Data_Science_Youtube_ETL.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "collapsed_sections": [
        "87VSMWdWATL_",
        "mMarXAGVaYpn",
        "H_mOw4d61txe",
        "RtT83Eq01z8Y",
        "mG-SJ8nncr6V"
      ]
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "<h1 style=\"text-align:center;\"><font color='red' size=10><b> Before you start! </b></font></h1>\n",
        "\n",
        "This project is divided in 3 parts:\n",
        "\n",
        "**1) ETL, where we will Extract, Transform and Load data into an AWS RDS.  <font color='red'>-> you are here</font>**\n",
        "\n",
        "https://colab.research.google.com/drive/1w9bPc49joLrceMWAF_RlEgXo73s1Eeee?usp=sharing\n",
        "\n",
        "<br><br>\n",
        "\n",
        "2) Data Analysis: exploratory data analysis to identify key features\n",
        "\n",
        "https://colab.research.google.com/drive/1a_Etj5kwEaq5epwoV9TVPNS3ShRRd7wU?usp=sharing\n",
        "\n",
        "<br><br>\n",
        "\n",
        "3) Prediction models: model building and comparison.\n",
        "\n",
        "https://colab.research.google.com/drive/1Nbj6TM5HaK2krMRa9R1oiolDhDronYQB?usp=sharing\n",
        "\n",
        "<br><br>\n",
        "\n",
        "Summary of this project: https://colab.research.google.com/drive/1CUjP7SdFGldPjuSVSIbHAk9UPWYp_RYz?usp=sharing\n",
        "\n",
        "<br><br>\n",
        "\n",
        "\n",
        "A summary of the data can be visualized on this <font color='red'>**Power BI dashboard:**</font> https://app.powerbi.com/view?r=eyJrIjoiNTkzZjNmY2UtNmQ5Mi00MTJhLTliNzgtZGU2NzRlYzQ5NDA1IiwidCI6IjE2OGQ0MTM3LWQ2ZjYtNDVmOC1hYWE3LWQxYTcwMjMzMDk1ZSIsImMiOjR9&pageName=ReportSection4f69a4c8629ea033a165"
      ],
      "metadata": {
        "id": "bO2wUSavERh6"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Analyzing Youtube channels that I am subscribed to using its API.**\n",
        "\n",
        "I love watching Youtube. It's a very diverse streaming service (even before video streaming was a thing), with videos ranging from comedy to science and curiosity, short or long, from channels with millions of subscribers to small comunnities with just a few hundred. \n",
        "\n",
        "But content creators suffer from \"punitive algorithms\" (in their words) that increase or decrease their visibility based on a plethora of metrics, more often than ever unknown to them. From YT perspective it's actually understandable, since a lot of viewers are migrating to other platforms (Tiktok, Twitch and an ever growing streaming services options). They need to maximize the amount of viewers and time spent on each video to justify that Youtube is actually a good platform to place ads.\n",
        "\n",
        "With this in mind I decided to use its API to collect some data and try to predict a video \"view count\" at the time it was published. Some models were built using video title features, thumbnail color, day and hour it was published, and so on. If successful, it could be used for example to fine tune ad placement in terms of expected return it will give. If unsuccessful it can at least be improved in future iterations.\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "yQEFZpNHZhLB"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **0) Import libraries**"
      ],
      "metadata": {
        "id": "87VSMWdWATL_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#!pip install psycopg2\n",
        "!pip install python-dotenv\n",
        "\n",
        "# Import libraries\n",
        "import requests\n",
        "import pandas as pd\n",
        "import time\n",
        "import datetime\n",
        "import pytz\n",
        "import numpy as np\n",
        "import psycopg2 as ps\n",
        "from google.colab import drive\n",
        "import os\n",
        "\n",
        "# Regular expressions\n",
        "import re\n",
        "\n",
        "# Environment variables\n",
        "from dotenv import load_dotenv\n",
        "\n",
        "# Image processing\n",
        "import matplotlib.image as img\n",
        "import matplotlib.pyplot as plt\n",
        "from skimage import io\n",
        "from PIL import Image\n",
        "from io import BytesIO\n",
        "import seaborn as sns"
      ],
      "metadata": {
        "id": "A9Zx1pwlAUmZ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "40eda7ec-f766-42c5-a881-c431368d65be"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Requirement already satisfied: python-dotenv in /usr/local/lib/python3.7/dist-packages (0.21.0)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **1) EXTRACT**"
      ],
      "metadata": {
        "id": "FZ-3NAXKGgv3"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.1) Import API keys\n",
        "The API key is in an environment variable that needs to be loaded from a .env file. Since Google Colab won't persist files between sessions, you need to store the .env file on your Googl Drive and then import/copy that file from yout private Google Drive to this Colab Notebook."
      ],
      "metadata": {
        "id": "mMarXAGVaYpn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cyG_xNCIf5IA",
        "outputId": "ed169353-4179-493d-d7fa-eb2868391e4a"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!cp /content/drive/My\\ Drive/Profissional/Data\\ Science/Projects/Youtube\\ DS/.env .env # this should be *YOUR* path inside Google Drive to the '.env' file\n",
        "drive.flush_and_unmount()\n",
        "load_dotenv()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "B6xDL3Zviy57",
        "outputId": "1444fccb-7eac-4369-fb57-3ec30e026987"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Keys\n",
        "API_KEY = os.getenv(\"API_KEY\")"
      ],
      "metadata": {
        "id": "f8gL-Uqa1qrx"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.2) Methods to extract data from video uploads"
      ],
      "metadata": {
        "id": "H_mOw4d61txe"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def check_response_successful(response):\n",
        "    \"\"\" Checks if the response from a given request was succesful.\n",
        "    \n",
        "    Raise an exception if an error was found, or\n",
        "    returns True otherwise.\n",
        "    \"\"\"\n",
        "    try:\n",
        "        error = response['error']\n",
        "        error_code = error['code']\n",
        "        error_message = error['message']\n",
        "        raise Exception(f\"{error_code}. \\\"{error_message}\\\"\")\n",
        "    except KeyError:\n",
        "        return True\n",
        "        "
      ],
      "metadata": {
        "id": "DktW1YJBIyi-"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_video_details(video_id):\n",
        "    \"\"\" Get the video details from a video_id\n",
        "    \n",
        "    Returns the channel id, video title, thumbnail url, \n",
        "    upload timestamp(ISO 8601 standard), video duration (ISO 8601 standard), \n",
        "    view count, like count and comment count.\n",
        "\n",
        "    Utilize 1 quota.\n",
        "    \"\"\"\n",
        "    url = base_url+f\"/videos?id={video_id}&part=snippet,contentDetails,statistics&key={API_KEY}\"\n",
        "    video_details = requests.get(url).json()\n",
        "    response_sucessful = check_response_successful(video_details)\n",
        "\n",
        "    channel_id = video_details['items'][0]['snippet']['channelId']\n",
        "    video_title = video_details['items'][0]['snippet']['title']\n",
        "    thumbnail_url = video_details['items'][0]['snippet']['thumbnails']['high']['url']\n",
        "    upload_date_time = video_details['items'][0]['snippet']['publishedAt']\n",
        "\n",
        "    video_duration = video_details['items'][0]['contentDetails']['duration']\n",
        "\n",
        "    view_count = video_details['items'][0]['statistics']['viewCount']\n",
        "    like_count = video_details['items'][0]['statistics']['likeCount']\n",
        "    #dislike_count = like_count = video_details['items'][0]['statistics']['dislikeCount'] # from Dec 13th 2021, Youtube doesn't return dislike counts anymore\n",
        "    comment_count = video_details['items'][0]['statistics']['commentCount']\n",
        "\n",
        "\n",
        "    return channel_id, video_title, thumbnail_url, upload_date_time, video_duration, view_count, like_count, comment_count"
      ],
      "metadata": {
        "id": "fqJAHXX4qT_3"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_page_results(pageToken, channel_id, publishedAfter, publishedBefore):\n",
        "    \"\"\" Get search results from a given page that match the channel_id,\n",
        "    \n",
        "    Returns the response.\n",
        "\n",
        "    Utilize 100 quota.\n",
        "    \"\"\"\n",
        "    \n",
        "    url = base_url+f\"/search?key={API_KEY}&channelId={channel_id}&part=id&publishedAfter={publishedAfter}&publishedBefore={publishedBefore}&order=date&maxResults=10000&pageToken={pageToken}\"\n",
        "    response = requests.get(url).json()\n",
        "\n",
        "    return response"
      ],
      "metadata": {
        "id": "_-aSA_Z4GfJv"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_videos_from_channel(df, channel_id, publishedAfter, publishedBefore):\n",
        "    \"\"\" Get all videos from a channel (iterate all page results), \n",
        "    published after some date and before some date, and store them into a pandas\n",
        "    df.\n",
        "\n",
        "    Returns a pandas df with video id, channel id, video title, thumbnail url,\n",
        "    upload date and time (ISO 8601 standard), video duration (ISO 8601 standard),\n",
        "    view count, like count and comment count.\n",
        "\n",
        "    Utilize 100 quota per page result and 1 quota per video of that page.\n",
        "    \"\"\"\n",
        "\n",
        "    pageToken = \"\"\n",
        "    # Iterate through all page results\n",
        "    while pageToken is not None:\n",
        "        # API call\n",
        "        response = get_page_results(pageToken, channel_id, publishedAfter, publishedBefore)\n",
        "        time.sleep(1)\n",
        "        response_successful = check_response_successful(response)\n",
        "                     \n",
        "        pageToken = response.get('nextPageToken', None)\n",
        "\n",
        "        # Iterate within page to get data\n",
        "        for video in response['items']:\n",
        "            if video['id']['kind'] == \"youtube#video\":\n",
        "                # Get video description\n",
        "                video_id = video['id']['videoId']\n",
        "\n",
        "                channel_id, video_title, thumbnail_url, upload_date_time, \\\n",
        "                    video_duration, view_count, like_count, comment_count = get_video_details(video_id)\n",
        "\n",
        "                # Save into pandas data frame\n",
        "                df = df.append(\n",
        "                    {\"video_id\": video_id,\n",
        "                    \"channel_id\": channel_id,\n",
        "                    \"video_title\": video_title,\n",
        "                    \"thumbnail_url\": thumbnail_url,\n",
        "                    \"upload_date_time\": upload_date_time,\n",
        "                    \"video_duration\": video_duration,\n",
        "                    \"view_count\": view_count,\n",
        "                    \"like_count\": like_count,\n",
        "                    \"comment_count\": comment_count},\n",
        "                    ignore_index=True\n",
        "                )\n",
        "\n",
        "    return df"
      ],
      "metadata": {
        "id": "Dv2UKolW5ryO"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_videos_from_channel_list(df, channel_list, publishedAfter=\"\", publishedBefore=\"\"):\n",
        "    \"\"\" Get all videos from a channel list, published after some date and before\n",
        "    some date, and store it into a pandas df.\n",
        "\n",
        "    Features extracted from each video depend on get_videos_from_channel.\n",
        "\n",
        "    Returns a pandas df.\n",
        "\n",
        "    Utilize 100 quota per page result per channel and 1 quota per video.\n",
        "    \"\"\"\n",
        "\n",
        "    if publishedBefore==\"\":\n",
        "        d = datetime.datetime.now()\n",
        "        publishedBefore = d.strftime('%Y-%m-%dT%H:%M:%SZ')\n",
        "\n",
        "    df_tmp = df\n",
        "\n",
        "    for channel_id in channel_list:\n",
        "        try:\n",
        "            print(f\"Getting videos from channel {channel_id}.\\n\")\n",
        "            df_tmp = get_videos_from_channel(df, channel_id, publishedAfter, publishedBefore)\n",
        "            df = df_tmp # doesn't throw away previous results in the case of a bad request\n",
        "        except KeyError as e:\n",
        "            print(f\"Error: key not identified on response: {e}.\")\n",
        "        except Exception as e:\n",
        "            print(f\"Error: {e}.\")\n",
        "        \n",
        "    # Sometimes the Youtube API returns duplicate data, so we need to remove it -> https://stackoverflow.com/questions/72438701/youtube-data-api-search-returning-repeating-items\n",
        "    df = df.drop_duplicates(subset=['video_id'])\n",
        "    \n",
        "    return df"
      ],
      "metadata": {
        "id": "y3j5Jvqrcqxo"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.3) Methods to extract data from channels"
      ],
      "metadata": {
        "id": "RtT83Eq01z8Y"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def get_channel_details(df, channel_id):\n",
        "    \"\"\" Get channel details from a channel id.\n",
        "\n",
        "    Returns the channel title, country, total view count, subscriber count, \n",
        "    total number of videos and topic categories (a list of Wikipedia pages).\n",
        "    \n",
        "    Utilize 1 quota.\n",
        "    \"\"\"\n",
        "    \n",
        "    url = base_url+f\"/channels?key={API_KEY}&id={channel_id}&part=snippet,statistics,topicDetails\"\n",
        "    response = requests.get(url).json()\n",
        "    \n",
        "    channel_title = response['items'][0]['snippet']['title']\n",
        "    channel_country = response['items'][0]['snippet']['country']\n",
        "    total_view_count = response['items'][0]['statistics']['viewCount']\n",
        "    subscriber_count = response['items'][0]['statistics']['subscriberCount']\n",
        "    video_count = response['items'][0]['statistics']['videoCount']\n",
        "    topic_categories = response['items'][0]['topicDetails']['topicCategories']\n",
        "\n",
        "    return channel_title, channel_country, total_view_count, subscriber_count, video_count, topic_categories"
      ],
      "metadata": {
        "id": "KCaNk1HmwE1h"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_channels_from_channel_list(df, channel_list):\n",
        "    \"\"\" Get all channels from a channel list store it into a pandas df.\n",
        "\n",
        "    Returns a pandas df with channel id, channel title, channel country,\n",
        "    total view count, subscriber count, video count and topic categories\n",
        "    (a list of Wikipedia pages).\n",
        "\n",
        "    Utilize 1 quota per channel.\n",
        "    \"\"\"\n",
        "\n",
        "    for channel_id in channel_list:\n",
        "        try:\n",
        "            print(f\"Getting channel details from channel {channel_id}.\\n\")\n",
        "\n",
        "            channel_title, channel_country, total_view_count, \\\n",
        "             subscriber_count, video_count, \\\n",
        "             topic_categories = get_channel_details(df, channel_id)\n",
        "            \n",
        "            # Save into pandas data frame\n",
        "            df = df.append(\n",
        "                {\"channel_id\": channel_id,\n",
        "                \"channel_title\": channel_title,\n",
        "                \"channel_country\": channel_country,\n",
        "                \"total_view_count\": total_view_count,\n",
        "                \"subscriber_count\": subscriber_count,\n",
        "                \"video_count\": video_count,\n",
        "                \"topic_categories\": topic_categories},\n",
        "                ignore_index=True\n",
        "            )\n",
        "        except KeyError as e:\n",
        "            print(f\"Error: key not identified on response: {e}.\")\n",
        "        except Exception as e:\n",
        "            print(f\"Error: {e}.\")\n",
        "            \n",
        "    return df"
      ],
      "metadata": {
        "id": "MGGDSCLV14LB"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# url = base_url+f\"/search?key={API_KEY}&channelId=UCJLtfES4K2aYCLFz-i5teFA&part=snippet,id&order=date&maxResults=10000\"\n",
        "# response = requests.get(url).json()\n",
        "# response\n",
        "\n",
        "# url_statistics = base_url+f\"/videos?id={response['items'][0]['id']['videoId']}&part=snippet,contentDetails,topicDetails,statistics&key={API_KEY}\"\n",
        "# response_statistics = requests.get(url_statistics).json()\n",
        "# response_statistics\n",
        "\n",
        "#url = base_url+f\"/channels?key={API_KEY}&id=UCYO_jab_esuFRV4b17AJtAw&part=snippet,statistics,topicDetails\"\n",
        "# response = requests.get(url).json()\n",
        "# response"
      ],
      "metadata": {
        "id": "lZXLbnzkTWyN"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.4) Main\n",
        "Here we set the list of channels that we want to analyze, the start and end date, and then start the extraction process."
      ],
      "metadata": {
        "id": "mG-SJ8nncr6V"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "base_url = \"https://www.googleapis.com/youtube/v3\""
      ],
      "metadata": {
        "id": "Gba7YWHH5u2X"
      },
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "CHANNEL_ID0 = \"UCHnyfMqiRRG1u-2MsSQLbXA\" # Veritasium\n",
        "CHANNEL_ID1 = \"UCYO_jab_esuFRV4b17AJtAw\" # 3Blue1Brown\n",
        "CHANNEL_ID2 = \"UCJLtfES4K2aYCLFz-i5teFA\" # Tá Gravando\n",
        "CHANNEL_ID3 = \"UCQxWq7wL4HY40mqbr3f0Z2A\" # UMotivo\n",
        "CHANNEL_ID4 = \"UCBFLqK7PAP9DQ3JpIrWFI7w\" # Pipocando\n",
        "CHANNEL_ID5 = \"UCK4AOYSWvdrgtKDEDDB6mMA\" # São Paulo nas Alturas\n",
        "CHANNEL_ID6 = \"UC6107grRI4m0o2-emgoDnAA\" # SmarterEveryDay\n",
        "CHANNEL_ID7 = \"UCm2CE2YfpmobBmF8ARLPzAw\" # Caue Moura\n",
        "CHANNEL_ID8 = \"UCW8Ews7tdKKkBT6GdtQaXvQ\" # StrataScratch\n",
        "CHANNEL_ID9 = \"UCYJv-TfmSU0xUuKI7N0zkJw\" # Carla Quem Disse\n",
        "\n",
        "channel_list = [CHANNEL_ID0, CHANNEL_ID1, CHANNEL_ID2, CHANNEL_ID3, CHANNEL_ID4,\n",
        "                CHANNEL_ID5, CHANNEL_ID6, CHANNEL_ID7, CHANNEL_ID8, CHANNEL_ID9]\n",
        "\n",
        "#channel_single_test = [CHANNEL_ID5]\n",
        "#channel_list = channel_single_test\n",
        "\n",
        "#last date: 2018-11-01\n",
        "publishedAfter = \"2022-11-16T00:00:00Z\"\n",
        "publishedBefore = \"2022-11-29T00:00:00Z\""
      ],
      "metadata": {
        "id": "nouu7r-vS30g"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_videos_temp = pd.DataFrame(columns=[\"video_id\",\"channel_id\",\"video_title\",\"thumbnail_url\",\"upload_date_time\",\"video_duration\",\"view_count\",\"like_count\",\"comment_count\"])\n",
        "\n",
        "df_videos_temp = get_videos_from_channel_list(df_videos_temp, channel_list, publishedAfter, publishedBefore)\n",
        "df_videos_api = df_videos_temp"
      ],
      "metadata": {
        "id": "HpziKDue1tjR",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ca7d730e-7873-48ca-dd48-40d7c90f66c0"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Getting videos from channel UCHnyfMqiRRG1u-2MsSQLbXA.\n",
            "\n",
            "Getting videos from channel UCYO_jab_esuFRV4b17AJtAw.\n",
            "\n",
            "Getting videos from channel UCJLtfES4K2aYCLFz-i5teFA.\n",
            "\n",
            "Getting videos from channel UCQxWq7wL4HY40mqbr3f0Z2A.\n",
            "\n",
            "Getting videos from channel UCBFLqK7PAP9DQ3JpIrWFI7w.\n",
            "\n",
            "Getting videos from channel UCK4AOYSWvdrgtKDEDDB6mMA.\n",
            "\n",
            "Getting videos from channel UC6107grRI4m0o2-emgoDnAA.\n",
            "\n",
            "Getting videos from channel UCm2CE2YfpmobBmF8ARLPzAw.\n",
            "\n",
            "Getting videos from channel UCW8Ews7tdKKkBT6GdtQaXvQ.\n",
            "\n",
            "Getting videos from channel UCYJv-TfmSU0xUuKI7N0zkJw.\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_videos_api.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "rnsV0CWGHvlI",
        "outputId": "1b0287a3-d972-41b2-89b3-b10d553f1bfc"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "      video_id                channel_id  \\\n",
              "0  QiVgJPVKRGI  UCHnyfMqiRRG1u-2MsSQLbXA   \n",
              "1  esQyYGezS7c  UCHnyfMqiRRG1u-2MsSQLbXA   \n",
              "2  KuXjwB4LzSA  UCYO_jab_esuFRV4b17AJtAw   \n",
              "3  55VfDUU1pyg  UCJLtfES4K2aYCLFz-i5teFA   \n",
              "4  bo53RcrZmc4  UCQxWq7wL4HY40mqbr3f0Z2A   \n",
              "\n",
              "                                         video_title  \\\n",
              "0                  Fire conducts electricity #shorts   \n",
              "1  The world depends on a strange collection of i...   \n",
              "2                         But what is a convolution?   \n",
              "3  POSSO COLOCAR MINHA LINGUA NA SUA BOCA? : MAND...   \n",
              "4  🐍 GRIXIS CROBA! O novo MID-RANGE do STANDARD! ...   \n",
              "\n",
              "                                      thumbnail_url      upload_date_time  \\\n",
              "0  https://i.ytimg.com/vi/QiVgJPVKRGI/hqdefault.jpg  2022-11-23T20:12:06Z   \n",
              "1  https://i.ytimg.com/vi/esQyYGezS7c/hqdefault.jpg  2022-11-19T14:00:24Z   \n",
              "2  https://i.ytimg.com/vi/KuXjwB4LzSA/hqdefault.jpg  2022-11-18T16:00:39Z   \n",
              "3  https://i.ytimg.com/vi/55VfDUU1pyg/hqdefault.jpg  2022-11-27T23:04:02Z   \n",
              "4  https://i.ytimg.com/vi/bo53RcrZmc4/hqdefault.jpg  2022-11-28T17:56:10Z   \n",
              "\n",
              "  video_duration view_count like_count comment_count  \n",
              "0          PT46S     485727      41891           466  \n",
              "1       PT18M32S    5692797     156666          9516  \n",
              "2        PT23M1S     833605      46089          1846  \n",
              "3       PT13M56S     327198      26979           934  \n",
              "4        PT1H35S       8038        717            44  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-c26841c5-0d01-46d4-8101-a6ef5e5ce67d\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>video_id</th>\n",
              "      <th>channel_id</th>\n",
              "      <th>video_title</th>\n",
              "      <th>thumbnail_url</th>\n",
              "      <th>upload_date_time</th>\n",
              "      <th>video_duration</th>\n",
              "      <th>view_count</th>\n",
              "      <th>like_count</th>\n",
              "      <th>comment_count</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>QiVgJPVKRGI</td>\n",
              "      <td>UCHnyfMqiRRG1u-2MsSQLbXA</td>\n",
              "      <td>Fire conducts electricity #shorts</td>\n",
              "      <td>https://i.ytimg.com/vi/QiVgJPVKRGI/hqdefault.jpg</td>\n",
              "      <td>2022-11-23T20:12:06Z</td>\n",
              "      <td>PT46S</td>\n",
              "      <td>485727</td>\n",
              "      <td>41891</td>\n",
              "      <td>466</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>esQyYGezS7c</td>\n",
              "      <td>UCHnyfMqiRRG1u-2MsSQLbXA</td>\n",
              "      <td>The world depends on a strange collection of i...</td>\n",
              "      <td>https://i.ytimg.com/vi/esQyYGezS7c/hqdefault.jpg</td>\n",
              "      <td>2022-11-19T14:00:24Z</td>\n",
              "      <td>PT18M32S</td>\n",
              "      <td>5692797</td>\n",
              "      <td>156666</td>\n",
              "      <td>9516</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>KuXjwB4LzSA</td>\n",
              "      <td>UCYO_jab_esuFRV4b17AJtAw</td>\n",
              "      <td>But what is a convolution?</td>\n",
              "      <td>https://i.ytimg.com/vi/KuXjwB4LzSA/hqdefault.jpg</td>\n",
              "      <td>2022-11-18T16:00:39Z</td>\n",
              "      <td>PT23M1S</td>\n",
              "      <td>833605</td>\n",
              "      <td>46089</td>\n",
              "      <td>1846</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>55VfDUU1pyg</td>\n",
              "      <td>UCJLtfES4K2aYCLFz-i5teFA</td>\n",
              "      <td>POSSO COLOCAR MINHA LINGUA NA SUA BOCA? : MAND...</td>\n",
              "      <td>https://i.ytimg.com/vi/55VfDUU1pyg/hqdefault.jpg</td>\n",
              "      <td>2022-11-27T23:04:02Z</td>\n",
              "      <td>PT13M56S</td>\n",
              "      <td>327198</td>\n",
              "      <td>26979</td>\n",
              "      <td>934</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>bo53RcrZmc4</td>\n",
              "      <td>UCQxWq7wL4HY40mqbr3f0Z2A</td>\n",
              "      <td>🐍 GRIXIS CROBA! O novo MID-RANGE do STANDARD! ...</td>\n",
              "      <td>https://i.ytimg.com/vi/bo53RcrZmc4/hqdefault.jpg</td>\n",
              "      <td>2022-11-28T17:56:10Z</td>\n",
              "      <td>PT1H35S</td>\n",
              "      <td>8038</td>\n",
              "      <td>717</td>\n",
              "      <td>44</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c26841c5-0d01-46d4-8101-a6ef5e5ce67d')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-c26841c5-0d01-46d4-8101-a6ef5e5ce67d button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-c26841c5-0d01-46d4-8101-a6ef5e5ce67d');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_channels_temp = pd.DataFrame(columns=[\"channel_id\", \"channel_title\", \"channel_country\", \"total_view_count\", \"subscriber_count\", \"video_count\", \"topic_categories\"])\n",
        "df_channels_temp = get_channels_from_channel_list(df_channels_temp, channel_list)\n",
        "df_channels_api = df_channels_temp"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sKxi0BaRvh6c",
        "outputId": "eeaa0d30-3482-47ec-83eb-948df7c7ecff"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Getting channel details from channel UCHnyfMqiRRG1u-2MsSQLbXA.\n",
            "\n",
            "Getting channel details from channel UCYO_jab_esuFRV4b17AJtAw.\n",
            "\n",
            "Getting channel details from channel UCJLtfES4K2aYCLFz-i5teFA.\n",
            "\n",
            "Getting channel details from channel UCQxWq7wL4HY40mqbr3f0Z2A.\n",
            "\n",
            "Getting channel details from channel UCBFLqK7PAP9DQ3JpIrWFI7w.\n",
            "\n",
            "Getting channel details from channel UCK4AOYSWvdrgtKDEDDB6mMA.\n",
            "\n",
            "Getting channel details from channel UC6107grRI4m0o2-emgoDnAA.\n",
            "\n",
            "Getting channel details from channel UCm2CE2YfpmobBmF8ARLPzAw.\n",
            "\n",
            "Getting channel details from channel UCW8Ews7tdKKkBT6GdtQaXvQ.\n",
            "\n",
            "Getting channel details from channel UCYJv-TfmSU0xUuKI7N0zkJw.\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_channels_api.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "tiK3HWPN1o7H",
        "outputId": "cc06ddf5-626d-4661-dc51-2e298d0fbc1a"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                 channel_id channel_title channel_country total_view_count  \\\n",
              "0  UCHnyfMqiRRG1u-2MsSQLbXA    Veritasium              US       1891837444   \n",
              "1  UCYO_jab_esuFRV4b17AJtAw   3Blue1Brown              US        311356579   \n",
              "2  UCJLtfES4K2aYCLFz-i5teFA   Tá Gravando              BR        257307314   \n",
              "3  UCQxWq7wL4HY40mqbr3f0Z2A       UMotivo              BR         27752172   \n",
              "4  UCBFLqK7PAP9DQ3JpIrWFI7w     Pipocando              BR        689510386   \n",
              "\n",
              "  subscriber_count video_count  \\\n",
              "0         13100000         343   \n",
              "1          4840000         127   \n",
              "2          1600000         321   \n",
              "3           103000        2528   \n",
              "4          4320000        1343   \n",
              "\n",
              "                                    topic_categories  \n",
              "0  [https://en.wikipedia.org/wiki/Knowledge, http...  \n",
              "1          [https://en.wikipedia.org/wiki/Knowledge]  \n",
              "2  [https://en.wikipedia.org/wiki/Entertainment, ...  \n",
              "3  [https://en.wikipedia.org/wiki/Strategy_video_...  \n",
              "4  [https://en.wikipedia.org/wiki/Entertainment, ...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-2e95e228-51a5-4af6-a19d-b78812c23442\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>channel_id</th>\n",
              "      <th>channel_title</th>\n",
              "      <th>channel_country</th>\n",
              "      <th>total_view_count</th>\n",
              "      <th>subscriber_count</th>\n",
              "      <th>video_count</th>\n",
              "      <th>topic_categories</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>UCHnyfMqiRRG1u-2MsSQLbXA</td>\n",
              "      <td>Veritasium</td>\n",
              "      <td>US</td>\n",
              "      <td>1891837444</td>\n",
              "      <td>13100000</td>\n",
              "      <td>343</td>\n",
              "      <td>[https://en.wikipedia.org/wiki/Knowledge, http...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>UCYO_jab_esuFRV4b17AJtAw</td>\n",
              "      <td>3Blue1Brown</td>\n",
              "      <td>US</td>\n",
              "      <td>311356579</td>\n",
              "      <td>4840000</td>\n",
              "      <td>127</td>\n",
              "      <td>[https://en.wikipedia.org/wiki/Knowledge]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>UCJLtfES4K2aYCLFz-i5teFA</td>\n",
              "      <td>Tá Gravando</td>\n",
              "      <td>BR</td>\n",
              "      <td>257307314</td>\n",
              "      <td>1600000</td>\n",
              "      <td>321</td>\n",
              "      <td>[https://en.wikipedia.org/wiki/Entertainment, ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>UCQxWq7wL4HY40mqbr3f0Z2A</td>\n",
              "      <td>UMotivo</td>\n",
              "      <td>BR</td>\n",
              "      <td>27752172</td>\n",
              "      <td>103000</td>\n",
              "      <td>2528</td>\n",
              "      <td>[https://en.wikipedia.org/wiki/Strategy_video_...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>UCBFLqK7PAP9DQ3JpIrWFI7w</td>\n",
              "      <td>Pipocando</td>\n",
              "      <td>BR</td>\n",
              "      <td>689510386</td>\n",
              "      <td>4320000</td>\n",
              "      <td>1343</td>\n",
              "      <td>[https://en.wikipedia.org/wiki/Entertainment, ...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-2e95e228-51a5-4af6-a19d-b78812c23442')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-2e95e228-51a5-4af6-a19d-b78812c23442 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-2e95e228-51a5-4af6-a19d-b78812c23442');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **2) TRANSFORM**"
      ],
      "metadata": {
        "id": "3HmisldzGzZm"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2.1) Get number of words, questions marks, exclamation marks and ellipsis in a title\n",
        "For this we will use regular expressions (regex)."
      ],
      "metadata": {
        "id": "2XZVBqJfIB7Y"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_videos_api['n_words_title'] = df_videos_api['video_title'].apply(lambda x: len(re.findall(r'\\w+', x)))\n",
        "df_videos_api['n_question_marks_title'] = df_videos_api['video_title'].apply(lambda x: len(re.findall(r'\\?+', x)))\n",
        "df_videos_api['n_exclamation_marks_title'] = df_videos_api['video_title'].apply(lambda x: len(re.findall(r'\\!+', x)))\n",
        "df_videos_api['n_ellipsis_title'] = df_videos_api['video_title'].apply(lambda x: len(re.findall(r'\\.{3,}', x)))"
      ],
      "metadata": {
        "id": "w9ojA7KOHBUG"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_videos_api[['video_title','n_words_title','n_question_marks_title','n_exclamation_marks_title','n_ellipsis_title']].head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "eIVePkG4H68e",
        "outputId": "750bbefd-b839-4cd8-9c30-2f8b4c510688"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                         video_title  n_words_title  \\\n",
              "0                  Fire conducts electricity #shorts              4   \n",
              "1  The world depends on a strange collection of i...             13   \n",
              "2                         But what is a convolution?              5   \n",
              "3  POSSO COLOCAR MINHA LINGUA NA SUA BOCA? : MAND...             11   \n",
              "4  🐍 GRIXIS CROBA! O novo MID-RANGE do STANDARD! ...             10   \n",
              "\n",
              "   n_question_marks_title  n_exclamation_marks_title  n_ellipsis_title  \n",
              "0                       0                          0                 0  \n",
              "1                       0                          0                 0  \n",
              "2                       1                          0                 0  \n",
              "3                       1                          0                 0  \n",
              "4                       0                          2                 0  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-050bb94d-fd94-4c94-987d-49936de7ff2b\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>video_title</th>\n",
              "      <th>n_words_title</th>\n",
              "      <th>n_question_marks_title</th>\n",
              "      <th>n_exclamation_marks_title</th>\n",
              "      <th>n_ellipsis_title</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Fire conducts electricity #shorts</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>The world depends on a strange collection of i...</td>\n",
              "      <td>13</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>But what is a convolution?</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>POSSO COLOCAR MINHA LINGUA NA SUA BOCA? : MAND...</td>\n",
              "      <td>11</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>🐍 GRIXIS CROBA! O novo MID-RANGE do STANDARD! ...</td>\n",
              "      <td>10</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-050bb94d-fd94-4c94-987d-49936de7ff2b')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-050bb94d-fd94-4c94-987d-49936de7ff2b button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-050bb94d-fd94-4c94-987d-49936de7ff2b');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2.2) Transform video_duration to seconds\n",
        "Youtube API returns video duration in ISO8601 format. This is transformed to duration in seconds."
      ],
      "metadata": {
        "id": "jUSeWIsDJD9q"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def YTDurationToSeconds(duration):\n",
        "    \"\"\" Transform an ISO8601 duration (time interval) to seconds.\n",
        "\n",
        "    Refer to: https://stackoverflow.com/questions/16742381/how-to-convert-youtube-api-duration-to-seconds\n",
        "\n",
        "    Returns an integer containing the seconds.\n",
        "    \"\"\"\n",
        "    match = re.match('PT(\\d+H)?(\\d+M)?(\\d+S)?', duration).groups()\n",
        "    hours = _js_parseInt(match[0]) if match[0] else 0\n",
        "    minutes = _js_parseInt(match[1]) if match[1] else 0\n",
        "    seconds = _js_parseInt(match[2]) if match[2] else 0\n",
        "    return hours * 3600 + minutes * 60 + seconds\n",
        "\n",
        "# js-like parseInt\n",
        "# https://gist.github.com/douglasmiranda/2174255\n",
        "def _js_parseInt(string):\n",
        "    return int(''.join([x for x in string if x.isdigit()]))"
      ],
      "metadata": {
        "id": "uImct1G3Jc09"
      },
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_videos_api['video_duration'] = df_videos_api['video_duration'].apply(lambda x: YTDurationToSeconds(x))\n",
        "df_videos_api[['video_title','video_duration']].head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "hVyQl4r2LYko",
        "outputId": "8cafa8fd-5a59-4d64-8640-21f44665fe5b"
      },
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                         video_title  video_duration\n",
              "0                  Fire conducts electricity #shorts              46\n",
              "1  The world depends on a strange collection of i...            1112\n",
              "2                         But what is a convolution?            1381\n",
              "3  POSSO COLOCAR MINHA LINGUA NA SUA BOCA? : MAND...             836\n",
              "4  🐍 GRIXIS CROBA! O novo MID-RANGE do STANDARD! ...            3635"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-8011bbd2-709c-4be8-a2f8-d5a494ecb825\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>video_title</th>\n",
              "      <th>video_duration</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Fire conducts electricity #shorts</td>\n",
              "      <td>46</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>The world depends on a strange collection of i...</td>\n",
              "      <td>1112</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>But what is a convolution?</td>\n",
              "      <td>1381</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>POSSO COLOCAR MINHA LINGUA NA SUA BOCA? : MAND...</td>\n",
              "      <td>836</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>🐍 GRIXIS CROBA! O novo MID-RANGE do STANDARD! ...</td>\n",
              "      <td>3635</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-8011bbd2-709c-4be8-a2f8-d5a494ecb825')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-8011bbd2-709c-4be8-a2f8-d5a494ecb825 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-8011bbd2-709c-4be8-a2f8-d5a494ecb825');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2.3) Get the main color of the thumbnail in RGB format\n"
      ],
      "metadata": {
        "id": "ZaEQkg_SIdXX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def get_img_from_url(url):\n",
        "    \"\"\" Make a request for an url containing an image and returns an np array \n",
        "    containin the image.\n",
        "\n",
        "    Returns an np array containing an image.\n",
        "    \"\"\"\n",
        "    response = requests.get(url)\n",
        "    img = Image.open(BytesIO(response.content))\n",
        "    img = np.array(img)\n",
        "\n",
        "    return img\n",
        "\n",
        "# Test\n",
        "url = df_videos_api['thumbnail_url'][0]\n",
        "\n",
        "img = get_img_from_url(url)\n",
        "plt.imshow(img)"
      ],
      "metadata": {
        "id": "CqW09Z5gHC0A",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "9c74b391-51b7-414c-f41e-347282e8c368"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f8d69102790>"
            ]
          },
          "metadata": {},
          "execution_count": 22
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def remove_black_bars(img):\n",
        "    \"\"\" Remove black bars from the top and from the bottom of the image.\n",
        "\n",
        "    Returns an np array containing an image.\n",
        "    \"\"\"\n",
        "    h, w, c = img.shape\n",
        "    img = img[int(0.13*h):int(-0.13*h), :, :]\n",
        "\n",
        "    return img\n",
        "\n",
        "img = remove_black_bars(img)\n",
        "plt.imshow(img)"
      ],
      "metadata": {
        "id": "sr0pWlILHEON",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "3669d75b-3d3d-4884-a139-438966567df0"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f8d68b9f290>"
            ]
          },
          "metadata": {},
          "execution_count": 23
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAADdCAYAAABAMDLCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOy9XYwkWXbf9zv3RkRmfXT1zOzs7C61FCmbpD8k2DBkiIYlyyQMC5JsgW+E7RfaELBPehff/MpXAwYM74Ng6cGWBAOEZJqQZMsm/ESAkCDb+qJBUcvlLmd3dj67PjIj7r3n+OHeGxmZGVldNd3VU92Tp5GdVVmRETe+/nHu//zPOWJmHO1oRzva0d4sc1/0AI52tKMd7Wgv347gfrSjHe1ob6Adwf1oRzva0d5AO4L70Y52tKO9gXYE96Md7WhHewPtCO5HO9rRjvYG2oOBu4j8WRH5bRH5HRH55YfaztGOdrSjHW3f5CF07iLigf8P+I+B7wG/BfznZvZPX/rGjna0ox3taHv2UJ77nwB+x8x+18wG4K8Dv/BA2zra0Y52tKPtWPNA6/1DwO9Pfv8e8LPTBUTkW8C3yq9/XEQeaChHu4t57xnPgQDMz+j2J3q2tajVhUR4kTM6rueOJhio0TaOs5MFXdvgMIS8O9trssnv24M3LA8f8j4ICALTQwOo2va6rPxRBMwwy+up+yD1eNT1bY1990jZ5BjkT5Ll46x13WWcVhbJn+WfNZWfTVCz8srLgOFF6LqWxnsQw1Feks+b4ri8GRiiYiKICGaGk3o0y/bK6zHb9Nx9LpuehLk/ffH2oZl9de4PDwXuzzUz+zbwbQDnnLVt+0UN5UthczeilBu3aRouLi5wLk/kfJcw62GExvxSFTCHmoHJZJ22dbOb2cGbSu8zWUzD7H7MWUtDDAOdC/zRf+3H+KM/84d5uhAWBE5F6ZKycA6JibWLJL95dmgByWQwDANq0LYdvmnonNDi8AiNCB6HE2G1WmXQFFAxFPBeaBysh0CKiZtVT4gJ51uojxrnEHE4hDYoTsA5h3eCd+WAqWKWMFXMFDPjyiJRjSEl+pAIagxJGRL0KX8eYiIm5fp6RUiOkBpW0dMHRx+Mvr/BE/ixr73FN7/+NmcnjoWLtJLwYhnsmwVrOn7z//0X/MHH16SmI5lDnEPMISY450gpMQwDIYTxOnrZdtuD467beymjmhmGvYSVv4yHo6r+3qG/PRS4fx/48cnv3yyfHe0LsrmLyMwO3CQKLgIOTAAPSHVKESQDlUnxihNQb7jqP86MATDxdx3wxtu9w75E9bj2lGgD3//kiicfPeMnvvEW597RCLg0cIKncYnoFBMwEdQonqkQYiCo4psO13WYOJIqnYAT8AgdgsNxen5BMmWwRK+JoAlx4FVpxQFG6Adu+oFu6cA7FFDLx07MWIbsRXtnOGc0ziFiONN8oOs7hopiovlYi21mCVKPiZBUUIWoSkjKEIUYPCnl8+dkwaJpePrkgpNlg5NAa5EGRcURaFHrGNRjZjRiiBlJFdd2aMjjUdVX4rW/jAfGSxnhA5EKD338Hgrcfwv4aRH5I2RQ/8+A/+KBtnW0F7BdjxtAFcz5AuzlZdlrzxN4yrsBmimRTBRwG7gLkgHrBW3uph9nIV3H5eUVv//993nnScf50xPUjD4kOm8453DS4KVSHJliUTWG1QDiaBctjWuJqXjPDjA3ucmNxWJJUsUiDBYxS6RgRAPEow6CQkiGhYSoA99gKMkU1BDNMxwxwyVwYnjAOfACXgTnBURIYSCakcbX5MiXBxUi+YFklJeRLJFMMPM0jbBctnRdg5M8A7GUHxriff5+AfNKK5nlY+N0Qlcd7bWwBwF3M4si8peAv0t2+/6Kmf2Th9jW0V7cdsHdzKGWvb0M7C6/i8cqoJhgQp6mF2CXjAYHuXIDRO7orRxY7JA317aeGCNmENT45NNn/MH7H3Lq38MvHE2C1pS2yzMHUcFSIkUlpkRUJQ2KazyiDo/PXr0lMMXQ8tgSMAj9QDIlaERTAi2POmlRg5shcLka6EOisYBvHc5cBt+YiCmyMnBmGIqYgilehMYLbeNZtA2tNHgnRIOkkFSIakQ1QoIhQUhKSMaQjJS0LAdJjWSKakQ1IZZofIdHSaEHBjwB12TaqXLrKSU2p9LGWMQmsHC018EejHM3s18Hfv2h1n+0l2OqinOueGg6UhSq9dKQ0YOXSUBt/BtkCqUE9PaDl9Ol66PhLnY/vlWc4VuHWsT5lqTw/g8/5qTxNF99i3MR1ihNUpwKqkqIkSFEhhhJaoAjBSUMia41GtcBKfPfZUSKImTOOZoSU2SIgSFFBnMM6uiHgY8//pQf/OhTfNtyctLQkWhdAyh9iPTDwJAiqKKaME2gCe+gazyLruV0uWC5XNDSEDTTLkEhGkQV+qishsSQjCEk1jFlLjwJMWVwj5pImsG6dUbXLWgaAY1gCSWhDpoaONX8gKiee53dWPHcj9j++tgXFlA92uOxMQgqNUjqwPa58Qzc2aWrIJ61Fbb1z90CzDaHDgcW18kMYAroc+AeUs/JySkhaKYPfMvVqueDjy5598k5i6UnIPRqNBgpKkOI9CESQiKZ0S5aYkz0faTtjOVJC0Q0RRQhWebbEUGj0seBq/WKZ6sbbvo1VxEuQ0sIgaurK65WgbcW5wwmWDIk5YMYktHHxFUYUFU0RixFTBPeCYvWszQjOY+2LZ0oICRzRJRojmAwJGUdIkPM6xuiEpMS1ZHIFJGSUAwRY7HoOD1tWS4aOgynDtI291sf8qPah0q0HZH989oXpSg6gvuX3JxzI1hWkPfq8SU4WmV3mI4e4JaJIaJA5dJvD6iK7atlDgXn6md1fJVXn6dmjL5fZyWK6whRSOr5wQfPuDg5YfFjb9M1njWJU9/y6aefcH2zwrlMNTXdgsvrnj5Ekq745HLF+ZMnPDlfcH6yBNeQYiKFhBh4Jzy7Hvjws2d8+NlnPLu54nIwVpxu9r454ZPrNd2gtG3Cr2IOSmskpMTaIISEhog4o/MNIQVigHVcMSRjcf6EplnQ+JbQD6zXgZs+EBKshsDVqmdQR0wQU+bikwohGkOMRFUQ6LqG9772Fu++9QRPpJFMAbm2w7mqcAJTpe/X9P1ASmDOaJocf/BHuPjc9kUA/PFsHW3fko6BT0GRontvRlzeALhgiOlGa20bP2/fDpE2lpmfvU83QF5lmgfBXep2pQQQHWotGiPf/8FHnC8dJ19/G8XoV9esgqKuQUW4WfeE6zXJIKplimbd89HVDU8vTvjmN97jvXee4NSIN2vEhPUw8MGn13x8ucK6E7wJ6/6aT9dpMnpwKKc0JBSfAqaJFCNBldi1pJSDn1K/ph7vHFET1+uem/XAyemTTN9IIir0QRliYh0SQR3JhGhGsKz+CRHENYiHlNY0jePkrOX0tKVrDWIsM4QWVSsqnKyHT6qZljlw9o72+tgR3I+2bxqREHEOxEl+L7K4DKI6gmmGLz9qZeBw3ofJAZ9e5kFfze1564cCqiJWAApUC4ViHpLjo2cr3v/Rpzw5a3n3rRNCTPRANGO97rm6XrHqe2JSnG8R36KmDOGaH37yMc9uesyf8rWvfBUvpwyrNVchcmMta1mg5rgcej7rjeso08yAfMwaJZHwApYiMQaiKuKaTIHhMVNizGoZNQEVQh+4ulrxztt5f4ZkrIOyHiI3Q6CPRlAhqBFTVueoGkOC1rfg8jFpF44nFyecnS1oW4dToxFh2XqG6AiqmVNPiRgTMYb5E3i018qO4H60PfNmNKZ4wIvDe4fzgmpgA+yZikl4An5LBJl18PPg4ObUMiJ7WZoGuB1wn9JHMyseZw5JLU88FLx58C2fXd3wgw8/4vz8G/i2w5JxtVrx6bPLTF+kDJpt52hckz1gMfo+cvW9H3K1Vv7IN6/56ttfwWLko08uGaTD2lM+fvYZP/rkisu1YYtFHr1tZKLRBK8g3iG+xQE+KWa+PEAbTBOW8vGNGaUxFYZ1ZL0aQLKcs76GqIQIwTLAR4WYNO+7eVIyoinihcWiZbls8Y3RNOBCnnFpSqOyySz/HkIghHjva+Zoj8+O4H60PWtEWHrw3tG2DU3ryDJoX/iDBKKA0SfHZ/0OMB9IjhKR2YDqIaolFX5+Ghc4aC7LMrOEPGuzNSpJlQbj2fWaH30cefutE56eX+C6BUmuuQkDMRUNzGJBEkdMmauOMaHmCEPiX3z3fT788BlvnV/QiudmdYNrPNGUz1bXXPcJlQatYyYh5oquXcB5fNuyaD1SgpbXQ6JpstRRUyQO2bMP/Qon0HUtmpTry2ukEVb9miFkr98QTIwUbZQ9BtWsSfeLogBKdCcdy+UC5yGmAb84wXkHKTH0AXOWc9UsyyZTSsR4BPc3wR4JuFdP8GXbrnTvy2vepNQiEXRUN2aaxZxiLuWERzOctDRyyqKFk8Y47SKdS7TeEJdBPZmRknFpjt41pJRycg6AuJEjB6DWPSnJN7vmhDGpJj8AwEzpLBJwBDwqLb7xiEZ0WIEpzjeYa4g4nGWtuZjhLKIWgYRqYEgJJw2f9Q3f+WDNN+wJF2cnWHuBNitCWuOcp2kaQoiklKWRKSSGCEk9psYHn13xwUfPMM26eskDRwFzDhXHMKzxIogTGoFWjKXAmQinIixEaL3HxNCwpvWCd55osMLoUyIky+n+NKyt4ZPrQNMUuWMwQhJCgpCEqJmHjyao5gQmJxFLPd4pFyennC8WdJkdwkLM8ypTkqWSuwDJhJASIWmJdXicT0QciAccascGELfZoaDpUS0jDwDuNfnmaDgVxIQkGdwVyqGJGWhcwklOpjFZAEsaSZy4gacucep7lj7QiKGSwWWtCrbkqjknWCBq9vhyrRQ3Zjya1gQgI6V9r1AQGic0vsE5l+WBCXArbrRhoCVKi7iGBqOVrNmOJK4HQ7sz2qITd6Y4Ao0oSRLqFZoW7TpWsuAPnnmsSQxqaFwSOSeaZ+kbUoykoAgNi1LrxUTBe1Q8OEXLw03rTMKs8NVG0h4Flt2SJ8slT06WnHUNCxFOG+hEIAYIA2rKqfS5tIFrWLtEb5GUEiouyx4DDC7r8098JCZjCMIQpfDtENWh5nJWcSHDUrimkcT56ZJ3z045a1sWZpzQoENEUXCG73wRrua6QVldY+AacB5pjKQZ6A2P2dGjf549pkJqjwfcj/awJln/beLI/lcFJ18ceI8gODWUSJ9WdA6iQPINqQFtWpIjg3YsdU5Ci0suv4q37rzPyS9aapGUdEe1mue5ba78bZopqxhmQq0umQOOAZcGNKxJMYBvMbfMu1G/VzTa9b3+TVVHuuHy2TOIAQH6vmjNS/KOquK94H3W+XcYPlJqu5BnPWosu0WRDuZ9MzPEe/xiybLrOFsuOT/pWLYeC2tSv+Jm6ElhzdCvSaqcfe0tfNOQcJAE3yTaFpIoKVkuBGY9Q0jIoiZepU15AK0lAnRMNBLLdNTJacPTJ+c8eXJO58FrYIsUs1yXxqrc1dxepvLRXm87gvuXxmogNNeFGYuCmSAKoh4ngsvpOiiJgDCYY23gzCM0+JyNTsBYW6DHU1h40nRrZqgpGON7Bu19m45Kq/dZtqPlYWQFnNP6iotOeO/r73G1Hkg9DKbF208FqDevafZtSnmEV9eBOAw0Pq/XkYORMUZMFe9bXNPifEsrijSVR5cxgevtp2/hkG1wN8led0poCKzXK9YpgCTicEMYeoRE23i65YK26zJ3nupeS67xkkA1ZookGuIjLfmBmZJmuWaqDzRKhjBj6YfGC09OFrx1cc7ZySKfoS0+rNKVdexsAfsR2t8MO4L7l8RMrNR1ybTCWDPGSuEqk1y0ygzfBHyjmPf0IlxGT5+E65hlkQiEqPRBWaVIb5EQ4+gZO9OsmCkJUKY1EWoUZ2yZykYz7xBqJRc1l+kQcbkOTL+mTYGf+5N/ip//D/99fvXX/i7/x2/+P7hTD7YB9FRAfuqBVs9dNZfZRRXvXCnT6xhiIsWEqdF2kgHeO1oHXSr125Oipkgy3E2PGKWgVt6xGBOr1UBKkZgCMfUkC/gGFssG7wVFaE86nrx1gZZYxRCVfogMSQkhEmLmvmOy4mALqz7LE1Op0V6el8WDZywBgSlPzxreeXrOxdkJC5/pOHC5Ts70mkA25QW2Zk4PEf862qu2RwPuc1PBOYXEoSljTZ3f+/zFh/ZmmJS65VWfbuVDM5yBN8Uz4BhoG8M1CXAE82hoWeERKFx6pgxCyPVUopuCe6Z2vPejc1j7Whya7bsxtT97lCkrxEl4UilWJpYg9nztKxf8/H/ws/yxf/2n+V9+7dfHglu7dIzqBqCmv5sZNE0BNnCWA6IhKZoqbSGIb/DesRBhEWFY9/RDT3+9IoXAWhzOyJUdi/eO5vLI4ozOQec95gXphPe+/i4nT5ZcXj0jkehOFqySEVOuNTOEwBCNISpDSCXbVBDncOIYYmTDquegc6W7GKtMZhrs4uyMi7MFy8YVrb3m+EHaJJ9ltc3m+GwA/vB5OtrrZY8C3KsHMrXbpG+3LbvhWbMKA9lPX5/e+IfW+abZJgN0s59OhOVygQ6XPDk9QYeBk85QSywWSuuFflBCVNa94aQDMlCqZe84GURCPn6S6QuFkQLZ3arp/nEWU8QiQ0rjOTKgjxFpW5Kk3FmpEd46u+Bf+fEfw4YVN5eXLDvPKg6YRiiNLURk5P93ZZkiuaqiWcKJEFMqXYigbTxCBtjr1YrlYsFpmxN/sFxDp3MOYiKsepxaabZRM3Udje/AgzRCu2y4GW5ILvGf/Pk/y8nFKf/r3/k1Pv70Yz67/JS1tsRUwFxz8lJSyw1NnFDr5gc12sn+1JlCUsVKWWLMODs54enTp/z4eycsO4e3rIwpoh6kaZDSBKTGGMxSyQ0QUkqEGIilUqbWTFWzo5f0GtqjAPdqcwkru7YlsZt8D7anlvl3h3A4y/HLFkAyqV5ZpWgSGiNeAt7d8BM/ccEf/3d+hrOzyO9+53f5Z7/9IaqJmyExpEVOFFKHmmZaoKSuq0B5it5hDPPnb/cfQJIWJ76sV2lQnixbnp629KsbwioHCb13RN1f5+775rwXcsrGsAMGNIWmiGoMQXFeWUug9Z5m2dAuGiwkbIiwXOAVvObwtLN8gJ1vi7wU/MJDrwQL/PTP/BQX7zzlf/+Nv8e6X+HbllWfATrEUtkx5fIH2avOVEoea57Z1MGb5eQjjQlNkabxNN5zdtJxcX7CohEaEq5c1042WbP57Oda72Z5WpU7PrlMPY3ZaIVS+3LcGm+kPQpwF3IPT+fc+Dpkc+34ppTMFLBTInPKMw+MadDtywDumwl5ucnNECIxDnivfO29BX/hL/xJ/tyf+fd4+sT4x//k/+Z//tX/k3/4jz5gFcC7lpu+NPLQqoTJOuk7X0WHKLX6qiBUQdi3qOSVWxxwmnjrfMlbZ0u+/+mHDOuAM6XxnhQP0wm7TkNmUXIdekqJYyg1ZSp4IjjnubaIWzQ8OTmnaztcVKwPtDgaK+Bu5SUO1y7yjIaIayFdR+I68uTinLfefopqYnWzZnmWS/jWxhu59nqmo6yMyQo9tQkwl9LK5cmUUv550TWcLhdcPDnj/GRBI2s8OoJ5DgTPKYpyyQEzLdRZ6X7ljm76m2CPAtwhe+RN0+SMvdKseY4+aZrmIBe/+3JObgX3EMKXAtiBSq8XHrzAhiQaZ2iCH//DF/zpP/1v8s1vdqT+Q/6tf+Mb/LN/+nX+wT/4INMV4rG+FOUit5Eb7c6HUOade6lev5QnT3n3udWdBzREGhJvXyw4Oz/JcsgUcE64rUZ8zW6tL+ccWpUmkgMRTqS0wMvNpRNpLDec1FDXQGqx1tE2uZeox+fjWePTlssIuJMFGgMaEolEbwHxkl8OGudom9yQw1mLqRZaqjSoNitAO+16lVvnFWnMmKylZrTecX56ypPzU85OT1kuFni7Rki4kudRHmNo0pK9m8sUqCkWczMSncxwbnOujvb62KMCd+89bdvSNHlYc+AO+3z8oWCsd1lGN0f3pIlsblrP/E21Sfw0W1HOmCldC195Z8HFBcBndO0aawfOzxxPngjxRni2Ki3qaAq/ntvwmcFdU1syfM/QMuWfK+UJXAF6cc0Y/kuaufHTRUPTtQUQczu6lOKtgfb985+pD6wkI5XvVvmmlWsDM9QZliK6hl4HTn1Ha0K3PMve9QTc1QkqSm+RPvUkHbhc3aApEFIcq2uG3kAGkuuKrDHlxhqpFA2rwc4aiEaykrFGO1WzIkmMrms5PT3h9PSURdfiSg9WZ1kVlPVHpfGG5qYjGdwZP0MVc27LETra62+PA9wlexeY5JogybaBaGL5s33Qn6ohNnXAKUAx/b7tAfuXwUYqddSD5PZuAK6FPgV++1/+LoM94d2Llu/+/nf5gx/9gLe+KnC1wC49gyaIuVGFmsMoFJnlZhLU/28Fh8PUzN5nWtZbngdRoVm2JTpYlCDiGJFq9/uFrts0ISmURBlFhfnqgMeiFlEEZ5l+0qQ5qzMZcUikLtH5hiEmvBouZrCVlNuER4QwRNbrNZoCnz275r2vntK99YTTrzzFLxckEcQcfchyx1ASlgYltz6RnDlan8gZ7kvwU7UcF6VxcLpsOFu2nLTgiDi1wqcXialoUUpZ8dYnvWPLZ5TjkMpxUMvrqGGUql062vNt2oPgtgflrYIRarntsoxsfq6S4mqHHGB4LOBu5LrWKdL3mxT2F/cg5i/IYRj2PnvTvZVEvSY8kjtmAo4kiZskPEtL/tG//IhnziGy4rvf+R6/d73i9BvnrD4cWHDDSVC4aVlFT0gOTcvCqITtgDW3Hc8ZEK4vK2BS+F/6NeY82gg0nqtVg5ydoeIJ0pAs4XG00uQm1WzfWLCRQVaar9IOzsk4mnqVxJQhX3ybVTUiMESaZGgrJBxrEaIJN2Q9fQiBYRgIIaBSlC6pI6wFZx0r7Xl6fopctPR+4FmM3JjjzJ+RdIGKokRM8hkyteK957H4coC8DyQdSJqlq62DZQdPz4SLpeK4IQ0BoWU1Oi31WG/ug10RQfRufOj3YvRq9JoYkiIOvBgqMdM84sktkY82ZyIyUsre51pFUxzbBfyaBb1tRtJhdsa5LxgxPvrws4PjeRzgzj61EkLYk9O9LHs5D47XzMZp/pQrz82anYf1jfH9737MZ599TAzPGNYr0FO6xtO4Gxo/gK4KM1DS3cmZNA93KOtN4cp2hJOTE3zTEGIgJR0v+ocyo1SY1Jyk5FzuRtX3/VhBsV6rJoY5j6gQY36Mmhlt09K0La7czKaGahr57/oAUq1lGArgTmevZog4vEs0Al0jnJy0nJyc5G5Shd4JAzkgOjfrLeuZJpaNyVACpjKe291zcLS7Wdu2I8VcRSL1fRfc53HIELfYFwGU5Lx9hd9rAO67tpuI8rLsSwfqUyvB1M0vZbKncHMFH36w4voaNAUwYVgbV9c3XF1G4tAWLXkGgKwpTw947+fgq5FvgkQOkJ+fn+O7lmEYiCFtPJoHGoVZlg1KUoyYYwIuJ3Bp6VqUqY78IIgh4MyjyW3q03QdTbfAFTFAfVAkzRmvSROxBjnZlGAYZ+CVNyJ7hIvWsewazs6WLJZLQMY6M/kBNO+8zCnKRjlo/bvtMlwbzU2tR3O0wzbnkM557lNPfHtZxuqou+A+jQ/exaF5NOA+l5hUA6tHe4km9e41BEfTLtAU+PBH13z26cDZuSfFm9w1SB19rwzBs+oVTU2Wz1lWXRixBEhfvrpiQ+0IzuXArfcF3NsM7iGm0s2o7NMDAI+RFTMW0yhZdLUGA4X794IXh1rCUp5SO5/BvfGOrutoFx2uaQApCUSJVBLBMrAryayIbzKIVuYVchZsvSe6ruX0ZMHJcoGXhpQUTPC+Kad3PpZk5Umhms//WB6iJEsV2fuO1y9HUL+HTb31qcdebTcGNCcOqTPSNxbcHzK56EvtvQObu7fI5CyyujYuY8/1pZBiwIuUKbvgmo7r6wCSvXczASsccZHtPYSVepC4kvjkvbAstEwMkaQ6dmB6KKsBVrMqJQSSbuSCddYguSyD9x4nucha17SgDV3X4JtMy1SdfVIlSWmQYYqixKJVNdl41JQ6mj6n4+Gcp207usWSpmlRg5giXoTGdzgx+n5grj5MPUxTr12BGrutz8hRAVQf3OLAjlz7XSyEkGeaKY2cO2wCn1Op6dwMy8xQ2+fcX2twr9X7gL33l227Ee3685tgB0/8Vmn74rlhDCHixVjdBJqmY72KeFkiSKYOLOsoYnQgkGLWglcixLn9KeStSoB7nFNxjsblphhN08AQeeerX0W6jn6A9SogT1q8b/ClyfNdrp2p13TbeCvdU66WLHc0Q5BRZbI95YamMVrXgjQgQkxKt1jSNR7vctML71vMlCEMG8bFeRpXdPXZEd/RhBXOFo8mGPpIK0K7bAvwKllclB8AZq7QOhsw2ACMG1MLYkykWDKxpEGTse6HcZuUwPt9XfdD1+GbLLWsxziVMhohhK3AfrXn7b+RJnEsGT+dLnGX4/gowH1X3TD9/GXbPM/1Zl5sWzamr9d9ze6ZaUJaIYRI27SkJDhpMPHZO9fMCZsWmdyoOpSDwH4bWN5z0JgpMcTCgDjOzs9zWVyyN9NA6Vq0CSJ+Hqfg4JgRVPwI3iJujE3vBci84B00vsVwpSQBOaDqXGk27nHOoxoIo3ctRXboMqVSgN2m93VJSDKDlJRhCLReStu+2kAkp5dRtfKVU7eaoJW3Vc9bXrdDUy74ppJVazGUp4tUHVOl3o7VIp9nuxTMPD122/VpINOyD/uc/F3h6lGA+9FehZWAGEbtfwpGLt1iOK84vz11r+8ZICQrPJJhlrvzSPES7wrudXp5V1NVgiZUBUFZNDKWn/De5WbPLoPtQ0VUKx/NzstMR618jQ84Mufqncecy+WTvctFz1yukum8w0nW7G8al1Teu2juJ2KX8hzNx9QquBsJzXp65+hQvBjeTR4I7AZOM9cukh/QUteJKw+VErdNmfPd7JnbAfqHo8DeBHtxR1HIfYq37fOQGC8E7iLyHeCSLMBCTBEAACAASURBVKSNZvbvisg7wN8AfhL4DvCLZvbJi2znaC/BrJTgnebLu+KTNeCbmHuSmiKaSjGpyQtKp6KMAlmW50sW8L4SYHYIMwGkW4dcJYIuNxEZTaQ07wZpPOynLbxUy0TMBOSlZJLaFtyXmjXTQGX+ve06fONBlMY7xDG2xcsTDhnpE5X88KygvrmnC3ebIJkS1HIZZI0kEVovNE2uTS9SH847wVObTuen5yuXVVbNNE0MSs5ALnSMyBHT72AvhwGYFwa8cnAv9vNm9uHk918G/r6Z/YqI/HL5/S+/hO0c7YWsRM2E/J9kqHK+vFzC+YiRcgu9VKo/FmrELKeq1wxPweFcU2Il6YU894PTVCvqA5+pg2FIDMNQVCOOphHMeR4SeWpGr1KrVdZCXtNPsokqQQPiAhY8Yrlee9O0+SBPPPekoM42vPgI8HVl4yOYMbRsuYlJSpqfDqkEQbxgjQNrkGajxoDtzO1xj4SiQLJS80YgUfT3mgvueSbeeqVm9qQ0R9uxlwLwM+uQLarmbvYQtMwvAD9Xfv6rwG/wHHDfZNNNJ6Ob3w22rqkvBUf+IFan1Tb+uhW3mbx0/Fd6n5rlxsp1EZESmJt6gZvrcnOONifO7BZue1ctRS7o5ZwH70gpjyHGmD33poC+yw+q22wugH7bMvt/owAmIFJa2u2WKytBVk3gIsTMfyfVnLzkHFiiljmr6pS8r+Wz6iDvDcE2Hxa5ZP5yrvzotFTncfn8eD8VJ0zvr7IfFEmMSKk841DJpY5DqU5ZxwTkSpQjlffq7IsoDXJXbLmtltELjoDDwWvbvMnzvfkXBXcD/p7k4uD/vZl9G/iamb1f/v4D4GtzXxSRbwHfyr+A1fJTBSCESZNj21xW8kjSn+9LL3zR5lizdXOW+91wxCDcrM5AToHsoQftSZpyFmaKqBrONzTTwy8ByBTJdMVSJIGwARebuqS7Y9tpRFG9zUbanP2pwnroOW9bTt+6AAdPvvqU4DsGMYZW0Dh96GwHtWotoZoaPvVs6zIwf051BMlCaVRqqUrcrNZDN5CUPeGYQB1L54mU9qUq4FtiFGIvdI2jTzvlq3cukzpPEDOSxbxdKRy5tCCO5HKP2yE6XMrjPzvxeF8pGCs9VGr1GBtbJYooIRnSLBmc8PFNzyexYVg0hJQQMxoGPIkWK+WXu71jdMheFOhe9X3zMu7pWuLihczmvi9bd85dOiG+KLj/KTP7voi8B/xvIvLPp380MyvAv2flQfBtAOfGJpob4c8Xj4e32os+4V+1yfjfvlUA3uhot5Ng7kK3vHQzclegpFDoH+8lywnNWK/zwycD9UZD/LzxvSwp3u4DAsoDoLxjm3dNuXgXXcf5+QnLheeqT3e4xjcz19ETx7IOntxYBC3bRNFS1TFV6kyU8jyoa6EqMbbmVSXYmjSV7byc4/Mm28Pu3wwtM//xrfZC4G5m3y/vH4jIrwJ/AvihiHzDzN4XkW8AH7zINo720FagYaux9HaD6dtq8dz3Ir/P8ikm1BRHyfprSllgMy4vL4kx0DiHVrf9wPZGqeKkvsdDmaqOjaoNN9aP0ZRwmvuyiqO0uLvjg7/QYnlmkj15ydERsFyX3tcetCKEqLn1X50FF++9/m6Fd89g77MqqfTEzQ+GYz3359mh6+ixOHPwAqmFInImIk/qz8CfAf4x8LeBXyqL/RLwt150kEd7OBvlzoXPDiHkxJYC9Fnu6LYojbsETw9t6z7rEJd5fapmu7aEU+PqckVKpdHGLTfUoe296L7M2TgDqp67ZsXM8uQkl57u16xWPTGUNoU7UsWDL9jEPUpQtcZCUslcjDXT1QpQj684/pxKOeCktkmWslLfPSZiDLm++0vw3N90q/fF7usx2Yt47l8DfrXcEA3wP5rZ3xGR3wL+poj8ReD3gF988WEe7aGsKljqK8ZI5c3ngO9FPZM7ezuSC25ZjCW5JhGiI8XcT+/6piclEOcwPexr7oL2XYKqn9+K3NBK0w/LD6V3330XOTlBP/uY1c0KcHQLj93cw3OvQkzbeO5C1sp7LTMaEcSKEsq2PXcH4FypUZOpGiGLYsyMmCIxpqyBf1wY9SitgvuuPVQl289jnxvczex3gX975vOPgP/oXivbCYDV923e9/ONc35zm2n69PfnLf+62CyAHlg2X4wyeu5mVjIY91vU7QYqP48dytibzUzWLIWszZxNjTAMaBgIw4AZxBhzQNc2DTum11BNBa9tHL33s8HUre1ObMqtT98PLlv+7qUoVszoFgtASClxdfWMphHEnSCr/uD1vr/uPM1OVelCCbhOBBSefN5SBfDJywRIuXsTYjmPDWiNXJUy5E5R3jmsHK8c0NUqxDw4toeyh7rvDh3n27LXp9d/luLma+kx92J+FBmqVeVwJ2B/weO3e8HMqSfm/v662GGvdH55VRs1z/UCPQTqu977bRfzHADmxefBfW/8BXSY6NgNiGHAQhiDqCGEfJMVj2k6tnrTVUqp7tfudXabWubQVHtuHWPqUtl3K7zHerXKef0i9P0Nzitq/h6zoULNVNW9QVXeuzJuE8GcoCKIFlVPTb2S0g7QlZMgE2mkGTFlz11EcN6jI7gbYvKFkTQPed/dFYh3r+M5Km96rdVlHoM9CnA/2utth7zx+r55wb2ezoXDFpc9eCFnUJoavnFZOj6ud9926aQ6ZZ56WdPxHfLG5wB+uo4xUQjFSknlsWAXcHNzjZVZUVax3E8NNi5an3uSlZUlB6qWkhnV8LmxdvXeM8Dn/rRussb82sqqzdMO7D7n6Etq9Zw/Vq8djuB+tBe026a49f0u4D5PJWWPyDW5ZRmqhOK1N43P4F7KJByyKbjvlkzd9dyfty9T2+j4Jw8HV6orWpErls+HIaAx0q/XrG5WOGlYdB33qZtQpc8iE2je+ZlKv0j5pRzFTK4wNh7fWi/b5yc/gPTl8qBvoFV1GWyuj8fisVc7gvvRXtiex6NvUuBhDtxvC3DmTNWmlPzNvUpNc40W7yDYtm57d71znvt0fHPjPLQfB8dYXq4Eomt8QMv41+s1KQbiEBiGXO+bezaiMdjyzutnW68Jdsvk/611TD8TGGsHlU5OOpldHO2wpZTGeM5jpW4fN7g/vplOsfsO7L7LP8SFctegWEWI3WU/X1DtLvz83PJ1m1ZQzbncVzT3kSx15F1d/rZA5Hws567gnm/ceVpmdz2Uyoww9Yjh5uZmbH/nfW4ZuPWg2R3/3KGffFz3WLag/tAqhErU287v47K2WeeGk5+hqGaG9mWwW52PSYzqsdnjAXdr2HTrrcqHx3op5fDWnWxsa3dXe5jyCrlsw/66cz/UXHRqc/vm5g+QS8FWcKtNHsYG2aal7d6c/MtN1jd9zYxtJ6hd3xvXIM5l/XbKteUvL58ViiI3oXYLj1NHCnFSDWU/+Dv9/Lbg+f5nHpGd42Y5wWrjR5dtRUUMongGJ3QnHUmbrEdfrUmrG87Pl/z+zTO0EdoOrq/XWaFioFr5fZm8csE3RxivIyOXYVXJVEuS0qWpvFo3INNHhoETh5dp6YB8Tq/oGEQJLoBX2kZKJ6FyfRelswHiGhx+K/DeNM2D9Dp+aLsrGIvkMtNz9F2l5h6rPRJwl8n7YW/kUdldh3if3XnAZ9md9eUFgyuMVDCcgmXlZm8/X3f97IBZfStcfdH8xRiJ/UAIETNomobGGnQYxu/cBuzj6ncCqAcDqtTaj9Oh2fjX3WWnPnXlxEMIoz5/VB5NG2ZI2c7e5ifkygxXYlaomMnsoTb52F2VWdbe71r22jeFwsY7cVzB/HWzO+t5nezzeNlz19Rj9Nan9kjA/WiPyaqcb/ZvB4KRDzeOOonLT50YE6v1inU/bIH7cE/SYNrWcdzePfZl7vuHiIvVakUMYZJoJRO5aQWJqbTuzsPIAecyllvHLwdiCuyfy8OBZcb9e12B/ctkR3A/2r7ZDA9c//QKvTZTQ0tjCzCcMWbRDkMk98euNMvdx/KyMlT3ciYOfH21WjEMAyfOxu/5xs+C+31sLn4wGdz2sludnzamsq0imq53Zovsqo4em7b7aBs7gvvR9qx6c7N/e5XgXiN9kwBg9W5VLQOWWZFH2p1Jn5cBRIconDmrahm6je++myhGoWXM7jcDgRlwP7B/c9S4uvlg8+wxstefkvky2aMA9y/yQnno7d4WwHtRe1XHbNeje1XnawN8mY7piu698U2uqkhWz6jBlBK5i259N+vwvuM6OF423niWF9aCbBFZLqm1fGRCzWC1DnjRzk+CtDvaxoNWvWiRnKWKVYy//YuadCy8dkjSNx7HrXjLl8d2u1m9Lvv/KMB9ahvP7OGzvnY9lbtO1x+LrvWhjs+c/O+LyMJz3mEIIpmK6Rrohx7nHGGIoyfZth02BDC98w1oE0AbtzdXamC2ccL8NVB6GoEVxYnGMdllvV7D+YKmcajmmUbTNPlaNyXGRNs2YLkcxKiUGcHZ7eH0nLQTQM3DTop8/nz/mCRLaNmfqfpj/M6W5LMoq3aWedPtMRUDu489OnB/bHYQJF6Tp/frbKqaG8n5onO3DJS1NPFI17ykczEPVlMv+vPZMAxcXl5yc9IgTmjbh5XQVQXM3ucztMw46dlbh239bKM052ivix3B/Tl2BPcvzkwtN6Ku1RyTkmJkCAPDEMpCL297DwnuN9fXrNYnpcDZvCzxZZmNAfF96eSe3RHY7xPTONrjsCO432K3TvGP4P6gloO6OUnKWW76rMEIMTIMA8MwjHHDl3Um7gPu96GohiEQUyLFxHo9sF4nFouHm+pnXryI3aWK3g/Efg4g9rw80viSMDFvhB3B/Tk2d0McvZhXY6alcBibLFlNpWtQ2mmt94K58Yf5431wv2/sIcYs2RQnrNeB1SrSnCceKht523OfvB/yUe4QrM1xqZc+1KM9oD0ScDdE+5y151xWP0iuTmclTGXlyvL3Ata6pI2vXGnQaDyoQVJoW6ExwcL+1b8JSrnSxKKm4M9ohnW/6bFhuUHyHc172O0prmqzCSiOZvaGndMzmyRsq2RCkd+Ra303tDS+y1x2SthEk31bzft7KWcM5gBNRyd2CqRG7CLeGnQd6ZyDALFvseYcW56QHFhKtCESdgK+t9VKP5QuPx8oV/ZKTUhRolTvtm7HKdFFRD0LM5wKUYR24XgWej7u1/j2hI4bTliMtcOE3HWqNk6Zl1T62XMtMx9qQwmS3kH/LzJez9NzmTty7R6JfIxzm7/8Dww5FIjW/c8PPUTdXP2eSQOW7ZXk7d7FRPcfaofKOKM6U1Uk72tWQTlcyUdITlFJdRHqtZsnOBUn8is3L3/1QdlHAe6Nd5ydtvT9wBDjmAq9WzCpXIX3XPuuS5c9EN84nFme6od8E8z5UQI4Kd9xDik/q84A3Uz9FBk/v+tQ4/4u2v4Fmm3/BrTJ/9ur2B1FrmDYNC2mRZ2kCVNFSZjtV7x7OWqZuz+aTQwnLvf1NEEUUjKigorkNHsBZ9tB1Zc63pnhTl2GsiFG50FGd2Ss0GNiBE1EU1zT4J3H4yYP8Y3k8z4DyWzLDGUkSlFDPn/3ZhyVw3GmAmGWW/05y5Uv5+4bE8FmANtsXnWVZsCv6v93B7Hr/NxmMnvv5Kqde6Y2+7AsVxNYyoFqy317x7vP6n+G9xXkbbOvpUH5q7bHAe5Ny9tvv81HH3/C6rJHzXANvGgg65DVJg8i0Lb5IjppWhZ++3AIWY4nCM5PmuDavKzMVPeB3A57invfh9KgeOdzO3Ax2kwhMPINtGsqitXeauUCBUdtQK2aj3vbeoake+VMD3o7vCzQv5vFmOjXa1LSoid3yCgdfJxxkBSVVDTo3uUmI88H81djc1muh85nqZgzPkzVStbw7Mx0fj3OOXC7mk42hcq2tjefj7X79dtM3P48aJR17piX0oN2byWTc1WA3GH4nVQELFOJWILs4x9MbHsV9kjA3fPuu++y7geurntSvBsX+HlNFWI02hYunp7y9ttv8+RkycLPTyMrsD2vw/khEL9P1bxc0nbe8963GZ/JmFViqKQJuBf9tLpcsTAqP/zhj4hphXce1ZQfajue+9zYbgP9h7AQAjerFSlFqOflkXd0DhG0PCy9d0XH/jhsSx+/k6yza66AuRVgV8vU5qFZrM5WTnUzvjjM6TTn/Jk8zgM7MzvmeRiZvZsOeNjeu1L9tOxzXffMirXglgOc8zjnUIz4ZaVlRBxnZ2c8eXLOR588I8RMTTxUAKdeu12XZwzvvfcep53HW9hb1rnsGTpx2XsvbeNnPWkO+WJ399DGlPvdb99Rx2bkGcSuqauzCtm8zOGbljBEPv7kQ7SP+EZYDT0x7PeKbNv2lQL5nKUYczp/2s7y/FyM3SuyFDMAyngNyYM5Lve1XU/9ds8dfAXzwmWLQTMHcsKdH2IiILPNSw48ZKa5Xc+x2jR87/OZcr2tc/id61uAtm3HZLSU8ixMLW3HtkqnFOd8Louh0xZ8+oWc70cB7kgOKJ2enLJcLuj7+OAT1qYRLi6ecHFxwXK5xBEwjeUcTM5EbVwgRg5VOnYXmdqLnsPxhhjv/92538YO3YSzdKJUNnEK7gIkfGN0nePk5AlnZ2e8S0u3OMtLlztDRPj000/H2YuWlnfr9Zq+7w+ObXuqf3iSOrcvNTU/B6byd2NKY6u9reU+p+0GYQ+O4xYbvzdZzGxD/ana2EEqNxk54AIccGhs5HTvNp5MiTxnkbnxUzORN9dhPndsftZNfEMse69d14zjHuWpzqHNviPQNs2eg1Adh7qCuneHmmA07exOzF5bc86Ic26zvYl1rimtCPfHprZRaaWYGEJPDIGNMMGBuBHY1+uey2fX3IR19vS/gNrvjwPcLatB2q7j/OyU65sVQ9AHe9iZwcnJgqdPL1gsFpgZySKicc9bVYuF6ihd5Xeoiqkduhjvk748BdPnlTnI9Mtd7+Ipz2XkiaPRti0x9rz9zlt85Z13WS6XIB24xR4Hu1gsODk5oes6Li8vef/99+n7fj5ANukvOQX3xh++5Ha35zoHMY/Rm2ECw9CzXq9zUE2UGAPOlRnFc9Q89efdErlTYNsrKXDL8Z+enzz9juOUPaUEZjRNLnL2ySefcLP6egm45Sbf9XqxcQyHucjdcgm3BY1ND3ARB/ahWvY0M5h773aOGyQzGnEsFi2IoCnRtZ6ua/fuG/EO6fYBrWnbWQBdLJeFY9+soynJa7vmPHv7553D37F14W6phXG9txy2+pCIMdL3PUPfk2IsKjqPE4+IQxW6bsl63fP9732f4Uc9Qe8sqXip9ijAPaXE5eUl19dXrPv1bHDlZZpIvni995gZfd/jbKAhFQ9ASiBGtm6q7eDinCc9o3QR9i7mW0Z2C7jPXB5FFbG/f3Njk0nMcYz/j3zpzfUNV5ffzXymazGaPbCtNXhSSgzDwGq1QlXpum5ve4cTgu5n0wqVVgJvIQSS5q5dh4Jjj8msBODrsaz87WPgZsw2Lf9OT0955513uLi44OLiYuy0VEFWUsBbRBCub6559uwZ/brHeV+oJhkpJ+cczu/vny/1dKYmFJqkeN/iXBEz+Nl7R2TGWTrgCOmBy3A2UFtopjmrIgznDO/zzN+V2NWm0UoJNKeIacI7R+scYUYk8Srs0YB79mxWrFbrrHt+wOu+XIe58UMBKKcDrdMdEAfvm/HnqnGvNMGuZZ37XMDxblOy/W1Q1Cw2G7TyzX5iSR3r3tgkd9spS1HbBK36IfOICS6fXbFaDeA8Ubc9WsiNMYDSC9SzWCxGsN/flxcHd60a5VHGacQQ6Pt+rApZAf9VFzW7j23418zZqhpJs6Lii7YQQq602TQsFgvOz8+5uLjg/Pwc7z3e+xHgTxaexhs/+vBDwkp58s5TLsRNlFXlei0BRT8TUK302q7VZ53UBwS5oNr8eZ1xsUuT713TykJOx4Dg3D705ZnD3NgmqC8Kojhh1PebaolzOTQl+hAZhoBgtI3HxXTXppwv1R4JuCtXV1es+54YIuLKk/IBtxlC5PLqkn7IPSC9RdqiYR89ZhG6tqXqbWu1vprMtGvzSRf3U5N470YvRKRKqebBXVzaczWq3G5vbGMiWOXa8xRytVoj4jF1tO0JISgmHpNJTRHZtNob6SpVhmEYgX5vbLN3yZ0PQ17cFNSNGn2zzbQ4xTR67ukL8ozuamaGpjTSHrkW/eN4GNXzmVLi008/5erqqlSrnJ7TvIxvlLY1+n5guVjwUz/1r3L+5Ak2DFvrEhGcKS7N5GHs6wUYZZMlcUHrOg6h7WyS3uzH0yKW062RdF88McXw6dJaKDTI5ZtjHDBVRBnPZ3YwhBgSIeQSzxoDYg9HLz/PngvuIvJXgP8U+MDM/lj57B3gbwA/CXwH+EUz+0TyHf3fAH8euAH+SzP7h8/bhlkGihAiYLRNRvcQH+amFYG+H/jkk09p2wxODYov0/sKxhnENrRMpWLuw7mbzWeXHrK2zY2qtzNDdZaqMgK7V/QhPnEEd5sEVBEEX55Hjq5bcH29yoz8hJfeBXlVxXufuXDvZ/dvlhq681GYfKdunyxBi4US2mzTsuf0OFiOg6bluNWSCvoFKSh2bfeaNcsAZhPl2EiLScJ5ZQgJFUglOctcUSsVOtOQElic69mqs8HefK1vYkEgxKgH8j72BReHZqzpHk7iHC1jZGbBicOotfkHLIIryX8x5pLNqkYYEut1TwhFWZPm8sVfjd3Fc/8fgP8W+GuTz34Z+Ptm9isi8svl978M/Dngp8vrZ4H/rrzfaoYwpAxA4pvi2ZTUZsBReJqRL77rXVEvDNt6NytT5QSpyByTKilFRChTrtJZPk3q61VFAPNVQeYz6pjnhCfT06kiYcjirXyxVt5/9GwqS5GDb4eo/NmpbDORYpQbUQ1Wa6PthJPTZaagukAKDk1udn1V3gWZJ80zmEIhFTKzBsayyKOCWPZg3CF+XAq/Xv9VF885LFkBExgsMVgCn+MN3hydeG4skOZmN3MB0prammEoS9rMyg3MSAPlB9p8RuRtwe7iAmB4zBY4H/jssufZKiDLFuuE5A2N+SEZhpD5aueeq3KZPmjr7zOD20c02wRKc+BPiSnSth7ntmNLh9YrJnjpaCTg1EMw0moYk3mEnM2dV6BomqFJxnjD9hEbj7HYePxSijm2srUfEHU1e2zmz0c7Q+HMfr3Mum3nI9uiHVNKGStCj8aQaUE1UglE36wGYmkgEywx+ITOEv8Pb88FdzP7v0TkJ3c+/gXg58rPfxX4DTK4/wLw1yyfvd8UkbdE5Btm9v7ztqMISnnqm5ZuNVABfnPQ7yMpmn8Uq0LTlKCPeJxrcNQCVdlD9z4HS5pm/0IXwN2LRDtcY2PbOwbTjexPypwya+s3gFPBr94EW7tXKIC9I2E5Nb4cglz6o2CcbxrEg0pEGiDZVqbebpDXez9KIrM3X8Gy8q71IWc1EkX1yNyhTIDxxto8uKzU6IhmmJPsuZsSTEEMJzklxuNmZ0gHlS7iyv7lYykjh1tJMJvwB7dIKOZWPf6fXYCcnB+4Wg2s+oC0HukElclsqH5jDpQP2K0xhpksSxuDz7I5PyYlu3R/+bmHV4qKRWhwtOKxoCQLOaHMatmFfA6SJkK820Oxzhbuup8h3cwuNztD1nZv/w5tz2bUZ0bOrdgcv/pwCtRSIalcKmoQkmKSm8wMmhgMHjiEeNA+L+f+tQlg/wD4Wvn5DwG/P1nue+WzPXAXkW8B3wJGoIgxbnmGr9KaxtFKtwVaB7lyM4gvPsZKc0wvTKGqBjaee8XJCurVq40xslcxxmoQdtuSsA3uAA6WJ46mrR2B7j+BNAwVw+FyXZVxelxnSTby94jdGljaEC0bXN1bpjTINlPEUUoPMLa02xqb2ew5fNU3mhnEEAmxUh1zFUxe1Vg2ctCmaTg9W+C9G5ug1HvwkD4cNtfhzc0NXiCFIe+RbcA9y4vnHjK3PdzvZkn7veu+JhntLRv3Nf+73vj2GOZmG6mAu5b3zePbrJSWs/wSB+ayS1qFCjaZNL9Ke+GAqpmZ3KeSz+Z73wa+DdB1rdWgTk5xz8u8ygPifEMz8UgrwFcPfsvUsOZFwb3SLLYFTM5vvN8a1B1jRdW7t8zjuWEu6WY+w1B8M4J+EcpkJYErQSy0BCnvQ3sx0ilagtFT1swAc+WB5Mpdccu664yiMjdzV1VKiTAMGaDGWcKUrpoMTQ4lJd0+jpdtmnLDjlAajDgns1nEr8Km4N51HV23wHsZVTN1mTkbj6dt5MuoonEf3PN1MVP7yOavz0OO1J6Xj83Gmg45hYdyTOZjZodmelKouk0cKtPFZcZISdgtXnydqW/iCF+MfV5w/2GlW0TkG8AH5fPvAz8+We6b5bNbrXrJ+YmaP3v1We6FWsjuYHkJzrd7MCBecHNZcvew3Wle9RraIhnPaoGZvpm66eLTdn7+Rpz5yPluVEBYoZ3zJNQIKdEPA1jKFMt9LXNIIEKNUEh1a5i+bqE7pQI0o/JhjoBLKeXyFNgYF8kPtLvLIavk/1VZspyhWqmAQw+dV2FTWWvbttzcXBPjtnJkLiAPk5kmYJoVbikGck6RITlsn7djSprhmu9TfG4+rmHg5imc+THv02qZipyb0fl5IYBNxQX5Gm1ci3e55HYyKzEsYYhxPApqjqYJz42jPJR9XnD/28AvAb9S3v/W5PO/JCJ/nRxI/ewufHu1mh33qjz2aXKOmmTurJ5EKmu6jwReHH5Wuz4z8ANAMm04raPaw2ja/NlWEtPoOeSbypNB3Tf3iD9Yg9UqklueO5hEhhDZcrtvW9UkoJcHXnnbKt/UQlVPp7llmdmb55CnOPm61cNgY/anlIXqDTe3voOSzHtcY4eAZ+/zQ+s0iCnmBiPjd28Z3wPY7nhr7CSlfbXMrevIS6FmDCEQh4Fl60fP9ikJdQAAIABJREFUfUq8zcqFDyirDslp58Bd/L4y5hCN2nZz25sfm4hHDuQe7Dli6hBzo/+Sk4KFxlyOD1nObm/USn7D7Gof1O4ihfyfyMHTd0Xke8B/TQb1vykifxH4PeAXy+K/TpZB/g5ZCvlf3WUQMUaurq5wTlgs/I43+/JNpFaGzByjc47YedKiJG34Bm+5qpta2PfcEcIEfDcvN4Nd8xcu4kcw2FyTRmKoP+Za1AXVxou8XpT3xASnDbVEcPXcRSCkWJp0XLJYnND3EU3z2lznMjcbQmC5XNK2LavVDd3CgckorszNCirgbjweMMS1e7EDs0JN7Xy+PD1jdRPK55TU/c2DMQRITQJyTRBlA2C3XTu73OohhchG8TPHw+6XKjCxXGTOHILDWX6p1AzgK/5/9t4sVpYsO8/71t47IjPPdIeqO9TUVc0eyK6ubk7NZosUzJYoWaIpiRZgGxJsmRZkyw+ULcM2YFsv9osAP9gyIFgmIEOCKMAwJdoSLBAU2iIpSjZpkt1UU1QP7Lmr61bVrTufKTMjYu+9/LAjIuPkiTz3nHvucG5V/IWszBsnMjKGzH+vWPtf/yrrlJJzDiep8tO0kteVu0xTzHYc9JFcc14b6ep4PGY8HlOUs2NLWSE5uFZVmvPYWF8n5jmZrdNMMaDR00z6G3u8yuVVxLxKkZSPD9uANMVWy7Au9KRGpHPNY6eGxB0i94O59k6VsVdClYg7hqTuixGsy8CYOtce8JXi/ZPpYnUctcyfXfGnH+9ZV4GfOe1OPWp0g84mGvQhUJSKMRFrI9bbNlW0DGsMeVu5evhxaP1jmwYpfQ04WLHdk8DGSFvR0aRlmlSIBEJQYlFRlp7M5YQlpUOjMHDOtaZhxhjGeUbebRJUk3hW55eCD3ifeoiqgsnyNtLWNg+zyODEDu9WZYURIc9yBGU2j7UjZGoyzVLOfRmrAoTuj/0oYodeXl+NwxkAAPIM7m2nAqHRaASkHDx27VibXVz6U34HrG0j+KpKqZiT+B5pnX5Ixls5k8kEYsRKHbXHADFNyCIgK3KXx/0u9xN+U5ktB9ZpFFyHtiGHb52bO750cQ2mVqLFYIjLaUntpmUWqUPB1F3TbH23Emvdf5KAhqiJ9Js7zCeAM1GhehYQghJr4zBjFuXUvkcVY0TIjE23d7WqBVZ/aU9mk/to7lasymKCss63R4HKB0aTEUXpWd9YwxhLWVYpn9hBc2yNkqJR+Dgj6LxcDG6mlktWNWmEkKSVoRYd+ogs3eJKbG7p0+smst4vK4zJkWDJM4cx0t45pEnB/vZsXfSR++POuSPK/t6Uoixqv/z+Qp6HhaMmRCF9H4uiqN08w7FPRpOubCZj19bWkm85kVT+H9DgF4NqT4n/SfZ5ZbCULduENH5PKyZlew+w0dtomi+Kigbp94pXDtxVtnd+mqJz1YW1R1WFWgYaKKvaa+pJhO0M5N5C0TpqVLoJsma07kIiFBpqbTILlcgKnOzS9uSk2/89OJxqS+5dtYwC03nBaDymqko2Nze5d3e7dxvWWoqiQETaW2BRJUdSWiU2PTah1KI9FKH+oolBjFn8UDrPsfOa+t+FprsY7z2j3LVeNlVVEbW2ijDdxPxhrJxwfoy/t6pSjDU46+p0Tm05+4gEM6vIvfGAsdZ2PGWk1+BrxYbbu8iGTG0d4NTTLIC2aTPtIbXewVb7NeorC8Xq44CD1eS9uxy6nkpL+6Fp7oCasIO3vQVHXclym2JUD+qThDRGvFeCQuU9VYhUPuCD1jNNq6o7Hi0Gcq+R8uW17weLbi99XcGQRPCLqOEhKiC051a2Jb3TIKSJzg6MCGItISof+9hrfPHLX8Jlo7qv6uHPCyG0XjJJa65sjMeYssIag8sstu4R2qhXotb5yPpH4bWxuK1vcRFCDEhs8prpWVSxzpFnOWXwxKgUVcX+fplUJ3Xk3s3t957O3mvyeH9qMUo7MBmbGtOUlTxWcm+8gBozsJZQ5QST8p0UVqOLTx0O9FDkHjXSV9XQb0nQX1TUW/CEUpT77W8OpK2w7bvUwaf5oKVPbMUb3UhcY0afGU03dRfrvKGRgEjyC6pCTESuYFxGFUKyaADENAqcx0/vZ4Lc00lOfhQK7fPpz4es3EZ7uZsL3/QUpVOxWD9JfR/fFBQZA8YdtOdNx7Fqh0+QltHDl+Rkk8v9X3J0MTy0ae1U4knmcv7df+/P8Nf/+t/g+jtvpoFESF90taimik7vS7IM1AfW1ww//KmP8sPf/xrPrSeP983NDSZbW2SjjLKYE3yVDLN8VdcwKNMqX9wQ11FTUZRUvmplnkn3H/EuMBqNqcpA8JHd/W2K2R7v3LiemnoHBzh8LE6U5oiq7eh9v/yvdG7Luteg/3oI4rP6jj116lEJBFFKDyEKucsY58J8VhLCGDRNpMaQrHSTcq9OodUz30oa7Lqf2NztH5hmrV/YWv6XcsvNszIZRfIRrK9H0k/fYYJnvu+hLgibFkoVwOY2VY0LqQiNNOltosG59LfpvAS0TrPVpnn1oK66SHx0kX7by/M59KpJeuWtCrv7yYOqe3yJN/oGte4J6pymTv682Qe06ueL5d+3KsaCNbVxWP11ikBGTPl2kc6g8iTi9jNC7gAaF5K2dA4fxn1zf3Xi8u9Z6xxFkvElWKH9QqcfUsq/tGuYDlkexSzavx+r0WcQdvzIfdVAkHo4Li2Pac8ya/noR7+bl195juvvvI4SECsQsjTYhJqwfEWWwayY8/FPvJ//5Gf+bT7+kfdjil2MCOub62RbG22LViTWv+aQvv0q+DCp8/IGsXX3mlD/XaAWr6eJO6r6+CNaqxOuvXGTv/bXfhYNYExGFQVtIrxjMrzG2DoPwtEEnybkDjau6J7rLkQFQ5bqBySg4oGA5ILajJ3dGeVshi/2ISi72ztUZcSow6AYEUaZ4AQyBxsTw4VzW6xvrHNuYyPZ8o5HjEZj8jxnMhmzNpngXEaWZ2RZjnOW8XryizEm2eeKMThr2dhcI88cWZ7URRqVe2+9xdvX3kFNRhUd/99nv8D/81u/S6EejyVKaq4eRSGCqEGMxUfl7s7eypRKCqyOX8TUt6yvZ2+K8vvrO0515yzJMuG4dSMSpbawaPggPZV1O8XuV+rJUPsZIvcBjxo1+XV5rI44RCKf+cxnmE33U6oDc9CKQBQ05dldJkgxZzwes7O9x2/91md5ZiRkmWNtbcJ4PEJMmpxOzUQakgEVoVTBWEuWNU0gXJujzTKXCCrLEGuRaNNEl9bpA7HM9veYTfcpyhnjicP7AsUvwrdj4Kg0zqNAVaXPunPnDj/08Y/wr3/6R9k6t8XUXcZIkheOMsMoc0zGI3JnEQPjUcbmxjrj8YhQFqn5Q5aRZVk6d/X5S023LVibPMbzFNm2EqRmjsG5FNlEhaqEomDn5vu4feMO03lk7h33ZiX//LNfWAql0/2BtRZnFg1uiiI5ufYPjt3I9SBOm758VH18H1fNwePCQO7vFTRRRvPD6twxzueeX/3Vf8qNG+8Qg2IzSzygElIQxVmH93OiKl/96hv8wi/8Q/x0m/NOGY8Nee5aH/40eZdaollbe+CnDg6tqqbRdzeVpotG5OkOyc4mNJlbtQbEcWdnxjtvX8PImKA+NRbRsq3aPdapaInvMB7FDzzPLb6qKMuST33qU/zIj3yCzILaCRjTOT+pACztX2jTHGgk7k+TmojmbsOjBJQ5sYxUISw84zXlpBvtd9PlKNTa/BACVVFQFQW3bt3h1u0dXn/9Otdv7XHtzbewpmk+YQFDU+QZgxI0tNsIIdTZrRXzGqccP1dNqN5PDXTaz3y3YCD39wyk84gHFhsj5HnOfF5gjMVXsSMBajP0aG3jurmesbuzx3e+/TZXLkyIBCRmVIVSlDNC8GiTBqpbAaYtRaqq7ETZ6Q8xaC1FTQUfzWu7b7HOIJlQRWVaecgznr1yhd05VNVtVGC87tivSnocZvvRUeo0ss6VyoyHAAH29wvs1fMURcHunbvcu/sO73v1A6BQFRWxLInFnBA9fjZl984ddne2qYoSBWazvZYstSbOqEoInqqqKIuSokzNbubl7kKxUhNhCKnfrNbvmc8L5vOSO/sFRRBu397j7o5nfypM53PUSW0DIe0xxBgJtRS28V1K17lH7adCn4HEKlPAVcVNfcuP3/nrZDjJNh5VgeXDxEDuD4InPLg/2BcrTRgnhjA0Puupa5Ph6tUX+OIXvwqkVIfLure+taDLCKKGsqzI1DAZn+fSpctcXYeLz1xklOdEDYiBPHNAJGogxoXbpy+KVnGTHk3Vq28VGMF7oip5MWY8GeFGjmk15/bOPW7v7FHhuLe/y3hs2N0vyUdNX87jzksAHOwwtXxuH4gsmnRPM0evifR8CFg7Yn19nV//9V/ni7/727z9xrd49cPPAqnV3Ww6Z7pf4asK70vm8xllWSQ7XQCy9kMadUfTEUwkIhIxJu17jPMU5bdqrgUZN5F8yl0rXh1FVEYba1xeX+PGzRniilbJtAgIwNZ3YSJyoEJ3VVbmJJH7qvN9XIfY5ev4IBjI/T2G3gpUWUSzj3M/TmPLkMrXu5F46gUZQiTLLJ/97c9hrcH7QJZl6UelKU0gJqk3inKOeo9IJHOWt968jQ2Ke/48O7s3GY9HrK1PmKyNCSODdRZjR9hMyE3KC2ebG23kCSkKbasLlyLTUUjzq1EC8zBn/eKceO0NvvPWDdAUjRoBX1fLWg66Azbl9t1zpposArr54Obvy5rpPvuBo5RRqqnln4qk7kT1IY1HebKpVeWb3/gG+CnTnTvM7l7D1v0Csiy1PVSU8WhEPsrJx+sU84KiLFNTjVqjbq3FWEOe5WR5RgwBlznObW2xubnJ+uYE55KNhsscrq6m9t4TYqirhtNgetcr1+/ss70zZ2cn8PbN71BWYLLau7YREWhyQYwxdCy6j3BwXfEVfRgEehZSJ2dhH+6HgdyPwKoL+Lgv68PQ0MuByD3WChBa4prPy+SroYLUPn+LNyd56Hg8IlZC8BWhEna3C14vbnLrzbfwdal10rsbrMvqpieCdYI1gogllqOFVK5+FmMP2YAKMJGkXogmYpxic8HmMF5fT52iIjjjEJNy/bGTXlmO5LrnMEno+nO5R/1o7y95Pchpi8jdt01fXnzxRb7rfVe5d+st/vinU5Oy6XRatxCMlJXHGItYS1FV7O1Omc2mGJtqB1KD7UAMEWMN1qafsLOW0XjMKM9Z31zHiGllpWWM+Mqzu7tLiKSIvU5/7Ubl7m7g5s19btzc587dWT2gZSB20Re67pDU5/l/Ep47CSmeZQI9y/vWYCD39wyk81ynZTAsHB07f0OXlA5pHecsQR1EJUYo5pHoK6aSzMZSFG7IMgP41jm5aVuIGspydKB4RJPOlMNDphLLvdTwIKYJXbFw6comH/zgJiFYjGZYMTg7QqSq0xOrPX7uh5WkfcrfcQix9hkBl2VkLiME2N5LLdpu354xL0t8rOcbxBAV5mXF3v4+s9kcFUMMydOn8sk+OIaQhuoQKauS6XTKfF7gq4oQU6MK7zUV2IRIiD7psZvPiVAKFD5QlbFdZl2O0nQwqoOBZaXVgDOPgdzfM+hOkDYkXhO7ms7fZcUtdZK/ESOCwYhDsIRoCCYDSSX2dpThRhnSpq5im8ZSMUg+opXpabJEKMuyv+3ZaBMjipGK4OeUZWDuLZFRnbpxmGixwWLwh9NnJ8Cyfr2daDWnN36y1jKbp4nI17/9bb70L3+Hnbu7fO5zXydG5frNe8zmFcZaAoagyeGyKWcPMdYFN82AuCgUEklN1cUYqtJTlooxtQthm+ZKaahYW9QuakkAienOyhqiOtAMkWxR4COe2omIgd2fLgzk/p5CR/+4NFl2MFKXpXVT1BZ8qO1sLSKpclXVEkk2vhZBNEM0dbVCYk1EobYfUEKeUi0KdUWh4omE5W+iwmxqMUaTf4mJSO6xeY51E4xEUJuKm0KSCC6bSJ20XWN3MrV93ZyqU0BjakKT5xnRC1VZkuWO29sejXDjdsneLDAaKQFDVRtRqRhUkjY9hNiIi9qBqzmnTk1q7q050USi2IXctCl/VyW2zUKaQxI0zpOMNOQYkxOi4IPiMgfStH5sTsSApwkDuT8Q+tIIPO451hOiW4fdJFLrphrp18vBYzood1AUsaZu2i0pn6vJNwabYaxBMovLHTbPW6lh06EmatpGiIebDWvbE+qgNjobGYKPVF4xWDIDxtlUKm8jmJiISRf700WfCmZxaMvrNutoG+m2k9jx6Fz8Ytu1j70sPqa5gTECZRVSisZaRAzObKIR7GgPyhnB5qnkPyiBxvCqHrCyOp3VXpU2LF+4BokgLunZm79JrUNXTcZWHdf9NAhKRowVIUScq69xc8dFaK/N4nsy4GnBQO4PhPSjWMZJCmkeO6Q6FH0JgFGSW6NC6Dbq7sog6n/ZRiF/sMDJ2grrHC7LasL1tRYaUMWI1o1HBBeTn3lX0eKDb0vNu0ZO1m2DzYAxxBGqJdF7Cr9LsDsw2kexBAW/Gw7o3LsTq82/DyqNDp4Lrfez+7c2NaH9qqllFQ71OaIh0cS4VFXAbgrIiMrkbAfIXMZsvoeIwxslWkOlJkXspj7j4kHqegEdY/oCiu4NVv1smord7mqSCsr63m9qZ83IPC1ySyYY2nzQCfWNA54oBnIf8EjQNCg5aJWaUigN7ifrbPx8tH5ujNsOTMZy9DbOAoSFRDOlSI6v3x4w4EExkPuAR4JVssJupHs/3X5r1tYQfLc6SGvb1tikds4uRBqFSl1ZKSxSJwMGPCIM5D7gkaDr6HeQyJOevZuWWYWG3NsHi4nERm+d7HvPMrUDIi25J9+c5EA4YMCjxEDuAx4JvPcH/E0aEndu1Ev6fWjTMnTseev30aRl6ola2r+eTcSY5jOMpSX7AQMeJQZyPyYaEjpKQ/2g1gAnxXEqJY+Lhe5Z2zZsJ8kHt7K8uml20+O0u4/dHLn386XPX7Rl625PhFSkoxbRxgqhSfekfY6quCxHY4Y1sZbPL01w9h7zya9T9z3dfp0HJ26lNz3Uzj/Ur62xVFSNLie9Pyo2M/jQmLadPXRTbamXrrQdjc7oLr+nMZD7CXE/0njUWCnve/AtHtheX+n+cdAl6aaJdp8fTl/E2rxvuXlG1KbLz4I2F5YFtaFC3dhDjDSy7APnpW8gPO15Wx6ImtcHioM6qPcgFW212aWuMkfayoOzjocZWAx4tBjI/T2Ohp+6dyQnN3c67H3TN2GaiH/1JOtxEOsJVGctYurG3Ee8ve9YmurT4+Dh3B0t2s4tNP6n3uyAAUdiIPf3OE5Tst+ga//aXQaH0zJ9GZ+TfHaK8iOZc1hj8D7WLfz611+l2Hmc9q6NXl6jtoOT6ultDQYMOAoDub/HkeYs5VDkfhJC61t31aRpH7k3KZzjIGpMjTysbbvec0RS4+xE7ovirIHcBzwODOT+gFj1o3/cucj7SgkPrLy8QvM/bStK+yxd+z7nOGS8nJrRttLxiPe0e9R57hZG1oltY5K/eWpDt3o/V0G6+yE975PFeoc2WZ+37vLGckCXV+vsdrrrkCTd1NRIuTWMad/QV0H76IaBvvmJ1Su3K3LIrmJIM5053JfcReRvA38CuKGqr9XL/jvgPwJu1qv9FVX9pfpv/w3wF4AA/Keq+plHsN9PFI2y5HF+3kkHjcOTfc2lbiLd5BkiAiGUjMeOsgqU8ykijrad+9J+dLct0nif2APLFpH/0kMA20MgxiyqUOuCpcajpeuYIECGwUZwAcaSnBCJgeiFvivSTPA2r5OMsv+cBVVUDv7ZYcnVtBX4jUNPqMcUrfdTm3PbtBXUdAatgkEQDRSzGVtra0iIOCzWrhNjxNnUtUopMFYJsavrz1jVaPphoP+arhok6ya47UHbegZZHuX4M+ABcZzI/e8A/zPwd5eW/0+q+j90F4jIq8CfAT4KPA/8soh8WJtGiwOeIJZNSB6OT8jyIHKS9R8EsXaSNNZgGmXPEWmZXpLqKmfq09CelfrFIn5WgixF5gci7f733Rf1oJNklUvXpmf7Zx9P1c6+J3A4PFuCqv5z4M4xt/dTwM+raqGq3wK+DnzyFPs3YMABNB4t1rq6O9EDaNbRtsp1Ec0vOoY2EbdViAKVVbxJj2CUIEqs7R4FxWh6yAnnKUIIbdu8RdrmRIcyYMBK3Jfcj8BfEpHfE5G/LSIX6mUvAG901rlWLzsEEfmLIvI5EfncoJkdcFw0mnjnbB25N4H4CYiVWpbYmdzUmMhZlAMPFcXbiLeRYCLRRFSa5hURIWI0PfpLmFbvQxqkFuQ+/A4GPEw8KLn/LPAB4PuAt4H/8aQbUNW/qaqfUNVPPA39CAecDegBtYxpSfEktKi6FLnXE5xGFxG7jeAiiC6idJXac55F1C51Wqh5fYKdSF4zdbPp5QnaAQNOiwdSy6jqO81rEflfgV+s//km8FJn1RfrZUdiuXPOk4hgVt3en0SD/bjsB7o4rmqlXnvF3wWp5YHGmLpD0sn2Yflxkgnn5XL+A/tWV3YSNTXksEqMAWsdeZ5jjCAqWFlEv0ddh+azWiuDOuo3CrnLEFXKosSZNHiU0WOiYI1JChtNOnsrZtGlqZ6kbSZabWPqrovJSWMMzrnWiqBbG1D51KnKpUaz1G4LzWYBRdSwPISdpjZhGa0884RzKAPOLh4ocheR5zr//NPAF+rX/wj4MyIyEpH3Ax8CfvuY23ygCsmzhCdB7Ec97gul9mmRdnLSGHPiadbTFkE1hNvo3Q8RfEvY9eAfFecSuVsrGDHt+5tBonlf735ai1iDNQZj0vsFKOclTixjlzMyjpFYJmKZKLgqYIoSqTzWx1oFkyJ7tOlrujBKo1ULNVYJqeVdluWp+Cp0mpVUAY3Nteiki5q7BOUQsfcd38PAkwhQBjwaHEcK+b8DnwaeFZFrwH8LfFpEvo/07f028B8DqOoXReTvA18CPPAzg1JmwMOEapJUZpnDOpe04kaQeAL1T6Pvr3uEiyhGUnu66AMSIsEHfFSsVdadSXc2IgRVApoieEmN6BbpGkHEdFrZGYSINZaoqaPSaDTCZRkhCNZJS+LNeNbYOQwYcFrcl9xV9c/2LP5bR6z/V4G/epqdGjBgFZL0UcjzHOfS1zdZA3P8OdVaRy+RJNEWMCJkmYEQya3DWEcoKyxKrgJqiAJeFC9KpQGPElv5orbWxM1/QPtaROpUUpZSM9pJX6m2LfCawWvAgNNiqFAd8FQhkV8i9yxzNZmeLLvYDAYpTVNrzUWxRrAoBghlhbOWDWsR7wlBU/ckUcTV+X2h1sDXbaSbbdYxu6kj9xgiLreMRqN0p6FgbVo3zTMp1rBQ/gzkPuAhYCD39xR0xetTblUPWwWv9Kc5sKxT2did9D2C3BK3C846bB3u3rdqvm8/Gk+DppxUEiFba9AqMMpzfugHfpDv/+CHWFehjIF37tzmS9/4Kt946xq+9O3uIovdT1F69wiF0E4AZ0CaGLa2mYvtpmWWyrGOcYn6j43lXugPCU2lcd8J7/vAYZR6khjI/SHiqCYeTx6evh9g02xB1QJZnUTonyZZNdHZvO4agDWTpMv+MjHWxUKmk5lWkNg0QknLEMFLDpIBLnFKEMoyMpKc9cxxYTIm8xFLBTFr1TJ9nvfdfbDRpAhaDGlaM6ASQErAM8nh8oU1/sM//yP80E/8GH5zzJa7SNj1zPc9v/5/f4a/8bP/C5//8tvka3CvcoRqjFHDxHiwFuMMxlqsVYzLKGOF9/uMzIiJlkxiJMsKylhS+TnRCpUKwYwwYpN9gUZsracP2F6u7FV4rVh39TU9bnMWJbmKKEioH3HFZwlgj7ndAY8CpyliGrCEB1KsPE5Iz+PAn3sWNn875nHdbyBbFZUub7PJU9N97uyfNQbnLAt3xYNKq6P3dVHGlGAAg4ht7wZu3pzyhd/7fQyWUT7B5WPGGxc4d/UFfvyP/QSvfvhDjDMIJUQfUO9xo1F3TxZH0sp8lSxzZM6mStjmJkWo7yDS8TXK+dNFvn0X+/BDOHxtV563dpcW5gwrP27AE8dA7gOeOI6ScxrTIZuGBElSyCzLWgvh7npdSWUvaYnWs6mNIYxBscRgqKqkm/dB+Be/8/tMtyNj+wzEdZQ19m7s8M/+8S/ztS9/Ay2BAOuZZZRbJq4/UpW6klZEWJtMGI3HB/epTuW0xzhgwEPAkJYZcCbQRIrLvUnbiFZo9ejJfqAm9wC4Rke+2oe+mbxMZKqkVEQnulZL8AEJBnWWzMKN63ts35hz8argq8CdN67zCz/3d/iFf/D3ePudG+QOfASJERcqTKh6P9uIoarTTmtra0wmY5o7EjHNAJYklPIIHSAHvLcwkPuAJ47lNEA3qm3khUjK01NHwc45nHNE7aRwlvLtXXQnfQ+nFQQwjPI1JtZS7tzFErh1b8rNN+/xXa+u8eZXvsVnf/23+dLvfYX1fML6OOflDz7Dd736Ee7tBz7za59jVsx6I28x0vrkj8djRvlocZwcjOBPJOkcMOAInClyf9qr457WfReB+Xx+oFH1kziWgySfJl9TLb4uCoMEYgw45w4U/HR9yZsG3c3xdLffbLt+U/s3VZjO50ie2ve98tzzvPqhDzLJNoCc7RvbnN+8yF/6y/85u9vX+No3/wUvf/B5fvDH/jW++foNvvL7/yVf/851ghxMzTR5fefqlJG1KEqWCS7LqIgPbBr2ULpE1QVZg+3Auw9nity75NL1mnla8DSSe6Ny2dvbo2ls8ajJfVWk3uxPeoaAh2iwjcKmVtLMZ3Nc5mjI3hjBGcd0NifLsgP73teMQqM5MDAsdkQ5f/4cf+iH/gif/tQP84Mff40X3/8yhMj6ZMwP/8EfZbI1RudX+dgnXqGMe5T7O4RiigZFNCKzcyGDAAAgAElEQVTGtVLB1m3SKnkmxEhreDaZGFyWUfiqtS6gPr6TnPnTXqemcvbMigAGPDDOFLkPePdj2UOoGUh6B/IVPDOdTYkhsrY2wfuUx35m/SJ3vv16u+2yLFuTrv4NdzeuKJ4sM/zEH/+j/MxP/zkunVvHEnAjxeicS5fPMTo3RrJIrCrEKfhI6ecUxT7GKEUVe9V/IoaqUoqiYn19jUuXLvH2m9coQ0DR2hPHJoXMkJYZ8JAwkPuAx4ZVeXWgn9x7iRKqqqKsSqx1KDCZTCjG4wNpmMYrvRfakHsjFlNEIs9dfZZP/vD3ce7CiN3dm1T7O6zvr/PMiy+zcWEEk0j0M0qdUcQp4tJdTj7OeebiBvLN1bwcojKfl+RZzoXzF7DW4YsqmbYZg7WGGIeoecDDw0DuAx4r+rTUK+da+sjdACgxRLIsI3OCD57pdNrxeNcjonaAZAHQVlpKBALGKltbObP5XXx1D2TKzv4uG/sTJpsXieU9ZvNd9qtd1JUQK2bzCk/kueevMHJfZdbzaSEGQJjPS3wIZHmGaqTyHpVkA2yN4nUI2wc8PAzkPuCx4lCxUj0Bepjc+0lOY6qqdc6Rra+zNoHdnV3uaNEqaKqqIs9zQggrctLL1TbJsnc2m3L7zg38+y+BVCAlIXj2pncZrU/Ym1bszncpdQ+Nc5xTquAJMeDynCpK74AUY0yuk8FT1BPX3oe0f9YiYjDWQJQhKzPgoeFskLuQSpmbEnkxdVTVVwZveVSVHmd5Mun0+2Z6WUOBGIXgIctHoAEfyhNvfXkytj8aF6zJaTonadQkZRTbc0UFU62B3QV3i6gjvN8icAEvU8x4Tpxt40vFhDFxlKL1VUqZA/tqa4+aWlduNE1j7u3ss79fsF9FqiowykaIHbE788yu3SbE5BODEbCKN4Fiv2R+z/PGN95EA+AMiK8LpSxgyewGxRxme57prEBsgdeKqhSiSdfWiCHPHFghVLVCyHukVgs157j7WDUBuupOqOt1/0DQ2lJADPXBglY8bbWQj7Ixz1malD4b5A4kUu+8hn4Of0RhzVm6KH04/b6teH9HnbIo+X9wdL1cDu1BW2ufPnAhAZTDxKMg0aFGgSpVkCJEtVQ+YqwyWsuwRogFkNMOLKs+v7NpGnMtqXdMELyvqCqPjxEfI0YFq0LpI6WWaVRAwUTUe6IWVGVkf3/Kvbs7zVHSGmwRAYtgEQzBp+bemTVJ0tk0IKlvIkQMFgEraIiISdF8u0KvrPNxotkPDj7TfX5v4GlQxp0hch/wbkIf+QhyKPpcNvrqYtHVqP4xxfTvqvIYYzi/eY7xSJjuBR5G9BhjpChKQgiEGKkqJaoQxUKsO3ugYCoqP8eHKahjb7rP3qxIVsF9x1G7PoamuXfucK7/mEUMIjWxazIOVuVQ5erTQC4DniwGch/wSNBP2P3kvgpNu7rFemkg8D7l3be2thhPDLOHFMWGEJjNZnjvUz48eoKmuDsF403kXlKFkqoqiNFTlmVbgdp/HIm0Yz0HkBqN9K8rRrDWohIRNQjJO97HhXz0cdQiDHj6MZD7gEeCleQeD5N7H0kptMTejdyNdYAhhEA+GpE5i3MOkdMXvMUQKYqCEAIxRkKsMGqI0qRH6shdKtSkLkzzYo4xhtHIrTbOre9OQoxojPXE76rIPfnNKKToHUkk3xbVLu50BnIfcBTOHLnrql4Aj/xzV/9QzkIufrnS8mGhd3s9xZsPY9vLZfbdYqaV22grNpOvjJFFezojgsscxlisEYj+UM59daOO5c9J687n83ZyONZFRlFCbYOQEvVCmuSMwLyqUGOYbKytcMGHqGmS1/sKBbIswxghhlj72h8+4a3njDFoBGtSjr6pB3iQ6u1HORismlwc8ORwJsg9lZYvvghPKiI5y1/Q5S5H981THxPL24kxphzxg+/qkegOUkeVvAtgrAVrUWtqiYsl+oixoSW38SSR6ngyZm931qplQgit5v3wxvtHL2ssOzt7SapYeVSTMkZjAE2DiZA4vijmFOUUBWbFjM2trZVRifeeibPs7ZSIGNbX18nyjP39PbYuriEGjFEkgg8h5doVbK2iwQi5OGKMVFVFVVWEEBCR1mOne937zumR34tTXuwmTXRos2fkt7OMsy6eeFg4E+Q+YMBxkXLOyV6gIfAkyzvBNuq7gOa1NAOnCD4EvPctGUaNRI2I1u3wSB4yURUV8DEQFayzK294rLUp1VMXV507d47JeILqXmde4XhBTeO75FYl7QcMqDF8QwY8XRAIHqbTKd57rDWAX0mMvRGaSZK+1MC68YKP9WStp6qq5JRIUrhEjQgh5b7rtEzQiAKV94ncs9U/JWst0VeAEENga2uLyWRSDyBx8YxdKfVtjq8h9kbH/7SZ6w14fBjIfcBTBUEIgaRqCQFr7JHNoFdN7C5SKCnnLbXW3odE7s5EECWqEojJ8VFCisw1kXskVahGLNY5jKzi5kVXqd29PUQE61wdrXeKk2JE64YkRx2PtXZxNzCQ+4AVGMh9wFOFpmXdvCiITVpmFaWuyqvWKRhBOrG7thO1vvKYXDGkCtqoqUm1Ua3HhLr6tU7jKKYl916dez2haoxh+949yrI8IPE8pBpa3m1ttPKHPXkGDFiFM0PuTS5R9YiJsDOOoyY5HwVOFrV1eoZ2kDhKSF3sI2IUCX3rHSw4Wrbu7T6vhiKiNbFqsyS9d4nRVCCYXSAiYQLqMFqQup0qt95xbO+eo3Jj9u2blCGlP5r9dM6t/B6lfhq1xLJWxFQxUEbPrCyoRLAmQ41QVYGoBiESpapz9AZjtyiLKaqe0hesbY0olfT9FTAuVbKGGFCtcBiqoNy9VzIvcmADY4TReMQ8VsRQYazpl1MKSaFTT+g2/xOJWEvnDqCxk9ADktMGvfYD7Z2C1JJTQSU1ID989U6Gk3w/T22N8JjxNEzInilyX35+2gh+lc/H2YDSl79YNK1I0am0NfkrtrJUUdoMysdBSkesGgAPE0GQClFB1GHUgFZ1CkXZvqfs7jmCySjdHB/X2wGnaUCyquFL2v2DGvqokRA9RVUQgCgGFUOoJZBKbNM3goA6qtISY0bUgrX1ERglaMSIJvdK0+TSQUyacJ3NI0UhWLuGtdLup6K13cCqc8fS9UvX01g5QO5QqzY7VHxUFbBgDv9NgZ7lfdfovYqngdzv+6sUkZdE5J+KyJdE5Isi8pfr5RdF5J+IyNfq5wv1chGRvy4iXxeR3xORH3jUBzHgyaEh9+XHqi9/lmWHHstmY61SJcb2EernhmDm8zl7+3uIAWs5Os2xhAProPV2U3VoVXliSJ/TyEK722pez+dzvPetDHAyGbM26T9mVxdaZVlyrJzNprjMYW3ypm8Go4dx7lNO3rR5+eQVb5+6yHjA6XGcK+6B/0JVXwU+BfyMiLwK/NfAr6jqh4Bfqf8N8BPAh+rHXwR+9qHv9YAzg1XkflLcj9ybh8ZkQzCdTtnZ3k6FTLZ/cFh513TgM7rWBoaiKJMCpv68JrJv9qd53WjNm8/I85zJJFt1ktqJ0P39Pe7cuYu1BucsRVEQo2JM7RV8ghu9Vee+IXZr06AykPt7E/dNy6jq28Db9etdEfky8ALwU8Cn69V+Dvg14L+ql/9dTd/63xSR8yLyXL2dAe9C9BHHKnIN4fCU4yrf9aQvlzQdQMpKNImB2WzGzu5uLQ20aLkg6u6jN1UGtQRRFqkMVYxYyrIiBE+MDiTWOezDRTrddFSjt3cuAw7bJWtscu/Kzs4Ot2/fRqNinWNalGhmH6gTUzcNdeDuQsxBFU5noBzw3sGJcu4i8grw/cBvAVc6hH0duFK/fgF4o/O2a/WyA+QuIn+RFNljzNnPX501nJVcft/EKvSTuKrivT+0fJX3epMTT0wViUqdcU+Rc1EUrcSwCXmPlZZpUy20kXkqWk1qmZSWSfly0f6Bq3uH0jZ2b77HUqtb6smLGGMrW5zNZm0zcmsMIZadIqpVZ3k1uue+JW+TKlyPW628vM6TyCc/jO/z05AHf5w4NrmLyAbwfwL/maruLNkFqMhRauPDUNW/CfxNAOfs2WCqM4w+e4aTfJkfxgnuK2lviKtRphy3tLshoobwm+i4fb907ghkoQcvvSeE9F5rbV14pPWyxSRq8+ir5KyqiqhNmb+h6X8UvGc+n7M/nbK5OSYEj4ghemnnBtr0UGe/vU8WxOPxGCOpT2oVZ2k+wFgym2PFEJ2jmFfM53PW1tZwLmM8ht15RSRL6ZOonITnuqS4IPrDjVlSzj9LdsaNMdqKOybV+7UpfPgYyP3h41jkLiIZidj/N1X9B/Xid5p0i4g8B9yol78JvNR5+4v1sgGnwEq1wzG/0Kf97az6nIbcmr83hNwXNapqG9H3RdgHBi2FsioxGAwWKw5rDbnL8NWszXlnWUZVCSIWV3di6qYg+kmjb5lg6sKg5NxoCTE5RYLt3c++c2SaAiM1WJMse5tTZ4whxIqyLFibTEAMIUTW1ibszDxlVXEaAVt7jeq7kANH3DMIDHh34zhqGQH+FvBlVf1rnT/9I+Cn69c/DfxfneX/fq2a+RSwPeTb372IMVKWJWVZUlVV+9xEhceZJAUOTAQ2KY8m2hyPx6ytrbG+vs762hprkzEiwnw+r6P8g5OIzURiczdx6NFzHFLvQ6h179Ymkm1y5cdJ9wiC7U5u2iZ1I6hGnLX4KrC/v4+1hszleB8YjcaApOKmAQMeEo4TJvwo8OeAfyUiv1sv+yvAfw/8fRH5C8DrwL9T/+2XgH8D+DowBf78Q93jAWcOJ7ml7hYaNY+GkGER+YsIo2yEFYszGc5kZGKgqtCYiHs+n5M7YTyy7JaHB5GVRLxqd/WwV/py96gje7OihBgJUVJVawAhYNF2ArWaBfb39rHWMhrnteQyAJ28+5CkHPAQcBy1zP/L6qmeH+9ZX4GfOeV+DXhK0Ej8jpOWAdoceJcwsyyrG24czLlLlhpYE9MEYYwRiWmb8/mcqqy4cG6L0SgnzAKBg3cGqwednjwzKWXknMNI8l5vjif4w+mjXhVOVLyv8FWk0goTUz7fuBGo4qyj9IHpbIYxhsk4x4iwP52iMZ2bgdcHPCycmQpVjVJL05LjR1o4fNUfJbqRc5d0+6qDVykqGjLuqke6ufUuYS9XszbbaLTY3e2ICDGkiDeESBU8GgLEAKrs7RXMokC+yVwzymqfoIdlgf355eQK2TSubpdIwNkMEYv3SfdehQpVg48BaRwco2KdRWxaHkldmYwVQvBM9z3GOVyeIziqEMiNpdSKQuDmnqe0a2STMSqgoYKo4AUxESMOapVOTE3+SOUmA47CUXMh70WcCXJXhRgTwaMmicjeoxfkSUAkqUFaSd8K64dlwuySdpMr7yP2hrillhouF980n7c8EIR5U51a4X3Sn4dQkeUZu4Xn7l7JM5VlrjlV2O2NBXp/8GIwIlCTe2McBslLpiwDGgERVASsEOnU9xuYl0WSY1YlYg3qwTiDMRC9YDTHunWyLKMo5tg1QxFn+Mzx9m7FnQLMaISPinWKE8WrhwDGSu0EZOsHPCxyfzf/rvqu9bv5eO+HoWxtwLGxTL4NSS9XQR6Vg+/bRlc+2Uj0vPep21FRUJZl27Q6PSf/9u3tHUIInD9/HmtOL90zxrC3t8ve7h5RI7u7u8nTvZYPdh+QJJXNXU+MqYGHGGF9fZLsguuuUFmW1Rp4g2pkPpsznc4Zj9ZQhKqKuCxjNBqntnosEkfaeZwWR91FDXj34UxE7gPOProR0DJBpJ6g5r4Tjt33LnvQHMrRK4gxWGkiEEtyDAUjQpYZtrfvcfPmzbow6nT01/Qsnc/nzOZzrLF475Olr/YbwjVEP51OqaqKzc0NVK8To28Lq5p1jBdCVIoiUFUeweCyEc5lzOYlCEl62Sloak6HPlh90+FjXDGgDnh3YiD3AcdGlxCaqM9a2xb49FWl9m2jS+rNNrudhdqJVpskj2JcLdtWxuMRPlSIUXZ3d/jCF77A7u4+VahdFR8QUSPz+ZzLz57DWUvlK8bjMRojYg7/TJrCJRFhb2+PK1eu8OlP/xhv39jn2rXr5NmIqIGyLAAoy4ixkGWW4Cvu3tul3J0hJiOqp6wiKhbFNO67KR9P5LAh8snRl0pr/n1Wqp0HPFwM5D7gxOibTO3mzWG1oqTvDqB53ZB6o5F3ctg/3lqLeBiNMsqyYDabrfyso6AdZ2NVrXXuwksvvsj6xjq3bt1iMnHM5jOMPWwI1gxCu7u7XLt2jfPnz/Paa6/x6he/wRtvXCdEj6pBrCHPc2bzfSaZ49y5C8zne3z7W29gQqSslFjPM+WjnFhFYu2p3nRtfVjayFUEvzgnJ698fhrwXs3FnwlyX56EO64nxoDjQ8SwbBaeIuZFtOyca0vq+9Bcl67VQCM77BLHsgth8z4RaVU5y+/rrtfdTorcQSR1QWr2EyLj8ZjRKAdgZ1ocMg7rbrO/YjamiVMFiUmS+fIrL+Nchvdz3nrrBiEGkFT5ur29zXQ6BeD8+fMURcE3v/lNyrLk2rVrnLtwCe8LRJQsS5O1zqb0jnMZMShV6bEm4/XXr2FCStUgGSCUVUwTqdJ46IA2rh4P6efQNyB3U2nNs3PuWFW5Zx3vZQXNmSD3BsMt4qODwKGS9LTosJfJ/a7BcoS+TACrIkDVfuOwbhS5IGKotO58ZBKxi4AxpKbYonQHK2cd0RyuJG3UOQcIXho3yM7+CaxN1rh48SKbmxvs75fcu3cP4yzO5RRFUdvzJr+a69evUxQFzjnW19e5efMmX/7yF1MzEoF5UZI7IXiw1iFGalWY1pWwoEFaNYwKSQosqSVIpI7epXE3O/KSnAjN+b7fdV4u6BrwdOFMkfuApxNd+WQ3Su9bry8v3zexqqoUvqgj9wW5WytkWbLj7ZK7sfaAE+KREXz7LlkMeKrMi4L9/X3m8zmT8YSXX3mZPM9558YtdnZ2WiLf2tqiLEtu3LjBdDolxsjt27cR6xCTYZ2QZznFrEp+8y5Pza8hkXedaYmtzPGgKqaN3k/mxTdgwAEM5D7g1FieCD0qB94XBXZNvrqEHH1TraqkLI+iaupJxqYnbH2X4Ma9dyZ9dgeqmsgz1rcCMWncR6McQZjN5mRZzrPPXmI0HvGdN97kxo0bbG5utiqfq1evIiJ85zvfYX9/P92RiPDy+17iQx96iy9+4VuM85wQFq6RWkvlGw09JPlmc0ZiE6VLInfoRO4DBpwQA7kPOBWOym/3EXyflWxjQQsHvd0drt7OYt0YI4Q2vqUlPvHA4buH5dyyiBA1VZmKSOugaKzlhRde4Nz5c8xmU6BgMnHk+SZNkdf58+ex1lKWJbPZDFvn040xrK+v46uK8WTEJz/5Ca69cYtiXpJnGVUV6s9NdgopNSOImE7kngYYldhG7EPkPuA0ODPkPuT1VuOoyZ/TnjfVWooYmuhYOI571YLQY/I9N4Aqqh4jMeXF0USkTfrDWCwjNIaUZkHRWBFjQEKyFbBQN6KGLKZoWKIulC1RiR6aZt5pstVQEUFqrX0dIYuAE8H7lCePBJKsMOBsfYjGUJUl48mYP/Ajn+TiMxe5e/cdtndmKR1kLaNRxtWrV7h8+RLz+by2GnbkeUYInslkzLlzW9y6c5evfuX3+b7v/QR/6k/+Uf7xL/0q031P5gTVANFgrEMwqV2grdNKtcxTAFElEtvJ48UJP8kEYH+0r3Wj76jpkbb/3vzdPUq+OSuTtWeG3Bt0J3vezYT/MIpITnZ+pJ8gNEWMZRkQcQgOSC6F9/vclL+OILWrYZoGxNo06RlUUyMNFXAZRhyxWkOIGDwaZoRiioQSJ5AJTEawuWZYm2RcWR/VaRkLYlExzKvI7n7B7v6M/VlJUSqoYeYsWIcYiyr4EPAxYL2Q52P2ywpVy2RtDeY7GAl1O71IVRacf/4C3/WBF9nd3cFkOUbg7m7B3vR18szy0kvPE0JgfX3MeDxmMpngfcHW1jrOOTY3N1GUb7/xFtff/A5/7I/+EYrplF/8xV9Og4vNyd0azmSoSnKCzMp20OpeGbvMuQrUw97x4A+t2ty5dOsJ0kNpirTeS4VNj4pbztL5O3PkPuBpQyqrp0krkIQsvlKqEPEhEKKAB2OUzAesFUI5I1QzcgcXzuVc2Mp49uIGz125wJVLF9jaXOOVy6OUqxaT5ILiKCple3fGO7fu8PbbN3njrXe4d3fKte2KwlcEn8YrI+AwGMDiMeoRscx27zJxysVnLjIajVhfX+PChXN8z0e+m/X1dfb2dimKOSF45sWUUQYXzq8BSR64trbGhQsX2NnZoSxLnn/++XYS9tKlZ7l55w5f+eqX+YEf+H4+/r0f5Td+47e5dfMuYhQf5lRVSZ6PGU/GFPHx+rf3pc8GvHsxkPuAU0GURZ9RIaVmMKl03yvex5QiiWBMJNMpfh7JXeTy1XXe99KzfOCVF3jpuWd49uIG57fWyJ2i0XNufDdFQqaJ3C0+CM+FjA9+4Bz7+89z/cZN7ty5w7/8WuDNt+9w/Z1b7O5XWCfYfExRzSFGxtZSVCXf9/GP8ckf/AHe9+IL5HnG2toE1cBonLG3t4e1BmOFb37r22SZ5aMf+SCbmxtt05DJZALQNiYZjUZsbW1x584dbt66xe07N9nenvHVr32Zj3zPa/yhP/wj/PN/9hu8c/0uZRVZXxtjnKMod8E9Pm+X5fmQdgL8oRgbDDiLGMh9wCkhiJrknqhNtjw9J+KPiEZcjGQoa1TkG4YPvPIsH3/tu/me736Fi+fXGbmIocKZAmNiytvLXkoVYFAsgsUamxJHAvmGsLV2Dl45x/tevMDrb9zgK1/7Ol//1hvc2Z5RhDlZbtgv5hRe+OSnfpCf/vP/AS+98BKZc+zt7SIC+9Nd9vZ28b7CueRhM5vt88or38NHXv0IGoq2U1QIgaIoyLLU83Q+n3P58mUuXbpEPs7Z3t8mhIh1MJ44furf/Alee+17+K3f+h0+99nf5fr1bbyvmKzlCJPHeqX65KEny+UPeJowkPuAU0EAoymLnnTbafJUjGAMZCYisSQTZWyVl5+xvPqRl/i+732NF1+4zNbGiBDmhGofjSVGA7kxZM6ivtGx14pwVRxKRChDgKBkYjHW8MJ5y6WNF/jAc1t84coWv/vFr/DtN/eYhYhEw4/9wR/gT/9bf5rLz57j5o3rTCYT5vM5xia1jrWGtbUx586vsT/d5mMfe42Pfew1nrl4ETQ17rh37x7z+ZzxeFzn39eZzWbcu3ePCxcu8OEPf4hnrlxgPivIspztnbuoRt7/gZf42Mc/yo//kU/z937+/+Dzn/8yZeUZPeZfX59ktUmnDXj3YSD39zC6lgBdSwFrLLEnH9ubo62F28YkYq9CmsyMwUOoyKkjdie8ePkcP/7Jy7z/u17ihRcuMsoV/C6GktxWiAsYAkIFATIzSXOJCsRGJ57kiy5arAoaFfWKKe6xOVpn4/I6zrzM2tjw7LM3+Ny/eovLVzf5k3/iD3P+wiaz2R4us1S+rIujkg1viBXnL2wRQsUHPvBdVFXJ2vqEsioR9cm7vUge7lmWsba2xuXLlymKgps3b7Kzs8PFZy9wOb/CbDajKJIH/Z27t9jf3+Pch8/xoz/6w7z88ov8yq/8Gj/3c/+Q/f19Ll682Fa+Nj1c++SiJ0GyFz66oAuaivDTE3ufdfCQz3/yGMh9wGGVhBz2IVz9Y1UMIaVNat22V0U14lCsCaxZ5ZXnnuETH/sg738+cnHN4sIUSrBGERMQPELAGEUanbdu1uSuC1l7jBDBaFquMWnERzpDyopocy5uWF798Pu4/PxzbF08x9aVF3n+hat4gTKUBJ9SSdY23b8qrBWcM3ivjCcjrBNi9EBGWZbs7OxQFAXee/I8Z2NjA2MMo9GIqqrY29vj3r0dLj93hclkA+9D60E/nxe8/vq32dnZ5tKlS/ypP/UTbG5t8vM//0944403WFubkGVZm8Pvs2g4KfosIR7lJOqqiuQBTw4DuQ84JRTBJ425CFGSllosCIHcwbPn1vjgy1d59UMvcW70DrmFUM5QA3ZksSalcVKHpJjSOgJB81p9o9TeAjXxx1qJn/TgqDIeW0IMFHFOJjlb62Ps+iY/uHWerSvvA2vwQQkiFEXJKMuxNiPGpNEfZck/JoQKJGIMOGdrX3bI8/T3nZ0dNjc3OX/+PJBIbW1tjb29PW7fvoPLJ6ytTRiPx+R5skF2dsSdO7dbf5qrV6/yB3/0R3juuY/yi7/4i/zSL32GZ5+9SJ7nD0VKt0zifemYIS3z7sdA7gNOBSEghORDTiJoNUnzTqzYWB/x8kuXeO7yBrkpGec5zhgIMVWJ5slBsiF3aSP31GqOZs4v1stjbT0QFaOa/NYBKyViQMVSxYAPJUEMz16+itvc4O7+FG9yKnFkWSpMGo1zQkhpGWPgzt3brTnZaJQzHo/rASWlrhpfnDxPTpRNlL2+vs4LL7xAiDCfBfb3trFul42NDba2NmuzszQBu729i/eR0WiNF198mZ/8yZ/kN3/zt2vNeWiNyU4D1UXU3HX4XJWmGaj93YmB3AecGiIBEV/LFSG1mPPkuePiM1tcvXyerXUHYQZBMU4wztWeMQ4jFmpzsPYZUPENU6U8TIyIRCAkbWXU+t9KmM/T4GBHWKsYASsWMYabd+4wZ0xwQhErrm6ex9Ulqs5ZvI9UVdUak4VQoeqIGtr21Nvb21RVxXPPPcf58+e5d+/eAZfI9fV1jMm5t1tx8+YN9vZ2yfN7nDu3iXMZIobx+lpbOHT79l32p2OstVy5comvfe1rbG1tkef5qdMyq5w6+x7JLfRUHzfgjGIg9wGnQ5uK0VrHAkYjGjwbGznPnttga33E2sgxyS1VNSfLczLjECu1z7yp+UUX7geAGF3k2VuL35ictyNsEAsAABSWSURBVDSmVhZ1Gb0b5agKXoVYbyaqcufmTb55Y5dzz70fGTv2y8iF3OOiYmMi+Bg8RVGwvraOxsjOfE5lPKqRyWiMsY6iqFhbW+fSpSvMZgX70xlVVRFjZGMjw7lUZLW5McEax73xXXZ3trl27U2m0ynnz5/j5Zdf5sqVS4gx3Lr1Lb71rS/xyisvc+H8eTKX4StPWZTtncFyxlpOZBXQWPpKcqSMy9bOx2f0VRbOA842BnJ/xHi3/CBWHUc0jrm1aDQQLaMorGlkZIQXMsdlqzzjAlumROcFcd0y91NwI0YuRwVCjBhqoy+kNtUS1uJG7ZLSWPUm35VAIKKoaG24pdyRAoxBMVReqGLKrb/51j2+8e1bfHD8AjH3fOkbb/DZ25/nk9/7MV588QXefutN8tzxzMVzXD5/hVu3bnDjO/e4cvUywUR0LVJGz/ZexXObl5hXBh8yfBxhswl3bt5kZ2+b97//IsYY9qc7qdipHPHiix/h9u3bXL/+Njs7u3z1q1/nxRffh6py/a13+Pxv/h4vX36WV65c5l8UBePJhLKskNrXnSbVBUB17GtlTYZqY8gWCEGJUZDaYlioTSmX/Ws6UNU2FdUoeIwxbQPzxoStaQ5+3O/MMMn6+DCQ+2PAu5ngRZpb+45LSv3aOYvLMkzjFbRyu+l9ybO9aQEHKjVpdCZUkSaKP/iQpe2JKBoDs/097t6+xa0b71DZKb//hX/FW1/7Bi9cvgTA5z//O2TO8Af+wCc5f+H/b+/sYuW6rgL8rX3+ZubO/fH/X9I4TaIGJy0phDZVEUVIiFAhEFIfWiHah0rloUiNhIQakBA89IEHWkACRFERDy0UUCtRtZWq0lQpRdAmbhLHsWPX8U8c+/o6tu+duXd+z9578XDOzB3bY8eOf+Zy7/6k4zlnn3N891lnzjpr1l57rTqHXn2FbrfF5m2zJEmFrJKycOYCZ8+eo16fQcTQbreZn59ny5YtdDpdlpaWiKKYSqVCp9OhWq1y+vRper0+KysreA/vetfDvP7661y6tEi1WiXPLSdPnqTZbFKpVqlkSZEsLUkYryq5YWN74HO/OkJGrnLD3O0cK0G53z2Ccg/cMjIwA40UlZN8kdo2TVMqWYaUFt+1Y6qlLKMnlyv3QfzjULEPkqGPpvst2qWs5iEIRgqbNzZCJY2I8Bh1xOLprTTptFZIkpTFRpNXjxxhZqbGL7z/vRw8fJD9Lz3Pww8/hKTQo0NsMxYWzjM/f46tW7fR7fZYWFjgwEsv8cijj6KqnDp5kjy3bNm8mZWVFQCe/cEP2LFjB71ul127d/OhX/oQU7UpcmtxzpMmaWHBnztHq9Umzx0mzknSDO9vNdOnv46CD2wU3jK5hYjcKyLfF5FDIvKKiHymbP9TETkjIi+Wy4dHznlaRI6JyBER+bU7eQGBySNmoJALN4IpFW2SJKRZNpwgdS0FMzj3cuUuRbbJ0QFUXLl+xYIr3i2mKHJtRDCixBHUspQsMcTiqSSGLBIiEdR5siRGxVOpVWh1Vnj50Ms0Ww02b99EUom42LjA8/uf5/nn9rOy3MLmjpXlFieOn+To0de58OZFbO5otTo0G8v0ejm9Xk6n0+PihQbnF96k2+2zstzCe9ixYxetlTZGIqanZ6jX6zjnqFQy4ji6LJf9rXKtyJjAxuFGMhdZ4A9UdR/wBPBpEdlX7vuCqj5WLt8GKPd9FHgEeBL4WxG5tSl3gTWLlNbywNouQhpl6I9NkwQjZiR/+HiMkTHKvRxEHf0cumT86qesRtgIiimX1AgzUxXqWQx5B+23icWBOlrNS9RrGVvmZnj3ow+zc9c2Fs6fZdPWObbv3Erf9zj1xkme2/8cp06dJkkyVAVjYvLckeeeJMmYmZmjWp1CJCJJUqrVKZIkwzlDo9FCVYiihBMnTuE9LC016fctzikPPPAAe/fu5Z49e8oB2duTcjco9gDcgFtGVeeB+XJ9WUQOA3uuc8pvAV9V1R5wQkSOAe8D/uc29PeuM+7hWC95rwfXMBgQGwyYjXOOX1ldafT6fRmuKKXVHhlDJckwJhq6Y0xpmXvvgOiyv61qhpEcRahg4TP3JsJ7Lc9ZraoEOnRdFGXvInzfgVe8OtQpBkNsPFNZzJ6d29AYHJZtm6bo7NiE0R6x5FRTwfVbeNtjdrrGIz/zMEkak6QVGs0mURwjYpibm8OYiFptiiyrMDc3Rb1eZ3p6mmq1RpZVmJ3dxKZNhvn5eWq1jDwvJi3ValO02x1EBGuLGauNRjFbdevWrZxbWBhemzEG/NV1Zq93D68MfXSOUm5+eM9uVsEPzhsMnI7e98tewFd8F8b9P4PP0e/PuGMCt5eb8rmLyF7gvcCPgA8Cvy8iHweep7DuFykU//+OnPYGY14GIvIp4FNvq9cT5lpf0v+PjCqHYZHqtyjUcflDPZhIZIZtIoY0LWZ3Wmfxuvo1s9YRx5dPrIGiYISi4Ab7FIeUL45SkSjliyBCKMrVOQ9ePGj5EvAUE5yIiUpXzK6tc3S0SlSt88iD97F7ps7OzRVMvszOzVNcWjjDGyeqvOeRfdx///2kJqW93GGl2WF2ejNndJH73vFOdu7YxcLCebyDudlNOKuoF6Zq09SnZoijBGMMzcYySRzT6/bYuWMP+/Y9ypbN22g2Vpid2YTNHT/Z/yL9VhFOeeToUfK8R1at39K9HNxH57UMfVy9j2/n+zrupXAzyv16brgbPTbw9rnhhNIiUge+Bjylqk3g74AHgMcoLPu/uJk/rKpfVNXHVfXx9aIoA6usPvxFnLX3HmfdMIzu8uRVg0HAgRuhWDcmw0gGGuOdwVqD+ghDimpSTDTyEd4VFj6qqDpUPeIt4i2xOhLxVGJlbirj3p1befAd26jHPbqNs9y7fYaacfj2Cu9+8F1F8rFsmtdePQ79mHfsuZ+87zl9+g3a7S55bqnXp+n3cy5eXMQ5JcsqNJvLnDkzT6OxzMsvv8LS0jLNZod6fZo9u+/h0qVFnn32v7DW0WyucPz4SY4ePcYzzzzD4cOvlr+edDgL9kYZ64IJejLADVruIpJQKPavqOrXAVR1YWT/PwDfLDfPAPeOnH5P2RbYQMjAhaKK92BdUbxD7OVFsAfumtU8JwOlr4im4B1CRJalVKtV0jTFec/K8jLtdgdb1llV7a/OZtWyALYWlZhiKax643OySNG4j/Y7aD9nrjZNKzVUjDCVpnhJ0K5y/NVTGJtSSWq0Wj2+9a3vcvjwUXbv3kmj0eD48dcB4YEHHqRSqeKcp1abwpiIEydO0evl5Lnj6NGfMjX1LAcOHERVeeqpn2d+fp5Op0+axlhrSZIYE8VFvVlVioK0b821k4IF7R64AeUuhfn1JeCwqn5+pH1X6Y8H+G3gYLn+DeCfReTzwG7gIeDHt7XXgTXP6M927x3OOqx4jBtMrilDH4fpaVfdMYWbBvq9vPBvz84yNzdLfXqaJEmxeU6z2WBxaYmlxUXanbKYNR71RVoCVBB1JFHCzFSFHhWSLCG3HkyO9StIJPRbl8jE0V9ucOHsOXbs3supU+c4cvA41eoU/VxQLWarHjhwmGPHjiECee5pNJbJc8vs7By12hTbt2/nxImTOFdM9BGB48dPcvbsPO12n02bZvjyl7/C4uIicSyIE6IoIopjnAfrfZkP/8blPH7wNPwSDtyY5f5B4HeBl0XkxbLtj4CPichjFE/jSeD3AFT1FRH5N+AQRaTNp1X15n5rBibLWMNPrrVjZP/I1hUTZbx6XDlYWgzaXjHAplf7+uMoZWZ6ju3bd1Kr1zEi5HmOqmF2dgtZWsNIgl68SLffKgdlV2etemfJ+45IMrIoIjIQGTDax/ZWMHGFfqdPLauz3FrhtaNHmZvbyemTpzm/cIEs6/DmUoNez5JlCZVKTJ5b0jQhjiN6vT7GRGzbtp3l5WUuXlrihz/8b6x19PuOTZtmcM5iopiZmQTnHPv3/4Q4jlHvqWcJrhywVC3GF1ZdlGP80uNu1SCendXPwGQZN34wCdfzjUTL/JDxpsC3r3PO54DP3UK/AhNldNr7KsJIRGvpAUEjRNNh6KIai5eclu3SjyJcltIxOc538T6i7uaomilijdC+RSLBRKYIEPER3gveRYAwt2OK2Zk5jE+wLUNEiuvF+NwhosQqzOocFkuv0yty0RiH911y26Xveqgovt/B9ZbJW+ehk5O0G9TUk3c7ZN7T6TTxXri0NM/iygItv4SNHV6WwaZoBD1nkTgiiiM80O3nRP0ez73wAp08R0Q4dOgQrxw5ggBpLaVr82EEjJYPd1KtDl+THeJVMQsQjVPgZXZHro4mLtIxKF7LcQ0tBpmNXjljd7V4x40yDGVN0/KFquU4yM3Vfb1y0HW9D7LezPXdacIM1cA1kKu3xn0/dTipfWT3qi/Yqy98yVDEcZvSqi8HToswx8Ki1zIapkgnYIp0wEmERoIKqAhRmhDFMXm/h3qI0oSp6Slq3Sq9fhvbtzhv8c4VOVS0zCbptHDXOFs+gFGRp0YLd4h1Qsd2sc4ipgjV7OU5hpGwV1ktYZJlCe12hxdffInTp0+TJAmtVmuYg2WwrMbsXy6joh7J9V0wg13XSsx7WQgkq1b86NlvR7HfLm5UYYdUBXeGoNwDt4Er/bwyjIJx1mMjTxoZhKgYKxTFi8N5xTqPUVO4JIhBEsQYEIjSGInLtAaRIY5TIMZbSz/v4tUTZTFTUZ2pZoXuShNn+4AF54gixVkP1iHeYLwgzhdTnCQaKva+9fRzwfpiwlGlWsUkMa7TL9MMX/2YqCppmqDquXjx4tDFNCi4Margr5VY67ZIfsTXvjpQfXN++8D6JCj3wG1ikPulCH8sapMKziouUtQIYmIkKhKCefWodxjnkcIsRwSMico0AhCLQZxi817h1kgi8EK/06W13ER9ThQVueNdP6ff7qLiiZOoGFP1gliP2MJ6V2/AepQIK0KuSt9Dz3p6VjBZhTjNSLMKSVwU9RAT4cbkxFHVoZU+jPwZmYg0arnfMYmPGUwtom2CZg8E5R64VQRWPcWjoYwGvME5wTlBvSksc1NY4r7MxW69li6KYqaplDndRQxVjTAdy0pzhW7bIj7BO7C9nG6nhRGPMYrL+3TyHu1Wl2o9I8tSnHc4tag35SCulGngBacxuWixeEvfeXKvxCbGqZBbhysnU4kpEqFdSRRFOFfUSR34ocdVUbrTvtaBxT74daCqaDDcAwTlvqG50uK8fGKRjj12wOpAGWXOl8Hs1sJj7pwUA31a5HkvZpYa8tJydqo46yCOyxzuBlMWvFCJUDFELU/PrdA836C52KHfdWjhosflfcBhxOGcpVOJaPeVxEcQVZDEYfMOThxEpuiTCkQRThSrjlw9XWfoWKXdc1QrSpymIIJTMFHEtTwqAws9iiKiKBrmOR/IaqB0b3YA8hp3qiwUfqXsL58tOow+GhM1c72Bvmvd29HrvF2upZtJhfB20iasRSY1STMo9w2O954kSYDV1ALGmLEzJUeV/+CzOB5WrXdBveKs4m1hKauawoJ2Ec5D7sB6h3WWKI6JE0NsMtK0Bhpjc3C5543jZ+h1LK1ml3Yrp9PKwQqRiVBniSMhTQsr9UK3Rx55fNsSTQumMg1iMKaCiRyu6+hZS08tjU6XRqdDq91lqdliqdWhm3se2HUfDkNWqZGmKe2exTqLMclY2Q1yrwwU30B2ozH+tyVVxUiY6OBlMarUB22rk5oG/7w118uTNHqfB79UbpVxf+9ayc1u5ti1wlrKOyVrQVAi8ibQAi5Mui9rjK0EmYwjyGU8QS7jWc9yuU9Vt43bsSaUO4CIPK+qj0+6H2uJIJPxBLmMJ8hlPBtVLrfDIRgIBAKBNUZQ7oFAILAOWUvK/YuT7sAaJMhkPEEu4wlyGc+GlMua8bkHAoFA4Paxliz3QCAQCNwmgnIPBAKBdcjElbuIPCkiR0TkmIh8dtL9uZuIyD+KyHkROTjStllEvisiPy0/N5XtIiJ/XcrpgIj83OR6fucQkXtF5PsickhEXhGRz5TtG10uFRH5sYi8VMrlz8r2+0XkR+X1/6uIpGV7Vm4fK/fvnWT/7zQiEonICyLyzXJ7w8tlospdRCLgb4BfB/ZRFADZN8k+3WX+CXjyirbPAt9T1YeA75XbUMjooXL5FEUN2/WIpSi2vg94Avh0+Z3Y6HLpAb+iqj9LUbf4SRF5Avhz4Auq+iCwCHyyPP6TwGLZ/oXyuPXMZ4DDI9tBLtfKLHc3FuADwHdGtp8Gnp5knyYgg73AwZHtI8Cucn0XcKRc/3vgY+OOW88L8B/Arwa5XCaTGvAT4P0UMy/jsn34PAHfAT5QrsflcTLpvt8hedxD8cL/FYpazhLkohN3y+wBTo9sv1G2bWR26Gpt2nPAjnJ9w8mq/Mn8XuBHBLkMXA8vAueB7wKvAUuqastDRq99KJdyfwPYcnd7fNf4S+APgUF2sy0EuUxcuQeugxbmxYaMVRWROvA14ClVbY7u26hyUVWnqo9RWKrvAx6ecJcmjoj8BnBeVfdPui9rjUkr9zPAvSPb95RtG5kFEdkFUH6eL9s3jKxEJKFQ7F9R1a+XzRteLgNUdQn4PoW7YU5WS0WNXvtQLuX+WeDiXe7q3eCDwG+KyEngqxSumb8iyGXiyv054KFyZDsFPgp8Y8J9mjTfAD5Rrn+Cwuc8aP94GR3yBNAYcVOsG6TIl/ol4LCqfn5k10aXyzYRmSvXqxTjEIcplPxHysOulMtAXh8Bnil/8awrVPVpVb1HVfdS6I9nVPV32OByASY7oFrK9MPAUQr/4R9Puj93+dr/BZgHcgq/4Ccp/H/fA34K/CewuTxWKCKLXgNeBh6fdP/vkEx+kcLlcgB4sVw+HOTCe4AXSrkcBP6kbH8n8GPgGPDvQFa2V8rtY+X+d076Gu6CjH4Z+GaQS7GE9AOBQCCwDpm0WyYQCAQCd4Cg3AOBQGAdEpR7IBAIrEOCcg8EAoF1SFDugUAgsA4Jyj0QCATWIUG5BwKBwDrk/wApqxwM6pJEcgAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def get_main_color_rgb(img):\n",
        "    \"\"\" Get the RGB values of the most frequent color of an image.\n",
        "\n",
        "    Returns a pandas series containing the red, green and blue values.\n",
        "\n",
        "    \"\"\"\n",
        "    # Get width and height of Image\n",
        "    width, height, channels = img.shape\n",
        " \n",
        "    # Initialize Variable\n",
        "    r_total = 0\n",
        "    g_total = 0\n",
        "    b_total = 0\n",
        " \n",
        "    count = 0\n",
        " \n",
        "    # Iterate through each pixel\n",
        "    for x in range(0, width):\n",
        "        for y in range(0, height):\n",
        "            # r,g,b value of pixel\n",
        "            r, g, b = img[x,y]\n",
        " \n",
        "            r_total += r\n",
        "            g_total += g\n",
        "            b_total += b\n",
        "            count += 1\n",
        " \n",
        "    return pd.Series([int(r_total/count), int(g_total/count), int(b_total/count)])\n",
        "    \n",
        "r,g,b = get_main_color_rgb(img)\n",
        "print(f\"r: {r}, g: {g}, b:{b}\")"
      ],
      "metadata": {
        "id": "fyDql_hNHJZ3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4e38d3b2-6667-4281-ece9-4b46195caff4"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "r: 22, g: 19, b:14\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_videos_api[['thumb_red', 'thumb_green', 'thumb_blue']] = df_videos_api['thumbnail_url'].apply(lambda x: get_main_color_rgb(remove_black_bars(get_img_from_url(x))))"
      ],
      "metadata": {
        "id": "Tn-PoLR3HK3k"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_videos_api[['video_title','thumbnail_url','thumb_red','thumb_green','thumb_blue']].head()"
      ],
      "metadata": {
        "id": "A-I5ntGxHNUb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "f8eb6b4a-cd1c-420c-b6aa-5bb81f79c045"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                         video_title  \\\n",
              "0                  Fire conducts electricity #shorts   \n",
              "1  The world depends on a strange collection of i...   \n",
              "2                         But what is a convolution?   \n",
              "3  POSSO COLOCAR MINHA LINGUA NA SUA BOCA? : MAND...   \n",
              "4  🐍 GRIXIS CROBA! O novo MID-RANGE do STANDARD! ...   \n",
              "\n",
              "                                      thumbnail_url  thumb_red  thumb_green  \\\n",
              "0  https://i.ytimg.com/vi/QiVgJPVKRGI/hqdefault.jpg         22           19   \n",
              "1  https://i.ytimg.com/vi/esQyYGezS7c/hqdefault.jpg        125          113   \n",
              "2  https://i.ytimg.com/vi/KuXjwB4LzSA/hqdefault.jpg         35           30   \n",
              "3  https://i.ytimg.com/vi/55VfDUU1pyg/hqdefault.jpg        104           96   \n",
              "4  https://i.ytimg.com/vi/bo53RcrZmc4/hqdefault.jpg         57           96   \n",
              "\n",
              "   thumb_blue  \n",
              "0          14  \n",
              "1         110  \n",
              "2          29  \n",
              "3         100  \n",
              "4          96  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-29a022b1-024a-46fc-aa9a-cba131abeab8\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>video_title</th>\n",
              "      <th>thumbnail_url</th>\n",
              "      <th>thumb_red</th>\n",
              "      <th>thumb_green</th>\n",
              "      <th>thumb_blue</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Fire conducts electricity #shorts</td>\n",
              "      <td>https://i.ytimg.com/vi/QiVgJPVKRGI/hqdefault.jpg</td>\n",
              "      <td>22</td>\n",
              "      <td>19</td>\n",
              "      <td>14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>The world depends on a strange collection of i...</td>\n",
              "      <td>https://i.ytimg.com/vi/esQyYGezS7c/hqdefault.jpg</td>\n",
              "      <td>125</td>\n",
              "      <td>113</td>\n",
              "      <td>110</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>But what is a convolution?</td>\n",
              "      <td>https://i.ytimg.com/vi/KuXjwB4LzSA/hqdefault.jpg</td>\n",
              "      <td>35</td>\n",
              "      <td>30</td>\n",
              "      <td>29</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>POSSO COLOCAR MINHA LINGUA NA SUA BOCA? : MAND...</td>\n",
              "      <td>https://i.ytimg.com/vi/55VfDUU1pyg/hqdefault.jpg</td>\n",
              "      <td>104</td>\n",
              "      <td>96</td>\n",
              "      <td>100</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>🐍 GRIXIS CROBA! O novo MID-RANGE do STANDARD! ...</td>\n",
              "      <td>https://i.ytimg.com/vi/bo53RcrZmc4/hqdefault.jpg</td>\n",
              "      <td>57</td>\n",
              "      <td>96</td>\n",
              "      <td>96</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-29a022b1-024a-46fc-aa9a-cba131abeab8')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-29a022b1-024a-46fc-aa9a-cba131abeab8 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-29a022b1-024a-46fc-aa9a-cba131abeab8');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2.6) Get the number of words in the thumbnail (TO DO)"
      ],
      "metadata": {
        "id": "A2e27hp0TDch"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2.7) Get the number of human faces in the thumbnail (TO DO)"
      ],
      "metadata": {
        "id": "cqcWcKaeTH-p"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **3) LOAD**"
      ],
      "metadata": {
        "id": "kXWhg1pxGwnB"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 3.1) General methods to connect to DB"
      ],
      "metadata": {
        "id": "0nx5ZZRVBEQK"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def connect_to_db(host_name, database, port, username, password):\n",
        "    \"\"\" Connects to a database and returns the connection.\n",
        "    host_name = typically an IP from AWS\n",
        "    dbname = the type of the db (in my case, Postgres)\n",
        "    port = the port that gives access to the db\n",
        "    username = username of the db\n",
        "    password = password of the db\n",
        "\n",
        "    Returns a connection to a database.\n",
        "    \"\"\"\n",
        "    try:\n",
        "        conn = ps.connect(host=host_name, database=database, user=username, password=password, port=port)\n",
        "    except ps.OperationalError as e:\n",
        "        raise e\n",
        "    else:\n",
        "        print(\"Connected!\")\n",
        "    return conn"
      ],
      "metadata": {
        "id": "U8dilP7WalOj"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# create table\n",
        "def create_table(curs, command):\n",
        "    \"\"\" Creates a database table.\n",
        "    curs = cursor for the postgres table\n",
        "    \"\"\"\n",
        "    curs.execute(command)\n",
        "    print(\"Table created!\")       "
      ],
      "metadata": {
        "id": "g2bhLF3YloyL"
      },
      "execution_count": 28,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 3.2) Methods for \"videos\" table"
      ],
      "metadata": {
        "id": "FvLscAodLu4S"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def update_videos_row(curr, video_id, channel_id, video_title, thumbnail_url,\n",
        "                    upload_date_time, video_duration, view_count, like_count, \n",
        "                    comment_count, n_words_title, n_question_marks_title,\n",
        "                    n_exclamation_marks_title, n_ellipsis_title,\n",
        "                    thumb_red, thumb_green, thumb_blue):\n",
        "    \"\"\" Update a specific video_id entry of the db table\n",
        "    \"\"\"\n",
        "\n",
        "    \n",
        "    query = (\n",
        "        \"\"\"UPDATE videos\n",
        "        SET video_title = %s,\n",
        "            channel_id = %s,\n",
        "            thumbnail_url = %s,\n",
        "            upload_date_time = %s,\n",
        "            video_duration = %s,\n",
        "            view_count = %s,\n",
        "            like_count = %s,\n",
        "            comment_count = %s,\n",
        "            n_words_title = %s,\n",
        "            n_question_marks_title = %s,\n",
        "            n_exclamation_marks_title = %s,\n",
        "            n_ellipsis_title = %s,\n",
        "            thumb_red = %s,\n",
        "            thumb_green = %s,\n",
        "            thumb_blue = %s\n",
        "        WHERE video_id = %s;\"\"\")\n",
        "\n",
        "    vars_to_update = (video_title, channel_id, thumbnail_url, \\\n",
        "                      upload_date_time, video_duration, view_count, \\\n",
        "                      like_count, comment_count,\\\n",
        "                      n_words_title, n_question_marks_title,\\\n",
        "                      n_exclamation_marks_title, n_ellipsis_title,\\\n",
        "                    thumb_red, thumb_green, thumb_blue, video_id)\n",
        "    curr.execute(query, vars_to_update)"
      ],
      "metadata": {
        "id": "1N6VVzsqw-ue"
      },
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def update_videos_db(curr, df):\n",
        "    \"\"\" Update the videos table for all video_id, or if the entry doesn't exist\n",
        "    append it to a pandas df that will be returned.\n",
        "\n",
        "    Returns a pandas df with new videos.\n",
        "    \"\"\"\n",
        "\n",
        "    tmp_df = pd.DataFrame(columns=[\"video_id\",\"channel_id\",\"video_title\",\\\n",
        "            \"channel_title\",\"thumbnail_url\",\"upload_date_time\",\\\n",
        "            \"video_duration\", \"view_count\",\"like_count\",\"comment_count\", \\\n",
        "            \"n_words_title\", \"n_question_marks_title\", \\\n",
        "            \"n_exclamation_marks_title\", \"n_ellipsis_title\", \\\n",
        "            \"thumb_red\",\"thumb_green\", \"thumb_blue\"])\n",
        "\n",
        "    for i, row in df.iterrows():\n",
        "        if check_if_video_exists(curr, row['video_id']):\n",
        "            update_videos_row(curr, row['video_id'], row['channel_id'], \\\n",
        "                                row['video_title'], row['thumbnail_url'], row['upload_date_time'], \\\n",
        "                                row['video_duration'], row['view_count'], row['like_count'], \\\n",
        "                                row['comment_count'], row['n_words_title'], row['n_question_marks_title'], \\\n",
        "                                row['n_exclamation_marks_title'], row['n_ellipsis_title'], row['thumb_red'], \\\n",
        "                                row['thumb_green'], row['thumb_blue'])\n",
        "        else:\n",
        "            tmp_df = tmp_df.append(row)\n",
        "\n",
        "    return tmp_df"
      ],
      "metadata": {
        "id": "V9I9rnxyngbt"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def check_if_video_exists(curr, video_id):\n",
        "    \"\"\" Check if a given video_id exists in the database.\n",
        "\n",
        "    \"\"\"\n",
        "    query = (\"\"\"SELECT video_id FROM videos WHERE video_id = %s\"\"\")\n",
        "    curr.execute(query, (video_id,))\n",
        "\n",
        "    return curr.fetchone() is not None"
      ],
      "metadata": {
        "id": "_ZMT2NE2vmJ4"
      },
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def insert_into_videos_table(curr, video_id, channel_id, video_title, \\\n",
        "                            thumbnail_url, upload_date_time, video_duration, \\\n",
        "                            view_count, like_count, comment_count,\\\n",
        "                            n_words_title, n_question_marks_title, \\\n",
        "                            n_exclamation_marks_title, n_ellipsis_title, \\\n",
        "                            thumb_red, thumb_green, thumb_blue):\n",
        "    \n",
        "    \"\"\" Insert a new video entry into the database.\n",
        "\n",
        "    \"\"\"\n",
        "\n",
        "    insert_into_videos = (\n",
        "        \"\"\"INSERT INTO videos (video_id, channel_id, video_title, thumbnail_url, \n",
        "                upload_date_time, video_duration, view_count, like_count, \n",
        "                comment_count, n_words_title, n_question_marks_title,\n",
        "                n_exclamation_marks_title, n_ellipsis_title,\n",
        "                thumb_red, thumb_green, thumb_blue)\n",
        "            VALUES(%s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s);\"\"\")\n",
        "\n",
        "    row_to_insert = (video_id, channel_id, video_title, thumbnail_url, \\\n",
        "                    upload_date_time, video_duration, view_count, like_count, \\\n",
        "                    comment_count, n_words_title, n_question_marks_title, \\\n",
        "                    n_exclamation_marks_title, n_ellipsis_title, \\\n",
        "                    thumb_red, thumb_green, thumb_blue)\n",
        "\n",
        "    curr.execute(insert_into_videos, row_to_insert)"
      ],
      "metadata": {
        "id": "rebCc5DE1LaF"
      },
      "execution_count": 45,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def append_from_videos_df_to_db(curr, df):\n",
        "    \"\"\" Insert a list of videos from a pandas df in the database.\n",
        "\n",
        "    \"\"\"\n",
        "\n",
        "    for i, row in df.iterrows():\n",
        "        insert_into_videos_table(curr, row['video_id'], row['channel_id'], \\\n",
        "                                row['video_title'], row['thumbnail_url'], \\\n",
        "                                row['upload_date_time'], row['video_duration'],\\\n",
        "                                row['view_count'], row['like_count'],\\\n",
        "                                row['comment_count'], row['n_words_title'], \\\n",
        "                                row['n_question_marks_title'], row['n_exclamation_marks_title'], \\\n",
        "                                row['n_ellipsis_title'], row['thumb_red'], \\\n",
        "                                row['thumb_green'], row['thumb_blue'])"
      ],
      "metadata": {
        "id": "9LOWBmsb2JT2"
      },
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 3.3) Methods for \"channels\" table"
      ],
      "metadata": {
        "id": "Gr3IS3PfMGiy"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def update_channels_row(curr, channel_id, channel_title, channel_country, \n",
        "               total_view_count, subscriber_count, video_count, topic_categories):\n",
        "    \"\"\" Update a specific channel_id entry of the db table.\n",
        "    \"\"\"\n",
        "    query = (\n",
        "        \"\"\"UPDATE channels\n",
        "        SET channel_title = %s,\n",
        "            channel_country = %s,\n",
        "            total_view_count = %s,\n",
        "            subscriber_count = %s,\n",
        "            video_count = %s,\n",
        "            topic_categories = %s\n",
        "        WHERE channel_id = %s;\"\"\")\n",
        "\n",
        "    vars_to_update = (channel_title, channel_country, total_view_count, subscriber_count, video_count, topic_categories, channel_id)\n",
        "    curr.execute(query, vars_to_update)"
      ],
      "metadata": {
        "id": "2SwCaznvMI-A"
      },
      "execution_count": 47,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def update_channels_db(curr, df):\n",
        "    \"\"\" Update the channels table for all channel_id, or if the entry doesn't exist,\n",
        "    append it to a pandas df that will be returned.\n",
        "\n",
        "    Returns a pandas df with new channels.\n",
        "    \"\"\"\n",
        "\n",
        "    tmp_df = pd.DataFrame(columns=[\"channel_id\",\"channel_title\",\"channel_country\",\"total_view_count\",\"subscriber_count\",\"video_count\",\"topic_categories\"])\n",
        "\n",
        "    for i, row in df.iterrows():\n",
        "        if check_if_channel_exists(curr, row['channel_id']):\n",
        "            update_channels_row(curr, row['channel_id'], row['channel_title'], row['channel_country'], row['total_view_count'], row['subscriber_count'], row['video_count'], row['topic_categories'])\n",
        "        else:\n",
        "            tmp_df = tmp_df.append(row)\n",
        "\n",
        "    return tmp_df"
      ],
      "metadata": {
        "id": "CbnJi6PgM2Zw"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def check_if_channel_exists(curr, channel_id):\n",
        "    \"\"\" Check if a given channel_id exists in the database.\n",
        "\n",
        "    \"\"\"\n",
        "    query = (\"\"\"SELECT channel_id FROM channels WHERE channel_id = %s\"\"\")\n",
        "    curr.execute(query, (channel_id,))\n",
        "\n",
        "    return curr.fetchone() is not None"
      ],
      "metadata": {
        "id": "Q3X6rlxINs3T"
      },
      "execution_count": 49,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def insert_into_channels_table(curr, channel_id, channel_title, channel_country, view_count, subscriber_count, video_count, topic_categories):\n",
        "    \"\"\" Insert a new channel entry in the database.\n",
        "\n",
        "    \"\"\"\n",
        "\n",
        "    insert_into_channels = (\n",
        "        \"\"\"INSERT INTO channels (channel_id, channel_title, channel_country, total_view_count, subscriber_count, video_count, topic_categories)\n",
        "            VALUES(%s, %s, %s, %s, %s, %s, %s);\"\"\")\n",
        "\n",
        "    row_to_insert = (channel_id, channel_title, channel_country, view_count, subscriber_count, video_count, topic_categories)\n",
        "\n",
        "    curr.execute(insert_into_channels, row_to_insert)"
      ],
      "metadata": {
        "id": "BYvTC2AxN5i0"
      },
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def append_from_channels_df_to_db(curr, df):\n",
        "    \"\"\" Insert a list of channels from a pandas df in the database.\n",
        "\n",
        "    \"\"\"\n",
        "\n",
        "    for i, row in df.iterrows():\n",
        "        insert_into_channels_table(curr, row['channel_id'], row['channel_title'], row['channel_country'], row['total_view_count'], row['subscriber_count'], row['video_count'], row['topic_categories'])"
      ],
      "metadata": {
        "id": "mqAFPjViOgqj"
      },
      "execution_count": 51,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 3.4) Methods to add or remove a column from a table."
      ],
      "metadata": {
        "id": "sRbyJEl__et6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def add_column_to_table(curs, table_name, column_name, column_type, default=\"\"):\n",
        "    \"\"\" Add column to a table if it doesn't exist.\n",
        "    curs = cursor for the postgres table\n",
        "    table_name = the name of the table\n",
        "    column_name = name of column\n",
        "    column_type = type of column\n",
        "    default = default value that the column will take\n",
        "    \"\"\"\n",
        "\n",
        "    if default!=\"\":\n",
        "        default=f\"DEFAULT {default}\"\n",
        "\n",
        "    add_column_command = (\n",
        "        f\"\"\" ALTER TABLE IF EXISTS {table_name} \n",
        "            ADD IF NOT EXISTS {column_name} {column_type} {default}\n",
        "        \"\"\"\n",
        "    )\n",
        "\n",
        "    curs.execute(add_column_command)\n",
        "    print(\"Column added!\")\n"
      ],
      "metadata": {
        "id": "PiOOMgdBWkAZ"
      },
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def remove_column_from_table(curs, table_name, column_name):\n",
        "    \"\"\" Remove a column from a table if it exists.\n",
        "    curs = cursor for the postgres table\n",
        "    table_name = the name of the table\n",
        "    column_name = name of column\n",
        "    \"\"\"\n",
        "\n",
        "    remove_column_command = (\n",
        "        f\"\"\" ALTER TABLE IF EXISTS {table_name} \n",
        "            DROP COLUMN IF EXISTS {column_name}\n",
        "        \"\"\"\n",
        "    )\n",
        "\n",
        "    curs.execute(remove_column_command)\n",
        "    print(\"Column removed!\")"
      ],
      "metadata": {
        "id": "4S4q2awX4mJ-"
      },
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 3.5) Connect to AWS RDS and load the data."
      ],
      "metadata": {
        "id": "kCzIwteDn8eH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "host_name = \"datascience-youtube.cuxnfdexw55l.us-east-1.rds.amazonaws.com\"\n",
        "database = \"postgres\"\n",
        "port = \"5432\"\n",
        "db_username = os.getenv(\"db_username\")\n",
        "db_password = os.getenv(\"db_password\")\n",
        "conn = None\n",
        "\n",
        "conn = connect_to_db(host_name, database, port, db_username, db_password)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EiYp8AB0Z0VG",
        "outputId": "71ded563-582a-48de-87cf-eeb19f7a4da4"
      },
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Connected!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "curs = conn.cursor() # allows us to run SQL commands and fetch results"
      ],
      "metadata": {
        "id": "K1wpuMZblmib"
      },
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "videos_create_command = (\n",
        "        \"\"\"CREATE TABLE IF NOT EXISTS videos (\n",
        "            video_id VARCHAR(255) PRIMARY KEY,\n",
        "            channel_id VARCHAR(255) NOT NULL,\n",
        "            video_title TEXT NOT NULL,\n",
        "            thumbnail_url TEXT NOT NULL,\n",
        "            upload_date_time TIMESTAMPTZ NOT NULL DEFAULT CURRENT_TIMESTAMP,\n",
        "            video_duration INTEGER NOT NULL,\n",
        "            view_count INTEGER NOT NULL,\n",
        "            like_count INTEGER NOT NULL,\n",
        "            comment_count INTEGER NOT NULL,\n",
        "            n_words_title INTEGER NOT NULL,\n",
        "            n_question_marks_title INTEGER NOT NULL,\n",
        "            n_exclamation_marks_title INTEGER NOT NULL,\n",
        "            n_ellipsis_title INTEGER NOT NULL,\n",
        "            thumb_red INTEGER NOT NULL,\n",
        "            thumb_green INTEGER NOT NULL,\n",
        "            thumb_blue INTEGER NOT NULL\n",
        "        )\"\"\")\n",
        "\n",
        "create_table(curs, videos_create_command)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "d7hONOhJ46f7",
        "outputId": "d1521303-5371-45a3-a0c3-fbdaf5893357"
      },
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Table created!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "channels_create_command = (\n",
        "    \"\"\"CREATE TABLE IF NOT EXISTS channels (\n",
        "            channel_id VARCHAR(255) PRIMARY KEY,\n",
        "            channel_title VARCHAR(255) NOT NULL,\n",
        "            channel_country VARCHAR(255) NOT NULL,\n",
        "            total_view_count INTEGER NOT NULL,\n",
        "            subscriber_count INTEGER NOT NULL,\n",
        "            video_count INTEGER NOT NULL,\n",
        "            topic_categorieS TEXT\n",
        "        )\"\"\")\n",
        "\n",
        "create_table(curs, channels_create_command)"
      ],
      "metadata": {
        "id": "kbUO9V_OmdOC",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cf464a54-56b1-4983-d9e4-033335dd9230"
      },
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Table created!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# add_column_to_table(curs,\"videos\", \"upload_hour\", \"VARCHAR(10) NOT NULL\",0)\n",
        "# remove_column_from_table(curs,\"videos\",\"upload_hour\")\n",
        "# add_column_to_table(curs,\"channels\",\"total_view_count\",\"INTEGER NOT NULL\",\"0\")\n",
        "# remove_column_from_table(curs,\"videos\",\"video_duration\")"
      ],
      "metadata": {
        "id": "ITp220iQKdY9"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "new_videos_df = update_videos_db(curs, df_videos_api)"
      ],
      "metadata": {
        "id": "pu3o-2v-LT9k"
      },
      "execution_count": 59,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "new_channels_df = update_channels_db(curs, df_channels_api)"
      ],
      "metadata": {
        "id": "Z4Rnb-emOyDF"
      },
      "execution_count": 60,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "append_from_videos_df_to_db(curs, new_videos_df)"
      ],
      "metadata": {
        "id": "qfNfk3Uv2sCw"
      },
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "append_from_channels_df_to_db(curs, new_channels_df)"
      ],
      "metadata": {
        "id": "927RIe_aO6Nc"
      },
      "execution_count": 62,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "conn.commit()"
      ],
      "metadata": {
        "id": "pTCgro4GmjPG"
      },
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#**Conclusion**\n",
        "\n",
        "We have succesfuly used the Youtube API, extracted the data that we wanted (and we can always revisit this if we want to analyze other things or create more complex prediction models), transformed it (by applying regular expressions, changing some units and checking the main color of the thumbnail) and finished it up by loading the data in an AWS RDS running Postgres. \n",
        "\n",
        "In the next section we will analyze all this data!\n",
        "\n",
        "Let's go:  https://colab.research.google.com/drive/1a_Etj5kwEaq5epwoV9TVPNS3ShRRd7wU?usp=sharing"
      ],
      "metadata": {
        "id": "TDIIxbv-EsYO"
      }
    }
  ]
}