - val_root_mean_squared_error: 1814.7371\n",
            "\n",
            "Epoch 00002: val_root_mean_squared_error did not improve from 1793.54895\n",
            "Epoch 3/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 4354917.5340 - root_mean_squared_error: 2086.8079 - val_loss: 2986780.2500 - val_root_mean_squared_error: 1728.2305\n",
            "\n",
            "Epoch 00003: val_root_mean_squared_error improved from 1793.54895 to 1728.23047, saving model to /content/models/weights-03-1728.23.hdf5\n",
            "Epoch 4/40\n",
            "271/271 [==============================] - 2s 6ms/step - loss: 4230296.5993 - root_mean_squared_error: 2056.6683 - val_loss: 2958804.0000 - val_root_mean_squared_error: 1720.1174\n",
            "\n",
            "Epoch 00004: val_root_mean_squared_error improved from 1728.23047 to 1720.11743, saving model to /content/models/weights-04-1720.12.hdf5\n",
            "Epoch 5/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 4157899.8566 - root_mean_squared_error: 2038.8270 - val_loss: 3062253.0000 - val_root_mean_squared_error: 1749.9294\n",
            "\n",
            "Epoch 00005: val_root_mean_squared_error did not improve from 1720.11743\n",
            "Epoch 6/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 3961441.9164 - root_mean_squared_error: 1990.1617 - val_loss: 2697171.5000 - val_root_mean_squared_error: 1642.3068\n",
            "\n",
            "Epoch 00006: val_root_mean_squared_error improved from 1720.11743 to 1642.30676, saving model to /content/models/weights-06-1642.31.hdf5\n",
            "Epoch 7/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 3660914.5193 - root_mean_squared_error: 1913.2695 - val_loss: 2808646.5000 - val_root_mean_squared_error: 1675.9017\n",
            "\n",
            "Epoch 00007: val_root_mean_squared_error did not improve from 1642.30676\n",
            "Epoch 8/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 3300697.1489 - root_mean_squared_error: 1816.7017 - val_loss: 2239328.2500 - val_root_mean_squared_error: 1496.4385\n",
            "\n",
            "Epoch 00008: val_root_mean_squared_error improved from 1642.30676 to 1496.43848, saving model to /content/models/weights-08-1496.44.hdf5\n",
            "Epoch 9/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 2881752.5046 - root_mean_squared_error: 1697.4441 - val_loss: 2226759.2500 - val_root_mean_squared_error: 1492.2330\n",
            "\n",
            "Epoch 00009: val_root_mean_squared_error improved from 1496.43848 to 1492.23303, saving model to /content/models/weights-09-1492.23.hdf5\n",
            "Epoch 10/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 2451129.5092 - root_mean_squared_error: 1565.4338 - val_loss: 1525454.0000 - val_root_mean_squared_error: 1235.0927\n",
            "\n",
            "Epoch 00010: val_root_mean_squared_error improved from 1492.23303 to 1235.09265, saving model to /content/models/weights-10-1235.09.hdf5\n",
            "Epoch 11/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 1969934.1213 - root_mean_squared_error: 1403.2452 - val_loss: 1290135.6250 - val_root_mean_squared_error: 1135.8414\n",
            "\n",
            "Epoch 00011: val_root_mean_squared_error improved from 1235.09265 to 1135.84143, saving model to /content/models/weights-11-1135.84.hdf5\n",
            "Epoch 12/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 1476964.7629 - root_mean_squared_error: 1214.9472 - val_loss: 920586.1250 - val_root_mean_squared_error: 959.4718\n",
            "\n",
            "Epoch 00012: val_root_mean_squared_error improved from 1135.84143 to 959.47180, saving model to /content/models/weights-12-959.47.hdf5\n",
            "Epoch 13/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 1092272.1477 - root_mean_squared_error: 1044.7993 - val_loss: 555146.3750 - val_root_mean_squared_error: 745.0815\n",
            "\n",
            "Epoch 00013: val_root_mean_squared_error improved from 959.47180 to 745.08148, saving model to /content/models/weights-13-745.08.hdf5\n",
            "Epoch 14/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 727297.0634 - root_mean_squared_error: 852.3220 - val_loss: 377416.0625 - val_root_mean_squared_error: 614.3420\n",
            "\n",
            "Epoch 00014: val_root_mean_squared_error improved from 745.08148 to 614.34198, saving model to /content/models/weights-14-614.34.hdf5\n",
            "Epoch 15/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 448110.5457 - root_mean_squared_error: 668.8598 - val_loss: 249808.1562 - val_root_mean_squared_error: 499.8081\n",
            "\n",
            "Epoch 00015: val_root_mean_squared_error improved from 614.34198 to 499.80807, saving model to /content/models/weights-15-499.81.hdf5\n",
            "Epoch 16/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 254318.9627 - root_mean_squared_error: 503.6228 - val_loss: 127733.4531 - val_root_mean_squared_error: 357.3982\n",
            "\n",
            "Epoch 00016: val_root_mean_squared_error improved from 499.80807 to 357.39816, saving model to /content/models/weights-16-357.40.hdf5\n",
            "Epoch 17/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 128397.7925 - root_mean_squared_error: 357.9693 - val_loss: 63287.9727 - val_root_mean_squared_error: 251.5710\n",
            "\n",
            "Epoch 00017: val_root_mean_squared_error improved from 357.39816 to 251.57101, saving model to /content/models/weights-17-251.57.hdf5\n",
            "Epoch 18/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 63689.8951 - root_mean_squared_error: 252.1178 - val_loss: 29701.0098 - val_root_mean_squared_error: 172.3398\n",
            "\n",
            "Epoch 00018: val_root_mean_squared_error improved from 251.57101 to 172.33981, saving model to /content/models/weights-18-172.34.hdf5\n",
            "Epoch 19/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 32989.0406 - root_mean_squared_error: 181.5171 - val_loss: 19415.9434 - val_root_mean_squared_error: 139.3411\n",
            "\n",
            "Epoch 00019: val_root_mean_squared_error improved from 172.33981 to 139.34111, saving model to /content/models/weights-19-139.34.hdf5\n",
            "Epoch 20/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 22069.0571 - root_mean_squared_error: 148.5232 - val_loss: 12758.2031 - val_root_mean_squared_error: 112.9522\n",
            "\n",
            "Epoch 00020: val_root_mean_squared_error improved from 139.34111 to 112.95222, saving model to /content/models/weights-20-112.95.hdf5\n",
            "Epoch 21/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 18612.0266 - root_mean_squared_error: 136.3553 - val_loss: 11550.6982 - val_root_mean_squared_error: 107.4742\n",
            "\n",
            "Epoch 00021: val_root_mean_squared_error improved from 112.95222 to 107.47418, saving model to /content/models/weights-21-107.47.hdf5\n",
            "Epoch 22/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 16988.4492 - root_mean_squared_error: 130.3313 - val_loss: 10680.5205 - val_root_mean_squared_error: 103.3466\n",
            "\n",
            "Epoch 00022: val_root_mean_squared_error improved from 107.47418 to 103.34660, saving model to /content/models/weights-22-103.35.hdf5\n",
            "Epoch 23/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 17371.6001 - root_mean_squared_error: 131.7863 - val_loss: 9843.1416 - val_root_mean_squared_error: 99.2126\n",
            "\n",
            "Epoch 00023: val_root_mean_squared_error improved from 103.34660 to 99.21262, saving model to /content/models/weights-23-99.21.hdf5\n",
            "Epoch 24/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 16881.3732 - root_mean_squared_error: 129.9090 - val_loss: 10887.2939 - val_root_mean_squared_error: 104.3422\n",
            "\n",
            "Epoch 00024: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 25/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 17473.8385 - root_mean_squared_error: 132.1606 - val_loss: 10114.1387 - val_root_mean_squared_error: 100.5691\n",
            "\n",
            "Epoch 00025: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 26/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 16914.0928 - root_mean_squared_error: 130.0387 - val_loss: 10723.4814 - val_root_mean_squared_error: 103.5542\n",
            "\n",
            "Epoch 00026: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 27/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 16775.2191 - root_mean_squared_error: 129.4862 - val_loss: 9872.7119 - val_root_mean_squared_error: 99.3615\n",
            "\n",
            "Epoch 00027: val_root_mean_squared_error did not improve from 99.21262\n",
            "\n",
            "Epoch 00027: ReduceLROnPlateau reducing learning rate to 0.00031622778103685084.\n",
            "Epoch 28/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 17030.6776 - root_mean_squared_error: 130.4905 - val_loss: 10011.9346 - val_root_mean_squared_error: 100.0597\n",
            "\n",
            "Epoch 00028: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 29/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 17640.0318 - root_mean_squared_error: 132.7853 - val_loss: 10122.8252 - val_root_mean_squared_error: 100.6123\n",
            "\n",
            "Epoch 00029: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 30/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 18056.3656 - root_mean_squared_error: 134.3531 - val_loss: 10014.6377 - val_root_mean_squared_error: 100.0732\n",
            "\n",
            "Epoch 00030: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 31/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 16784.6754 - root_mean_squared_error: 129.5186 - val_loss: 10037.8838 - val_root_mean_squared_error: 100.1892\n",
            "\n",
            "Epoch 00031: val_root_mean_squared_error did not improve from 99.21262\n",
            "\n",
            "Epoch 00031: ReduceLROnPlateau reducing learning rate to 0.00010000000639606199.\n",
            "Epoch 32/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 16493.2746 - root_mean_squared_error: 128.3917 - val_loss: 10283.9785 - val_root_mean_squared_error: 101.4100\n",
            "\n",
            "Epoch 00032: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 33/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 16975.9861 - root_mean_squared_error: 130.2576 - val_loss: 10075.5078 - val_root_mean_squared_error: 100.3768\n",
            "\n",
            "Epoch 00033: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 34/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 17155.9948 - root_mean_squared_error: 130.9505 - val_loss: 10267.4824 - val_root_mean_squared_error: 101.3286\n",
            "\n",
            "Epoch 00034: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 35/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 17403.2395 - root_mean_squared_error: 131.9081 - val_loss: 10680.4199 - val_root_mean_squared_error: 103.3461\n",
            "\n",
            "Epoch 00035: val_root_mean_squared_error did not improve from 99.21262\n",
            "\n",
            "Epoch 00035: ReduceLROnPlateau reducing learning rate to 3.1622778103685084e-05.\n",
            "Epoch 36/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 17223.3312 - root_mean_squared_error: 131.2052 - val_loss: 10091.2793 - val_root_mean_squared_error: 100.4554\n",
            "\n",
            "Epoch 00036: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 37/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 16820.3487 - root_mean_squared_error: 129.6879 - val_loss: 10101.2227 - val_root_mean_squared_error: 100.5048\n",
            "\n",
            "Epoch 00037: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 38/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 17781.8012 - root_mean_squared_error: 133.3305 - val_loss: 10220.8379 - val_root_mean_squared_error: 101.0982\n",
            "\n",
            "Epoch 00038: val_root_mean_squared_error did not improve from 99.21262\n",
            "Epoch 39/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 17284.4048 - root_mean_squared_error: 131.4490 - val_loss: 10167.7393 - val_root_mean_squared_error: 100.8352\n",
            "\n",
            "Epoch 00039: val_root_mean_squared_error did not improve from 99.21262\n",
            "\n",
            "Epoch 00039: ReduceLROnPlateau reducing learning rate to 1.0000000409520217e-05.\n",
            "Epoch 40/40\n",
            "271/271 [==============================] - 1s 5ms/step - loss: 16831.5801 - root_mean_squared_error: 129.6882 - val_loss: 10166.5654 - val_root_mean_squared_error: 100.8294\n",
            "\n",
            "Epoch 00040: val_root_mean_squared_error did not improve from 99.21262\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 372
        },
        "id": "npQbm1jflgyg",
        "outputId": "8e165ab6-5ad1-43a1-d0d8-16d45d786aec"
      },
      "source": [
        "# 5. Analysing model's performance using train-cv (loss and rmse) plots\r\n",
        "plt.figure(figsize = (16, 5.5))\r\n",
        "plt.subplot(1,2,1)\r\n",
        "plt.plot(hist.history['loss'], label = 'train_loss')\r\n",
        "plt.plot(hist.history['val_loss'], label = 'cv_loss')\r\n",
        "plt.xlabel('epoch')\r\n",
        "plt.ylabel('loss')\r\n",
        "plt.legend()\r\n",
        "\r\n",
        "plt.subplot(1,2,2)\r\n",
        "plt.plot(hist.history['root_mean_squared_error'], label = 'train_rmse')\r\n",
        "plt.plot(hist.history['val_root_mean_squared_error'], label = 'cv_rmse')\r\n",
        "plt.xlabel('epoch')\r\n",
        "plt.ylabel('rmse')\r\n",
        "plt.legend()\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x396 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YhcwZFLUlg9o",
        "outputId": "4f22b7f1-04f2-4906-8ab6-3514b16b9dfb"
      },
      "source": [
        "# 6. find best model - based on (metric we care about) best val_rmse \r\n",
        "print('epoch :', np.argmin(hist.history['val_root_mean_squared_error']), ' | ',\\\r\n",
        "      'val_rmse :', hist.history['val_root_mean_squared_error'][np.argmin(hist.history['val_root_mean_squared_error'])])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "epoch : 22  |  val_rmse : 99.21261596679688\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZWIH-J7RlhEV"
      },
      "source": [
        "# 7. load best model\r\n",
        "model.load_weights('/content/models/weights-23-99.21.hdf5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CD7N0tGVBOWX",
        "outputId": "53955bcb-beaf-4f10-dced-4ee8d04b8263"
      },
      "source": [
        "# 8. Predictions\r\n",
        "idx = 145\r\n",
        "x_test = [dense_x_cv[idx].reshape(1, -1), lstm_x_cv[idx].reshape(1, 15, 2)]\r\n",
        "y_pred = model.predict(x_test)\r\n",
        "y_true = y_cv[idx] \r\n",
        "\r\n",
        "print('y_pred :', y_pred, ' | y_true :', y_true)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "y_pred : [[3.9101667e+03 3.8252997e-01]]  | y_true : [3.979e+03 2.200e-01]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o6ykO2ZPeHtH"
      },
      "source": [
        "tensorflow.keras.utils.plot_model(model, show_shapes = True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "12oZ8cznDGtT"
      },
      "source": [
        "# 5. Error Analysis (Future improvements)\r\n",
        "\r\n",
        "### Observations:\r\n",
        "    1. 'Memory_Used' target feature has real (integer) values greater than 512 i.e., very large values \r\n",
        "    2. 'CPU_Used' target feature has real (float) values ranging between 0 to some small values (e.g. 20) i.e., it has very small float values as compare to 'Memory_used' feature\r\n",
        "\r\n",
        "### Conclusion:\r\n",
        "    1. Better we train 2 different models for 2 different target features.\r\n",
        "    2. Because difference between both target feature is very huge, we are not able to make clear sense that on which feature, model is performing well.\r\n",
        "    3. We need to employ baseline model for better comparison i.e., Train simple ML models like linear regression and some complex models like Random forest, GBDT etc. to compare performance of models with Deep learning models.\r\n",
        "    4. We need to experiment with more time series featurization techiques e.g. \r\n",
        "        i. Fourrier transform\r\n",
        "        ii. Simple moving average\r\n",
        "        iii. Exponential weighted moving avg\r\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hZKNtGakf6yy"
      },
      "source": [
        "# 6. TASK - II : Predictions per group\r\n",
        "### 6.1. Pre-processing and data cleaning -"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "07rh1kxhhnNc"
      },
      "source": [
        "# 1. making group data model ready : basic featurization : lag features\r\n",
        "grp_lstm_x_train = []\r\n",
        "grp_dense_x_train = []\r\n",
        "grp_y_train = []\r\n",
        "grp_lstm_x_cv = []\r\n",
        "grp_dense_x_cv = []\r\n",
        "grp_y_cv = []\r\n",
        "for i in path_lst[0]:\r\n",
        "    instance_df = create_df_instance(i)\r\n",
        "    window = 15\r\n",
        "    lstm_x = []\r\n",
        "    dense_x = instance_df[['instance', 'Memory_Allocated', 'CPU_Allocated']].values[window:]\r\n",
        "    y = []\r\n",
        "    data = instance_df[['Memory_Used', 'CPU_Used']].values\r\n",
        "    for i in range(len(data) - window):\r\n",
        "        if i+window+1 <= len(data):\r\n",
        "            lstm_x.append(data[i:i+window])\r\n",
        "            y.append(data[i+window])\r\n",
        "    lstm_x = np.array(lstm_x)\r\n",
        "    y = np.array(y)\r\n",
        "\r\n",
        "# 2. Train test split : time based spliting\r\n",
        "    train_set = 0.70\r\n",
        "    lstm_x_train = lstm_x[:int(len(lstm_x) * train_set)]\r\n",
        "    dense_x_train = dense_x[:int(len(dense_x) * train_set)]\r\n",
        "    y_train = y[:int(len(dense_x) * train_set)]\r\n",
        "\r\n",
        "    lstm_x_cv = lstm_x[int(len(lstm_x) * train_set):]\r\n",
        "    dense_x_cv = dense_x[int(len(lstm_x) * train_set):]\r\n",
        "    y_cv = y[int(len(lstm_x) * train_set):]\r\n",
        "\r\n",
        "    # collecting data in one list\r\n",
        "    grp_lstm_x_train.append(lstm_x_train)\r\n",
        "    grp_dense_x_train.append(dense_x_train)\r\n",
        "    grp_y_train.append(y_train)\r\n",
        "    grp_lstm_x_cv.append(lstm_x_cv)\r\n",
        "    grp_dense_x_cv.append(dense_x_cv)\r\n",
        "    grp_y_cv.append(y_cv)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Lqvw-ARWqWWB",
        "outputId": "849410f4-ab35-469a-e339-0de6dbde02bb"
      },
      "source": [
        "# 3. Creating final group data\r\n",
        "grp_lstm_x_train = np.concatenate(grp_lstm_x_train, axis = 0)\r\n",
        "grp_dense_x_train = np.concatenate(grp_dense_x_train, axis = 0)\r\n",
        "grp_y_train = np.concatenate(grp_y_train, axis = 0)\r\n",
        "grp_lstm_x_cv = np.concatenate(grp_lstm_x_cv, axis = 0)\r\n",
        "grp_dense_x_cv = np.concatenate(grp_dense_x_cv, axis = 0)\r\n",
        "grp_y_cv = np.concatenate(grp_y_cv, axis = 0)\r\n",
        "grp_lstm_x_train.shape, grp_dense_x_train.shape, grp_y_train.shape, grp_lstm_x_cv.shape, grp_dense_x_cv.shape, grp_y_cv.shape, "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((448130, 15, 2),\n",
              " (448130, 3),\n",
              " (448130, 2),\n",
              " (192082, 15, 2),\n",
              " (192082, 3),\n",
              " (192082, 2))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 143
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D-PhNZhIOid-"
      },
      "source": [
        "### 6.2. - EDA After data cleaning"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 700
        },
        "id": "2QHUrUpsOnBs",
        "outputId": "094c06b7-f1aa-4c2a-ae9a-b00552de495c"
      },
      "source": [
        "# 1. plotting\r\n",
        "plt.figure(figsize = (16, 12))\r\n",
        "plt.subplot(2, 2 , 1)\r\n",
        "plt.plot(np.concatenate((grp_y_train[:, 0], grp_y_cv[:, 0])))\r\n",
        "plt.ylabel('Memory_used')\r\n",
        "\r\n",
        "plt.subplot(2, 2 , 2)\r\n",
        "plt.plot(np.concatenate((grp_y_train[:, 1], grp_y_cv[:, 1])))\r\n",
        "plt.ylabel('CPU_usage')\r\n",
        "\r\n",
        "plt.subplot(2, 2 , 3)\r\n",
        "plt.plot(np.concatenate((grp_dense_x_train[:, 1], grp_dense_x_cv[:, 1])))\r\n",
        "plt.ylabel('Memory_Allocated')\r\n",
        "\r\n",
        "plt.subplot(2, 2 , 4)\r\n",
        "plt.plot(np.concatenate((grp_dense_x_train[:, 2], grp_dense_x_cv[:, 2])))\r\n",
        "plt.ylabel('CPU_Allocated')\r\n",
        "plt.show()\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x864 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "axOus9_rT5-P",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 728
        },
        "outputId": "0c007f1a-6167-4de2-f0d0-7d07ef8c5534"
      },
      "source": [
        "# 2. Understancding outliers in 'Memory_used' with PDF\r\n",
        "plt.figure(figsize = (16, 12))\r\n",
        "plt.subplot(2, 2 , 1)\r\n",
        "sns.distplot(np.concatenate((grp_y_train[:, 0], grp_y_cv[:, 0])), hist = False)\r\n",
        "plt.title('Memory_used')\r\n",
        "\r\n",
        "plt.subplot(2, 2 , 2)\r\n",
        "sns.distplot(np.concatenate((grp_y_train[:, 1], grp_y_cv[:, 1])), hist = False)\r\n",
        "plt.title('CPU_usage')\r\n",
        "\r\n",
        "plt.subplot(2, 2 , 3)\r\n",
        "sns.distplot(np.concatenate((grp_dense_x_train[:, 1], grp_dense_x_cv[:, 1])), hist = False)\r\n",
        "plt.title('Memory_Allocated')\r\n",
        "\r\n",
        "plt.subplot(2, 2 , 4)\r\n",
        "sns.distplot(np.concatenate((grp_dense_x_train[:, 2], grp_dense_x_cv[:, 2])), hist = False)\r\n",
        "plt.title('CPU_Allocated')\r\n",
        "plt.show()\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7AAAALICAYAAABGuxzaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeZgcdbX/8U9Vdffsk8mErBBkSwQTDfBLiCyGRQkBQgbCepELFzeUwIgQBETELXAVxPsgIIrKcvV6lUtMoqjIkgVZgpgoSFiULQmQdWaSzNrdVfX7o7t6emZ6tqS7qot+v57Hh0xPT9VJi37n1Dnf8zVc13UFAAAAAECRM4MOAAAAAACAoSCBBQAAAACEAgksAAAAACAUSGABAAAAAKFAAgsAAAAACAUSWAAAAABAKJDAAvDVD37wAy1cuDDoMAAAABBCJLDAbjjhhBM0depUNTU19Xj99NNP1wc/+EFt3LgxoMgAAMBAfvvb32r+/Pk67LDDdMwxx+gzn/mMnn/+ef3gBz/QlClTdNhhh2n69Ok677zztHbtWkn9P3z94Ac/qLffftvvvwJQ0khggd2099576+GHH858/eqrr6qjoyPAiFKSyWTQIQAAUJTuvfde3XTTTfr85z+vp556SsuXL9f555+vxx9/XJJ08skna+3atXrmmWd0+OGH6/LLL5frugFHDSAbCSywmxoaGrRkyZLM10uWLNHpp5+e+Toej+s73/mOjjvuOB111FH62te+ps7OTknS6tWrNWvWLN1zzz068sgjdcwxx+ixxx7TypUrddJJJ+mII47Q3Xff3eNaixYt0jHHHKNjjjlGixYtUjwe73GtH//4xzr66KN13XXXae7cuXriiScyP59IJDRz5kytW7eu37+Pd51sJ5xwgp5++mlJ0gsvvKD58+fr8MMP11FHHaWbb745876//e1vOu+88zR9+nTNmzdPq1evznxvw4YNuuCCC3TYYYfp4osvVnNz87A+ZwAA8mHXrl26/fbb9bWvfU2zZ89WZWWlotGoTjjhBF1zzTU93huNRnXGGWdo69ate7xuZa+lUs9qbldXlxYuXKiZM2dq+vTpOvPMM7Vt2zZJ0kMPPaSTTz5Zhx12mD7+8Y/rf//3f3tc95577sn8XvDggw/2qAYP9DsIEHYksMBuOvTQQ9Xa2qrXX39dtm3r4Ycf1rx58zLfv/XWW/Xmm29qyZIl+tOf/qQtW7bozjvvzHx/27Zt6urq0qpVq9TY2KivfvWrWrZsmR566CH94he/0F133aUNGzZIkn74wx/q73//u5YuXaply5bpxRdf1F133dXjWjt27NDy5cv1rW99Sw0NDVq2bFnm+ytXrtSYMWP0oQ99aLf/vosWLdKFF16oNWvW6NFHH9XJJ58sSdq8ebMuueQSfeELX9Bzzz2na665Ro2NjZn26oULF2rKlClavXq1Lr30Uv3mN7/Z7RgAANhda9euVVdXl0488cRB3xuPx7V48WKNHz9e9fX1BYvpN7/5jVpbW7VixQqtXr1a3/jGN1ReXi5JGjVqlH70ox9pzZo1uvnmm3XzzTfrpZdekiStWrVK9913n+699149+uijPR4cS4P/DgKEWSgT2Ouuu05HHnmk5s6dm5frHXLIIWpoaFBDQ4M+//nP5+WaKA1eFfapp57SgQceqLFjx0qSXNfVr3/9a33lK19RXV2dqqurdckll/RoOY5EIvrCF76gaDSqU045Rc3NzbrwwgtVXV2tSZMm6aCDDtKrr74qKbVfZ8GCBRo1apTq6+u1YMGCHgmqaZpqbGxULBZTeXm55s2bp5UrV6q1tVWStGzZsh7J9e6IRCJav369mpqaVFVVpUMPPVSStHTpUs2aNUvHHnusTNPU0UcfralTp2rlypV699139eKLL+qLX/yiYrGYZsyYoRNOOGGP4gAAYHe0tLRo5MiRikQi/b7nj3/8o6ZPn65jjz1WL730ku64446CxhSJRNTS0qK3335blmVp6tSpqq6uliQdd9xx2nfffWUYho444ggdffTRev755yVJf/jDHzR//nxNmjRJFRUVuvzyyzPXHMrvIECY9f+/4CI2f/58XXDBBX3aPXZXeXm5li5dmpdrobQ0NDToggsu0MaNG9XQ0JB5vbm5WR0dHZo/f37mNdd15ThO5uu6ujpZliVJPZ62esrKytTW1iZJ2rJliyZMmJD53oQJE7Rly5bM1yNHjlRZWVnm67Fjx+rwww/XI488ohNPPFGrVq3S9ddfv0d/10WLFun222/XySefrH322UeXXXaZjj/+eL377rv64x//qOXLl2fem0wmNXPmTG3ZskW1tbWqrKzsEft77723R7EAADBcdXV1am5uVjKZ7DeJnTNnjm699dY+r1uW1WfGRCKRkKQBE+LBNDQ0aNOmTbryyiu1c+dOzZs3T1/60pcUjUa1cuVK3XnnnXrrrbfkOI46Ozs1efJkSanfC6ZOnZq5zvjx4zN/bmpqGvR3ECDMQpnAzpgxo8+U1/Xr1+sb3/iGmpubVV5erm9961s68MADA4oQpWLvvffWPvvso5UrV2rRokWZ10eOHKny8nI9/PDDmarsnhgzZozeffddTZo0SZL03nvvacyYMZnvG4bR52fOOOMMPfjgg7JtW4ceeuigcVRUVPTYH2Pbdo8py/vtt59uu+02OY6jP/3pT2psbNTq1as1fvx4NTQ06Nvf/nafa77zzjvauXOn2tvbM0nsu+++mzNeAAAK6bDDDlMsFtNjjz2mOXPmDOtnJ0yY0ONBrSRt3LhRkUhkSOtr9pDHrVu3Zv4cjUZ12WWX6bLLLtPGjRv1uc99Tvvvv78aGhrU2Nio73znO/r4xz+uaDSqSy+9NDNQasyYMdq8eXPmOtkPhvP9OwhQbELZQpzLDTfcoBtuuEGLFy/WNddco2984xtD/tmuri7Nnz9f55xzjh577LECRon3o0WLFun+++/vUWU0DENnn322brrpJm3fvl1Saq/ok08+uVv3OPXUU/XDH/5QTU1Nampq0p133qnTTjttwJ/5xCc+oXXr1umBBx7oMVyqP/vvv7+6urq0YsUKJRIJ/fCHP8wMipJSrcJNTU0yTVO1tbWSUq3L8+bN0/Lly/Xkk0/Ktm11dXVp9erV2rRpk/bee29NnTpVP/jBDxSPx/X888/3+QUAAAA/1NTUqLGxUd/85jf12GOPqaOjQ4lEQitXrtR3v/vdAX/2Yx/7mN544w0tWbJEiURCLS0t+v73v6/Zs2cPWoE9+OCD9fvf/16JREIvvviiHnnkkcz3nn32Wb366quybVvV1dWKRCIyTVPxeFzxeFz19fWKRCJauXKlnnrqqczPzZkzR4sXL9brr7+ujo6OHnMxTNPM6+8gQLEJZQW2t7a2Nq1du1Zf/OIXM695v3j/6U9/0u23397nZ8aOHauf/vSnkqTly5dr7Nix2rBhgy666CJNnjxZ++67rz/BI/T6+3fl6quv1p133qlzzjlHzc3NGjt2rP7t3/5NH/vYx4Z9j0svvVRtbW2Zfaxz5szRpZdeOuDPlJeXa/bs2Xr44YeHNLCipqZGN954o7761a/Ktm195jOf0bhx4zLff/LJJ/Wf//mf6uzs1IQJE/T9739f5eXlGj9+vO666y7dcsstuuqqq2Sapj7ykY/o61//uiTpe9/7nq655hrNnDlThx56qE4//XTt3Llz2J8BAAB76lOf+pT22msv3XXXXVq4cKGqqqo0ZcqUzLE6/Rk1apTuuece3XLLLVq0aJHKysp07LHH6stf/vKg97ziiit05ZVX6ogjjtCMGTN02mmnqaWlRVJqCOONN96ozZs3q7KyUqeccooaGhoUiUT01a9+VVdccYXi8biOP/74HjMkjj32WP37v/+7LrzwQhmGoUsvvVRLlixRLBaTlN/fQYBiY7ghPdxq48aN+vznP6/f/e53am1t1Zw5c/TnP/95j6977bXX6rjjjht2awlQjO644w699dZbOffzAACA94fXX39dc+fO1YsvvrhHe3KBMHhftBBXV1drn3320R/+8AdJqY3qr7zyypB+dseOHZlqbVNTk9asWaODDjqoYLECfmlpadFDDz2kc889N+hQAABAnj366KOKx+PasWOHbrnlFh1//PEkrygJoazAXnnllXruuefU3NysUaNG6fLLL9dHP/pRff3rX9fWrVuVTCZ1yimn6LLLLhv0WmvWrNGNN94owzDkuq4uvPBCnX322T78LYDC+fWvf62bbrpJ8+bN0ze/+c3M68uWLdONN97Y5/0TJkxgvD4AALvpsMMOy/n6Pffco+nTpxfknp/+9Kf1t7/9TZZlacaMGbrxxht7DHgE3q9CmcACAAAAAErP+6KFGAAAAADw/he6RnnHcWTbhSsaW5ZR0OsXCnH7i7j9Rdz+KrW4o1GrANGUlkKtzWH8d5GY/RPGuInZP2GMO4wxS4WJe6C1uSgS2Pvuu08PPvigDMPQ5MmTdfPNN6usrCzne23bVUtLe8FiqaurLOj1C4W4/UXc/iJuf5Va3KNH1xQgmtJSqLU5jP8uErN/whg3MfsnjHGHMWapMHEPtDYH3kK8efNmPfDAA3rooYf0u9/9TrZtM0wGAAAAANBH4AmsJNm2rc7OTiWTSXV2djJBDQAAAADQR+AtxGPHjtWnPvUpHX/88SorK9PRRx+tY445JuiwAAAAAABFJvAEdseOHXr88cf1+OOPq6amRl/84he1dOlSNTQ05Hy/ZRmqq6ssWDyWZRb0+oVC3P4ibn8Rt7+IGwAAFKvAE9inn35a++yzj+rr6yVJs2fP1tq1a/tNYBnilBtx+4u4/UXc/iq1uBniBABAeAS+B3bChAn6+9//ro6ODrmuq2eeeUYHHnhg0GEBAAAAAIpM4BXYadOm6aSTTtIZZ5yhSCSiQw45ROeee27QYQEAAAAAikzgCawkNTY2qrGxMegwAAAAAABFLPAWYgAAAAAAhoIEFgAAAAAQCiSwAAAAAIBQIIHNYjuuOuJ20GEAAFDyXJc1GQDQFwlsll+ueUcn/+BJua4bdCgAAJS0Ff/ariO/+4Q6EiSxAIBuJLBZtuzq0jstnWpqTwQdCgAAJW3zri61ddlqowoLAMhCApvFdlKV140tHQFHAgBAaXPS3VB0RQEAspHAZrFdL4HtDDgSAABKm/dQ2SF/BQBkIYHNkkyvkhuowAIAECgvcaUCCwDIRgKbxaGFGACAouC1EFOBBQBkI4HNQgsxAADFobuFmAwWANCNBDYLQ5wAACgO3ppM/goAyEYCm8VbLHd0JrWjg6N0AAAIild5tclgAQBZSGCzJLM22mzcQRsxAABBsTNDnIKNAwBQXCJBB1BMHFeKRUzFk47eaenQlHE1QYcEAEAgTjjhBFVVVck0TVmWpcWLF/t6f4c9sACAHEhgs9iOqw/UV+qfW1o5SgcAUPLuv/9+1dfXB3Jvr3WY/BUAkI0W4iy246qqzNKY6pg2MIkYAIDAeLt6qMACALJRgc2SdF1ZhqF96iq0sZkKLACgtH3605+WYRg699xzde655/b7PssyVFdXmdd7R6OWJKmqujzv1y4kyzJDFa8UzpilcMZNzP4JY9xhjFnyP24S2Cy24yoaMVVXEdUrW1qDDgcAgMD88pe/1NixY7V9+3ZdfPHFOuCAAzRjxoyc77VtVy0t7Xm9f3tn6jSAnbs61NISnl9X6uoq8/5ZFFoYY5bCGTcx+yeMcYcxZqkwcY8e3f8sIlqIs9iOq4hpqrosotauZNDhAAAQmLFjx0qSRo0apRNPPFEvvPCCr/f3jrZzHFqIAQDdSGCzOK4ryzRIYAEAJa29vV2tra2ZPz/11FOaNGmSrzF4e1/JXwEA2cLTk+MD2/ESWEtx21VX0lFZhBwfAFBatm/frgULFkiSbNvW3LlzNWvWLF9jsBniBADIgQQ2S9JxFTEN1ZSlPpbWrqTKIrGAowIAwF8TJ07UsmXLAo3Bax0mfwUAZKO8mMV2XJnpFmJJ2kUbMQAAgci0EIsMFgDQjQQ2i92rAttGAgsAQCBsJ/VPKrAAgGwksFlst3sPrEQFFgCAoHgVWJspTgCALCSwWZx0BbY6swfWDjgiAABKk5fAUoEFAGQjgc2SdLqP0ZGowAIAEJSkwx5YAEBfgU8hfuONN/SlL30p8/WGDRvU2Nio//iP//A9FtuVrF5TiAEAgP84BxYAkEvgCewBBxygpUuXSkqdNTdr1iydeOKJgcTinQNbETVlGSSwAAAExckMcSKDBQB0K6oW4meeeUYTJ07U3nvvHcj9vQTWMAxVlUXYAwsAQEBsKrAAgBwCr8Bme/jhhzV37twB32NZhurqKgtyf0euopapurpK1VZE1eWqYPfKNysdd9gQt7+I21/E7a+wxo3cuoc4kcECALoVTQIbj8f1xBNP6KqrrhrwfbbtqqWlvSAxJG1XpmGopaVdlRFTTbs6C3avfKurqwxNrNmI21/E7S/i9tfuxj16dE0BosGectKlV5v8FQCQpWhaiFetWqUpU6Zor732CiwGO32MjiTVlEfYAwsAQEC8xJUKLAAgW9EksA8//LBOPfXUQGOwnVQFVpKqYxG1xtkDCwBAEJhCDADIpSgS2Pb2dj399NOaPXt2YDE4ritXylRgq8sj2tVJBRYAgCDYDntgAQB9FcUe2MrKSq1evTrQGLyF0vIS2Jil1jgJLAAAQfDWZSqwAIBsRVGBLQa9E9iasojauuxMCxMAAPCPl7iyDgMAspHApnnnzUWsdAW2LCJXUjv7YAEA8J2dOUYn4EAAAEWFBDYtVwVWEpOIAQAIQPcxOmSwAIBuJLBpmQTWm0JcZkmSdpHAAgDgO8dliBMAoC8S2LQ+Q5wyFVhaiAEA8Jud2QMbbBwAgOJCApuWTK+QkV4JLBVYAAD8xzE6AIBcSGDTvCe87IEFACB4XgsxFVgAQDYS2DQ7U4FNfSTeHlgSWAAA/Mc5sACAXEhg07yFMp2/sgcWAIAAcQ4sACAXEti0pNuzAhu1TJVFTPbAAgAQgO4WYhJYAEA3Eti03lOIpdQ+WFqIAQDwX/cQp4ADAQAUFRLYNO8Jb3YCWxWzaCEGACAAVGABALmQwKbZvY7RkaRYxFTCdoIKCQCAkuXtgSV/BQBkI4FNy9VCHLVMJRwSWAAA/JZ0qMACAPoigU1L5khgY5ahhM3CCQCA3xxvD2zAcQAAigsJbFquCmzEooUYAAC/ua6bSVxtDoIFAGQhgU2z3b57YKMmFVgAAPyWvfTSQQwAyEYCm+Ztde3ZQkwFFgAAvzlZVVf2wAIAspHApiVzTCGOsgcWAADfZSetPEYGAGQjgU2zM+fAdn8kEaYQAwDgOzsrgXWpwAIAspDApmWGOHUXYNkDCwBAALKfHTPDCQCQjQQ2LdcU4liEPbAAAPgte/IwFVgAQDYS2DRvv41lZbUQU4EFAMB32S3EPEcGAGQjgU3LPcSJCiwAAH7LHuLUfSIsAAAksBk5W4gtQwk23wAA4Cu7xzE6AQYCACg6JLBp3UOcuhPYiGXKdlzOoAMAwEfZSSt7YAEA2Uhg03JVYKPpP7MPFgAA//Q4B5YlGACQpSgS2J07d6qxsVFz5szRySefrLVr1/oegzcwovceWEnsgwUAwEc9W4jJYAEA3SJBByBJixYt0sc+9jHdfvvtisfj6uzs9D2G7Aqsl656CWySCiwAAL7JrrqSwAIAsgVegd21a5f+8pe/6KyzzpIkxWIx1dbW+h6HnXMKcerPcSqwAAD4hiFOAID+BJ7Abty4UfX19bruuut0+umn6/rrr1d7e7vvcXgtxFaOBDbhkMACAOCX7HNgGeIEAMgWeAtxMpnUunXrdMMNN2jatGn69re/rR//+Me64oorcr7fsgzV1VXmPY5oLPVRxKJW5vojasolSRWVZQW5Zz5Zlln0MeZC3P4ibn8Rt7/CGnexsm1bZ555psaOHasf/ehHvt6bIU4AgP4EnsCOGzdO48aN07Rp0yRJc+bM0Y9//ON+32/brlpa8l+hbWuPyzQkx+m+fqIrKUna3tKu+mjgxeoB1dVVFuRzKTTi9hdx+4u4/bW7cY8eXVOAaMLvgQce0IEHHqjW1lbf7+04VGABALkFnpWNHj1a48aN0xtvvCFJeuaZZ3TggQf6HkfS6dk+LHGMDgCgNG3atEkrVqzIzKfwm91jiFMgIQAAilTgFVhJuuGGG7Rw4UIlEglNnDhRN998s+8xOK4ry+iVwHp7YBniBAAoITfddJOuvvpqtbW1Den9+d7eU7mjK/PnSNbWnjAIYyt7GGOWwhk3MfsnjHGHMWbJ/7iLIoE95JBDtHjx4kBjsB23bwXWO0aHx78AgBKxfPly1dfXa+rUqVq9evWQfibf23t27OzI/LmzKxmqlvYwtuCHMWYpnHETs3/CGHcYY5YKE/dA23uKIoEtBrbj9jhCR+pOYDlGBwBQKtasWaMnnnhCq1atUldXl1pbW7Vw4ULdeuutvsXgMIUYANAPEtg023Vl9ttCzOIJACgNV111la666ipJ0urVq/Wzn/3M1+RV4hxYAED/Ah/iVCySuVqIzXQLMRVYAAB803OIExksAKAbFdg0J+ce2NTXcSqwAIASNHPmTM2cOdP3+/Y8Rsf32wMAihgV2DTb7X+IE1OIAQDwT3bVlQosACAbCWxa7iFO6T2wbMABAMA3XuNTxDTYAwsA6IEENs12XPXKXzN7YBNJKrAAAPjFayGOWiYVWABADySwaTmHOEWowAIA4DcvaY1YBntgAQA9kMCm2Y4rq/cxOiZ7YAEA8JvtJbCmQQUWANADCWya46pPBTaS3gObZAoxAAC+8c6BjVmmeIQMAMhGApuWa4iTaRiKmIbiVGABAPCNk152Uy3EPEQGAHQjgU1L5jhGR0pNIk5QgQUAwDfdLcQmU4gBAD2QwKalphDnSmBNJR0qsAAA+CV7iJNDBgsAyEICm2bnmEIsiRZiAAB85i27USqwAIBeSGDT+ktgY5ZJCzEAAD7yKrDRiCFXrMEAgG4ksGmO23eIk+TtgaUCCwCAXxz2wAIA+kECm5bMcQ6sJEUsU0lWTwAAfOMdoxNlCjEAoBcS2LSBWojjSSqwAAD4xUtgIxYVWABATySwaf1PITaUYPUEAMA33rIbNY1MOzEAABIJbIbturJyfBpR01CSPbAAAPgmcw6sZZLAAgB6IIFNc5zcQ5wilqk4U4gBAPCNd/arZRq0EAMAeiCBTUsOeIwOFVgAAPziuK5MQ7IMhjgBAHoigU2zXeVMYKOWwRRiAAB85K3JpiEqsACAHkhg0/ob4hQxqcACAOAnJ70mmyYVWABATySwaXY/e2CjlsEeWAAAfGS7qbPZTYM9sACAnkhg0wY6B5YpxAAA+Md2XJmm0i3EZLAAgG4ksGne097eIpahBBVYAAB847ipAU4GFVgAQC8ksGn9VWCjlqk4FVgAAHyTmkKcGuLEHlgAQLZI0AFI0gknnKCqqiqZpinLsrR48WLfY+i/hZgpxAAA+CnVQsweWABAX0WRwErS/fffr/r6+kDu7biuXKmfFmKmEAMA4CfHdWUZksEeWABAL7QQK/WkV+rnHFgz9fTX5hEwAAC+8M6BtUxD5K8AgGxFU4H99Kc/LcMwdO655+rcc8/t932WZaiurjKv9+6I25KkqsqYLMvscf2aqjJJUmV1uSpiVl7vm0+94w4L4vYXcfuLuP0V1rjRV+YcWMOgAgsA6KEoEthf/vKXGjt2rLZv366LL75YBxxwgGbMmJHzvbbtqqWlPa/3b4snJUmJeFK27fS4vp1IfW9bU5tqyovi48qprq4y75+LH4jbX8TtL+L21+7GPXp0TQGiwZ7w5lKkWoiDjgYAUEyKooV47NixkqRRo0bpxBNP1AsvvODr/QdsIbZSH1HCYR8sAAB+SE0hFhVYAEAfgSew7e3tam1tzfz5qaee0qRJk3yNIZPA5hjiFE0ntfEkCSwAAH6wXWUdoxN0NACAYhJ4T+z27du1YMECSZJt25o7d65mzZrlawzdFdi+3/MqsBylAwCAP5x0CzEVWABAb4EnsBMnTtSyZcsCjSE5UAXWSr2WsFlAAQDwg+12D3Fi9QUAZAu8hbgYeMXVAffAchYsAAC+6N4DyzmwAICeSGA12BAnrwJLAgsAgB8cR4pkWoiDjgYAUExIYNWdwEYGrMCyggIA4AevhdgwDDlksACALCSwkpLp9iRzoD2wHKMDAIAvbMeVaaanEEtyaSMGAKSRwGqQFmIz9RHFqcACAOALx3VlGd0PllmBAQAeElgNbQ9skj2wAAD4wnaUbiFOfU0XMQDAQwKr7gmHA08hZvUEAMAPjptqIfbWZVqIAQAeElhlDXHKuQc2ncCyBxYAAF+kWoiNTAsxFVgAgIcEVlJyKMfoJFk9AQDwg+2kzoHtbiFmDQYApJDAqrsCa+b4NKjAAgDgL8dNPVTursCSwAIAUkhglTpvTpKsXC3E6aose2ABAPCH7brpBDb1NfkrAMBDAivJK65GBhziRAUWAAA/pFqIDRlUYAEAvUSCDqAYDGkPLBVYAECJ6Orq0ic/+UnF43HZtq2TTjpJjY2Nvt3fcVN7YL11mSFOAABPXiuwl112mVasWCEnZPtF7QGO0YlkWojD9XcCAEDavbU5Fovp/vvv17Jly7RkyRI9+eST+tvf/lbAKHtynN4txGSwAICUvCaw559/vn77299q9uzZuvXWW/XGG2/k8/IFkxnilGMPrGEYilqG4lRgAQAhtDtrs2EYqqqqkiQlk0klk8lMO68fbFe9Woh9uzUAoMjltYX4qKOO0lFHHaVdu3bpd7/7nS6++GKNHz9eZ599tubNm6doNJrP2+WNPUALsZSqwiZDVlUGAEDa/bXZtm3Nnz9f69ev1/nnn69p06b1ew/LMlRXV5nXuCvKo4qkjweorilXXW15Xq9fKJZl5v2zKLQwxiyFM25i9k8Y4w5jzJL/ced9D2xzc7OWLVumpUuX6pBDDtG8efP017/+VUuWLNF///d/5/t2eeElsLmGOEmpQU5JKrAAgJDanbXZsiwtXbpUO3fu1IIFC/Taa+dYQK0AACAASURBVK9p8uTJOd9r265aWtrzFm/CdpRMJCWl1t6WHR0qC8mD5Lq6yrx+Fn4IY8xSOOMmZv+EMe4wxiwVJu7Ro2v6/V5eE9gFCxbozTffVENDg+6++26NGTNGknTKKado/vz5+bxVXg20B1byKrAksACA8NnTtbm2tlYzZ87Uk08+2W8Cm2+Omzraztvawx5YAIAnrwnsOeeco2OPPbbHa/F4XLFYTIsXL87nrfIq00Lcz/6eiGkwxAkAEEq7szY3NTUpEomotrZWnZ2devrpp/XZz37Wj3AlpaYQW6Yhb1nmGTIAwJPXIU7/9V//1ee1c889N5+3KIjB9sBGLVMJVk8AQAjtztq8ZcsWXXjhhTrttNN01lln6aijjtLxxx9fqBD78M6BtTgHFgDQS14qsFu3btXmzZvV2dmpdevWZVp9Wltb1dHRkY9bFNRgFdioZbAHFgAQKnuyNh988MFasmSJH2Hm5CWw3S3EgYUCACgyeUlg//znP2vx4sXatGmTbr755szrVVVVuvLKK/Nxi4IafA+syRRiAECohHltTrUQK6uFmAwWAJCSlwT2jDPO0BlnnKFHHnlEJ510Uj4u6avBW4gNJajAAgBCJMxrs3cOrMk5sACAXvKSwC5dulQNDQ165513dO+99/b5/sUXX5yP2xTM4OfAUoEFAIRLmNdmx3FlmoZMKrAAgF7yksB6e2na28N3bpE0eAsxFVgAQNiEdW12XVeuJMuQDPbAAgB6yUsCe95550mSLrvssnxcznfdQ5xyfz9iGupIUIEFAIRHWNdm73mxZRqZB8tUYAEAnrweo/Pd735Xra2tSiQSuuiii/TRj35US5cuzectCiI17bD7SW9vUYsWYgBAOIVtbXbSD5VTe2BTr5G/AgA8eU1gn3rqKVVXV2vFihXae++99eijj+qnP/3poD9n27ZOP/10XXLJJfkMZ8iSTv/tw1L6GB0mSAAAQmh31+agZLb1GEbmwbIj1mAAQEpeE1jbtiVJK1as0Jw5c1RTUzOkn3vggQd04IEH5jOUYXFct98zYKVUC3HCpgILAAif3V2bg+Jt6+kxxImHyACAtLwmsMcdd5zmzJmjl156SUceeaSamppUVlY24M9s2rRJK1as0FlnnZXPUIbFdtwBK7ARy2SIEwAglHZnbQ6St9/VNMQxOgCAPvIyxMmzcOFCfeYzn1FNTY0sy1JFRYXuuuuuAX/mpptu0tVXX622trZ8hjIstuMqMlALsUkLMQAgnHZnbQ6SN3LCMjhGBwDQV14TWEl644039M4772RaliTp9NNPz/ne5cuXq76+XlOnTtXq1auHdH3LMlRXV5mXWDPXjFqyLFN1dZWZf2arrIjKdt283zefcsUdBsTtL+L2F3H7K6xx+2E4a3PQvD2wpmlwjA4AoI+8JrBXX321NmzYoIMPPliWZUlKTfbtb5Fcs2aNnnjiCa1atUpdXV1qbW3VwoULdeutt/Z7D9t21dKS3zPt2jsTMiW1tLSrrq6yz/XdpKN40sn7ffMpV9xhQNz+Im5/Ebe/djfu0aOLe0/onhru2hw0x+0+2i5zjA5DnAAAaXlNYP/xj3/o97//fb/H0fR21VVX6aqrrpIkrV69Wj/72c8GTF4LZfA9sIaS7IEFAITQcNfmoNkcowMAGEBehzhNmjRJW7duzeclfeG4gySwpqkE58ACAEIobGtz5hidrBZi9sACADx5rcA2Nzfr1FNP1Uc+8hFFo9HM63ffffegPztz5kzNnDkzn+EM2aBDnCxDCduV67qheYINAIC0Z2tzEDJDnEyjewoxz5ABAGl5TWAvv/zyfF7ON7Yz8DmwUSv1PduVIuSvAIAQCdvanBnilD2FmD2wAIC0vCawRxxxhN555x29/fbbOuqoo9TR0dFj4mGxSjquzAGaqaPpbyZtRxHT8ikqAAD2XNjW5uxzYA3OgQUA9JLXPbC//vWv1djYqK997WuSpM2bN2vBggX5vEVBDFaBjaQrsAkGOQEAQiZsa3N2C7GVOUaH9RcAkJLXBPYXv/iFfvnLX6q6ulqStN9++6mpqSmftygIewhDnCQpySYcAEDIhG1tztlCTP4KAEjLawIbi8UUi8UyXyeTyXxevmAcR4MOcZKowAIAwidsa7OTlcAaVGABAL3kdQ/sjBkzdPfdd6uzs1NPPfWU/ud//kcnnHBCPm9REMlBK7DpBJYKLAAgZMK2NjuOd4yOqMACAPrIawV24cKFqq+v1+TJk/WrX/1Kxx57rK644op83qIgbGfgBDZqpT4mKrAAgLAJ29qcdLrPgTU5BxYA0EteK7CmaeoTn/iEPvGJT6i+vj6fly4o23EVifSfy3stxEkeAQMAQiZsa7O31LIHFgCQS14SWNd1dccdd+jnP/95Zp+KaZq64IILdNlll+XjFgU1WAXWayFO2rQQAwDCIaxrs1dttdgDCwDIIS8txPfdd5/WrFmj//u//9Nzzz2n5557Tg8++KDWrl2r++67Lx+3KCjHdQcc4hShhRgAEDJhXZttx0u2lXm4TAUWAODJSwK7dOlSfe9739PEiRMzr02cOFG33HKLlixZko9bFFRykHNgoyYtxACAcAnr2pxdge1uIWb9BQCk5CWBTSaTOffV1NfXF/24fmkILcSZY3RoIQYAhENY12av2ckys1uIAwwIAFBU8pLARqPR3fpesbAdNzPpMJeomW4hpgILAAiJsK7Ndo8pxOnXyGABAGl5GeL0yiuv6PDDD+/zuuu6isfj+bhFQdmuK2uAVD4zhZg9sACAkAjr2pzZA2t0H6PDECcAgCcvCezLL7+cj8sExnaGNsQp6dBCDAAIh7CuzdkVWMNgiBMAoKe8tBCH3VCP0WEKMQAAheUNbIoYRqY7igosAMBDAqvUwIiBEthMCzEVWAAACirZYw8sFVgAQE8ksEpXYIcyxIkKLAAABZV9DiwtxACA3khgNZxjdFhBAQAopFznwNJCDADwkMBq8ATWq8DSQgwAQGF5FdhIVgsxx+gAADwksEofozNQCzHH6AAA4Itk+lmxaWZXYIOLBwBQXEhgNYwpxFRgAQAoqOwW4u49sGSwAIAUElgNnsBaHKMDAIAvss+B9bqjyF8BAJ6ST2Ad15WrgY/RMQxDUcvIjPYHAACFYfc4Rif1GhVYAICn5BPYzEI5wB5YKdVGnLBpIQYAoJDsHC3EpK8AAA8JbNaT3oFELZMhTgAAFBgVWADAQEhg3aElsBGTFmIAAArNS2BNQ5ljdJihCADwkMAOowJLCzEAAIWVOtouNX/CoAILAOglEnQAXV1d+uQnP6l4PC7btnXSSSepsbHRt/sPaw8sFVgAAArKdrofKhtGqo2Yx8cAAE/gCWwsFtP999+vqqoqJRIJnX/++Zo1a5YOPfRQX+7fXYEd+H1Ry2APLAAABdb7aDvDMORSgQUApAWewBqGoaqqKklSMplUMpnMTB30Q3KIFdioZSrJJhwAQAl477339OUvf1nbt2+XYRg655xzdNFFF/lyb9t1M3tfpdReWBqgAACewBNYSbJtW/Pnz9f69et1/vnna9q0af2+17IM1dVV5u3eu9KLYk11merqKmVZZs7rl0UtuWbu7xWD/uIudsTtL+L2F3H7K6xxFyPLsnTttddqypQpam1t1Zlnnqmjjz5aBx10UMHv7TiuImZ2AksFFgDQrSgSWMuytHTpUu3cuVMLFizQa6+9psmTJ+d8r227amlpz9u9W1o6JEldnQm1tLSrrq4y5/UNV+pIv6cY9Rd3sSNufxG3v4jbX7sb9+jRNQWIJtzGjBmjMWPGSJKqq6t1wAEHaPPmzb4ksLbbq4VYqX2xAABIRZLAemprazVz5kw9+eST/Saw+ebtgY0MOoWYY3QAAKVn48aNevnll33rjrIiliLparplmbJMQ7EyKzTV9TB2AoQxZimccROzf8IYdxhjlvyPO/AEtqmpSZFIRLW1ters7NTTTz+tz372s77dPzmMc2A7EjwCBgCUjra2NjU2NuorX/mKqqur+31fPrujOjoTMqRMV1TqtWRougLC2MEQxpilcMZNzP4JY9xhjFkqTNwDdUcFnsBu2bJF1157rWzbluu6mjNnjo4//njf7t99YPrgQ5x2dSX9CAkAgMAlEgk1NjbqtNNO0+zZs327b+8WYstkDywAoFvgCezBBx+sJUuWBHb/7mN0aCEGAECSXNfV9ddfrwMOOEAXX3yxr/e2HVdW1pJsiCnEAIBug5x++v7nDKOFOMEUCQBACfjrX/+qpUuX6tlnn1VDQ4MaGhq0cuVKX+7d+xxY0zAyazUAAIFXYIOWGeI0SAtxxDKVsFlAAQDvf9OnT9err74ayL1tt+dDZcOQyF8BAJ6Sr8Amh9pCbNJCDABAoaVaiHtWYG0yWABAWsknsEPfA2vSQgwAQIH1bSEWQ5wAABkksK43hXjg90WowAIAUHC9pxCn9sAGGBAAoKiQwHp7YAcb4mQxxAkAgELr20JMBRYA0I0ENp2TDqWFmAosAACFZTuuzB5DnKjAAgC6kcAO6xgdl6fAAAAUkNOnhVgcowMAyCCBHfIQJ6PH+wEAQP7ZjtvjaDvDMDhGBwCQQQLreEOcBjtGJ/VR0UYMAEDhJB1XZtZvJ1RgAQDZSGCHMcRJkhI2iygAAIXiuOoxxIk9sACAbCSwQ2whjqQfByccJhEDAFAovc+BtQyDCiwAIIME1hviNFgLcboCm6QCCwBAwfROYA1D7IEFAGSQwA65AptuIaYCCwBAwdhu73NgqcACALqRwA55CnG6hZgKLAAABdO7AmsaEisvAMBDAjvcFmImSQAAUDB9W4gNjrADAGSQwA5ziFPSpoUYAIBC6dtCzB5YAEA3EtihJrAcowMAQMH1bSFmDywAoBsJrDvEPbAMcQIAoOBy7YFl5QUAeEhgvQrswPlrZogTx+gAAFA4tusq+5myYRhyqcACANJIYB1XlpFaIAfSfYwOiygAAIXiOOpbgWXpBQCklXwCm3Qkc5D2YSlrCjFDnAAAKBjbdTMPjSUqsACAnko+gU1VYAdPYCNeCzGPgQEAKJik48rMnkIsiWfHAABPySewjusOOsBJymohZg8sAAAFk2sKsSvWXgBASsknsLbTs1WpP7F0BTae5DEwAACF4OQ4GcA02QMLAOhGAjvECmxlzJIktSfsQocEAEBJ8k4GiPSuwLIHFgCQVvIJbO+9Nv2pSiewbfFkoUMCAKAkeQls9rpsiAosAKBbySewvffa9CdqmSqLmGrrogILAEAh2LlaiA0j01oMAEAk6ADee+89ffnLX9b27dtlGIbOOeccXXTRRb7df6gJrJSqwrZSgQUAoCC8CqzV4xgdkcACADICT2Aty9K1116rKVOmqLW1VWeeeaaOPvpoHXTQQb7c33GHNsRJkqrLIlRgAQAokEwCm7UspyqwAQUEACg6gbcQjxkzRlOmTJEkVVdX64ADDtDmzZt9u/9Qz4GVUhXYtjgJLAAAheCdVNfnGB0qsACAtMArsNk2btyol19+WdOmTev3PZZlqK6uMm/3NCxLsaiZuaZlmf1ev64qps6kk9f758tAcfdnZ0dCtRXRAkU0NLsTdzEgbn8Rt7+IG0HprsBmJ7AMcQIAdCuaBLatrU2NjY36yle+ourq6n7fZ9uuWlra83bfzq6EXKf7mnV1lf1ev8w0tLU9kdf758tAceeyvrlDZ9/7F91x1oc1Y9+RBYxsYMONu1gQt7+I21+lFvfo0TUFiAa7IzOFuNceWCqwAABP4C3EkpRIJNTY2KjTTjtNs2fP9vXeQz0HVvJaiN8fQ5w27+qU40qPvLw16FAAAJDUPayp9zmwVGABAJ7AE1jXdXX99dfrgAMO0MUXX+z7/R1Hwxvi9D7ZA9uRcCRJK1/friS/GQAAikCynxZiKrAAAE/gCexf//pXLV26VM8++6waGhrU0NCglStX+nb/5HArsF3J98VC2plIJeItHQn9/Z0dAUcDAEB/LcRGZrgTAACB74GdPn26Xn311cDuP7xzYCOyXakz6agiahU4ssJqz6okP/HaNv2/iXUBRgMAQO5zYL0K7H8+9k+1x21985SDgwoPAFAEAq/ABm1Yx+iUpZLWtq7ufbCr32rWXX9+syCxFVJHMtVCfNg+I7TiX9s4JB4AEDhvLerZQmyoI+Hody9t1rpNu4IKDQBQJEhgHbdHq9JAqmOpgnVrVvXy4XWbdf9zGxRPJ4Rh4bUQf2LyXtrSGte7OzoDjggAUOq8CmzvIU67upLqSjra1fX+GKQIANh9JLCOO+QhTrkqsO/uSE3zfSdkCWBHwpZlSHtVxST1bCkGACAIycwe2O7Xspukdr1P5lAAAHYfCaw7jBbiHBXY93amEtf1zeE6M7Ej4ag8aqk8vZe3I0ECCwAIljcUv3cLceqfUsJ21RWyjicAQH6RwA5jiFO1V4FNJ7DxpKOtrXFJ0vrmjsIEWCAdCVsVUSszjKqTXwgAAAHrb4iTJB21f70k0UYMACWOBHaYU4il7hbi93Z2ymtk2tASrgS2M2GrImqqImpmvgYAIEh2jnNga8oiqi2P6OOT95Ik7ewkgQWAUhb4MTpBs11XQ8xfVRVLVSu9FmKvfThqGSGswKZbiCNW5msAAIJku30rsBceMVGnf2S83tzeJklqpQILACWNBHZYQ5x6VmC9yb3T9h6h9U1h2wObaiEupwILACgSuVqIve0u21q7JFGBBYBSRwvxMFqII6ah8oiZ2QP77s4uRUxDh+8zQlta46EahNTdQpyuwLIHFgAQsFwtxJ6a8qgk9sACQKkjgXU15ARWSlVh2+LdFdhxtWXar75SkrQhRG3EHQknXYFND3EKUfINAHh/cnK0EHtq0oMUd1GBBYCSRgLrDP0YHSm1D7a1q3sP7Pjacu07skJSuAY5dSRslUctxSxDpkECCwDodt111+nII4/U3Llzfb1vMkcLsacmvY2HCiwAlDYS2GG0EEtSda8K7IQR5ZpYl0pgwzTIqSPdQmwYhiqiFkOcAAAZ8+fP109+8hPf7+sNccq1LEes1OR8ElgAKG0ksMNMYKtiltq6bHUmbDW1J7T3iHJVxiztVRULVQLbmW4hlqSyiKnOJBVYAEDKjBkzNGLECN/vm2uIU7aasggtxABQ4phC7A6/hbipPa5300fojK8tlyRNHFmhjSFpIXZdN9NCLIkKLABgj1iWobq6yj2+Tll5TJJUX1epuroKWZbZ47p1lTF1Om5e7lUovWMOgzDGLIUzbmL2TxjjDmPMkv9xk8DuTgtxl633dqTG+U8YkUpgR1fF9PLmXQWJMd+6ko5cSRWRVAG+PGqyBxYAsNts21VLy54fJ7erLbW2trV2qkWpRDX7ulVRU02tXXm5V6H0jjkMwhizFM64idk/YYw7jDFLhYl79Oiafr9HC/HutBDHbb2RPlDdS2BHVkbV3JEoSIz51pmutlZkVWA7qcACAALmtRCb/XRGVZdFOAcWAEpcSSewjuvK1e4do/OnV7bqkLHV2qsq1e5UVxFVa5ethF38iWBHer+rl8CWR61QnWELAHh/cgbZA1tbHlErQ5wAoKSVdAI70IHp/amOWXJc6ZUtrZpzyJjM63UVqQPWW0JQhfWS1fJouoU4YpLAAgAyrrzySp133nl68803NWvWLD344IO+3NebQhzpJ4GlAgsAKOk9sINNO8ylKn0OnWVIJx3cncCOrEwlsM3tCY2uLstjlPnXkauFOFn8lWMAgD9uu+22QO47WAtxbXlEbXF72Nt/AADvHyVdgR3owPT+VMdSSd8RHxipUen2YSlcFVhvYFN3AssQJwBA8AZ7sFydfohMGzEAlK6STmAdd/gJbG15avHMbh+WuiuwYUhgOzIJrNdCTAUWABA8e5B12VuDd5HAAkDJKukEdnf2wE6fWKdvnPxBzT64VwJb0d1CXOy8FuJyhjgBAIpI97qc+/s1ZSSwAFDqSGAlRYbxKUQsU6d8aGyfARO15VEZUiiO0unI0UKcsN1MSzUAAEGwHVemIRn9PFiu8SqwDHICgJJV0glscpBhEcNhmYZqyyOhaCHu7N1CnE5k2QcLAAiS7Q68rYcKLACgpBPYwfbaDNfIymioWoizK7ASCSwAIFi24w64rSeTwFKBBYCSVdIJrJOeW5S3BLYiGooKbEfCliGpLN077SWyXmILAEAQBjsep4YhTgBQ8ko6ge3eA5ufBLauMhaaPbDlUTOzx6g8nch2JqnAAgCC47gDJ7CVUUuWQQILAKUs8AT2uuuu05FHHqm5c+f6fu9kvluIK6JqCUULsZ2pukrde2CpwAIAgpQcpIXYMAxVl0W0kxZiAChZgSew8+fP109+8pNA7r07x+gMpK4ioh2dicz5ssWqI+H0SmDZAwsACJ7tuDIHeahcWx5RKxVYAChZgSewM2bM0IgRIwK5t5fADrZYDlVdZUyOK+3sKO6FtbNXBZY9sACAYuC4br9nwHqowAJAaYsEHcBwWZahurrKvFyrYldckjSipjxzTcsyd/v6e4+qkiTZUStvMQ7VcOJOuFJ1RSTz/jHpxNWMFXfcxYS4/UXc/iJuBMV23EHnUlCBBYDSFroE1rZdtbS05+VaO3d2SJI62+OZa9bVVe729cvSrcPrN+/SqKi/xe3hxL2rI6GyiJl5f6Ijlcg37ejI22c7VHvyeQeJuP1F3P4qtbhHj64pQDTYHckhtBDXlEW0aWeXTxEBAIpN4C3EQUpmWojzc726yqgkFf0kYoY4AQCKkeMOPpeipjzCFGIAKGElncB2pIcWZSdze2JkRSqBbWmP5+V6hZLaA9v9X333HliGOAEAgjPYObBSqgK7qyspt8gHJgIACiPwBPbKK6/UeeedpzfffFOzZs3Sgw8+6Nu9W7tSCVt1LD+d1HUVYanAOpmqqyTFLEOGpM4kFVgAQHCGmsAmbFddrFkAUJIC3wN72223BXZvbwhEdVl+KrCxiKmqmKXmIj8LtncLsWEYqohaHKMDAAiU7Q58DqyUaiGWUmt4eZ46qAAA4RF4BTZI3Qls/vL4UVUxbW0t3hZi13X7tBBLqbNgaSEGAARpqBVYSdrJPlgAKEmlncDGbVmmobJI/j6GyaOrtW7TrrxdL98Stivb7bvvtzxqqZMhTgCAAA0pgU1XYHdxFiwAlKTSTmC7kqqOWTIGaVcajg9PqNGmXV3a1lacVdj2dJW1d9tVBRVYAEDAUi3EA7/Hq8AyiRgAShMJbB7bhyVpyrjUeYIvvbczr9fNlx3pAVMjynv+vSuiFkOcAACBcobRQkwCCwClqaQT2La4nfcE9oNjqmWZhv7xXnG2EbekE9iR6TNrPeURkyFOAIBAJR3RQgwAGFBJJ7CpCmx+JxiWRy1NHl2lfxRpBdabkOydWespj1rqYA8sACBAtuvKHGwKMRVYAChpJZ7A2nk7Azbb1PG1WrepVbZTfIesN2cqsLEer3OMDgAgaENpIY5apsojpnZSgQWAklTiCWxS1eWFSGBr1J6w9WZTe96vvaf6rcBGGOIEAAiW7bqKDJLASlJteSRzFB4AoLSUdgIbT00hzrep42slSc+vb8n7tfdUc0dCVTFLsV5HBzHECQAQtKQzeAuxlDq/nQosAJSmkk1gHddVW1f+hzhJ0sS6cn1oXI1uX/WGnvjntrxff080t8dV16v6KknlHKMDAAjYUFqIJSqwAFDKSjaBbY/bcqWCJLCGYegHZ07VwWNqdN1v1+nRV7fm/R67q6Uj0WcCsSTVlkeVsF21x0liAQDBsN2hJbBUYAGgdJVsAus9uS1EC7GUSgjvOOvDmjahVl/7/St6+s2mgtxnuJrbEzkrsHtVpYY6bWuL+x0SAACSJJsKLABgEKWbwKYrjYWowHoqY5ZuO2OqDhhVqS8vW6d3dnQU7F5D1dKR6DPASZL2qvYS2C6/QwIAQFI6gR08f1VNWUQ7SWABoCSVbALb5lVg83wObG/VZRHdevoUxZOOfvePzQW912Bc11VzPy3EmQpsKxVYAEAwbFdDqsDWlEXU1mXLcYvvuDoAQGGVbAK7K5PAFq4C6xlfW67p+9bpDy9vkRvgYtsWt5Ww3ZwtxKOraSEGAAQrVYEdQgJbHpEr0UYMACWoZBPY1q50C3Gs8AmsJJ3yoTF6Z0enXnh3py/3y8U7A7a+MtbnezVlEcUsgwosACAwQ90DW5N++LyLBBYASk4JJ7D+tBB7jp+0l8oipn6/bosv98uluSOVwNblaCE2DEN7VcW0lQosACAgzhCnEGcSWCYRA0DJIYH1oYVYkqpiER130Cg99tpWxZOOL/fszavA5hriJEl7VZfRQgwACExyGC3EEhVYAChFpZvAxm1ZpqGyiH8fwZxDxmhnZ1J/Wd/i2z2ztXSkktNcQ5yk1CCn7bQQAwACYjuuTCqwAIABlG4C25VUdcySMYQnvflyxL4jVV1m6fHXtvZ4fVtbXJ0Ju+D3H6wCO7o6pq0cowMACMhQW4hrqcACQMkq7QTWp/ZhTyxiataBo7Ty9e1K2qk24m2tXTr3vud16YMvKOkUdkJxc0dC5RFT5dHc+35HVcXU2mX7kkwDAJDNcV05rhQZwoNlb/3eSQUWAEpOySawbXHb9wRWkk6YNFo7O5N6fkOLXNfVdx7/l9q6knrxvV2699n1Bb13c3vuM2A9mbNg2QcLAPCZk36Iaw7hN5OqmKXKqKX1zR0FjgoAUGz8z+CKRKoC688E4mwf3W+kKqOWHvr7e1q3qVUr/rVdl39sf/1rW5t++uzbOmr/kZoyvrYg927uSOQ8A9aTOQu2Na596ioKEgMAALl4XUhDGeJkGIaO3H+k/vxGkxzXlenjdiAAQLBKtgLb2mVnhkD4qSxiatZBo7TiX9v1w6fe0rQJtTp/+j768scPUn1VTN9b/oZctzCtxC3tiZxnwHr2owJrmgAAIABJREFUqiqTRAUWAOA/bxfNUPbAStKsA0dpW1tcL29uzcv9XdfVM2816XO/+rv+87F/5uWaAID8K+kKbFUACawkLTz+QM390FiNrSnTPiMrFDENVZdF9NkjP6CbHv2nVr3epGMPGpX3+za1x3Xg6Kp+v79XugLLWbAAAL/ZXgV2iAnsUfvXyzKkVa9v15RxNXt076Tj6ut/eEWPvLJVZRFTazfu0JnTxmvS6Oo9ui4AIP9KtwIbT00hDsKIiqhm7jdS+42qVCRroT5t6jjtO7JCd/35zcxCni8vbdqlLa1xTR4ggR1RHlHUMrSNo3QAAD6zh9FCLEl1FVFN23uEVv1r+x7f91uPvKpHXtmqzx65r5Z+5ghVxSz97NkNe3RdAEBhlGQC67iu2rqCGeI0kIhp6NJj9tMb29v17z9fo1+vfVd/Wd+sjS0dcvawrfiB5zaouszSvKnj+n2PYRgaVRnTNo7SAQD4zHaHV4GVUm3E/9rWplc271LCdvSrNe/o2t+u0+/XbR7SRH3XdfW95a/r9+u26PNHf0CfO2o/jaqK6ZzDJujx17bqze3tu/33AQAURnFlcD5pj9typaJLYCXphEl76cY5k/U/f31Htzzxr8zr1WWWDt17hI6ftJcO23uEassj6kjY2toa1wfqK1Q3wDXfbmrX8n9u00VHTBz07zy6OkYFFgDgu+G2EEvS8ZP20t1PvaV///lajSiPaEdnUiPKI3r8tW26fdWbumH2ZB19QH2/P//z5zfqwb+9q0/+v3306Y9+IPP6+Yfvo/9d845ufeJfuv3MDw8rJgBAYRVFBrdq1SotWrRIjuPo7LPP1uc+97mC3s8yDVVETe07svgm7RqGoblTxunUD43VhpZObdnVpY0tHVq3eZdWv9WsP7/R1OdnTEM6dGKdPn7QXjrp4NEakTVpeNPOTt224nVFLUPnHb73oPcfU1OmtRt3qGWQicUoLu/t7NSPn35baza06Mj963XalLEFm2YN+OmfW1sVTzo6eGwNSYTP/F6bPVFr6P89TxhRrl9fPF2PvrJV6zbv0ryp4/TR/Ubq+fUtum3F67riN//QtAm1qquIqqrMUlUsookjKzS2OqY/vbpVj7+2TSd+cLQaj92/x3XrKqO68rgDtejRf+rHT7+lLxyzfz8RhJ/rumrrSiqedBSLlGRjHoCQMdxCjbwdItu2ddJJJ+nee+/V2LFjddZZZ+m2227TQQcdlPP9iYStlpY9b+npTNgqi5gyeu21qaurzMv1C8F1Xb28uVVvbm/Xjs6EKqKW6iujemVzq/78VrNe2bRLliHtN6pS42vL1dSe0Kubd0mGoS8cvZ8uOmLioPf4x3s7dcmv/q4PT6jVHWd+WBGrsItZMX/eAymWuJva4/rZs+u1+IX3ZEg6fGKd1m7coa6ko1M/NEafP3o/jastz7y/WOIeLuL2V9Bxt3Qk9MjLW7TsH5v02tY2SdLIiqhO/OBonX3oBO03qjLnz+1u3KNH79kAoPejoNbmh1/arOMmjVJVLPV8fU/+XexKOvrZ6vVas6FFbXFbbV1J7exKqrUr1Vo8ojyihg+P0+eO2k9l/SRu337kNS39xyadOmWsZkysU31VVDHLVFfSUcJ2FbWM1H9MU1HLUMQ0VVldpqaWdsVtR0nH1YjyqEZUROS6qfPYX9q0Sy9t2qV1m3bJcV19YGSlPlBfoX3qKmRI6kjY6kw46rIdWYZUHrU0qiqmyqiluO0onnQUtx11pf+Z+tqV47pyXclV6sF2dSyiypilqpil6rKIasojqimLqDb9zx2dSf3s2bf12Gvb1JV0VFMW0ReP3V/zpo7r87tRsQr6/6t2BzH7J4xxhzFmqTBxD7Q2B57Arl27VnfccYd++tOfSpJ+9KMfSZIuueSSnO/P1yLZnzD/i7P61S1a/q9tenVLqzbv6tKoqpgmj67S2YdO6JHEDOb36zbrxj+8qnE1Zaopjyhi/n/27jy+qSrvH/jnZmuBthSQpqDI80IQlUXwQZFFqoW2SCll9YERZRhRRLT44IYbKiOouA1ugMswjjL+BlRataM4lFVFFAE7io+KUmmBBqQb3ZLcm/P7o81t0yZpkjbJTft5v16+bJKb3O+9tDn3e8/3nCPBqNchGM2ZwaCDLDsA1DW6wdZWv+2N43azl7bZiRcOAZRW22A5a4UAkDEoEQtGnY/EuGhU2WT8bV8h3t5fBNkhcE4XE87pYoIkAUajHkJxQJKkoPx7Bov38x1e3v61tRy3N+GKW3EIVNTacbLCCtkhcLE5BhmDExEXZcDuX85gx5HfYVcEunc2om+3Tng8/WIkxEap72cC23a00jYHo03+vcqGotIaXGSOQbTR+2SOVtmBJ7f9jN2/nEFFrdxmMZzTxYTBvWJh0OnwW2k1jpXWwNrkb06vkyCEQEtzOkoATAYd9JIEZ96pOARqffgbjjLoMOmSBPRPjEPeYQsOFJWja7QBJoMOOkmCTkKT/0to3Hg0bkckl+clt8+7f5/k4Xn3MTfe3mDQQak/TslDXFqjbxRz4EJ7hKFuE9rq/olW22Bvh6fVmFvijDsmyoCH0i7EOV08L9vpK29tc9hLiC0WCxITGyYWMpvNyM/P97i90agP+sVGpF7MjB3cC2MH92r158xLisW8JPd32Un7egJ4dHo3PDp9aLhDIWozc8dd4NN2kfr9rTVaapvb+nN79gQu/i/ft3/phhFtun+tWpw8INwhEBH5hIMdiIiIiIiIKCKEPYE1m80oLi5WH1ssFpjN5jBGRERE1LGxbSYiIq0KewI7ZMgQFBQUoLCwEDabDbm5uUhOTg53WERERB0W22YiItKqsI+BNRgMWL58ORYsWABFUTBjxgwMGMBxGEREROHCtpmIiLQq7LMQExEREREREfki7CXERERERERERL5gAktEREREREQRgQlsI7t370ZaWhpSUlLw6quvhiWG5ORkZGRkIDMzE9OnTwcAlJWVYf78+UhNTcX8+fNRXl4OABBC4PHHH0dKSgoyMjLw/fffq5+zZcsWpKamIjU1FVu2bFGf/+6775CRkYGUlBQ8/vjjCLSC/P7778eoUaMwefJk9blQxOlpH62J+8UXX8RVV12FzMxMZGZmYteuXepr69evR0pKCtLS0rBnzx71eU+/K4WFhZg1axZSUlJw5513wmazAQBsNhvuvPNOpKSkYNasWSgqKvIr7pMnT+KGG27ApEmTkJ6ejjfffNPr+dDKOfcUt9bPudVqxcyZMzFlyhSkp6fjhRdeCHhfbXU8rYl72bJlSE5OVs/3Dz/8AEA7vydOiqJg6tSpWLhwoddzoZXzTcGjhfbYX56+7yJB0789rauoqEBWVhYmTpyIa6+9FgcPHgx3SD7529/+hvT0dEyePBlLly6F1WoNd0jN+HN9pyXu4n7qqacwceJEZGRkYPHixaioqAhjhM25i9npr3/9KwYOHIiSkpIwROaZp5jfeustTJw4Eenp6Vi9enXwAxEkhBBClmUxfvx4cezYMWG1WkVGRob4+eefQx7HNddcI86cOePy3FNPPSXWr18vhBBi/fr1YvXq1UIIIXbu3Cluuukm4XA4xMGDB8XMmTOFEEKUlpaK5ORkUVpaKsrKykRycrIoKysTQggxY8YMcfDgQeFwOMRNN90kdu7cGVCcX331lfjuu+9Eenp6SOP0tI/WxP3CCy+I119/vdm2P//8s8jIyBBWq1UcO3ZMjB8/Xsiy7PV3JSsrS3z00UdCCCEefvhhsXHjRiGEEG+//bZ4+OGHhRBCfPTRR2LJkiV+xW2xWMR3330nhBDi7NmzIjU1Vfz888+aP+ee4tb6OXc4HKKyslIIIYTNZhMzZ84UBw8e9HtfbXk8rYn7vvvuEx9//HGz7bXye+L017/+VSxdulTccsstXs+FVs43BYdW2mN/efq+iwRN//a07t577xWbNm0SQghhtVpFeXl5mCNqWXFxsbjmmmtETU2NEKLuu+e9994Lc1TN+XN9pyXu4t6zZ4+w2+1CCCFWr16tubjdxSyEECdOnBB/+tOfxNVXX90sJwg3dzHv3btXzJs3T1itViGEEL///nvQ42APbL38/Hz07dsXffr0gclkQnp6OvLy8sIdFgAgLy8PU6dOBQBMnToV27Ztc3lekiQMGzYMFRUVOHXqFD777DOMGTMG8fHx6Nq1K8aMGYM9e/bg1KlTqKysxLBhwyBJEqZOnRrwMV5++eXo2rVryOP0tI/WxO1JXl4e0tPTYTKZ0KdPH/Tt2xf5+fkef1eEEPjyyy+RlpYGAJg2bZoa9/bt2zFt2jQAQFpaGvbu3etX73dCQgIGDRoEAIiJiUG/fv1gsVg0f849xa31cy5JErp06QIAkGUZsixDkiS/99WWx9OauL2dby38ngBAcXExdu7ciZkzZwJAQP+2oT7fFBxabo+98ff7Tiua/u1p3dmzZ/H111+r8ZpMJsTFxYU5Kt8oioLa2lrIsoza2lokJCSEO6Rm/Lm+0xJ3cY8dOxYGQ92CK8OGDXNZ21oLPF2TPvHEE7jnnnu8tt/h4i7md955B7fccgtMJhMAoEePHkGPgwlsPYvFgsTERPWx2WwOW8Nz0003Yfr06fjnP/8JADhz5oz6JdezZ0+cOXMGQPOYExMTYbFYPB6Lp+3bSiji9LSP1tq4cSMyMjJw//33q6UxvsbnfL60tBRxcXHql2XjuC0WC3r16gWgbnmK2NhYlJaWBhRrUVERfvjhB1x66aURdc4bxw1o/5wrioLMzEyMHj0ao0ePRp8+ffzeV1seT6BxO8/3888/j4yMDKxatUotk9XS78mqVatwzz33QKera5YC+bcNx/mmtqel9jhQTb/vtKzp357WFRUVoXv37rj//vsxdepUPPjgg6iurg53WC0ym83405/+hGuuuQZjx45FTEwMxo4dG+6wfBKsa69Qeu+99zBu3Lhwh9Gibdu2ISEhARdddFG4Q/FZQUEB9u/fj1mzZmHu3LnIz88P+j4j49uqA3nnnXewZcsWvPbaa9i4cSO+/vprl9clSdLkHZmmQhFnW+1jzpw5+Pe//42cnBwkJCTgySefbIPogqOqqgpZWVl44IEHEBMT4/Kals9507gj4Zzr9Xrk5ORg165dyM/Px6+//hrukHzSNO6ffvoJS5cuxSeffIL33nsP5eXlQR9T6O/vyY4dO9C9e3cMHjw4iFERhYa372mticS/PVmWcfjwYcyZMwfZ2dno1KlTRIyTLi8vR15eHvLy8rBnzx7U1NQgJycn3GH5LVKuQxtbu3Yt9Ho9pkyZEu5QvKqpqcH69euxZMmScIfiF0VRUF5ejk2bNuHee+/FnXfeGfAcO75iAlvPbDa7lBZYLBaYzeawxAHUdb+npKQgPz8fPXr0wKlTpwAAp06dQvfu3d3GXFxcDLPZ7PFYPG3fVkIRp6d9tMY555wDvV4PnU6HWbNm4T//+Y/buD3F53y+W7duqKiogCzLzeI2m804efIkgLrG9+zZs+jWrZtfcdrtdmRlZSEjIwOpqakAIuOcu4s7Us45AMTFxWHkyJE4dOiQ3/tqy+MJNO49e/YgISEBkiTBZDJh+vTpHs93uH5PDhw4gO3btyM5ORlLly7Fl19+iZUrV0bU+aa2o5X2OBDuvu+0zN3f3t133x3usLxKTExEYmKi2rs9ceJEHD58OMxRteyLL77Aeeedh+7du8NoNCI1NTViJp8KxrVXqLz//vvYuXMnnnnmGc0n3seOHUNRUREyMzORnJyM4uJiTJ8+HadPnw53aF6ZzWakpKRAkiQMHToUOp0u4CpDXzGBrTdkyBAUFBSgsLAQNpsNubm5SE5ODmkM1dXVqKysVH/+/PPPMWDAACQnJyM7OxsAkJ2djfHjxwOA+rwQAocOHUJsbCwSEhIwduxYfPbZZygvL0d5eTk+++wzjB07FgkJCYiJicGhQ4cghHD5rLYQijg97aM1nF/KQF3pxoABA9R95ebmwmazobCwEAUFBRg6dKjH3xVJkjBy5Ehs3boVQN0srs7foeTkZHUm161bt+LKK6/064tUCIEHH3wQ/fr1w/z589XntX7OPcWt9XNeUlKizlZYW1uLL774AhdccIHf+2rL4wk07n79+qnnWwjR7Hxr4ffkrrvuwu7du7F9+3Y899xzuPLKK/Hss89q/nxTcGihPQ6Ep+87LXP3t/fMM8+EOyyvevbsicTERLUqZu/evbjgggvCHFXLevfujW+//RY1NTUQQkRM3EBwrr1CYffu3Xj99dexdu1adOrUKdzhtGjgwIHYu3cvtm/fju3btyMxMRHvv/8+evbsGe7QvJowYQL27dsHADh69CjsdntAHQZ+Cfo0URFk586dIjU1VYwfP1688sorId//sWPHREZGhsjIyBCTJk1SYygpKRE33nijSElJEfPmzROlpaVCiLoZRx999FExfvx4MXnyZJGfn69+1ubNm8WECRPEhAkTxLvvvqs+n5+fL9LT08X48ePFY489JhwOR0Cx/u///q8YM2aMuOSSS8RVV10lNm3aFJI4Pe2jNXHffffdYvLkyWLy5Mli4cKFwmKxqNu/8sorYvz48SI1NdVlxmZPvyvHjh0TM2bMEBMmTBB33HGHOiNbbW2tuOOOO8SECRPEjBkzxLFjx/yK++uvvxYXXnihmDx5spgyZYqYMmWK2Llzp+bPuae4tX7Of/jhB5GZmSkmT54s0tPTxYsvvhjwvtrqeFoT9w033KA+d9ddd6kzFWvl96SxL7/8Up0JVevnm4In3O1xIDx930WKxn97Wnf48GExbdo0MXnyZLFo0SJ1lnStW7NmjUhLSxPp6eni7rvv1uT3jT/Xd1riLu4JEyaIcePGqX+PztnrtcJdzI25W5kk3NzFbLVaxV133SXS09PF1KlTxRdffBH0OCQhglykTERERERERNQGWEJMREREREREEYEJLBEREREREUUEJrBEREREREQUEZjAEhERERERUURgAktEREREREQRgQksERERERERRQQmsERERERERBQRmMASERERERFRRGACS0RERERERBGBCSwRERERERFFBCawREREREREFBGYwBIREREREVFEYAJLREREREREEYEJLBEREREREUUEJrBEREREREQUEZjAEhERERERUURgAktEREREREQRgQksERERERERRQQmsERERERERBQRmMASERERERFRRGACS0RERERERBGBCSwRERERERFFBCawREREREREFBGYwBIREREREVFEYAJLREREREREEYEJLBEREREREUUEJrBEREREREQUEZjAEnVQycnJ+OKLLwAAL774Iu6+++4wR+RdJMRIRERERMHFBJY6vOTkZAwePBglJSUuz0+dOhUDBw5EUVFRmCJrvaqqKgwfPhwLFiwIdyguCTMREVFH0Pjma1FREQYOHAhZlsMclWeRECMRE1giAOeeey5yc3PVxz/++CNqamrCGFGd1jYgn376KUwmE7744gucPn26jaIiIiLSpg8//BDTp0/H8OHDMXbsWCxYsAD79+/Hiy++iEGDBmH48OEYMWIEZs+ejYMHDwLwXOEzcOBA/Pbbbz7td9myZbjkkktw6tSpNj0ef7FaiToCJrBEADIzM5Gdna0+zs7OxtSpU9XHNpsNTz31FK6++mqMHj0ay5cvR21tLQBg3759GDduHF577TWMGjUKY8eOxbZt27Br1y6kpaXhiiuuwLp161w+a+XKlRg7dizGjh2LlStXwmazuXzWq6++ijFjxuD+++/H5MmTsX37dvX9drsdI0eOxOHDh1s8ri1btmD27NkYOHAgPvjgA5/PR15eHtLT0zFixAjccMMN+OWXX9TXTp48idtvvx1XXnklRo4ciRUrVgAAjh07hhtvvBEjR47EyJEjcdddd6GiogIAcM899+DEiRO49dZbMXz4cLz22msAgEOHDmH27NkYMWIEpkyZgn379qn7KSwsxNy5czF8+HDMnz8fpaWlPsdPREQdz4YNG7Bq1Srceuut+Pzzz7Fjxw784Q9/QF5eHgDg2muvxcGDB7F3715cdtlluOOOOyCEaPV+q6ursXXrVsTGxvrV1hJRYJjAEgEYNmwYKisr8csvv0BRFOTm5mLKlCnq68888wyOHj2K7OxsfPrppzh16hRefvll9fXff/8dVqsVu3fvRlZWFh566CF88MEHeO+997Bx40a88sorKCwsBACsXbsW3377LXJycvDBBx/gP//5D1555RWXzyovL8eOHTvw5z//GZmZmS4N4q5du5CQkIBLLrnE6zEdP34cX331FTIyMpCRkeGSoHtz9OhR3HXXXXjggQewd+9ejBs3DrfeeitsNhsURcHChQvRu3dvbN++Hbt378akSZMAAEIILFy4EHv27MHHH3+M4uJivPjiiwCAp59+Gr1798a6detw8OBB3HzzzbBYLFi4cCEWLVqEr776Cvfddx+ysrLUUu67774bgwYNwr59+3Dbbbdhy5YtPsVPREQdz9mzZ/HCCy9g+fLlSE1NRefOnWE0GpGcnIz77rvPZVuj0Yhp06bh9OnTbXJz9NNPP0VcXBxuu+02n9taALBYLLj11ltxxRVXICUlBZs2bVJfUxQF69atw4QJEzB8+HBMnz4dJ0+eBAA8/vjjSEpKwmWXXYbp06dj//79AIDdu3dj/fr1+PjjjzF8+HD1Oubs2bN44IEHMHbsWFx11VV4/vnnoSiKup+nnnoKI0eOxPjx47Fr165Wnw+iYIvIBPb+++/HqFGjMHny5Db5vIsvvhiZmZnIzMzErbfe2iafSZHH2Qv7+eef44ILLoDZbAZQl5ht2rQJDzzwAOLj4xETE4OFCxe6lBwbDAYsWrQIRqMRkyZNQmlpKW688UbExMRgwIAB6N+/P3788UcAdeVNixcvRo8ePdC9e3csXrzYJUHV6XTIysqCyWRCdHQ0pkyZgl27dqGyshIA8MEHH7gk157k5ORg4MCB6N+/P9LT03HkyBGfem3/9a9/ISkpCWPGjIHRaMRNN92E2tpaHDx4EPn5+Th16hTuvfdedO7cGVFRURgxYgQAoG/fvhgzZgxMJhO6d++O+fPn4+uvv/Ya37hx45CUlASdTocxY8Zg8ODB2LVrF06cOIH//Oc/WLJkCUwmEy6//HIkJye3GDsREXVMBw8ehNVqRUpKSovb2mw2vP/+++jVqxe6d+/e6n1v2bIF6enpSE9Px6+//orvvvvOp/ctXboUiYmJ2LNnD1544QU899xz2Lt3L4C63uTc3Fy8+uqrOHDgAFatWoXo6GgAwJAhQ5CdnY2vvvoKkydPxpIlS2C1WjFu3DgsXLhQ7Wl2XlssW7YMBoMBn376qXqds3nzZgDApk2bsGPHDmRnZ+O9997DJ5980urzQRRshnAHEIjp06dj7ty5ze6oBSo6Oho5OTlt8lkUuTIzMzF37lwUFRUhMzNTfb60tBQ1NTWYPn26+pwQAg6HQ30cHx8PvV4PAGoD06NHD/X1qKgoVFVVAQBOnTqF3r17q6/17t3bZcxMt27dEBUVpT42m8247LLLsHXrVqSkpGD37t148MEHWzyenJwczJo1S/2Myy+/HFu2bGmx57ZpfDqdDr169YLFYoHBYEDv3r1hMDT/6vj999+xcuVK7N+/H1VVVRBCIC4uzuN+Tpw4gU8++QQ7duxQn5NlGSNHjsSpU6cQFxeHzp07q6/17t1bvftMRETUWFlZGbp16+a2fXL65JNPsHPnThiNRgwYMAAvvfRSq/d74sQJ7Nu3D8uWLcM555yDUaNGITs7G4MHD/b6vpMnT+LAgQNYv349oqKicPHFF2PWrFnIycnBqFGjsHnzZtxzzz3o168fAOCiiy5S39v4GuVPf/oT1q5di6NHj7ps4/T7779j165d2L9/P6Kjo9G5c2f88Y9/xD//+U/Mnj0bH3/8MebNm4devXoBABYuXIivvvqq1eeFKJgiMoG9/PLLm80Me+zYMTz22GMoLS1FdHQ0/vznP+OCCy4IU4QUic4991ycd9552LVrF1auXKk+361bN0RHRyM3N1ftlW2NhIQEnDhxAgMGDABQ14glJCSor0uS1Ow906ZNw+bNm6EoCoYNG9ZiHAcOHEBBQQFeffVVbNiwAUDdjMQ///wz7rvvPq8NfEJCAn766Sf1sRACJ0+ehNlshslkwsmTJyHLcrPPeO655yBJEj788EPEx8dj27Zt6vhYd3r16oXMzEw8/vjjzV47fvw4KioqUF1drSaxJ06ccHtuiIiI4uPjUVpa6rZ9cpo4cSKeeeaZZs/r9fpmkyba7XYA8NpeAnU3iy+44AJcfPHFAICMjAw8+eSTuO+++2A0Gj2+79SpU+jatStiYmLU53r37q323hYXF+P88893+9433ngD7777Lk6dOgVJklBZWemxFPrEiROQZRljx45Vn3M4HGrCeurUKfVnZwxEWheRJcTuPPzww3j44Yfx/vvv47777sNjjz3m83utViumT5+O6667Dtu2bQtilKR1K1euxJtvvunS8ydJEmbNmoVVq1bhzJkzAOrGrezZsyegfaSnp2Pt2rUoKSlBSUkJXn75ZWRkZHh9z4QJE3D48GH8/e9/d5lcypPs7GyMGTMGubm5yM7ORnZ2Nj788EPU1tZi9+7dXt977bXXYteuXdi7dy/sdjv++te/wmQyYfjw4Rg6dCh69uyJZ599FtXV1bBarfjmm28A1CXInTt3RmxsLCwWC15//XWXzz3nnHPUccAAMGXKFOzYsQN79uyBoiiwWq3Yt28fiouLce6552Lw4MF48cUXYbPZsH//fpeeWiIiosaGDx8Ok8kU0HVc7969cfz4cZfnioqKYDAYWrxhnJ2djcLCQowZMwZjxozBE088gdLS0hbHkiYkJKC8vFwdHgRAvVkMAImjbTzaAAAgAElEQVSJiTh27Fiz9+3fvx+vv/46/vKXv+Drr7/G/v37ERsbq05G1fRGb2JiIkwmE7788kvs378f+/fvx4EDB9RhUD179nSpbmKlE0WCdpHAVlVV4eDBg1iyZAkyMzOxfPlydcmQTz/9FJMnT27230033aS+f8eOHXj//ffx7LPPYtWqVW6/MKhjOP/88zFkyJBmz99zzz3o27cvrrvuOlx22WX44x//iKNHjwa0j9tuuw2DBw/GlClTMGXKFAwaNAi33Xab1/dER0cjNTUVRUVFLY7vsVqt+PjjjzF37lz07NlT/a9Pnz7NZlt2p1+/fnj66afx5z//GVdeeSV27NiBdevWwWQyQa/XY926dfjtt99wzTXXYNy4cfj4448BALfffjsOHz6MESNG4JZbbkFqaqrL595yyy1Yu3YtRowYgTfeeAO9evXCK6+8gvXr12PUqFFISkrCG2+8oZZmP/vss/j2228xcuRIvPzyyz4l7kRE1DHFxsYiKysLK1aswLZt21BTUwO73Y5du3Zh9erVXt971VVX4ddff0V2djbsdjvKysrw/PPPIzU11WsP7MGDB1FYWIjNmzerN4s/+ugjTJ48ucWhab169cLw4cPx3HPPwWq14v/+7//w7rvvqnNczJo1C2vWrEFBQQGEEPi///s/lJaWoqqqCnq9Ht27d4csy3jppZdckuAePXrg+PHjaluakJCAMWPG4Mknn0RlZSUcDgeOHTumlglfe+21eOutt1BcXIzy8nK8+uqrPp1vonCSRFvMHx4GRUVFuPXWW/HRRx+hsrISEydOxGeffdbqz122bBmuvvpqTJw4sQ2iJGo7L730EgoKCtyWPxEREVHdRId/+9vf8Ouvv6JLly4YNGiQuqzOb7/95rENPXDgAJ5++mkcOXIEUVFRSEpKwr333ouuXbt63Nfy5ctRWlqqzrjvlJ+fjz/84Q/47LPP8NZbb6n7LSoqwvjx4/H999/DYDCguLgYjzzyCA4ePIi4uDjcdNNNmDNnDoC62YHXr1+Pd999F6WlpejXrx9efvll9OzZEw899BC2bt2Kzp07Y968eXjnnXfw+OOPY/To0SgtLcVtt92GI0eO4LzzzsOWLVtw9uxZPPPMM9ixYweqqqrQp08f3HzzzUhPT4csy1i9ejVycnLQpUsX3HTTTVixYoUaI5EWtYsEFgBmz56NefPm4dprr4UQAj/++KPbwexNlZeXo1OnTjCZTCgpKcHs2bPxyiuvoH///sE+BCKflZWVYdq0aVi9ejUuv/zycIdDRERERBQWEXlrZenSpfjqq69QWlqKcePG4Y477sDTTz+NRx99FGvXroUsy5g0aZJPCewvv/yCRx55BJIkQQiBm2++mckracqmTZuwatUqTJkyxSV5/eCDD/DII4802753794uS/wQEREREbUXEdsDS0REREQUbMOHD3f7/GuvvaauhU5EocMEloiIiIiIiCJCxJUQOxwOKEpocm69XgrZvrSO56IBz4Urno8GPBcNIulcGI36cIcQ8ULZNjcVSb9r/uBxRY72eEwAjyvStLfj8tY2R1wCqygCZWXVIdlXfHznkO1L63guGvBcuOL5aMBz0SCSzkXPnrHhDiHihbJtbiqSftf8weOKHO3xmAAeV6Rpb8flrW1uF+vAEhERERERUfvHBJaIiIiIiIgiAhNYIiIiIiIiighMYImIiIiIiCgiMIElIiIiIiKiiMAEloiIiIiIiCICE1giIiIiIiKKCExgiYiIiIiIKCIwgSUiIiIiIqKIwASWiIiIiIiIIgITWCIiIiIiIooITGC9KCytxpkqW7jDICIioghXZZNx5PeqcIdBRBTxmMB6kfX/DmHNrl/DHQYRERFFuP934DjmbzwIhxDhDoWIKKIZwh2AlpVV2xFtYI5PRERErXOy3Ipa2YFqm4KYKF5+EREFitmZF1bZgbO1crjDICIiogh3prpuSFKVTQlzJEREkY0JrBe1soKzViawRERE1DrOOTWqbLyuICJqDdaweGGVHYBgQ0NEREStoyawVvbAEhG1BhNYD4QQsMkO2GUHHEJAJ0nhDomIiIgikBACJdV2AOyBJSJqLZYQe2CVHQAAAaCSZcREREQUoIpaGbKjbvZhjoElImodJrAe2BSH+jPHwRIREVGgnBM4ASwhJiJqLSawHtjkRgksZyImIiKiADnHvwJAJUuIiYhahQmsB7WNEtgKJrBEREQUoJIqu/oze2CJiFqHCawHjUuIOQaWiIiIAuUsIZbAHlgiotYK2izEVqsV119/PWw2GxRFQVpaGrKysly2sdlsuPfee/H9998jPj4ezz//PM4777xgheQXG3tgiYiIqA2cqbLBqJfQNdrISZyIiFopaD2wJpMJb775Jj744ANkZ2djz549OHTokMs2mzdvRlxcHP7973/jj3/8I5555plgheM3q8xJnIiIiKj1zlTb0b2zCTFRepYQExG1UtASWEmS0KVLFwCALMuQZRlSk7VUt2/fjmnTpgEA0tLSsHfvXgghghWSX6zsgSUiIqI2cKbKhh5dTOhiMnAdWCKiVgrqGFhFUZCZmYnRo0dj9OjRuPTSS11et1gs6NWrFwDAYDAgNjYWpaWlwQzJZ+yBJSIiorZwpsqG7p2N6GLSo5I9sERErRK0MbAAoNfrkZOTg4qKCixevBg//fQTLrzwwlZ+poT4+M5tFKFnhqgK9edahwjJPrVMr9d1+HPgxHPhiuejAc9FA56L8Nu9ezdWrlwJh8OBWbNm4ZZbbnG73datW5GVlYV3330XQ4YMAQCsX78e7777LnQ6HR566CFcddVVoQy93SmptuOSxFhU1Mr4vao63OEQEUW0oCawTnFxcRg5ciT27NnjksCazWacPHkSiYmJkGUZZ8+eRbdu3bx+lqIIlJUF/8u/tKIGANA12oAzZ60h2aeWxcd37vDnwInnwhXPRwOeiwaRdC569owNdwhtTlEUrFixAhs2bIDZbMbMmTORnJyM/v37u2xXWVmJv//97y4VUkeOHEFubi5yc3NhsVgwf/58bN26FXq9PtSH0S4oDoHS6roSYodDcBInIqJWCloJcUlJCSoq6noxa2tr8cUXX6Bfv34u2yQnJ2PLli0A6u4AX3nllc3GyYaLs4S4excTznIMLBERRZD8/Hz07dsXffr0gclkQnp6OvLy8pptt2bNGtx8882IiopSn8vLy0N6ejpMJhP69OmDvn37Ij8/P5ThtytlNXY4BNCjsxFdojgGloiotYLWA3vq1CksW7YMiqJACIGJEyfimmuuwZo1azB48GCMHz8eM2fOxD333IOUlBR07doVzz//fLDC8ZtzGZ2eXUwoPmsNczRERES+s1gsSExMVB+bzeZmSej333+P4uJiXH311XjjjTdc3tu4R9ZsNsNisXjdX6iG97jft7bL1Ytr63pc+/SMRbUDqLIp6Nq1U4s37LV+XIFqj8fVHo8J4HFFmvZ6XO4ELYG96KKLkJ2d3ez5JUuWqD9HRUXhhRdeCFYIrWJT6hLYc2JM+Pl0VZijISIiajsOhwNPPvkknnjiiTb5vFAN73FH6+XqBcXlAIBoCOgdDggBnDxdic4m7yXZWj+uQLXH42qPxwTwuCJNezsub8N7QjIGNhLV1vfA9uhsQoVVhhBCM+XNRERE3pjNZhQXF6uPLRYLzGaz+riqqgo//fQTbrzxRgDA6dOnsWjRIqxdu7bF95J/SqrtAIBu9SXEAFBlk1tMYImIyL2gLqMTyWyyAwadhPhORigOgRq7o+U3ERERacCQIUNQUFCAwsJC2Gw25ObmIjk5WX09NjYW+/btw/bt27F9+3YMGzYMa9euxZAhQ5CcnIzc3FzYbDYUFhaioKAAQ4cODePRRDbnnBrRBh1i6pNWLqVDRBQ49sB6YFMciDLqEBtdd4rOWnm3lIiIIoPBYMDy5cuxYMECKIqCGTNmYMCAAS7zUHgyYMAAXHvttZg0aRL0ej2WL1/OGYhbwV4/JCnKoEMXU0MPLBERBYYJrAdW2YEogx6x9eU+Z2tlmGOjWngXERGRNiQlJSEpKcnlucbzUDT21ltvuTxetGgRFi1aFLTYOhJnD6zJoEOX+hvhVeyBJSIKGEuIPahLYBt6YCus9jBHRERERJHGrggAgEmvQ5eo+gSWPbBERAFjAuuBTXYg2qBDnLOEuJZ3S4mIiMg/VsUBCYBBJyGmvqqr0sZrCiKiQDGB9aBZCTF7YImIiMhPdtkBk0EHSZIaSoiZwBIRBYwJrAdWxQGTUacmsBW1LPchIiIi/9gUB0z6usutzs5JnKy8piAiChQTWA+cJcQxjSZxIiIiIvKHTXHAqK9bR96gkxBt0HEZHSKiVmAC64GzhFivkxBl0MGmcB1YIiIi8o+tflJIpy5RBk7iRETUCkxgPbApDQ2OQSepswgSERER+cqmCBj1jRJYk55jYImIWoEJrAdW2YEoY+MElj2wRERE5J9mPbAmPXtgiYhagQmsB84SYgAw6nWwO9gDS0RERP5pPIkTUF9CzDGwREQBYwLrgU12ILq+B9aolyCzB5aIiIj8VJfASurjGJMeleyBJSIKGBNYD6yNSn6Meh3HwBIREZHfbLKAqVEJcbRRj1o7b4oTEQWKCawHVqWhhNigk1hCTERERH6rW0an4XLLyHk1iIhahQmsG7JDQHGIJj2wbGyIiIjIP41XNQAAk4FVXURErcEE1g2bXJesRhmdkzhJkNnYEBERkZ9scpMeWD3Xliciag0msG6oCayzB1Ynwe5gY0NERET+sSkORDVKYE16lhATEbUGE1g3rPUNS3R9AmvgJE5EREQUgLoe2IZZiOt6YAWE4HUFEVEgmMC6YVV7YBtKiHm3lIiIiPxlUxwusxA7k1mZk0MSEQWECawbzhJik1pCrGNDQ0RERH6zKQImlxLiup9Z2UVEFBgmsG6oJcRG5yzE7IElIiIi/yj1qxq49sDW/cyJnIiIAsME1g2rrABoKCHmGFgiIiLyl/Pmd9NJnBq/RkRE/mEC64a7WYhZQkxERET+cM6pYWQPLBFRm2EC64ZVrktWG0qIdbxTSkRERH5p6IFtmIVYHQMr88Y4EVEgmMC64SwhNjWahZg9sEREROQP55waRn3zWYi5vjwRUWCYwLrhLOtxlhAbdOyBJSIiIv84e1mj3JYQ88Y4EVEgmMC64RwD6zoLMRsaIiKKHLt370ZaWhpSUlLw6quvNnv9nXfeQUZGBjIzMzFnzhwcOXIEAFBUVIShQ4ciMzMTmZmZWL58eahDbzfc9cA2lBDzxjgRUSAM4Q5Ai2rVSZz0EIpDLSEWQkCSpBbeTUREFF6KomDFihXYsGEDzGYzZs6cieTkZPTv31/dJiMjA3PmzAEA5OXl4YknnsAbb7wBADj//PORk5MTltjbE2f1lssyOoa66whO4kREFBj2wLrRbBbi+rulHAdLRESRID8/H3379kWfPn1gMpmQnp6OvLw8l21iYmLUn2tqaniDNgicsxCb3E3ixMouIqKAsAfWjcZjYGsBGHTONdsEjPowBkZEROQDi8WCxMRE9bHZbEZ+fn6z7TZu3IgNGzbAbrfjzTffVJ8vKirC1KlTERMTgzvvvBMjRozwuj+9XkJ8fOe2OwA/6PW6sO27JabfqwEAPeI7qzF2q5brXos2eo1by8fVGu3xuNrjMQE8rkjTXo/LHSawblhlB0x6Sb0bbVTvljoAMIMlIqL24frrr8f111+PDz/8EGvXrsVTTz2FhIQE7NixA926dcN3332HxYsXIzc316XHtilFESgrqw5h5A3i4zuHbd8tKSmrAQDYamxqjNYaKwCgtKLGa9xaPq7WaI/H1R6PCeBxRZr2dlw9e8Z6fI0lxG5YZYfreBV1ynuW+xARkfaZzWYUFxerjy0WC8xms8ft09PTsW3bNgCAyWRCt27dAACDBw/G+eefj6NHjwY34HbK7m0SJ5YQExEFhAmsG1bZgShDQ0+rUVc/BpYTLhARUQQYMmQICgoKUFhYCJvNhtzcXCQnJ7tsU1BQoP68c+dO9O3bFwBQUlICRalbD72wsBAFBQXo06dPyGJvT5ouywc0XkaH1xRERIFgCbEbNsWBqEYTLhj0DWNgiYiItM5gMGD58uVYsGABFEXBjBkzMGDAAKxZswaDBw/G+PHj8fbbb2Pv3r0wGAyIi4vDU089BQD4+uuv8cILL8BgMECn0+Gxxx5DfHx8mI8oMjkncXLtgXVeUzCBJSIKRNAS2JMnT+Lee+/FmTNnIEkSrrvuOsybN89lm3379uG2227DeeedBwBISUnB7bffHqyQfGZr2gPrLPdxsLEhIqLIkJSUhKSkJJfnlixZov780EMPuX1fWloa0tLSghpbR+FMUqP07npgeVOciCgQQUtg9Xo9li1bhkGDBqGyshIzZszAmDFjXNagA4ARI0Zg/fr1wQojIFbZoY57BQCjjj2wRERE5B+1B9bQ6JrCZWJIIiLyV9DGwCYkJGDQoEEA6taa69evHywWS7B216Zkh4DBzd1SjoElIiIiXzlvfLv2wLKEmIioNUIyBraoqAg//PADLr300mavHTp0CFOmTEFCQgLuu+8+DBgwwOtnhWKtOUmvQ5RRr66nFN+1bkrqqM5RHWZ9paY60tpSLeG5cMXz0YDnogHPBRFgVRyQAOh1DT2wOkmCQSexhJiIKEBBT2CrqqqQlZWFBx54oNkacoMGDcL27dvRpUsX7Nq1C4sXL8ann37q9fNCsdZcrVWu35cDZWXVsNbYAACl5dUoK4sK6r61qr2tLdUaPBeueD4a8Fw0iKRz4W2tOaLWsNcvy+dcV97JpNexB5aIKEBBXUbHbrcjKysLGRkZSE1NbfZ6TEwMunTpAqBusglZllFSUhLMkHwiOwQMusZjYLlmGxEREfnHpjjUdV8bM+ol2GQmsEREgQhaAiuEwIMPPoh+/fph/vz5brc5ffo0hKhLCvPz8+FwONTF08NJcQiXch8jl9EhIiIiP9mUuh7Ypox6HewOXlMQEQUiaCXE33zzDXJycnDhhRciMzMTALB06VKcOHECADBnzhxs3boV77zzDvR6PaKjo/Hcc881K7MJh6Y9sIb6HliZy+gQERGRj2yy67ryTia9xBJiIqIABS2BHTFiBH788Uev28ydOxdz584NVggBkx0O1wSWPbBERETkJ5si1JUMGjPqdbDJvKYgIgpEUMfARiql6RhYTnlPREREfrLJ7kuITQZO4kREFCgmsG7ITcfAOidx4ngVIiIi8pHnSZx0sDGBJSIKCBNYN2RFwOBm0XGZjQ0RERH5yOMkTjqJN8WJiALEBNYNRTQtIeYyOkRERFpUWm1TVzTQGpssYHIziZPRoIOdy+gQEQWECawbstJ0FmKOgSUiItKa4+U1GLN6B/YWlIY7FLc8lRCb9BJLiImIAsQE1o2my+ioPbAs9yEiItKMbwrLYVcEispqwx2KW55KiE16Hau6iIgCxATWDaXJJE56nQSdxDGwREREWpJ/vAIAUGmVwxyJezaZkzgREbU1JrBuNF0HFqhrbHi3lIiISDvyT9QlsGe1msB67IGVeFOciChATGDdaFpCDNSNg2UJMRERkTaU19hxtKQaAHC2VqMJrIceWINeBxtvihMRBYQJbBMOIeAQgEHnemrqemB5t5SIiEgL/nOyQv1Z0z2wbidx4jUFEVGgDOEOQGuU+l5WfbMSYgky75YSERFpQv6JCuh1Ei5MiNFwAitgMrhZRoezEBMRBYw9sE3I9QlsszGwOgmyg40NERGRFnx7vAIDE2JgjovWZAmx4hBQHMJjDyxLiImIAsMEtglnL6uhycLjBk7iREREpAlCCHxffBZDesUiLtqgyR5YZ4mwpwRWcQg4BK8riIj8xQS2CbWEWGpeQsxJnIiIiMLPKjtglR3oGROFuE5GTS6jY5XrE1g3sxA7b5LzxjgRkf84BrYJZ5lw0x5Yo44TLhAREWlBrb2uPe5k1MGh0+GsVYZDCOik5uNNw6WhB7Z5TM5eWbviQJSbBJeIiDzjt2YTHsfAchInIiIiTaiRFQBAtFGPuE4GOARQbVPCHJUrq+K5B9ZYn8ByIiciIv8xgW1C9jALsUGvg52TOBERUYTYvXs30tLSkJKSgldffbXZ6++88w4yMjKQmZmJOXPm4MiRI+pr69evR0pKCtLS0rBnz55Qhu2TGntdstrJqEdctBEANFdGbJfrrifcj4Gtu8awybyuICLyF0uIm2jogW2yDqxOQi0bGiIiigCKomDFihXYsGEDzGYzZs6cieTkZPTv31/dJiMjA3PmzAEA5OXl4YknnsAbb7yBI0eOIDc3F7m5ubBYLJg/fz62bt0KvV4frsNppqZRCbExqu5SpqJWRmJcOKNyZfU2iZPBWULMyi4iIn+xB7YJzyXEHANLRESRIT8/H3379kWfPn1gMpmQnp6OvLw8l21iYmLUn2tqaiDVjx/Ny8tDeno6TCYT+vTpg759+yI/Pz+k8bektnEPbKe6HlitzUTsvGYweikhZmUXEZH/2APbhOKhhNiol9TkloiISMssFgsSExPVx2az2W0SunHjRmzYsAF2ux1vvvmm+t5LL73U5b0Wi8Xr/vR6CfHxndso+pbpLJUAgJ7dusBkrEsGhUEf0hhaYiqrBQB079qpWVzxsdEAgKhOUR5j1ut1mjqettIej6s9HhPA44o07fW43GEC24SnHlgDZyEmIqJ25vrrr8f111+PDz/8EGvXrsVTTz0V0OcoikBZWXUbR+fZ72U1AADZakO3Ll0AACfPVIU0hpaUltfFaK22NYvLZrUDAErKqlHW2f2lWHx8Z00dT1tpj8fVHo8J4HFFmvZ2XD17xnp8jSXETciKh2V09BLHqhARUUQwm80oLi5WH1ssFpjNZo/bp6enY9u2bQG9NxzcTeKktRJi58oF7kqI1UmceGOciMhvTGCbUITnZXTYA0tERJFgyJAhKCgoQGFhIWw2G3Jzc5GcnOyyTUFBgfrzzp070bdvXwBAcnIycnNzYbPZUFhYiIKCAgwdOjSU4bfIOQY22qBDbJQBErQ3C7EzOTXqvK8DS0RE/mEJcRPOO6bNx8DqOAaWiIgigsFgwPLly7FgwQIoioIZM2ZgwIABWLNmDQYPHozx48fj7bffxt69e2EwGBAXF6eWDw8YMADXXnstJk2aBL1ej+XLl2tqBmKg8SzEeuh0ErpE6VFRq60EVp3Eyc0sxOokTqzsIiLyGxPYJjwto2PQsYSYiIgiR1JSEpKSklyeW7JkifrzQw895PG9ixYtwqJFi4IWW2vV2BVIAKLqy3Njowya64F1Xk8Y9c17YJ3PsQeWiMh/LCFuQuEyOkRERJpWY1cQbdSpS//ERhk01wNrk1vugbXxxjgRkd+YwDYhe1lGx84SYiIiorCrtTvQydhQ1hwbrb0eWLuXHliTmsDyxjgRkb+YwDbhaRkdo04HxSHgEExiiYiIwqmuB7ZRAhtlwFmrEsaImnNWbZnc9MCaWEJMRBQwJrBNeCohdi6rI7Pch4iIKKxq7Ao6GRsuYepKiO1hjKg557wZTa8nAE7iRETUGkxgm5AddXdD3c1CDAB2B++WEhERhZP7EmJt9sC6GwNrMnAZHSKiQDGBbUL2cMfU+Zh3S4mIiMLLXQlxtV3R1HJ3doeAXmp+QxxoWBuWY2CJiPzHBLYJpX6Mq6HJHVOjWkLMxoaIiCicauwKOhlcS4gBoFJDMxHbZUezawknvU6CBM5CTEQUCCawTag9sFLzSZwAcCZiIiKiMKuVm5cQA8BZDc1EbHcItzMQA4AkSTAZdLDLvClOROQvJrBNqLMQ691P4sQSYiIiovCqm8SpIYGNidJgAqs43M5A7GTUSywhJiIKABPYJjwuo6PnhAtERERaUDcGtuESxjkjca2snYmc7IrD7QzETia9TlNjdomIIgUT2Cacy+g0m4VYx2V0iIiItKCmySzEzp9r7Nq5yWxXhNsZiJ0MOgk2lhATEfktaAnsyZMnccMNN2DSpElIT0/Hm2++2WwbIQQef/xxpKSkICMjA99//32wwvGZcxkdjz2wXEaHiIgobOyKA4pDuCSw0Ya6n612LfXACq8lxCaDjiXEREQBMATrg/V6PZYtW4ZBgwahsrISM2bMwJgxY9C/f391m927d6OgoACffvopvv32Wzz66KPYvHlzsELyiVw/7b0kcQwsERGR1tTUJ6mNS4ij1RJi7SSEdsXRbD6Nxox6Ha8piIgCELQe2ISEBAwaNAgAEBMTg379+sFisbhsk5eXh6lTp0KSJAwbNgwVFRU4depUsELyieIQ7tdsUxNY7TSOREREHY2zTNi1B7Y+gdVSD6zD4bWE2KRnDywRUSCC1gPbWFFREX744QdceumlLs9bLBYkJiaqjxMTE2GxWJCQkODxs/R6CfHxnYMWq96oh1GvQ3x8Z+jr/w8A3c7aAABRnUxB3b9WNT4XHR3PhSuejwY8Fw14LihYnD2wLgls/c9a6oG1KQImLz2wJr3EeTWIiAIQ9AS2qqoKWVlZeOCBBxATE9Pqz1MUgbKy6jaIzL2qajv0OgllZdWIj++s7stabQUAlFXUBHX/WtX4XHR0PBeueD4a8Fw0iKRz0bNnbLhDID/UqglsoxJitQdWOwmsrDhg8LqMDntgiYgCEdRZiO12O7KyspCRkYHU1NRmr5vNZhQXF6uPi4uLYTabgxlSixQhoJea3zE1qMvo8G4pERFRuDhLiKMb9cAa9DrodZLGltER6goG7hj1EoclEREFIGgJrBACDz74IPr164f58+e73SY5ORnZ2dkQQuDQoUOIjY31Wj4cCrIi3E66oC6jwzXbiIiIwsZdCTFQ1wurpR5Ym+LwOgtxXQ8srymIiPwVtBLib775Bjk5ObjwwguRmZkJAN4ZVHsAACAASURBVFi6dClOnDgBAJgzZw6SkpKwa9cupKSkoFOnTli1alWwwvGZ7HC/8Li6jA7vlhIREYWNuxLiusd6NbnVAlkR6gSQ7nASJyKiwAQtgR0xYgR+/PFHr9tIkoRHHnkkWCEERHYIDwls/SzE7IElIiIKG3ezEAN1S+loaRInu6OlMbASZCawRER+C+oY2EjkcRkdXd2pYmNDREQUPg3rwDYtIdZrahkdm+xoYRZilhATEQWCCWwTdT2wzU+LQV0Hlo0NERFRuNR4KCHWWg+s7BBe14E16nUclkREFAAmsE14LiHmGFgiIqJwc07UFG1oPomTVUM9sHbF/fWEk1EvcQwsEVEAmMA2IXsoIXY2QhwDS0REFD41dgVRBl2ztjraqNdUD2xLsxCb9DpWdRERBSBokzhFKk89sHqdBL3EMbBERBQZdu/ejZUrV8LhcGDWrFm45ZZbXF7fsGEDNm/eDL1ej+7du2PVqlU499xzAQAXX3wxLrzwQgBAr169sG7dupDH70mNXUG0oXliqLVldGTF4XUWYqOBJcRERIFgAtuE4nC/DixQt1A675YSEZHWKYqCFStWYMOGDTCbzZg5cyaSk5PRv39/dZuLL74Y7733Hjp16oR//OMfePrpp/GXv/wFABAdHY2cnJxwhe9VjexoNoETAERpaBkdxSGgCHidhdikl+AQnm+cExGReywhbkJWBPSS+4bEqJdYQkxERJqXn5+Pvn37ok+fPjCZTEhPT0deXp7LNldeeSU6deoEABg2bBiKi4vDEarfau1KswmcgPoeWI2UEDt7VlsqIW68LRER+YY9sE3IDgcM+uZ3doG6pXTY0BARkdZZLBYkJiaqj81mM/Lz8z1u/+6772LcuHHqY6vViunTp8NgMOCWW27BhAkTvO5Pr5cQH9+59YH7QBZATLRR3Z9er0N8fGfEx0TBKjtCFoc3Z2vtAIC4mCiP8cTGRAEAOnWJQnxnU7PXncfV3rTH42qPxwTwuCJNez0ud3xKYG+//XbMnDkT48aNg87NEjPtieJhGR3Aueg4e2CJiCh0gt0G5+Tk4LvvvsPbb7+tPrdjxw6YzWYUFhZi3rx5uPDCC3H++ed7/AxFESgrq27z2NypqLHDqJPU/cXHd0ZZWTUkxYEau4LS0ipIHiqpQqW02gYAsFtlj+dFsckAgN9LqoD6nxtzHld70x6Pqz0eE8DjijTt7bh69oz1+JpPLeEf/vAHfPjhh0hNTcUzzzyDX3/9tc2C0xpPsxAD9WNgHeyBJSKi0AmkDTabzS4lwRaLBWazudl2X3zxBdatW4e1a9fCZDK5vB8A+vTpgyuuuAKHDx9ugyNpGx5LiOvHxVo1UEZsq7/ZbfI2iVN9CbGNN8aJiPziUwI7evRoPPvss9iyZQvOPfdczJ8/H7Nnz8Z7770Hu90e7BhDyttkCkadxEmciIgopAJpg4cMGYKCggIUFhbCZrMhNzcXycnJLtscPnwYy5cvx9q1a9GjRw/1+fLycthsdT2IJSUlOHDggMvkT+FWY1fQyc0kTs6ZibUwDtY53MjIMbBERG3O5zGwpaWl+OCDD5CTk4OLL74YU6ZMwTfffIPs7Gy89dZbwYwxpLwmsHqOgSUiotDztw02GAxYvnw5FixYAEVRMGPGDAwYMABr1qzB4MGDMX78eKxevRrV1dVYsmQJgIblcn755Rc88sgjkCQJQgjcfPPNmkpgrbLD/TI69b2ytXYF6GQMdVgunDe7vS2j4+yd5Y1xIiL/+JTALl68GEePHkVmZibWrVuHhIQEAMCkSZMwffr0oAYYaoqXEmKjXoLMWYiJiCiEAm2Dk5KSkJSU5PKcM1kFgL/97W9u33fZZZfhww8/bH3gQWKVHTC5XQe2rlc2UnpgDWoJcfjjJSKKJD4lsNddd12zRtBms8FkMuH9998PSmDh4q0H1sBZiImIKMQ6UhvsC5viQJTBTQlx4x7YMHMuuedbDyyvK4iI/OHTGFjnwuaN/c///E+bB6MFsuLwUkLMMbBERBRaHakN9oVVdrhdX1XtgbWHPyG01/cCG73MGm1kDywRUUC89sCePn0aFosFtbW1OHz4MISoS94qKytRU1MTkgBDTRHeS4irbWxoiIgo+DpiG9wShxCwKwJRhubttNoDK2uhB7Y+gXUTp1PDJE68MU5E5A+vCexnn32G999/H8XFxXjiiSfU57t06YKlS5cGPbhwkBVv68DqIDuar9VGRETU1jpiG9wSW33PptseWKOGemCdkzh56YHlLMRERIHxmsBOmzYN06ZNw9atW5GWlhaqmMJKdggYPIxZMXAZHSIiCpGO2Aa3xFluGxUxy+h4WQe2vneW68ASEfnHawKbk5ODzMxMHD9+HBs2bGj2+vz584MWWLh4m4WYkzgREVGodMQ2uCXW+uQ0yk1i2NADq4ESYnUZHS9jYHXsgSUiCoTXBNY5xqa6ujokwYSb4hAQgPdJnLiMDhERhUBHa4N94Uxg3S+jo6EeWEfLy+g4e2dtGoiXiCiSeE1gZ8+eDQC4/fbbQxJMuDnXePWWwMq8U0pERCHQ0dpgXzhLiL2Nga3RQg+s7MMyOvUJN2+MExH5x6dldFavXo3KykrY7XbMmzcPV155JXJycoIdW8gpLSawOo6BJSKikOoobbAv1BJiN+vAmvQSJERODywncSIiCoxPCeznn3+OmJgY7Ny5E+eeey7+/e9/44033gh2bCEn1zc4nsfASmqjREREFAodpQ32hU1NYJu305IkIdqo09YYWA/XE0CjdWA1kHATEUUSnxJYRalrDHbu3ImJEyciNjY2qEGFS0MJsedldNgDS0REodRR2mBfeBsDCwDRBr26TTg1zELc8hhYXlcQEfnHpwT26quvxsSJE/H9999j1KhRKCkpQVRUVLBjCzm1hNjDmBWOgSUiolDrKG2wL9RldNyUEAPQXA+sycsYWJ0kQa+T1GMiIiLfeJ3Eyenuu+/GggULEBsbC71ej06dOuGVV14Jdmwhp/bASh4SWJ0OivC+1A4REVFb6ihtsC8altHx0ANr1GtjDKzifUiSk0nP9eWJiPzlUwILAL/++iuOHz+uljIBwNSpU4MSVLjIivceWOfzMhNYIiIKoY7QBvui5RJiHWrt4U9gbYqAUS9B8nBD3Mmk5/ryRET+8imBveeee1BYWIiLLroIen1d2Y4kSe2u8fRlFmKg7s5qlIfGk4iIqC11lDbYF84JjzyV5tb1wIa/hFh2ONwu9dOUUa9jCTERkZ98SmC/++47/Otf/2rxTmKkc5YQe+pddc4mKLPch4iIQqSjtMG+cCZ70Z7GwBp0KK+VQxmSW3ZFeLwZ3lhdCTETWCIif/jUjThgwACcPn062LGEnXMZHc89sJLLdkRERMHWUdpgX7RYQmzUa2ISJ5vi8DoDsZNBr4ONN8WJiPziUw9saWkp0tPTMXToUBiNRvX5devWBS2wcFBaWEbH4CwhdrCxISKi0OgobbAvfBoDq4FJnGTF4XUGYieOgSUi8p9PCewdd9wR7Dg0oaGE2P3rXLONiIhCraO0wb6wKQ7oJc+VUlpZRsemCPWmtzdGzkJMROQ3nxLYK664AsePH8dvv/2G0aNHo6amxmUmxPZCbqEH1qhrmMSJiIgoFDpKG+wLq+zwuAYsUDc21qqBHli74lBventj4iRORER+82kM7KZNm5CVlYXly5cDACwWCxYvXhzUwMJBXUanpTGwvFtKREQh0lHaYF9YZYfH8mEA6FTfAytEeNtp2SF8m4XYwBJiIiJ/+ZTAbty4Ee+88w5iYmIAAP/1X/+FkpKSoAYWDrLwPgtxwxhYNjZERBQaHaUN9oVN9j62NNqohyIaKqrCxSY7PFZzNWbUSZzEiYjITz4lsCaTCSaTSX0syy1PUX///fdj1KhRmDx5stvX9+3bh//+7/9GZmYmMjMz8dJLL/kYcvCoPbAeGkfnMjocr0JERKESSBvcXtkUB6KNnkuInWu014R5HKzdIXwuIWYPLBGRf3waA3v55Zdj3bp1qK2txeeff45//OMfSE5O/v/t3XlgVOW5P/DvmS2ZyTKTdRIgCfsmAVxQFGtqIMRroCiE3tpiW6u9VysKxWutvUqV9qrdUKy/q1Irtmq9bbGUatqigiXgBioY2VSWgQSSAZJMtslsZ87vj1mSkMzMmT3L9/MXmTkz856TMOc853ne5w36mqVLl2LFihW47777Am5z2WWX4dlnnw1vxHEkhlxGh3NgiYgosSI5Bw9Xdpc7aGmu1hvcdjvdyExN1Kj6c4lupGlCX2KpGcASEYVNVgb2v/7rv5CdnY3Jkyfjj3/8I8rKyrB69eqgr5kzZw70en1MBpkoPV2Ig8+B5TI6RESUKJGcgwGgtrYWlZWVqKiowMaNG/s9v2nTJlx//fVYvHgxvvWtb+H06dP+57Zs2YKFCxdi4cKF2LJlS0z3Jxqh5sDqfAGsI7kZWE8XYhkZWBVLiImIwiUrA6tQKLBgwQIsWLAA2dnZMfvw/fv34ytf+Qry8/Nx3333YdKkSTF770jI7ULs4t1SIiJKkEjOwaIoYt26ddi0aROMRiOqq6tRXl6OiRMn+reZNm0aXn31VWi1WvzhD3/AL37xCzzxxBOwWCx46qmn8Oqrr0IQBCxduhTl5eWD4qa0Q3QjJUhgqNV4AlhrskuIRbe/aisYZmCJiMIXNICVJAlPPfUUXnrpJX9HP4VCgRUrVmDlypVRffBFF12EHTt2IC0tDTt37sSdd96JN954I+TrlEoBBoMuqs8ORJPqWSA+26CFwaCDUqno81nZdtG7nSZuYxisLjwWIxmPRV88Hj14LHrwWEQvmnNwXV0dSkpKUFRUBACoqqrC9u3b+wSwc+fO9f979uzZ+Nvf/gYA2L17N+bNmweDwQAAmDdvHnbt2hWwp0UiOVzBS3P9GdgkB7CeLsShM7BqhQDHIFj2h4hoKAkawL7wwgv4+OOPsXnzZv9JsL6+Hg899BBeeOEFfPvb3474g33dFAGgrKwMDz/8MFpaWkLeXRZFCRaLNeLPDaaj0w4A6Oq0wyIABoOuz2d1Wx0AAEtHd9zGMFhdeCxGMh6Lvng8evBY9BhKxyIvLyPZQxhQNOdgs9mMgoIC/89GoxF1dXUBt9+8eTOuueaagK81m81BxxrPm8u9OSUgTavu81m9b5bkdXrO0wqNKqk3UFxuCboLxjmQjDQNXG5pwO2G602g4bhfw3GfAO7XUDNc92sgQQPYrVu34vnnn+8TVBYVFeEXv/gFvvOd70QVwJ47dw65ubkQBAF1dXVwu93IysqK+P1ioaeEOMQcWM5XISKiOIvnOfjCzzlw4ABeeumliN8jnjeXe+u2uyC4+35W75slot0JADjXak3qDRS7yw3J5Q45BrfLDYc48HZD6SZQOIbjfg3HfQK4X0PNcNuvYDeXgwawLpdrwIxodnZ2yDb+a9aswZ49e9Da2oprrrkGd911l/81N910E7Zt24ZXXnkFSqUSqampWL9+PQQhdLlNPIUMYL2Pcw4sERHFWzTnYKPRiKamJv/PZrMZRqOx33bvvvsunnnmGbz00kv+pXqMRiP27NnT57WXX355pLsRUw7RjRRV4GuFQVNCLAbvluyjUQpwihIkSUr6NRAR0VARNIBVq9URPQcA69evD/r8ihUrsGLFiqDbJJroDr4OrMq/jA4zsEREFF/RnINLS0thMplQX18Po9GImpoa/OpXv+qzzaFDh7B27Vo899xzyMnJ8T9+9dVXY/369WhrawPgmRO7Zs2aKPYkduwuN1JUgdeB7b2MTjI5RLesLsTqXtcVmiCBORER9QgawB45cgSXXHJJv8clSYLD4YjboJLFl1lVBrgLymV0iIgoUaI5B6tUKqxduxa33XYbRFHEsmXLMGnSJGzYsAEzZszA/Pnz8fOf/xxWqxWrVq0CABQWFuKZZ56BwWDA9773PVRXVwMA7rzzTn9Dp2QLtYyOdpBkYJ2iJKsLsS9L6xCD7xcREfUIGsAePnw4UeMYFPwlxAFOOr5ldNjynoiI4i3ac3BZWRnKysr6POYLVgFPk6hAqqur/QHsYOIIUZqrVgpQKgRYk7gOrCRJ8rsQ+3tr8LqCiEgu3u7rxVdCHOic4ysHcrGEmIiIKKHckgSnKAWdAysIAnRqZVIzsL5pRnLXge39GiIiCo0BbC8utwSlQgjYSEEheO7sOt28U0pERJRIvvVSg82BBQCtWpHcANZ7jRBuCTEREcnDALYXl1sK2IHYR60QeKeUiIgowezeADbUXFGtWgmrI3kBodPlzcCGuJ4AuDwfEVEkGMD24nJL/pNJIGqlgnNViIiIEsyXpUwJcZ7WaZJbQmwX5QXaADOwRESRYADbi1N0+xs1BaJSCP5mT0RERJQYdtklxMkNYHtKncOZA8sAlohILgawvbhEORlYgScaIiKiBAunhHhQZGBlNXHyXHMwA0tEJB8D2F6cbnfAJXR8VEoF56oQERElmENmYOiZA5v8DGw4JcS8riAiko8BbC9OUQrZdIFNnIiIiBKvpzQ31BzY5HYh9o9TTgZWxRJiIqJwMYDtxSm6Q7a9VysVcHEZHSIiooQKbw5s8s7T4TVx8pUQ88Y4EZFcDGB7kdeFmBlYIiKiRAtrGR2nCElKzrk6nBJifxMnF2+MExHJxQC2F6fohipkF2Iuo0NERJRoPcvoBD9P6zRKiG4paTeb5Y4TYBMnIqJIMIDtxSm3CzGX0SEiIkqocDKwAJI2D1buOIFeTZx4XUFEJBsD2F48c2BDB7Au3iklIiJKKLnrq2rVnueTFcD6S4hDXE8ALCEmIooEA9hePBnY0E2cOAeWiIgosewyu/v6MrDWZAWw3muEUIE20JOBZQkxEZF8DGB7cbrdUIVYRkelEOBiqQ8REVFCOWR299VpvCXESVoLVu44gZ4sLW+MExHJxwC2F/kZWN4pJSIiSqTw58Am51wdzjqwSgWbOBERhYsBbC8umXNg2WyBiIgosRyiG0qFELJSKtklxHbRDYXQE5wGIwgCNFyej4goLAxge3G6JahDLKOjVijYxImIiCjB7C63rKymTp3kEmKXGxqlAoIQOoAFPBllZmCJiORjANuLU5SgCpGBVfFOKRERUcLZXW5Z80q1muQuo+NwuWU1cPJJUSlhdyVnrEREQxED2F6cotvfETAQtVIBp5t3SomIiBJJbmCoGwQlxHICbZ8UlcI/v5eIiEJjANuLyy2FngOrYAaWiIgo0ewyA9jBsA5sqJvhvaWoFLAlqeEUEdFQxAC2F6fohipkBlbgHFgiIqIEc8iokgIAlVIBtVJIXhfiMDOwqczAEhGFhQGsl+iW4JY8GdZgVEoFRMmzPRERESWGTeYcWMBTRpysJk5ym035eAJYzoElIpKLAayXb23XkOvAegNcFwNYIiKihLG73EiVGcCmqpVJmwPrCCPQBjxNnGzMwBIRycYA1ssXkIZeB9ZzyJwsIyYiokGstrYWlZWVqKiowMaNG/s9v3fvXtx4442YPn06/vnPf/Z5btq0aViyZAmWLFmC22+/PVFDDsrmFJGqDiMDm6wANtwSYjVLiImIwqFK9gAGC19Aqgq1Dqw3wHWxkRMREQ1Soihi3bp12LRpE4xGI6qrq1FeXo6JEyf6tyksLMSjjz6K559/vt/rU1NTsXXr1kQOOSRPEyelrG21muQFsHaXG2ka+ZdX7EJMRBQeBrBevs7CoTKwviZPXEqHiIgGq7q6OpSUlKCoqAgAUFVVhe3bt/cJYMeMGQMAUIS4cTtY2MIoIdapFUmbAxtuBtbThZhzYImI5BoaZ60E8AWkcpbRAcCldIiIaNAym80oKCjw/2w0GmE2m2W/3m63Y+nSpfjqV7+Kt956Kx5DDJvcZXQA3xzYJHUhdrmhCXEt0VuKSskMLBFRGJiB9fJnYEOWEHMOLBERDW9vv/02jEYj6uvr8a1vfQuTJ09GcXFxwO2VSgEGgy6uY7K73NCnp/T7HKVS0e8xfZoGp9tscR/TQJxuIEOnkf3Z+vQU2EW3rP0aDobjfg3HfQK4X0PNcN2vgTCA9XLJLCH2Pe9kF2IiIhqkjEYjmpqa/D+bzWYYjcawXg8ARUVFuPzyy3Ho0KGgAawoSrBYrJEPWAabU4Qguvt9jsGg6/eYCkCHzRX3MQ3E5hSBAcYZkEuEzelGa2sXBKHnGmSg/RoOhuN+Dcd9ArhfQ81w26+8vIyAz7GE2MtXQqwKsYyOr8mTixlYIiIapEpLS2EymVBfXw+Hw4GamhqUl5fLem1bWxscDgcAoKWlBR9//HGfubPJ4BLdcLkl2SXEGSkqdNhdkKTE32wOfxkdz7YsIyYikocZWC+5TZx8zzs4B5aIiAYplUqFtWvX4rbbboMoili2bBkmTZqEDRs2YMaMGZg/fz7q6uqwcuVKtLe34+2338avf/1r1NTU4NixY/jxj38MQRAgSRK++93vJj2A9a2TmqqW14U4S6uG3eWGzeWGVuZrYsUuyp+rC/Tsk93llr1/REQjGQNYL9+c1lBzYH0nJQfvlBIR0SBWVlaGsrKyPo+tWrXK/++ZM2eitra23+suueQSvPbaa3EfXzh82Um5gaFBqwYAtFqd0OoTFxS63BJEtwRNiGqu3piBJSIKD0uIveTOgfW18Le52PKeiIgoEXznXLnL6Bh03gC22xm3MQ3EdzM8nAxsiv+6ggEsEZEccQtg77//flx55ZVYtGjRgM9LkoSf/vSnqKiowOLFi3Hw4MF4DUUWuXNgU7zlPbYktecnIiIaaewRlBADgMWa2ADWN051GBnYnhJi3hgnIpIjbgHs0qVL8dxzzwV8vra2FiaTCW+88QZ+8pOf4KGHHorXUGTpWUZHXgaWpT5ERESJ4btpLDezmeXNwFoSnIH1TS9iEycioviJWwA7Z84c6PX6gM9v374dN9xwAwRBwOzZs9He3o6zZ8/Gazgh+efAhrhr6rtTyhJiIiKixAi7hFibnBJih6+EOJwMrK+EmJVdRESyJK2Jk9lsRkFBgf/ngoICmM1m5OfnB31dvBZL16R6TnbZWTr/+w+0ILDSu52gUo6YxYKBkbU4cig8Fn3xePTgsejBY0GxFG4JcZpGCZVCQGuSSoiZgSUiip8h14U4Xoult3XYAADdnTZYvOedgRYE9q3/2tphG1aLBYcy3BZHjgaPRV88Hj14LHoMpWMRbLF0GhzCLSEWBAFZOjUs3Y54DqsfXwY2nC7EqSrOgSUiCkfSuhAbjUY0NTX5f25qaoLRaEzWcPzruoYqIVYpFVAqBJb6EBERJYg/AxtGZtOgVcPS7YrXkAbk8C/3E7yfRm/sQkxEFJ6kBbDl5eX461//CkmSsH//fmRkZIQsH46nnjmwoU86qSoFTzREREQJYnN6spPhLE9j0KqHVAkxryuIiOSJWwnxmjVrsGfPHrS2tuKaa67BXXfdBZfLcyf0pptuQllZGXbu3ImKigpotVo88sgj8RqKLC6ZGVjAMweHpT5ERESJYQtzDizgWUrncHtHvIY0oIiaOKk5B5aIKBxxC2DXr18f9HlBEPDjH/84Xh8fNt86sKGW0QE8d0tZQkxERJQYkZQQZ+nUie9CHFEG1re+PG+MExHJkbQS4sHGtw6sUkYAyxJiIiKixIm0hLjTLvqbLyaCPYImThqlAAHMwBIRycUA1sspSlArBQiCjABWreSdUiIiogSxu9xIUSlknaN9fGvBWhKYhe1p4iT/8koQBGhUCgawREQyMYD1crndUCvkHQ5mYImIiBLH5nKHVT4MeEqIASS0jNju8lRzhVNCDHivK3hjnIhIFgawXr4MrBypap5oiIiIEsXuEsPKagI9GdhEdiKOZB1YwJOxZQaWiEgeBrBeTtEtqwMx4Fl0nCcaIiKixLA53WF1IAZ6MrCDvYQY8K1uwOsKIiI5GMB6Od3yM7ApLCEmIiJKGJt3Dmw4kjEH1i66IQBQyWgI2RszsERE8jGA9XKFk4FlCTEREVHC2F0iUlXhZWD1qWoISHAJscsNTZjNpgBfbw1eVxARycEA1sspSrLvmLKEmIiIKHFsTjdS1OFdsigVAjJTVQlt4uSIIFMMMANLRBQOBrBeYc2B9WZgJUmK86iIiIjIHkEXYsAzDzbRJcThNnACgBTeGCciko0BrFc4c2BTVUqIEuByM4AlIiKKN1sEJcSAZx5sSwJLiJ2iO+wldADfjXEGsEREcjCA9XKJbqjllhB7y5h4t5SIiCj+IikhBoAxBi1OtljjMKKBOVxupESUgVXAzjmwRESyMID1cooSVDJPOr75LWzkREREFH+RlhBPyktDi9WJ5i5HHEbVn90VWQaWqxsQEcnHANYr3BJiADzZEBERJYDNJUbUHGlibhoA4Oj5rlgPaUCOCOfAsjkkEZF8DGC9nKIbaoX8Jk4AOF+FiIgoziRJ8mRg1eHPgZ2U5wlgvziXoADW5UaKKrwldABmYImIwsEA1sslRpKBZQkxERENTrW1taisrERFRQU2btzY7/m9e/fixhtvxPTp0/HPf/6zz3NbtmzBwoULsXDhQmzZsiVRQx6QU5TglhBhF2INctM0OHquMw4j688uShGXEItuCS6RQSwRUSiqZA9gsHC63bLnwDIDS0REg5koili3bh02bdoEo9GI6upqlJeXY+LEif5tCgsL8eijj+L555/v81qLxYKnnnoKr776KgRBwNKlS1FeXg69Xp/o3QDQc7M4khJiwJOFTWQGNqISYnXP1KT0CF5PRDSS8FvSyylK8rsQq9iFmIiIBq+6ujqUlJSgqKgIGo0GVVVV2L59e59txowZg6lTp0JxwfSZ3bt3Y968eTAYDNDr9Zg3bx527dqVyOH34TvXRlJCDHgC2BMt1oRkNx2iO6JAO4XXFUREsjGA9XKKbqhldyFmCTERNBLbWwAAIABJREFUEQ1eZrMZBQUF/p+NRiPMZnPcXxsPvmqnSEqIAWBiXhqcogRTa3cshzUge4QZWAawRETysYTYyxVOF2KWEBMREfkplQIMBl1c3tts89wsztZrB/wMpVIR9LMvHZ8L4DOc6XLisonxGaOP0y0hXacJ+1hkZ2oBAGptz2tD7ddQNRz3azjuE8D9GmqG634NhAGsl1N0QyW3C7FvHVhmYImIaBAyGo1oamry/2w2m2E0GmW/ds+ePX1ee/nllwd9jShKsFiskQ02hHOtnvmrosM14GcYDLqgn52tEqBWCvjkZAuuKTHEZYw+XXYXlFL4x0J0uAAA51u7kKfxXGOE2q+hajju13DcJ4D7NdQMt/3Ky8sI+BxLiL2c4XQh9jVbYAaWiIgGodLSUphMJtTX18PhcKCmpgbl5eWyXnv11Vdj9+7daGtrQ1tbG3bv3o2rr746ziMOzD8HNsISYpVSgYm5aThkjm8nYpfbs9yPThP+XF1/bw1eVxARhcQMLDxrzIVVQswMLBERDWIqlQpr167FbbfdBlEUsWzZMkyaNAkbNmzAjBkzMH/+fNTV1WHlypVob2/H22+/jV//+teoqamBwWDA9773PVRXVwMA7rzzThgM8c1cBhPtHFgAmFGYiZqDZohuCUqZDRvDZfVmUdMiCGBTeF1BRCQbA1h4sq8AZDdxUikVUApstkBERINXWVkZysrK+jy2atUq/79nzpyJ2traAV9bXV3tD2CTrWcZnci6EANA6agM/Hn/GRw734XJ+emxGlofXQ7POCMJYH29NXhdQUQUGkuI4VkDFgBUYdyVTVUrWUJMREQUZz3L6ER+yVJamAkA+LSxPSZjGkhPABt+bsAXnDOAJSIKjQEsws/AAp5yH5b6EBERxZfN6TnXRlNCPFqfimydGp+eiWMAa/eUEEcyB9ZfQswb40REITGABfyLm8udAwswA0tERJQINm9WMpoSYkEQUFqYiU8bO2I1rH6szshLiLXe7HK3kzfGiYhCYQALz7ptAKCWuYwO4LkTbGOpDxERUVzFooQYAEpHZeJUazcs3c5YDKsfaxQlxBkpnte0e7O4REQUGANY9JQQq8LOwPJOKRERUTzZnCIUQnh9KgZSOsqzpuCBOM2D7bJ7A9iU8DOwKqUCOrUSHTYGsEREoTCABeD0lxCHNweWzRaIiIjiy+ZyI0WlgCBEF8BON2ZAIQCHmuJTRtzlvamtU0dW6pyRqmIGlohIBgawAFy+Jk7hdCGOoIS40+7CBydbw3oNERHRSNblECMqy71QqlqJwsxUnGzpjsGo+vM1cYpkDiwAZKaqmIElIpKBASx6ltEJJwMbSQnxlrpGrNz8KUwt1rBeR0RENFJ12V1Ij6AsdyAl2dq4nYOtDhEpKgVUYVxL9JaRokKHLT7zc4mIhhMGsOi9jE58M7ANFhsA4F9fnA/rdURERCNVp11Eekr0GVgAGJutw6nWbrglKSbv11uXQ4y4fBjwZGBZQkxEFBoDWEQ2BzZVrQg7A3um3RvAHm0O63VEREQjVafDhfQYlBADQEmWFjaXG2c77DF5v966HK6IGjj5eDKwDGAHA0mS8Pz7p+I2X5qIosMAFr2W0QkrA6sMu4lTkzeAPdjUAXMcTp5ERETDTYfNFbMMbEm2DgDiMg82+gysGm0MYAeFj+rb8PQ7Jjz//qlkD4WIBsAAFoDLl4ENYx3YlDBLiCVJQmO7HfPGZQMAdh5lGTEREVEonQ4xhnNgPQFsPObBWh0i0qIItDNTVbC73HBwhYOk2/SBJ3B9/2Qrl0wkGoQYwCLSdWAVEN2SP/gNxdLthN3lxhVjszAuW4e3OQ+WiIgopE577DKwOTo10jRKnGyNfQbW6hAj7kAMeJbRAcB5sEl2sKkDe05ZcNW4LNhdbrxn4uoRRINNXAPY2tpaVFZWoqKiAhs3buz3/F/+8hfMnTsXS5YswZIlS/DnP/85nsMJKKIuxCrPSUpuFvZMu6dkeFRmCuYUG3DY3BnmKImIiEYWl+iG3eWOWQZWEASMzdbFJQPb5XBFFcBmeoN0zoNNrt/vqUdGigrr/m0qMlNVrJgjGoRic0tzAKIoYt26ddi0aROMRiOqq6tRXl6OiRMn9tnu+uuvx9q1a+M1DFki6ULsO0l1yLwz7Jv/WpCZitx0DbocImxOEalRzJchIiIazjrtnvLNWDVxAjxL6Xx4yhKz9/PpcojQxSIDy6V0kkaSJHxYb0H55FzotWpcPT4bu463wCW6I14eyedEsxVvHDkLCcCcYgMuLTLEZtBEI1DcMrB1dXUoKSlBUVERNBoNqqqqsH379nh9XFR8zZjCycBm6TQAgFarvBNNoz8Dm4psndrz2m6epIiIiALpdHiykbEqIQY8S+mc7XTA6ojt3EZPE6fo5sACnhvjlBwtVifabS6Mz/HMlf7yxFy021zYd7ot6vfe+K4Jz71/Cr99/xQee+uLqN+PaCSLWwbWbDajoKDA/7PRaERdXV2/7d544w3s3bsX48aNw/3334/CwsKg76tUCjAYdDEdq9vbvGl0Xga0ve6eKpWKgJ9VbHQAAOyCvPG0eDO1RQWZGONdD9ahCPz+g02wYzHS8Fj0xePRg8eiB48FxUKn3RfAxq5aqSRLCwA41WrFVGNGTN7T5ZZgd7mjXkYHANpZQpw0vtJyXwB7mTdLesTciTnFWVG9d92ZdiyckoeJeWn4390mtHU7odeqoxsw0QgVtwBWjmuvvRaLFi2CRqPB//3f/+G+++7D73//+6CvEUUJFkts5640t3dDKQC2Lhvs1p4yYoNBF/Cz1KLnzm39uU5YjOkhP8N0thMFGSmwWKxIlTwZ31PmDpSka2KwB/EX7FiMNDwWffF49OCx6DGUjkVeXmyCGIo9fwlxDDOwxb2W0olVAGv1ZoqjmQOrT/UEM5wDmzwnmj3fWWO9fyMZqSpk69RRN/0yd9hxttOB0lGZmJibBgA40NThX5mCiMITtxJio9GIpqYm/89msxlGo7HPNllZWdBoPAHc8uXLcfDgwXgNJyirQ4ROo4IgyJ8Dm5MWXglxU4cdhZkpAIBs72tbrI4wR0pERDRydNhjX0I8Rp8KADjj7U0RC75y5GgC2HR2IU46U4sVOrUSxowU/2PFWVqcijKAPdDYDgAoLczA9IIMKATgwJn2qN6TaCSLWwBbWloKk8mE+vp6OBwO1NTUoLy8vM82Z8+e9f97x44dmDBhQryGE1RnBK3vtWolUlUK2UFoY7sNhZmek2aWt2SkRWbwS0RENBLFo4Q4Va1ETpoGpy2xC2C7vAGsLopmUyqFgDSNkiXESXSi2YqSbG2fhEYsAti6M+3QKAVMzk+HTqPEhNw0fNrIAJYoUnErIVapVFi7di1uu+02iKKIZcuWYdKkSdiwYQNmzJiB+fPn48UXX8SOHTugVCqh1+vx6KOPxms4QXXZXRHNW8nWqWVlYDtsLnTaRRR4M7CpaiXSNEoGsEREREF0OmLfhRgARutT0dAWu7Vgu2KQgQU882A72IU4aUwtVlxW3Lc7cHGWDn87YI5qPeIDjR2YaszwNwstLczEtiNn4ZYkKMKo/iMij7jOgS0rK0NZWVmfx1atWuX/9z333IN77rknnkOQJdLOgVk6jawAttFbpjTKW7bkea0aLV0sISYiIgrEl4FNi2EJMeAJYPc1RN9Z1icWc2ABz5xLZmCTo9PuwtlOh3/+q0+xt+lXvaUb0yKYM+0U3Thi7sDy2aP9j5WOysBf6hpxotmKCd45sUQkX9xKiIcSq0OMKAObpVPLKiFu6vAsoVPQa05Ftk6DFi6jQ0REFFCn3QWtWgGVIrZZqtH6VJg77HCK7pi8X08JcXQBbGaqisvoJMlJbwficQEC2FMtkWXsPz/bCYcooXRUT/A7ozATQM/cWCIKDwNYAF0OF9IjOOlk69Sy1nI978205qb3DmCZgSUiIgqmyy7GtIGTz2hDKiT0rNEerZ4S4ujGmpHCDGyynPAGsGNz+gawYwxaCEDE82APNHYA6AlaAc9STpmpKv9zRBQeBrDwnHgiOelk6TRosTohSVLQ7Zq9gWqOrme9r2yZ5cdERESRqK2tRWVlJSoqKrBx48Z+zzscDqxevRoVFRVYvnw5GhoaAAANDQ2YOXMmlixZgiVLlmDt2rWJHrpfp8MV8/mvADBa78mqnY7RPFhmYIe+E83dUCkEjDFo+zyeolKgMDMFJ1sjWxbM1GJFeooS+b2WTRQEAeNzdP6sLxGFJ6nrwA4WnmV0IsvAim4JHXYXMlMDL0bd3OWAQauGSqno81pLtxMutxTz0igiIhrZRFHEunXrsGnTJhiNRlRXV6O8vBwTJ070b/PnP/8ZmZmZePPNN1FTU4Nf/vKXeOKJJwAAxcXF2Lp1a7KG7xdN45xgxhg8PSli1Yk4VnNgM1PVzMAmyckWK4oM2gGvyYqzdBFnYE+1dqMkS9dvqcbiLC12H2+J6D2JRroRn4F1S5I3AxvZHFgg9HI45zsdyE3T9HksS6eBBMDCebBERBRjdXV1KCkpQVFRETQaDaqqqrB9+/Y+2+zYsQM33ngjAKCyshLvvfdeyIqiROuwizFdQscnJ02DFJUCp9tiFcCK0CgFf5fZSGWmqmB3uWF3xWZuLslXb+lGUZZ2wOd8S+lE8v/jVGu3fx5tbyVZOrRYnf5GZUQk34gPYP2Lj0dwhzdb6wlKQ5UCN1v7B7A5aWrvazkPloiIYstsNqOgoMD/s9FohNls7rdNYWEhAM/SdxkZGWhtbQXgKSO+4YYbsGLFCnz44YeJG/gF4pWBVQgCRmWmxiyAjXQq0oUyvPvKpXQSS5IknG6z+TPzFyrO0qLLIaI5zKlfNqeIpg77gAGsvzlUlGvMEo1EI76EOJp5K74MbKgg9HynAyVFfb+8snWegLalywnkhf3RREREcZGfn4+3334bWVlZOHDgAO68807U1NQgPT094GuUSgEGgy7g85GyOkVkZ6QEfW+lUhHRZ5fkpqGp3RaTcTslID1VFfV7FWR7llSRNOqI92uwG4z7dbbDBrvLjUmFmQOObXqRZ23YFqcbEwd4PtA+HWnyNGmaNsbQ7/mLSrIBAOfs4qA7Hj6D8XcVC9yvoW/EB7BW/yLpkc2BBYKXEEuShGarAzlpKX0e95cfdzMDS0REsWU0GtHU1OT/2Ww2w2g09tumsbERBQUFcLlc6OjoQFZWFgRBgEbjuck6Y8YMFBcX48SJEygtLQ34eaIowWKJfUOaDpsLGiDoexsMuog+25imxl5TC1pbu/rNTwxXa6cdWpUi6mOgdHuuSc6c68Ck/PS4HNNki/T3FU+HvGsCZ2uUA44tS+0pWDzSYMHkAbK0gfbp4CnPHNfcAd5XrwAE73uWlRii3YW4GIy/q1jgfg0NeXmB110e8SXEXf7GC+HH8gZfFjVIBrbd5oJTlPwlwz45vTOwREREMVRaWgqTyYT6+no4HA7U1NSgvLy8zzbl5eXYsmULAGDbtm2YO3cuBEFAS0sLRNETSNXX18NkMqGoqCjh++AUPXNB41FCDACj9Knocohoi0HTJKszsl4aF/JNNzrbyZvbidTg7UZ9YQdin4LMVCgVAhos4ZX7+sqDB5pbq1EpUKhPZQkxUQRGfAa2y+5buy38E49KIUCfqgqagfWvAXvBHNj0FCXUSiFkAygiIqJwqVQqrF27FrfddhtEUcSyZcswadIkbNiwATNmzMD8+fNRXV2Ne++9FxUVFdDr9Xj88ccBAHv37sWTTz4JlUoFhUKBhx9+GAZD4jNEvuY28WjiBPQEK/Wt3TBoA68kIEer1YmiAMFPOHzvcSrCJVsoMg0WGxQCUJiZMuDzKoWAUZkpqG8Nb870ydZu5KdrAk5TK/E2hyKi8DCAdUa3dluo9Vz9AWx63wBWEARkadVBs7dERESRKisrQ1lZWZ/HVq1a5f93SkoKnnzyyX6vq6ysRGVlZdzHF0qn9wZzvDKwY7M9c8VOtFhROiozqvc622nHZUXRB/mpas96oQxqEqvB0o2CjJSgXaRHG7Rhrxt8qmXgDsQ+xVlafHK6HZIkRV3GTjSSsITYe4c3LcI7vFk6ddAmTs3eANZXMtxbTpqGASwREdEAOh2+DGz8SojVSgGm5uiynV0OFzrtIowZA2fvwlWcrUM9A9iEOt1mw+gQGfQigxb1lvCW0jnVakVxVuCmOsVZOlidov9akYjkYQDrW0Ynwvb32Tp10DLg5gAZWMATwJ7nPBciIqJ+OmzxLSFWKQSUZOlwoiW6ANbcYQeAmAWwLCtNvAZL4CV0fMYYUtFpF9HWLW/OtKXbiTabCyXZgQPjEm929iR/30RhGfEBrH8d2ChKiM93OQLekTvf5YBWrRgwQM5L1+AcA1giIqJ+Ov2rBMRvttPYbB1ORJmBjXUAW2TQos3m4jrxCdJpd8HS7cQYffAMrH/OtMxGTr6bEEFLiLMZwBJFYsQHsF0OFzRKIei8h2Am5KWhyyEGXAy9ucuBnLT+2VcAyEtPQWu3Ew6XO6LPJiIiGq56mjjFL4Adl6PFmTYbbN5+GJEwt3sD2AANgMLlC3iiLW0meXzXb6EysL4GWw0y58GavJn9kiAlxMaMFKSoFDgZZRUA0UjDANYhRlw+DADTjZ6F3Q+bOwd8/nyXo18HYh9jeop/m5Fm47sm/HLH0WQPg4iIBimLd3qOXhvPADYNEhBVya65ww4BQF6Ac324fEuumM53xeT9KLjT3ozq6BAZ2FH6VAgAGmR2Ij5+3ooUlQKj9IEDY4UgYFy2Dsd5s4IoLCM+gO20uyJu4AQAE3LToFYKONzUMeDzQTOwGd713rzlRyNFi9WBF/bUY/P+MyyRIiKiATW225CZqorqJnMo47ydiE1RZMDMHXbkpmugirCS60Kj9alQCsCJZgawidBg8QSko0NkYFNUChgzUmSXEJ9o6UJJlhZKRfDuwuNyoi9jJxppRnwAa3WI0KkjD2DVSgUm5aXjkHngADZYBjbfm4E92zmyAthXP2mEU5QgSsCOL84nezhERDQINXXYURCjeaWBFGdpoRAQVQBh7rDHbP4r4LmuGKVPxUkGNQlxstWKbJ1aVqn6GEOqP+AN5USzFeNyApcP+4zL0cHcYfeXzBNRaCM+gO1yiEiLcn7NNGM6jpg74b6gkZPNKaLTLgbMwPoC2JHUyMnhcmPz/jO4alwWxmZr8caRc8keEhERDUKN7TYUZgbPikVLo1JgjEEbVSfiWAewgKeM+ARLiBPi2HkrJuSmydp2jEGLBhkZWKtDRGO7HeNzQr/veG+Qy3mwRPIxgHWIEXcg9pluzECXQ+w3h8Y3tzVQAJueokSqSjGiMrBvfX4OLVYnvn7JGCycko99DW0jroSaiIiCkyQJTe12FMSoMVIw0XQiliQJ5g67/4Z0rBRn6XCyxRrWmqMUPrck4Xhzl+wAtsigRWu307/EUyC+knR5GVjPZ3MeLJF8Iz6AtTpc0QewBRkAgMMXlBF/cc5z93R8gC8wQRCQn5GCsx0jJwP7zvEW5KdrcHmJARVT8iDBE9QSERH5dNhd6HKIcc/AAp4A9lRrN1xi+CsCtNtcsLncMc/AFmdpYXWIONMur1yVItPYbkO3040JMgJNABif69nuaIjs+HHv/GU5AewofSo0SoHzYInCMOID2Gi7EAPA2BwdUlQKHG7q24n4sLkDSoWAiUHu7OWna0ZUBvawuQMXFWZCEASMzdFhXI4O75lakz0sIiIaRBq9S9MUJiADO6MwAy63hAONA/eyCCbWa8D6zCkyAADeOc7zYzwdO+8JGuVmYKfke1ae+PzswCtP+JxotkKtFPxrxwajUggoydZFVcZONNIwgI1BCbFKIWBKfjoONLb3efywuRPjc3RIDdIkKj8jBedGSADbYXOh3mLDNO/SQwBQWpiBI+ZOlkkREZFfkzfzWJCADOylRXooBGDPqfCDxXgFsL4bvLXH2Ogwno6dl58pBYDcNA2ydWp8FiKAPd5sRUmWDqoQHYh9uJQOUXhGdADrEt2wu9zQRRnAAsDcsVk40NjhD0YlScLhpg5MN2YEfV1eegrOdTr6NYAajo6c9dzd7h3ATjNmwNLtRBPnwRIRkZcvA5uIObCZqWpMM2Zgz0lL2K+NVwALAPOnGfFRfRu708bRsfNdKMxMkdWBGPBM/Zqcl47PzwUvIZbbgdhnXI4OjW02dDtF2a8hGslGdADb5fB8UUTbhRgAFkz2zOfc8bnnbmlThx1tNhem9grWBpKfroHLLaHVu2D7cOYrsZ7aK6j3BbOHzcHvZhIR0cjR2G5DikqBLK06IZ93eYkBBxrbww4WzR12KBVCwGaN0Zg/NR8ut4R3T7TE/L3J43iz/A7EPpPz03HsfBecAeZM25wizrTZwgpgx+foIIGdiInkYgALRF1CDHjunk3I1fkbEh1u8mYbC4JnYHuW0hn+GcjD5k6M0qfC0OuCZGJeOpQKwX+8iIiImto9a8AKgrwSzGhdXpwFUQI+bmgL63X1lm4YM1KglFkqGo6LiwzI0qqx82hzzN+bAJdbgqnFKmupm96m5KfB5ZYCNl367GwnJCBo/5MLjfdue4Q384lkGeEBrOdOa3oMAlgAmD85D5+cbse5TjsOmTuhUgiYFOILLM9bdmQeAZ2ID5s7+pQPA0CKSoGJuWn9OjgTEdHIlYg1YHubOSoTKSoF9pyUPw/W5Zaw56QFlxXp4zImpULA1eOz8c6JFpYRx0FDazecooQJufIzpYAnAwsg4DzYj+o9N0EuHi3/76IkS4ssrRr7Tod3A4VopBrZAazdk4GNxRxYoKeM+E/7zuBgYzsm5KZBowp+iPPTPWVHwz0D225z4nSbDdMGmBM8zZiOw2zkREREXolaA9ZHo1Lg4jF6vHOiBaJb3rmo7kwbOuwuzBuXHbdxLb94FLocIv7v49Nx+4yRyrcUzoQwM7BFBi1SVYqA82D31lswKS8NBp388ndBEHBJkR4f17fxWohIhhEdwJ5s9ZR/jNKHbnMux7gcHS4t0uOFPfX4sL6tX7ZxINk6DZTC8A9gfXNcB5oTPK0gA+02F9e7IyIi2JwiWrudCc3AAsCSGQVosNhQc8gsa/t3jrdApRBweUlW3MY0zZiBsgk5ePmjBnTYEpOFPd9px/q3j+F/d5/wN6kajj6st0CnVvrXdpVLqRAwKS99wAys3eXGp2facZl3GaRwXDLGgKYOO6+FiGQY0QHsYXMn0lOUGGOI3UnyqWWlWH/DRfjKDCNumFkYcnulQkB+RgrqLcP7C+uT021QCBgwqPc3cmri3A8iopGuKYEdiHubPzkXpYUZeHq3SVY32HdOtGD2GL3sDraR+u5VJei0i/jDRw1x/RwAeP1gE5Y9/yH+vP8MfrenHkt+8wE27z8T989NNEnyNMe6vMQAtTL8S+GpxnQcMXfAdsHfyYHGdthdblwaSQDrLUX/uJ5lxEShjPgAdmp+OhQxbBKhUirwpQk5eLByCi4K0cDJZ+aoTHzcMLzLRt490YqLCjKQmdq/pGZibhpSVArs59wPIqIRz+TtxDpan9gMrCAIWFU2Hue7HHj+/VNBt21qt+HYeSuujmP5sM+U/HQsmJyL3+2txydxPE9+croNP932OaYa0/HHb1+GLbdejivGZuEXO47iA1P4a+QOZsebrWhst+OqCH9/107KQbfTjdpjfRtsfXjKAoUAXDIm/HnR43N0MGjVYTcSIxqJRmwA6xTd+OJcZ58lXZLl0iIDmrscONnSneyhxEWr1YFDTR0BTxRqpQKXFxuw82jzsA7iiYgotF3Hm5GeosR0mTeBY2nWaD2qpufjhT31eOYdU8Bz0j8OnwWAuM5/7e2HCyahICMFP/jbITTGocTU0u3Ef9ccQUFmKn51w0UoztJilD4Vjy6ajvE5abj/9cNosAyfaxTf0kSRBrCXFhmQn67B3w+d7fP4R/UWTMlPR0Zq+Fl5hSDg4jF6fFwf/nrERLHWYOnGkzuP45l3TPj7ITNcAZaNSpYRG8AeP2+FU5RkzVONN99ciQ+H6ZfWe6ZWSAh+ovjypFw0ddgDdvUjIqLhT3RLqD3WgnnjsiMq7YyFByqn4CszjPjt+6cGDBjfN7Xg2XdM+NL4bJRkx6aHRih6rRrrb5gBh+jGihc/xot762Hpjs368ZIk4eF/foYWqwOPLp7WpyRap1HiVzdcBEEAHqg5MuguYiP1zokWTMpLgzEjsjJ1hSDg36Yb8b6pBc1dnlUkTrV2o66xA3OKI58TfekYPc6024fVzQIaWkS3hKd3n8BXX/gQf/j4NJ5//xR+/I/P8N0/fjKo/i5HbAB7yLtsy0BdcRNtjCEV+ekafDRMA9h3T7QgW6cesIGTz5fGZ0MhAG9zvTsigqdzudUReh4iDS+fnGmDpduJL0/MTdoYVAoBDyycjLuvGYf3TK1YvulDPPSPI3jtQBOeeceEH752GONz0/CTqqkJW6cWAMbm6PCbr83GRQUZeLL2BCr+9z0s3vgB1v79CGoOmvvNx5Tr5Y9OY/fxFqwuGz/gNdEofSruXzAJB5s68FyI0uqhoNPuwv7T7VFnz6+fng9RArYdOQtJkvDYW18gVaXATZeMivg9vzwpFyqFgJc/jP98Z6ILOVxu/Oj1w3j+g3pUTMnD3267HO+uvhr/UzUVphYrbn7pY9SdaU/2MAEA8e08MIgdiUMDp0gJgoDLig1490Qr3JIU0zm5ySa6JbxvasW88dlB9ytLp8Hs0XrsPHoed8wbm7gBEtGg8sePT+O590/B0u1EmkaJtZWTUT45L9nDogTZebQZGqWAK8fFr7OvHIIg4OY5RaiYkoeN757EzmPNqDl0FgoBmFGYif+pmoo0TeIvoSbmpuHJZaU40NiOfQ1tONTUifdNrfjH4bP4f7tP4DtXFOPGmYVQKuTO8eE8AAAUWklEQVRdR+xraMNTu07g2km5WD47cOC1YEoe3jnRgk0fnEJpYSbmjU9M6XQ8/N/HpyG6JZRNzInqfcbnpGF6QQaefeckDp7twt5TFtw3fyJy0yNvPmbMSMGii4zYeqAJt1xRjPwIM8RE4bI6RNy79SD2nLLg+18ej69fOsb/3MKp+bioMAN3bf4Ud23+FE8snYGLI5jnHUsjNgN72NyBqcaMhN49DebSIgMs3U4cP29N9lBi6n1TK9psLlw1NvTJrmxiDo6dt+JU6+ApUSCixHBLEh7/1zH88u1jmJyXhruvGYex2Trc99phPPGv43BzfvywJ0kS/nW0GZeXZCUlOBxIQWYq1l43BdvuuBKvfPNSvL1yHn5702wUJHiJnwvNKMzEzXOK8OjiafjnHXPx9PKZGKNPxc+2H8V3XtmPz2VMxznY1IHvbzmA0fpUPLhwcsjroXvLJ2JyXjp++NohHGgcHFkYwNPTRG7/jBPNVjz/wSksnJKHGYWZUX/2/1RNxbzx2XjzsBmlhRlYOiv06hOhfOvyIrjdEl5kFpYSpK3biTs31+Gjegt+fN3kPsGrz2i9Fs98dRby0jW469VP8eZn55Iw0h5xDWBra2tRWVmJiooKbNy4sd/zDocDq1evRkVFBZYvX46GhsT8Z3WKbhw934Vp+cmf/+ozp9gzD3bzJ2fCamQkuiUcb+7C9s/PYe+pVpxpGzzL8ZzrtGPdts8wNlsr605n+aRcqJUC1v3zs4hLoYho6Gm3ObFmy0H84aPT+PeLR+HJZaW4eU4RfvO1WVg+exRe/qgBP3r9MOyu4TH/LlGiOQc/++yzqKioQGVlJXbt2pWQ8TpECWc77Lhuan5CPi8cKoWAiXlp0GmUyR5KPwpvFdez/z4L/1M1FY1tNtz80sd45M3PB1xjXpIkvPnZOdz96qfQa9X43+UzZTUd0mmUeGLpDOSma3D3qwfwz8Nnk9J4scHSjT/tO43vbzmARRs/wFVP7MYV63dhwf97F6v+8ile+OAUjp7r6jc2S7cTP9n2OXRqJe4pnxCTsYwxaPHIoml4e00ZnlxWGpMKujEGLa6bbsSrn5zBW0kOEmj4e+dEC25+6WN8frYTP/vKdCy6qCDgtvkZKfjN12Zhan46fvT6YTy58zg67YlZn/pCghSnbx9RFFFZWYlNmzbBaDSiuroa69evx8SJE/3bvPzyy/jss8+wbt061NTU4M0338QTTzwR9H2dThEWS3RZStEtYfVfDuDOL40N2oXYYNBF/Vnh+MX2o/jT/jNYNqsQt1xRjLx0TZ8vQ7ckweoQ0WDpxudnu/CeqRXvmVrQdcE8seIsLa4en41Zo/WYnJeG3DQNUtXRnXQHOhaSJEECIEmA5HkAEjzH99PGdjy924QvznXhdysuxvicNFmf89Zn5/Cj1w/jynFZ+I+rxmJyXlrSGnkEkui/i8GOx6MHj0WPUMfC5hRh7rCj9lgzNu8/g7OdDqy5dgKqZxX2ywT94aMGPP6v4yjMTMHXLx2DuWOzMCozFRpVbL4b8vKS3wsh1qI5Bx89ehRr1qzB5s2bYTabccstt2Dbtm1QKgOfR2JxbgY88xPDXVd1uP6/i3S/LN1OPPfeSWz+pBGSJGFOsQGzRulh0Klh7rBj7ykLDjV1YFJeGn6xZDpG68NrRHW6rRsP1BzBgcYOXFqkx4LJeZg1OhP56SlI0yihVAhBs7ly9kuSJNhcbrTbXDjXaYepxYpDTZ3Ye6oVJu+KDcVZWkwvyEBxlhZut4RzXQ58eqYdx5s9752bpkHpqEwYM1LQ7RDx5mfn0O0Use76qbhuWmxvksT6b7DV6sB/bT2EujPtWDarEPMn52JCbhq0aiVSVIqETTXj/y2P3qGS1OdxDPj4hU/KeU2wcCzw66U+j+v1OrS19d+v3q9xiG6cabPhUFMHtn9+DvtOt2NsthYPVk7BzFHyqhIcLjd+seMo/vppE/SpKlRdZMScYgPGZutiEnP4BDs3xy2A3bdvH5566in89re/BeC5mwsA//mf/+nf5tZbb8XKlStx8cUXw+VyYd68eXj//feDfvHF6iQpR6L/40qShF/XnvCXjSgFQNFrHotLlPr8EWfr1PjS+BxcPEaPCbk6dNpFHDvfhd0nWvBxvQUOsWdrAYAgeOb1CAAUvf4tCL2CUAQOTIG+24WiVgp46LopWBjm3fRXPzmDn711tM/nKL3jVSp6xpwsgiD4v2gGS1VjMofh+/sZbJKyJFOAgzFYDk8iD0mwvwvfjS6fiwoy8P0vj8es0YHn1HxwshW/efckPvE2kEhVKfDizZdgbLYu6rEOxwA2mnOwL1vr27b3doEk8tx8IV5kD6y+tRuvH2zCW5+f90/NUSoEjM/R4auzR2HxjALZc2UvJLol/HHfafx5/xk0WPpXfgnez1IqBCh6fRdceJ3h+XfPtYZvG/cA3x06tRIzR3nm384bl42irIED73Odduw+3oKP6i040NiBdpsLCgGYU5yF264sxoRceTfUwxGPv0GHy41fvX0MWz9thHjB8fAEsTH9uAH1vt6Jtws/JtCnygkm+70+kgB0BBifo8Oii4z494tHR3RD+DNzJ55914QPTrb6Y44UlQJ/+OalKA7w/zMcwc7NcZtkYjabUVDQk4Y2Go2oq6vrt01hoWe+gEqlQkZGBlpbW5GdHXi+pFqtTOjFRqIvbH5SPQs/qZ4V1XusjNFYkuX2BVNw+4IpyR4GEQ0ii/IysOiy4mQPY8iI5hxsNpsxa9asPq81m81BPy/R5+YLDcebEEB0+5WXl4FLJudjbQzH09uq6zKx6rppcXr3yOXlZWD6uMR3sY7H3+D6r1+C9TF/V6LYyMvLwNUzop/3HYnBVZtJREREREREFEDcAlij0Yimpib/z2azGUajsd82jY2NAACXy4WOjg5kZSW3dT4REdFQF805WM5riYiIkiVuAWxpaSlMJhPq6+vhcDhQU1OD8vLyPtuUl5djy5YtAIBt27Zh7ty5g2ZZGyIioqEqmnNweXk5ampq4HA4UF9fD5PJhJkzZyZjN4iIiPqJWxMnANi5cyceeeQRiKKIZcuW4Y477sCGDRswY8YMzJ8/H3a7Hffeey8OHz4MvV6Pxx9/HEVFRfEaDhER0YgRzTn46aefxquvvgqlUokf/ehHKCsrS/LeEBERecQ1gCUiIiIiIiKKFTZxIiIiIiIioiGBASwRERERERENCSM+gK2trUVlZSUqKir8i7f35nA4sHr1alRUVGD58uVoaGhIwigTJ9Tx2LRpE66//nosXrwY3/rWt3D69OkkjDIxQh0Ln23btmHKlCn49NNPEzi6xJJzLP7+97/j+uuvR1VVFe65554EjzCxQh2PM2fO4Oabb8YNN9yAxYsXY+fOnUkYZfzdf//9uPLKK7Fo0aIBn5ckCT/96U9RUVGBxYsX4+DBgwkeIQ1ncr+jh5ry8nIsXrwYS5YswdKlS5M9nIgN9P1gsVhwyy23YOHChbjlllvQ1taWxBFGZqD9+vWvf40vfelLWLJkCZYsWTIkv/MbGxtx8803+8/jv/vd7wAM7d9ZoH0a6r8vu92O6upqfOUrX0FVVRWefPJJAEB9fT2WL1+OiooKrF69Gg6HI8kjjSNpBHO5XNL8+fOlU6dOSXa7XVq8eLH0xRdf9NnmpZdekh588EFJkiTp9ddfl1atWpWMoSaEnOPx3nvvSVarVZIkSXr55ZeH7fGQcywkSZI6Ojqkr3/969Ly5culurq6JIw0/uQcixMnTkhLliyRLBaLJEmSdP78+WQMNSHkHI8HHnhAevnllyVJkqQvvvhCuvbaa5Mx1Ljbs2ePdODAAamqqmrA5//1r39Jt956q+R2u6V9+/ZJ1dXVCR4hDVdyv6OHomuvvVZqbm5O9jCiNtD3w89+9jPp2WeflSRJkp599lnp5z//ebKGF7GB9uvJJ5+UnnvuuSSOKnpms1k6cOCAJEmea5uFCxdKX3zxxZD+nQXap6H++3K73VJnZ6ckSZLkcDik6upqad++fdLdd98tvf7665IkSdKDDz7ovw4ZjkZ0Braurg4lJSUoKiqCRqNBVVUVtm/f3mebHTt24MYbbwQAVFZW4r333oM0TPteyTkec+fOhVarBQDMnj27z1qBw4mcYwEAGzZswHe/+12kpKQkYZSJIedY/OlPf8I3vvEN6PV6AEBOTk4yhpoQco6HIAjo7OwEAHR0dCA/Pz8ZQ427OXPm+H/nA9m+fTtuuOEGCIKA2bNno729HWfPnk3gCGm4kvsdTckz0PeD7zsBAG644Qa89dZbyRhaVEJ97w1V+fn5uOiiiwAA6enpGD9+PMxm85D+nQXap6FOEASkpaUB8Kzh7XK5IAgC3n//fVRWVgIAbrzxxmH9nTiiA1iz2YyCggL/z0ajsd8fttlsRmFhIQBApVIhIyMDra2tCR1nosg5Hr1t3rwZ11xzTSKGlnByjsXBgwfR1NSEL3/5ywkeXWLJORYmkwknTpzA1772NXz1q19FbW1tooeZMHKOx8qVK/Haa6/hmmuuwX/8x3/ggQceSPQwB4ULj1VBQcGwuHig5Av3fDXU3HrrrVi6dCn++Mc/JnsoMdXc3Oy/oZeXl4fm5uYkjyh2Xn75ZSxevBj333//kCqzHUhDQwMOHz6MWbNmDZvfWe99Aob+70sURSxZsgRXXXUVrrrqKhQVFSEzMxMqlQrA8D/fjugAliK3detWHDhwALfddluyh5IUbrcbjz32GO67775kD2VQEEURJ0+exIsvvohf/epXePDBB9He3p7sYSVNTU0NbrzxRtTW1mLjxo34wQ9+ALfbnexhEdEQ8Morr2DLli34zW9+g5dffhl79+5N9pDiQhAECIKQ7GHExE033YQ333wTW7duRX5+Ph577LFkDyliXV1duPvuu/GjH/0I6enpfZ4bqr+zC/dpOPy+lEoltm7dip07d6Kurg7Hjx9P9pASakQHsEajsU8JrNlshtFo7LdNY2MjAE+avqOjA1lZWQkdZ6LIOR4A8O677+KZZ57B008/DY1Gk8ghJkyoY9HV1YXPP/8c3/zmN1FeXo79+/fjjjvuGJaNnOT+PykvL4darUZRURHGjh0Lk8mU4JEmhpzjsXnzZvzbv/0bAODiiy+G3W4ftpUbwVx4rJqamgb8TiEKl9zz1VDk24+cnBxUVFSgrq4uySOKnZycHP80grNnzyI7OzvJI4qN3NxcKJVKKBQKLF++fMheCzidTtx9991YvHgxFi5cCGDo/84G2qfh8vsCgMzMTFxxxRXYv38/2tvb4XK5AAz/8+2IDmBLS0thMplQX18Ph8OBmpoalJeX99mmvLwcW7ZsAeDpNjt37twhefdJDjnH49ChQ1i7di2efvrpYT3PMdSxyMjIwAcffIAdO3Zgx44dmD17Np5++mmUlpYmcdTxIefvYsGCBdizZw8AoKWlBSaTCUVFRckYbtzJOR6FhYV47733AADHjh2D3W4fcif9WCgvL8df//pXSJKE/fv3IyMjY9jOB6bEkvP/cCiyWq3++fNWqxXvvPMOJk2alORRxY7vOwEA/vrXv2L+/PlJHlFs9J7b/9Zbbw3J35kkSfjv//5vjB8/Hrfccov/8aH8Owu0T0P999XS0uKvcrPZbHj33XcxYcIEXHHFFdi2bRsAYMuWLcPiOzEQQRquHYlk2rlzJx555BGIoohly5bhjjvuwIYNGzBjxgzMnz8fdrsd9957Lw4fPgy9Xo/HH3982F6YA6GPx7e//W18/vnnyMvLA+C5UH/mmWeSPOr4CHUserv55pvxgx/8YFgGsEDoYyFJEh577DHs2rULSqUSt99+O6qqqpI97LgJdTyOHj2KBx54AFarFYIg4N5778XVV1+d7GHH3Jo1a7Bnzx60trYiJycHd911l//u70033QRJkrBu3Trs2rULWq0WjzzyyLD9P0KJN9D/w6Guvr4ed955JwDP1IxFixYN2f0a6PthwYIFWL16NRobGzFq1Cg88cQTMBgMyR5qWAbarz179uDIkSMAgNGjR2PdunVD7mbdhx9+iG984xuYPHkyFApPfmvNmjWYOXPmkP2dBdqn119/fUj/vo4cOYIf/vCHEEURkiThuuuuw8qVK1FfX4/vf//7aGtrw7Rp0/DLX/5y2FZKjvgAloiIiIiIiIaGEV1CTEREREREREMHA1giIiIiIiIaEhjAEhERERER0ZDAAJaIiIiIiIiGBAawREQ0JN1///248sorsWjRIlnb//3vf8f111+Pqqoq3HPPPXEeHREREcWDKtkDICIiisTSpUuxYsUK3HfffSG3NZlM2LhxI1555RXo9Xo0NzcnYIREREQUa8zAEhHRkDRnzhzo9fo+j506dQq33norli5diq9//es4duwYAOBPf/oTvvGNb/i3z8nJSfh4iYiIKHrMwBIR0bDx4IMP4uGHH8bYsWPxySef4OGHH8bvf/97mEwmAMDXvvY1uN1urFy5Etdcc01yB0tERERhYwBLRETDQldXF/bt24dVq1b5H3M4HAAAURRx8uRJvPjii2hqasKKFSvw2muvITMzM1nDJSIioggwgCUiomFBkiRkZmZi69at/Z4zGo2YNWsW1Go1ioqKMHbsWJhMJsycOTMJIyUiIqJIcQ4sERENC+np6RgzZgz+8Y9/APAEtEeOHAEALFiwAHv27AEAtLS0wGQyoaioKGljJSIiosgIkiRJyR4EERFRuNasWYM9e/agtbUVOTk5uOuuuzB37lw89NBDOHfuHFwuF66//nqsXLkSkiThsccew65du6BUKnH77bejqqoq2btAREREYWIAS0REREREREMCS4iJiIiIiIhoSGAAS0REREREREMCA1giIiIiIiIaEhjAEhERERER0ZDAAJaIiIiIiIiGBAawRERERERENCQwgCUiIiIiIqIh4f8DrL74eq56IfkAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 1152x864 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "g9Da52TpCZaC",
        "outputId": "734298ef-9cee-49b1-cb87-596fa2a0bee3"
      },
      "source": [
        "# 3. finding outlier threshold using percentiles - 'Memory_used'\r\n",
        "for  i in np.arange(90, 100.5, 0.5):\r\n",
        "  print(i, np.percentile(np.concatenate((grp_y_train[:, 0], grp_y_cv[:, 0])), i))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "90.0 65163.0\n",
            "90.5 65256.0\n",
            "91.0 65435.0\n",
            "91.5 65437.0\n",
            "92.0 65530.0\n",
            "92.5 65530.0\n",
            "93.0 65530.0\n",
            "93.5 65531.0\n",
            "94.0 65531.0\n",
            "94.5 65531.0\n",
            "95.0 65531.0\n",
            "95.5 65547.0\n",
            "96.0 65547.0\n",
            "96.5 65618.0\n",
            "97.0 65618.0\n",
            "97.5 65649.0\n",
            "98.0 65655.0\n",
            "98.5 65657.0\n",
            "99.0 65662.0\n",
            "99.5 191098.94499999995\n",
            "100.0 389103.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pIYH6zIFEBV7",
        "outputId": "54dd36ba-2165-47de-ca1e-feeb6a6d9569"
      },
      "source": [
        "# 4.  finding outlier threshold using percentiles - 'CPU_used'\r\n",
        "for i in np.arange(90, 100.5, 0.5):\r\n",
        "  print(i, np.percentile(np.concatenate((grp_y_train[:, 1], grp_y_cv[:, 1])), i))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "90.0 0.41999999999999993\n",
            "90.5 0.5\n",
            "91.0 0.59\n",
            "91.5 0.65\n",
            "92.0 0.74\n",
            "92.5 0.8899999999999999\n",
            "93.0 1.05\n",
            "93.5 1.16\n",
            "94.0 1.22\n",
            "94.5 1.27\n",
            "95.0 1.31\n",
            "95.5 1.34\n",
            "96.0 1.37\n",
            "96.5 1.44\n",
            "97.0 1.83\n",
            "97.5 2.43\n",
            "98.0 2.74\n",
            "98.5 3.15\n",
            "99.0 3.5\n",
            "99.5 4.0\n",
            "100.0 16.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yOI9qnEW-uSC"
      },
      "source": [
        "# Observations\r\n",
        "    1. Each instance has different set of values in it.\r\n",
        "    2. Hence we need to utilize instance information using OneHotEncoding technique.\r\n",
        "    3. 'CPU_usage' target features seems to be in range\r\n",
        "    4. 'Memory_used' target feature has very wide range and contain some extreme outlier points.\r\n",
        "\r\n",
        "# Conclusion:\r\n",
        "    1. Remove datapoints with 'Memory_used' values more than 65662.0 i.e., retaining 99% of the data points.\r\n",
        "    2. Remove datapoints with 'CPU_used' values more than 4.0 i.e., retaining 99.5% of the data points."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hHhvrHhHEyza",
        "outputId": "bd8dac80-7ec5-4e42-eb9f-7560654c04a5"
      },
      "source": [
        "# 5.1. Finding useful index\r\n",
        "train_idx = np.where((grp_y_train[:, 0] <= 65662.0) & (grp_y_train[:, 1] <= 4.0)) \r\n",
        "cv_idx = np.where((grp_y_cv[:, 0] <= 65662.0) & (grp_y_cv[:, 1] <= 4.0)) \r\n",
        "\r\n",
        "# 5.2. Outlier removal\r\n",
        "grp_lstm_x_train = grp_lstm_x_train[train_idx]\r\n",
        "grp_dense_x_train = grp_dense_x_train[train_idx]\r\n",
        "grp_y_train = grp_y_train[train_idx]\r\n",
        "\r\n",
        "grp_lstm_x_cv = grp_lstm_x_cv[cv_idx]\r\n",
        "grp_dense_x_cv = grp_dense_x_cv[cv_idx]\r\n",
        "grp_y_cv = grp_y_cv[cv_idx]\r\n",
        "\r\n",
        "grp_lstm_x_train.shape, grp_dense_x_train.shape, grp_y_train.shape, grp_lstm_x_cv.shape, grp_dense_x_cv.shape, grp_y_cv.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((443895, 15, 2),\n",
              " (443895, 3),\n",
              " (443895, 2),\n",
              " (189249, 15, 2),\n",
              " (189249, 3),\n",
              " (189249, 2))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 148
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "L4khv96p6OQR",
        "outputId": "c1846b8a-e7f5-403b-af44-df7b4107332b"
      },
      "source": [
        "# 6. creating OHE for instances\r\n",
        "from sklearn.preprocessing import OneHotEncoder\r\n",
        "enc = OneHotEncoder()\r\n",
        "enc.fit(grp_dense_x_train[:, 0].reshape(-1, 1))\r\n",
        "instance_train = enc.transform(grp_dense_x_train[:, 0].reshape(-1, 1)).todense()\r\n",
        "instance_cv = enc.transform(grp_dense_x_cv[:, 0].reshape(-1, 1)).todense()\r\n",
        "instance_train.shape, instance_cv.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((443895, 42), (189249, 42))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 149
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nJCIQF3D_hpr",
        "outputId": "c5fa31e2-c4f8-4c57-819d-d511990691f1"
      },
      "source": [
        "grp_dense_x_train = np.asarray(np.concatenate([instance_train, grp_dense_x_train[:, 1:]], axis = 1)).astype(np.float32)\r\n",
        "grp_dense_x_cv = np.asarray(np.concatenate([instance_cv, grp_dense_x_cv[:, 1:]], axis = 1)).astype(np.float32)\r\n",
        "grp_dense_x_train.shape, grp_dense_x_cv.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((443895, 44), (189249, 44))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 150
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2glnSX1isoWH"
      },
      "source": [
        "# 7. Modelling for Group data\r\n",
        "\r\n",
        "### 7.1. Model for CPU_Used"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0rhKhjzssr3W",
        "outputId": "cc22ef1d-cb0c-4fde-90ce-33fe7b12cf21"
      },
      "source": [
        "# 1. building a model - CPU_Used\r\n",
        "tf.keras.backend.clear_session()\r\n",
        "\r\n",
        "dense_ins = Input(name = 'constants', shape = (44,))\r\n",
        "\r\n",
        "lstm_ins = Input(name = 'Sequence', shape = ((15, 1)))\r\n",
        "lstm_out = LSTM(name = 'LSTM', units = 128, kernel_initializer = 'glorot_normal')(lstm_ins)\r\n",
        "\r\n",
        "concat = Concatenate(axis = -1)([dense_ins, lstm_out])\r\n",
        "\r\n",
        "dense_1 = Dense(units = 24, activation = 'relu')(concat)\r\n",
        "out = Dense(units = 1)(dense_1)\r\n",
        "cpu_model = Model(inputs = [dense_ins, lstm_ins], outputs = out)\r\n",
        "cpu_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "Sequence (InputLayer)           [(None, 15, 1)]      0                                            \n",
            "__________________________________________________________________________________________________\n",
            "constants (InputLayer)          [(None, 44)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "LSTM (LSTM)                     (None, 128)          66560       Sequence[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 172)          0           constants[0][0]                  \n",
            "                                                                 LSTM[0][0]                       \n",
            "__________________________________________________________________________________________________\n",
            "dense (Dense)                   (None, 24)           4152        concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "dense_1 (Dense)                 (None, 1)            25          dense[0][0]                      \n",
            "==================================================================================================\n",
            "Total params: 70,737\n",
            "Trainable params: 70,737\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3KzKhlAys1tb"
      },
      "source": [
        "# 3. tensorflow-keras callbacks\r\n",
        "# 3.1 To save best models\r\n",
        "!rm -r grp_cpu_models\r\n",
        "if not os.path.isdir('/content/grp_cpu_models'):\r\n",
        "  os.mkdir('/content/grp_cpu_models')\r\n",
        "save_model = tf.keras.callbacks.ModelCheckpoint(filepath = '/content/grp_cpu_models/weights-{epoch:02d}-{val_root_mean_squared_error:.2f}.hdf5',\r\n",
        "                                                monitor='val_root_mean_squared_error', verbose=1,\r\n",
        "                                                save_best_only=True, save_weights_only=True)\r\n",
        "\r\n",
        "# 3.2 Reduce learning rate if model is not learning\r\n",
        "reduce_lr = tf.keras.callbacks.ReduceLROnPlateau(monitor = 'val_root_mean_squared_error', factor = np.sqrt(0.1), patience = 4, verbose = 1, min_lr = 10**-6)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-SDWHGtCvyO7"
      },
      "source": [
        "# use RMSE for smaller values\r\n",
        "# use MAE for larger values"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pUi0kBO_s1td",
        "outputId": "283c8101-b6f0-4dc1-e922-71b2bbc0e487"
      },
      "source": [
        "# 4. training and evaluating model on cross validation\r\n",
        "rmse = tf.keras.metrics.RootMeanSquaredError(name=\"root_mean_squared_error\", dtype=None)\r\n",
        "opt = tf.keras.optimizers.Adam(learning_rate = 0.001)\r\n",
        "\r\n",
        "cpu_model.compile(loss = 'mean_squared_error', optimizer = opt, metrics = ['mean_absolute_error', rmse])\r\n",
        "hist = cpu_model.fit([grp_dense_x_train, np.expand_dims(grp_lstm_x_train[:, :, 1], axis = -1)], grp_y_train[:, 1], batch_size = 2048, epochs = 25,\r\n",
        "                 validation_data = ([grp_dense_x_cv, np.expand_dims(grp_lstm_x_cv[:, :, 1], axis = -1)], grp_y_cv[:, 1]),\r\n",
        "                 callbacks = [save_model, reduce_lr]) "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/25\n",
            "217/217 [==============================] - 4s 14ms/step - loss: 4817643.5631 - mean_absolute_error: 445.7028 - root_mean_squared_error: 1939.7788 - val_loss: 1.9065 - val_mean_absolute_error: 1.0847 - val_root_mean_squared_error: 1.3807\n",
            "\n",
            "Epoch 00001: val_root_mean_squared_error improved from inf to 1.38075, saving model to /content/grp_cpu_models/weights-01-1.38.hdf5\n",
            "Epoch 2/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 1.9609 - mean_absolute_error: 1.0507 - root_mean_squared_error: 1.3994 - val_loss: 1.0732 - val_mean_absolute_error: 0.7900 - val_root_mean_squared_error: 1.0359\n",
            "\n",
            "Epoch 00002: val_root_mean_squared_error improved from 1.38075 to 1.03594, saving model to /content/grp_cpu_models/weights-02-1.04.hdf5\n",
            "Epoch 3/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 1.1215 - mean_absolute_error: 0.7553 - root_mean_squared_error: 1.0581 - val_loss: 0.5251 - val_mean_absolute_error: 0.5241 - val_root_mean_squared_error: 0.7246\n",
            "\n",
            "Epoch 00003: val_root_mean_squared_error improved from 1.03594 to 0.72463, saving model to /content/grp_cpu_models/weights-03-0.72.hdf5\n",
            "Epoch 4/25\n",
            "217/217 [==============================] - 3s 13ms/step - loss: 0.6234 - mean_absolute_error: 0.5116 - root_mean_squared_error: 0.7890 - val_loss: 0.2547 - val_mean_absolute_error: 0.3277 - val_root_mean_squared_error: 0.5046\n",
            "\n",
            "Epoch 00004: val_root_mean_squared_error improved from 0.72463 to 0.50464, saving model to /content/grp_cpu_models/weights-04-0.50.hdf5\n",
            "Epoch 5/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.3913 - mean_absolute_error: 0.3365 - root_mean_squared_error: 0.6254 - val_loss: 0.1531 - val_mean_absolute_error: 0.2121 - val_root_mean_squared_error: 0.3913\n",
            "\n",
            "Epoch 00005: val_root_mean_squared_error improved from 0.50464 to 0.39132, saving model to /content/grp_cpu_models/weights-05-0.39.hdf5\n",
            "Epoch 6/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.3113 - mean_absolute_error: 0.2629 - root_mean_squared_error: 0.5579 - val_loss: 0.1229 - val_mean_absolute_error: 0.1896 - val_root_mean_squared_error: 0.3506\n",
            "\n",
            "Epoch 00006: val_root_mean_squared_error improved from 0.39132 to 0.35062, saving model to /content/grp_cpu_models/weights-06-0.35.hdf5\n",
            "Epoch 7/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.2888 - mean_absolute_error: 0.2486 - root_mean_squared_error: 0.5374 - val_loss: 0.1137 - val_mean_absolute_error: 0.1778 - val_root_mean_squared_error: 0.3371\n",
            "\n",
            "Epoch 00007: val_root_mean_squared_error improved from 0.35062 to 0.33713, saving model to /content/grp_cpu_models/weights-07-0.34.hdf5\n",
            "Epoch 8/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.2815 - mean_absolute_error: 0.2397 - root_mean_squared_error: 0.5306 - val_loss: 0.1098 - val_mean_absolute_error: 0.1702 - val_root_mean_squared_error: 0.3313\n",
            "\n",
            "Epoch 00008: val_root_mean_squared_error improved from 0.33713 to 0.33135, saving model to /content/grp_cpu_models/weights-08-0.33.hdf5\n",
            "Epoch 9/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.2712 - mean_absolute_error: 0.2318 - root_mean_squared_error: 0.5208 - val_loss: 0.1069 - val_mean_absolute_error: 0.1650 - val_root_mean_squared_error: 0.3269\n",
            "\n",
            "Epoch 00009: val_root_mean_squared_error improved from 0.33135 to 0.32688, saving model to /content/grp_cpu_models/weights-09-0.33.hdf5\n",
            "Epoch 10/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.2689 - mean_absolute_error: 0.2292 - root_mean_squared_error: 0.5185 - val_loss: 0.1027 - val_mean_absolute_error: 0.1592 - val_root_mean_squared_error: 0.3205\n",
            "\n",
            "Epoch 00010: val_root_mean_squared_error improved from 0.32688 to 0.32047, saving model to /content/grp_cpu_models/weights-10-0.32.hdf5\n",
            "Epoch 11/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.2588 - mean_absolute_error: 0.2230 - root_mean_squared_error: 0.5087 - val_loss: 0.1000 - val_mean_absolute_error: 0.1555 - val_root_mean_squared_error: 0.3163\n",
            "\n",
            "Epoch 00011: val_root_mean_squared_error improved from 0.32047 to 0.31626, saving model to /content/grp_cpu_models/weights-11-0.32.hdf5\n",
            "Epoch 12/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.2529 - mean_absolute_error: 0.2189 - root_mean_squared_error: 0.5028 - val_loss: 0.0963 - val_mean_absolute_error: 0.1511 - val_root_mean_squared_error: 0.3104\n",
            "\n",
            "Epoch 00012: val_root_mean_squared_error improved from 0.31626 to 0.31037, saving model to /content/grp_cpu_models/weights-12-0.31.hdf5\n",
            "Epoch 13/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.2436 - mean_absolute_error: 0.2133 - root_mean_squared_error: 0.4935 - val_loss: 0.0921 - val_mean_absolute_error: 0.1459 - val_root_mean_squared_error: 0.3035\n",
            "\n",
            "Epoch 00013: val_root_mean_squared_error improved from 0.31037 to 0.30352, saving model to /content/grp_cpu_models/weights-13-0.30.hdf5\n",
            "Epoch 14/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.2325 - mean_absolute_error: 0.2071 - root_mean_squared_error: 0.4822 - val_loss: 0.0877 - val_mean_absolute_error: 0.1408 - val_root_mean_squared_error: 0.2962\n",
            "\n",
            "Epoch 00014: val_root_mean_squared_error improved from 0.30352 to 0.29621, saving model to /content/grp_cpu_models/weights-14-0.30.hdf5\n",
            "Epoch 15/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.2187 - mean_absolute_error: 0.1992 - root_mean_squared_error: 0.4676 - val_loss: 0.0823 - val_mean_absolute_error: 0.1352 - val_root_mean_squared_error: 0.2869\n",
            "\n",
            "Epoch 00015: val_root_mean_squared_error improved from 0.29621 to 0.28688, saving model to /content/grp_cpu_models/weights-15-0.29.hdf5\n",
            "Epoch 16/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.2041 - mean_absolute_error: 0.1918 - root_mean_squared_error: 0.4518 - val_loss: 0.0768 - val_mean_absolute_error: 0.1292 - val_root_mean_squared_error: 0.2772\n",
            "\n",
            "Epoch 00016: val_root_mean_squared_error improved from 0.28688 to 0.27717, saving model to /content/grp_cpu_models/weights-16-0.28.hdf5\n",
            "Epoch 17/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.1875 - mean_absolute_error: 0.1829 - root_mean_squared_error: 0.4330 - val_loss: 0.0705 - val_mean_absolute_error: 0.1245 - val_root_mean_squared_error: 0.2655\n",
            "\n",
            "Epoch 00017: val_root_mean_squared_error improved from 0.27717 to 0.26552, saving model to /content/grp_cpu_models/weights-17-0.27.hdf5\n",
            "Epoch 18/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.1600 - mean_absolute_error: 0.1724 - root_mean_squared_error: 0.3999 - val_loss: 0.0635 - val_mean_absolute_error: 0.1199 - val_root_mean_squared_error: 0.2520\n",
            "\n",
            "Epoch 00018: val_root_mean_squared_error improved from 0.26552 to 0.25196, saving model to /content/grp_cpu_models/weights-18-0.25.hdf5\n",
            "Epoch 19/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.1318 - mean_absolute_error: 0.1600 - root_mean_squared_error: 0.3630 - val_loss: 0.0562 - val_mean_absolute_error: 0.1118 - val_root_mean_squared_error: 0.2371\n",
            "\n",
            "Epoch 00019: val_root_mean_squared_error improved from 0.25196 to 0.23709, saving model to /content/grp_cpu_models/weights-19-0.24.hdf5\n",
            "Epoch 20/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.1096 - mean_absolute_error: 0.1483 - root_mean_squared_error: 0.3311 - val_loss: 0.0522 - val_mean_absolute_error: 0.1072 - val_root_mean_squared_error: 0.2285\n",
            "\n",
            "Epoch 00020: val_root_mean_squared_error improved from 0.23709 to 0.22854, saving model to /content/grp_cpu_models/weights-20-0.23.hdf5\n",
            "Epoch 21/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.0963 - mean_absolute_error: 0.1399 - root_mean_squared_error: 0.3102 - val_loss: 0.0485 - val_mean_absolute_error: 0.1019 - val_root_mean_squared_error: 0.2203\n",
            "\n",
            "Epoch 00021: val_root_mean_squared_error improved from 0.22854 to 0.22029, saving model to /content/grp_cpu_models/weights-21-0.22.hdf5\n",
            "Epoch 22/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.0821 - mean_absolute_error: 0.1296 - root_mean_squared_error: 0.2866 - val_loss: 0.0462 - val_mean_absolute_error: 0.0992 - val_root_mean_squared_error: 0.2149\n",
            "\n",
            "Epoch 00022: val_root_mean_squared_error improved from 0.22029 to 0.21490, saving model to /content/grp_cpu_models/weights-22-0.21.hdf5\n",
            "Epoch 23/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.0735 - mean_absolute_error: 0.1244 - root_mean_squared_error: 0.2710 - val_loss: 0.0435 - val_mean_absolute_error: 0.0962 - val_root_mean_squared_error: 0.2087\n",
            "\n",
            "Epoch 00023: val_root_mean_squared_error improved from 0.21490 to 0.20866, saving model to /content/grp_cpu_models/weights-23-0.21.hdf5\n",
            "Epoch 24/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.0644 - mean_absolute_error: 0.1174 - root_mean_squared_error: 0.2538 - val_loss: 0.0422 - val_mean_absolute_error: 0.0971 - val_root_mean_squared_error: 0.2054\n",
            "\n",
            "Epoch 00024: val_root_mean_squared_error improved from 0.20866 to 0.20538, saving model to /content/grp_cpu_models/weights-24-0.21.hdf5\n",
            "Epoch 25/25\n",
            "217/217 [==============================] - 3s 12ms/step - loss: 0.0590 - mean_absolute_error: 0.1125 - root_mean_squared_error: 0.2430 - val_loss: 0.0402 - val_mean_absolute_error: 0.0939 - val_root_mean_squared_error: 0.2005\n",
            "\n",
            "Epoch 00025: val_root_mean_squared_error improved from 0.20538 to 0.20051, saving model to /content/grp_cpu_models/weights-25-0.20.hdf5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 361
        },
        "id": "mwOxUUDjs1ti",
        "outputId": "ffb002da-7b62-4705-dac8-902277011f0b"
      },
      "source": [
        "# 5. Analysing model's performance on group data using train-cv (loss and rmse) plots\r\n",
        "plt.figure(figsize = (16, 5.5))\r\n",
        "plt.subplot(1,2,1)\r\n",
        "plt.plot(hist.history['loss'][1:], label = 'train_loss')\r\n",
        "plt.plot(hist.history['val_loss'][1:], label = 'cv_loss')\r\n",
        "plt.xlabel('epoch')\r\n",
        "plt.ylabel('loss')\r\n",
        "plt.legend()\r\n",
        "\r\n",
        "plt.subplot(1,2,2)\r\n",
        "plt.plot(hist.history['root_mean_squared_error'][1:], label = 'train_rmse')\r\n",
        "plt.plot(hist.history['val_root_mean_squared_error'][1:], label = 'cv_rmse')\r\n",
        "plt.xlabel('epoch')\r\n",
        "plt.ylabel('rmse')\r\n",
        "plt.legend()\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x396 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nc7dMiLVs1tk",
        "outputId": "76a88c32-4722-4678-bec8-afb3537ac475"
      },
      "source": [
        "# 6. find best model - based on (metric we care about) best val_rmse \r\n",
        "print('epoch :', np.argmin(hist.history['val_loss']), ' | ',\\\r\n",
        "      'val_loss :', hist.history['val_loss'][np.argmin(hist.history['val_loss'])])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "epoch : 24  |  val_loss : 0.04020466282963753\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wieVyaePs1tl"
      },
      "source": [
        "# 7. load best model\r\n",
        "cpu_model.load_weights('/content/grp_cpu_models/weights-25-0.20.hdf5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fDOIchdKs1tm",
        "outputId": "07c83fed-e600-4bf3-d3ba-873d8b4f399e"
      },
      "source": [
        "# 8. Predictions\r\n",
        "idx = 104\r\n",
        "x_test = [grp_dense_x_cv[idx].reshape(1, -1), np.expand_dims(grp_lstm_x_cv[:, :, 1], axis = -1)[idx].reshape(1, 15, 1)]\r\n",
        "y_pred = cpu_model.predict(x_test)\r\n",
        "y_true = grp_y_cv[idx][1] \r\n",
        "print('predicted :', y_pred, ' | y_true :', y_true)\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "predicted : [[0.01807757]]  | y_true : 0.06\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NJyI-d1jtfGk"
      },
      "source": [
        "tensorflow.keras.utils.plot_model(cpu_model, show_shapes = True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rhW7knkF6Fii"
      },
      "source": [
        "### 7.2. Model for Memory_Used"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LsE-LD586Fik",
        "outputId": "39130519-0103-4925-98fa-3af3dac22886"
      },
      "source": [
        "# 1. building a model - CPU_Used\r\n",
        "tf.keras.backend.clear_session()\r\n",
        "\r\n",
        "dense_ins = Input(name = 'constants', shape = (44,))\r\n",
        "\r\n",
        "lstm_ins = Input(name = 'Sequence', shape = ((15, 1)))\r\n",
        "lstm_out = LSTM(name = 'LSTM', units = 128, kernel_initializer = 'glorot_normal')(lstm_ins)\r\n",
        "\r\n",
        "concat = Concatenate(axis = -1)([dense_ins, lstm_out])\r\n",
        "\r\n",
        "dense_1 = Dense(units = 64, activation = 'relu')(concat)\r\n",
        "dense_2 = Dense(units = 24, activation = 'relu')(dense_1)\r\n",
        "out = Dense(units = 1)(dense_2)\r\n",
        "memory_model = Model(inputs = [dense_ins, lstm_ins], outputs = out)\r\n",
        "memory_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "Sequence (InputLayer)           [(None, 15, 1)]      0                                            \n",
            "__________________________________________________________________________________________________\n",
            "constants (InputLayer)          [(None, 44)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "LSTM (LSTM)                     (None, 128)          66560       Sequence[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 172)          0           constants[0][0]                  \n",
            "                                                                 LSTM[0][0]                       \n",
            "__________________________________________________________________________________________________\n",
            "dense (Dense)                   (None, 64)           11072       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "dense_1 (Dense)                 (None, 24)           1560        dense[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "dense_2 (Dense)                 (None, 1)            25          dense_1[0][0]                    \n",
            "==================================================================================================\n",
            "Total params: 79,217\n",
            "Trainable params: 79,217\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IDngkKii6Fil"
      },
      "source": [
        "# 3. tensorflow-keras callbacks\r\n",
        "# 3.1 To save best models\r\n",
        "!rm -r mem_grp_models\r\n",
        "if not os.path.isdir('/content/mem_grp_models'):\r\n",
        "  os.mkdir('/content/mem_grp_models')\r\n",
        "save_model = tf.keras.callbacks.ModelCheckpoint(filepath = '/content/mem_grp_models/weights-{epoch:02d}-{val_loss:.2f}.hdf5',\r\n",
        "                                                monitor='val_loss', verbose=1,\r\n",
        "                                                save_best_only=True, save_weights_only=True)\r\n",
        "\r\n",
        "# 3.2 Reduce learning rate if model is not learning\r\n",
        "reduce_lr = tf.keras.callbacks.ReduceLROnPlateau(monitor = 'val_loss', factor = np.sqrt(0.1), patience = 4, verbose = 1, min_lr = 10**-6)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NZI5BvPR6Fim"
      },
      "source": [
        "# use RMSE for smaller values\r\n",
        "# use MAE for larger values or mean_squared_logarithmic_error\r\n",
        "# huber_loss - worked best"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3GsQkxda6Fio",
        "outputId": "aadc8bf7-f3cb-413f-94ee-673c6db2a5b7"
      },
      "source": [
        "# 4. training and evaluating model on cross validation\r\n",
        "rmse = tf.keras.metrics.RootMeanSquaredError(name=\"root_mean_squared_error\", dtype=None)\r\n",
        "opt = tf.keras.optimizers.Adam(learning_rate = 0.001)\r\n",
        "\r\n",
        "memory_model.compile(loss = 'huber_loss', optimizer = opt, metrics = ['mean_absolute_error', rmse])\r\n",
        "hist = memory_model.fit([grp_dense_x_train, np.expand_dims(grp_lstm_x_train[:, :, 0], axis = -1)], grp_y_train[:, 0], batch_size = 512, epochs = 30,\r\n",
        "                 validation_data = ([grp_dense_x_cv, np.expand_dims(grp_lstm_x_cv[:, :, 0], axis = -1)], grp_y_cv[:, 0]),\r\n",
        "                 callbacks = [save_model, reduce_lr]) "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/30\n",
            "867/867 [==============================] - 7s 6ms/step - loss: 10340.9674 - mean_absolute_error: 10341.3262 - root_mean_squared_error: 20636.2444 - val_loss: 12325.7383 - val_mean_absolute_error: 12326.0752 - val_root_mean_squared_error: 23626.9434\n",
            "\n",
            "Epoch 00001: val_loss improved from inf to 12325.73828, saving model to /content/mem_grp_models/weights-01-12325.74.hdf5\n",
            "Epoch 2/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 10002.1756 - mean_absolute_error: 10002.5347 - root_mean_squared_error: 20363.1348 - val_loss: 10307.3408 - val_mean_absolute_error: 10307.6582 - val_root_mean_squared_error: 21905.6445\n",
            "\n",
            "Epoch 00002: val_loss improved from 12325.73828 to 10307.34082, saving model to /content/mem_grp_models/weights-02-10307.34.hdf5\n",
            "Epoch 3/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 8363.2062 - mean_absolute_error: 8363.5704 - root_mean_squared_error: 18651.2674 - val_loss: 8976.3574 - val_mean_absolute_error: 8976.7168 - val_root_mean_squared_error: 20025.3320\n",
            "\n",
            "Epoch 00003: val_loss improved from 10307.34082 to 8976.35742, saving model to /content/mem_grp_models/weights-03-8976.36.hdf5\n",
            "Epoch 4/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 7599.6080 - mean_absolute_error: 7599.9824 - root_mean_squared_error: 17266.5456 - val_loss: 8254.8281 - val_mean_absolute_error: 8255.1455 - val_root_mean_squared_error: 19087.3340\n",
            "\n",
            "Epoch 00004: val_loss improved from 8976.35742 to 8254.82812, saving model to /content/mem_grp_models/weights-04-8254.83.hdf5\n",
            "Epoch 5/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 6731.3038 - mean_absolute_error: 6731.6680 - root_mean_squared_error: 15506.0355 - val_loss: 6834.7080 - val_mean_absolute_error: 6835.0454 - val_root_mean_squared_error: 16333.1240\n",
            "\n",
            "Epoch 00005: val_loss improved from 8254.82812 to 6834.70801, saving model to /content/mem_grp_models/weights-05-6834.71.hdf5\n",
            "Epoch 6/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 5687.3947 - mean_absolute_error: 5687.7580 - root_mean_squared_error: 13090.9161 - val_loss: 5229.0430 - val_mean_absolute_error: 5229.3838 - val_root_mean_squared_error: 12837.8750\n",
            "\n",
            "Epoch 00006: val_loss improved from 6834.70801 to 5229.04297, saving model to /content/mem_grp_models/weights-06-5229.04.hdf5\n",
            "Epoch 7/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 4968.5217 - mean_absolute_error: 4968.8879 - root_mean_squared_error: 11134.0441 - val_loss: 4591.6323 - val_mean_absolute_error: 4591.9766 - val_root_mean_squared_error: 11309.1133\n",
            "\n",
            "Epoch 00007: val_loss improved from 5229.04297 to 4591.63232, saving model to /content/mem_grp_models/weights-07-4591.63.hdf5\n",
            "Epoch 8/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 4596.5526 - mean_absolute_error: 4596.9148 - root_mean_squared_error: 10223.8054 - val_loss: 3959.0576 - val_mean_absolute_error: 3959.4163 - val_root_mean_squared_error: 9722.4854\n",
            "\n",
            "Epoch 00008: val_loss improved from 4591.63232 to 3959.05762, saving model to /content/mem_grp_models/weights-08-3959.06.hdf5\n",
            "Epoch 9/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 4227.9589 - mean_absolute_error: 4228.3265 - root_mean_squared_error: 9774.2965 - val_loss: 3580.9277 - val_mean_absolute_error: 3581.2622 - val_root_mean_squared_error: 8825.5430\n",
            "\n",
            "Epoch 00009: val_loss improved from 3959.05762 to 3580.92773, saving model to /content/mem_grp_models/weights-09-3580.93.hdf5\n",
            "Epoch 10/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 3698.6676 - mean_absolute_error: 3699.0320 - root_mean_squared_error: 10254.2779 - val_loss: 1860.9816 - val_mean_absolute_error: 1861.3419 - val_root_mean_squared_error: 5373.7358\n",
            "\n",
            "Epoch 00010: val_loss improved from 3580.92773 to 1860.98157, saving model to /content/mem_grp_models/weights-10-1860.98.hdf5\n",
            "Epoch 11/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 3182.7518 - mean_absolute_error: 3183.1192 - root_mean_squared_error: 9318.6193 - val_loss: 2111.9390 - val_mean_absolute_error: 2112.2822 - val_root_mean_squared_error: 6335.2495\n",
            "\n",
            "Epoch 00011: val_loss did not improve from 1860.98157\n",
            "Epoch 12/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 3285.6741 - mean_absolute_error: 3286.0421 - root_mean_squared_error: 11402.1735 - val_loss: 2155.7107 - val_mean_absolute_error: 2156.0295 - val_root_mean_squared_error: 6539.6938\n",
            "\n",
            "Epoch 00012: val_loss did not improve from 1860.98157\n",
            "Epoch 13/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 3189.7718 - mean_absolute_error: 3190.1368 - root_mean_squared_error: 11388.5398 - val_loss: 2108.2239 - val_mean_absolute_error: 2108.5828 - val_root_mean_squared_error: 6521.4614\n",
            "\n",
            "Epoch 00013: val_loss did not improve from 1860.98157\n",
            "Epoch 14/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 3176.5851 - mean_absolute_error: 3176.9462 - root_mean_squared_error: 11499.2324 - val_loss: 2095.8821 - val_mean_absolute_error: 2096.2263 - val_root_mean_squared_error: 6019.0063\n",
            "\n",
            "Epoch 00014: val_loss did not improve from 1860.98157\n",
            "\n",
            "Epoch 00014: ReduceLROnPlateau reducing learning rate to 0.00031622778103685084.\n",
            "Epoch 15/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2933.6540 - mean_absolute_error: 2934.0125 - root_mean_squared_error: 11236.4986 - val_loss: 2078.3604 - val_mean_absolute_error: 2078.7217 - val_root_mean_squared_error: 6476.5024\n",
            "\n",
            "Epoch 00015: val_loss did not improve from 1860.98157\n",
            "Epoch 16/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2895.6272 - mean_absolute_error: 2895.9855 - root_mean_squared_error: 11369.0010 - val_loss: 2225.7183 - val_mean_absolute_error: 2226.0857 - val_root_mean_squared_error: 6831.2070\n",
            "\n",
            "Epoch 00016: val_loss did not improve from 1860.98157\n",
            "Epoch 17/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2888.0117 - mean_absolute_error: 2888.3712 - root_mean_squared_error: 11055.6695 - val_loss: 2154.4487 - val_mean_absolute_error: 2154.7717 - val_root_mean_squared_error: 6757.3911\n",
            "\n",
            "Epoch 00017: val_loss did not improve from 1860.98157\n",
            "Epoch 18/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2870.7614 - mean_absolute_error: 2871.1209 - root_mean_squared_error: 11259.3891 - val_loss: 2118.1614 - val_mean_absolute_error: 2118.5154 - val_root_mean_squared_error: 6550.3198\n",
            "\n",
            "Epoch 00018: val_loss did not improve from 1860.98157\n",
            "\n",
            "Epoch 00018: ReduceLROnPlateau reducing learning rate to 0.00010000000639606199.\n",
            "Epoch 19/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2771.4377 - mean_absolute_error: 2771.7926 - root_mean_squared_error: 11149.2681 - val_loss: 2112.6443 - val_mean_absolute_error: 2112.9934 - val_root_mean_squared_error: 6578.9180\n",
            "\n",
            "Epoch 00019: val_loss did not improve from 1860.98157\n",
            "Epoch 20/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2774.1472 - mean_absolute_error: 2774.5039 - root_mean_squared_error: 11262.5601 - val_loss: 2070.6206 - val_mean_absolute_error: 2070.9600 - val_root_mean_squared_error: 6417.5010\n",
            "\n",
            "Epoch 00020: val_loss did not improve from 1860.98157\n",
            "Epoch 21/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2762.8170 - mean_absolute_error: 2763.1740 - root_mean_squared_error: 11135.8306 - val_loss: 2088.3354 - val_mean_absolute_error: 2088.6609 - val_root_mean_squared_error: 6477.5718\n",
            "\n",
            "Epoch 00021: val_loss did not improve from 1860.98157\n",
            "Epoch 22/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2761.1484 - mean_absolute_error: 2761.5045 - root_mean_squared_error: 11105.6630 - val_loss: 2121.2305 - val_mean_absolute_error: 2121.5903 - val_root_mean_squared_error: 6631.0776\n",
            "\n",
            "Epoch 00022: val_loss did not improve from 1860.98157\n",
            "\n",
            "Epoch 00022: ReduceLROnPlateau reducing learning rate to 3.1622778103685084e-05.\n",
            "Epoch 23/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2778.6455 - mean_absolute_error: 2779.0013 - root_mean_squared_error: 11277.8115 - val_loss: 2111.8416 - val_mean_absolute_error: 2112.1726 - val_root_mean_squared_error: 6610.6343\n",
            "\n",
            "Epoch 00023: val_loss did not improve from 1860.98157\n",
            "Epoch 24/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2768.9221 - mean_absolute_error: 2769.2759 - root_mean_squared_error: 11308.0922 - val_loss: 2095.7336 - val_mean_absolute_error: 2096.0625 - val_root_mean_squared_error: 6526.2651\n",
            "\n",
            "Epoch 00024: val_loss did not improve from 1860.98157\n",
            "Epoch 25/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2777.5326 - mean_absolute_error: 2777.8885 - root_mean_squared_error: 11299.5086 - val_loss: 2105.5039 - val_mean_absolute_error: 2105.8364 - val_root_mean_squared_error: 6578.2886\n",
            "\n",
            "Epoch 00025: val_loss did not improve from 1860.98157\n",
            "Epoch 26/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2779.7372 - mean_absolute_error: 2780.0917 - root_mean_squared_error: 11098.9177 - val_loss: 2172.2834 - val_mean_absolute_error: 2172.6147 - val_root_mean_squared_error: 6819.9043\n",
            "\n",
            "Epoch 00026: val_loss did not improve from 1860.98157\n",
            "\n",
            "Epoch 00026: ReduceLROnPlateau reducing learning rate to 1.0000000409520217e-05.\n",
            "Epoch 27/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2755.6103 - mean_absolute_error: 2755.9600 - root_mean_squared_error: 10567.0904 - val_loss: 2134.2534 - val_mean_absolute_error: 2134.5862 - val_root_mean_squared_error: 6768.1699\n",
            "\n",
            "Epoch 00027: val_loss did not improve from 1860.98157\n",
            "Epoch 28/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2771.7793 - mean_absolute_error: 2772.1285 - root_mean_squared_error: 10694.6178 - val_loss: 2124.7009 - val_mean_absolute_error: 2125.0320 - val_root_mean_squared_error: 6753.4663\n",
            "\n",
            "Epoch 00028: val_loss did not improve from 1860.98157\n",
            "Epoch 29/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2772.6819 - mean_absolute_error: 2773.0335 - root_mean_squared_error: 10791.8397 - val_loss: 2119.8469 - val_mean_absolute_error: 2120.1807 - val_root_mean_squared_error: 6717.4531\n",
            "\n",
            "Epoch 00029: val_loss did not improve from 1860.98157\n",
            "Epoch 30/30\n",
            "867/867 [==============================] - 5s 6ms/step - loss: 2657.2854 - mean_absolute_error: 2657.6398 - root_mean_squared_error: 10569.1697 - val_loss: 2130.1897 - val_mean_absolute_error: 2130.5205 - val_root_mean_squared_error: 6736.4404\n",
            "\n",
            "Epoch 00030: val_loss did not improve from 1860.98157\n",
            "\n",
            "Epoch 00030: ReduceLROnPlateau reducing learning rate to 3.1622778678900043e-06.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 361
        },
        "id": "x0CH4mh36Fip",
        "outputId": "7ab4dc1b-2a58-4f17-e319-5087d6ddbe4c"
      },
      "source": [
        "# 5. Analysing model's performance on group data using train-cv (loss and rmse) plots\r\n",
        "plt.figure(figsize = (16, 5.5))\r\n",
        "plt.subplot(1,2,1)\r\n",
        "plt.plot(hist.history['loss'], label = 'train_loss')\r\n",
        "plt.plot(hist.history['val_loss'], label = 'cv_loss')\r\n",
        "plt.xlabel('epoch')\r\n",
        "plt.ylabel('loss')\r\n",
        "plt.legend()\r\n",
        "\r\n",
        "plt.subplot(1,2,2)\r\n",
        "plt.plot(hist.history['root_mean_squared_error'], label = 'train_rmse')\r\n",
        "plt.plot(hist.history['val_root_mean_squared_error'], label = 'cv_rmse')\r\n",
        "plt.xlabel('epoch')\r\n",
        "plt.ylabel('rmse')\r\n",
        "plt.legend()\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x396 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lOyEP6GH6Fiq",
        "outputId": "8bedaae4-5fa5-42ee-ba9d-5107ecf208fb"
      },
      "source": [
        "# 6. find best model - based on (metric we care about) best val_rmse \r\n",
        "print('epoch :', np.argmin(hist.history['val_loss']), ' | ',\\\r\n",
        "      'val_loss :', hist.history['val_loss'][np.argmin(hist.history['val_loss'])])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "epoch : 9  |  val_loss : 1860.9815673828125\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3935HNe_6Fiq"
      },
      "source": [
        "# 7. load best model\r\n",
        "memory_model.load_weights('/content/mem_grp_models/weights-10-1860.98.hdf5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_-OhMA7PY9KI",
        "outputId": "a704b697-2646-4fa2-ee4e-7a1707a597c5"
      },
      "source": [
        "# 8. Predictions\r\n",
        "idx = 145\r\n",
        "x_test = [grp_dense_x_cv[idx].reshape(1, -1), np.expand_dims(\\\r\n",
        "          grp_lstm_x_cv[:, :, 0], axis = -1)[idx].reshape(1, 15, 1)]\r\n",
        "y_pred = memory_model.predict(x_test)\r\n",
        "y_true = grp_y_cv[idx][0] \r\n",
        "print('predicted :', y_pred, ' | y_true :', y_true)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "predicted : [[14626.854]]  | y_true : 16382.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U_jlZhzrtsIA"
      },
      "source": [
        "tensorflow.keras.utils.plot_model(memory_model, show_shapes = True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xoQRGeygpu4L"
      },
      "source": [
        "# 8. Final results - \r\n",
        "\r\n",
        "|  Models | train_rmse  | val_rmse  |train_mse |val_mse |train_mae|val_mae|\r\n",
        "|---------|-------------|-----------|----------|--------|---------|-------|\r\n",
        "| Instance_based_model  |131.7863   |99.2126   |-   | -  | -  |-   |\r\n",
        "| Group_based_model (CPU) |0.243   | 0.2005  |0.059   |0.0402   | 0.1125  |0.0939   |\r\n",
        "|Group_based_model (mem)   | -  | -  | -  | -  |3699.032   |1861.3419   |"
      ]
    }
  ]
}