03_Growth_Returns.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# SimFin Tutorial 03 - Growth & Returns\n",
    "\n",
    "[Original repository on GitHub](https://github.com/simfin/simfin-tutorials)\n",
    "\n",
    "This tutorial was originally written by [Hvass Labs](https://github.com/Hvass-Labs)\n",
    "\n",
    "----\n",
    "\n",
    "\"We shall fight the oppressors for your right to have babies, brother ... sister, sorry. What's the point of fighting for his right to have babies, when he can't have babies?! It is symbolic of our struggle against oppression! No, it's symbolic of his struggle against reality ...\" &ndash; [Monty Python's Life of Brian](https://www.youtube.com/watch?v=PObBA2wH5l0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Introduction\n",
    "\n",
    "Another common operation on financial data is to calculate changes over time, such as stock-returns and sales-growth. This tutorial first discusses Pandas' built-in functionality for doing this, but it does not seem to work so well for our data and purposes. We then introduce SimFin's helper-function `sf.rel_change` and demonstrates its features for calculating various kinds of stock-returns and growth-rates.\n",
    "\n",
    "If you know a better way of calculating this using Pandas' built-in functionality, then we are very interested in hearing it.\n",
    "\n",
    "It is assumed you are already familiar with Tutorial 01 on the basics of SimFin before you read this tutorial."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "# Import the main functionality from the SimFin Python API.\n",
    "import simfin as sf\n",
    "\n",
    "# Import names used for easy access to SimFin's data-columns.\n",
    "from simfin.names import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'0.3.0'"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Version of the SimFin Python API.\n",
    "sf.__version__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SimFin data-directory.\n",
    "sf.set_data_dir('~/simfin_data/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SimFin load API key or use free data.\n",
    "sf.load_api_key(path='~/simfin_api_key.txt', default_key='free')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Seaborn set plotting style.\n",
    "sns.set_style(\"whitegrid\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load Datasets\n",
    "\n",
    "In these examples, we will use the datasets for annual and quarterly Income Statements, as well as daily Share-Prices, all for USA."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset \"us-income-annual\" on disk (9 days old).\n",
      "- Loading from disk ... Done!\n",
      "Dataset \"us-income-quarterly\" on disk (28 days old).\n",
      "- Loading from disk ... Done!\n",
      "Dataset \"us-shareprices-daily\" on disk (8 days old).\n",
      "- Loading from disk ... Done!\n",
      "CPU times: user 16 s, sys: 1.16 s, total: 17.1 s\n",
      "Wall time: 15.7 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Data for USA.\n",
    "market = 'us'\n",
    "\n",
    "# Annual Income Statements.\n",
    "df_income_ann = sf.load_income(variant='annual', market=market)\n",
    "\n",
    "# Quarterly Income Statements.\n",
    "df_income_qrt = sf.load_income(variant='quarterly', market=market)\n",
    "\n",
    "# Daily Share-Prices.\n",
    "df_prices = sf.load_shareprices(variant='daily', market=market)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Select a subset of the tickers and data-columns we are interested in:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "tickers = ['AAPL', 'AMZN', 'MSFT']\n",
    "df_income_ann = df_income_ann.loc[tickers, [REVENUE, NET_INCOME]].copy()\n",
    "df_income_qrt = df_income_qrt.loc[tickers, [REVENUE, NET_INCOME]].copy()\n",
    "df_prices = df_prices.loc[tickers, [CLOSE, ADJ_CLOSE]].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ticker</th>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">AAPL</td>\n",
       "      <td>2008-09-30</td>\n",
       "      <td>3.749100e+10</td>\n",
       "      <td>6119000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-09-30</td>\n",
       "      <td>4.290500e+10</td>\n",
       "      <td>8235000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-09-30</td>\n",
       "      <td>6.522500e+10</td>\n",
       "      <td>14013000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2011-09-30</td>\n",
       "      <td>1.082490e+11</td>\n",
       "      <td>25922000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2012-09-30</td>\n",
       "      <td>1.565080e+11</td>\n",
       "      <td>41733000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         Revenue   Net Income\n",
       "Ticker Report Date                           \n",
       "AAPL   2008-09-30   3.749100e+10   6119000000\n",
       "       2009-09-30   4.290500e+10   8235000000\n",
       "       2010-09-30   6.522500e+10  14013000000\n",
       "       2011-09-30   1.082490e+11  25922000000\n",
       "       2012-09-30   1.565080e+11  41733000000"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_income_ann.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ticker</th>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">AAPL</td>\n",
       "      <td>2009-06-30</td>\n",
       "      <td>9.734000e+09</td>\n",
       "      <td>1828000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-09-30</td>\n",
       "      <td>1.220700e+10</td>\n",
       "      <td>2532000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>1.568300e+10</td>\n",
       "      <td>3378000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-03-31</td>\n",
       "      <td>1.349900e+10</td>\n",
       "      <td>3074000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-06-30</td>\n",
       "      <td>1.570000e+10</td>\n",
       "      <td>3253000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         Revenue  Net Income\n",
       "Ticker Report Date                          \n",
       "AAPL   2009-06-30   9.734000e+09  1828000000\n",
       "       2009-09-30   1.220700e+10  2532000000\n",
       "       2009-12-31   1.568300e+10  3378000000\n",
       "       2010-03-31   1.349900e+10  3074000000\n",
       "       2010-06-30   1.570000e+10  3253000000"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_income_qrt.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ticker</th>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">AAPL</td>\n",
       "      <td>2007-01-03</td>\n",
       "      <td>11.9714</td>\n",
       "      <td>10.42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-04</td>\n",
       "      <td>12.2371</td>\n",
       "      <td>10.65</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-05</td>\n",
       "      <td>12.1500</td>\n",
       "      <td>10.57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-08</td>\n",
       "      <td>12.2100</td>\n",
       "      <td>10.62</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-09</td>\n",
       "      <td>13.2243</td>\n",
       "      <td>11.51</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Close  Adj. Close\n",
       "Ticker Date                           \n",
       "AAPL   2007-01-03  11.9714       10.42\n",
       "       2007-01-04  12.2371       10.65\n",
       "       2007-01-05  12.1500       10.57\n",
       "       2007-01-08  12.2100       10.62\n",
       "       2007-01-09  13.2243       11.51"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_prices.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pandas Functions\n",
    "\n",
    "Pandas has a lot of functionality for indexing DataFrames by time and date, shifting a DataFrame backwards or forwards in time, and calculating percentage changes over time. Unfortunately, these aspects of Pandas are poorly documented, so it is unclear exactly how they are supposed to work and what their short-comings are. So we will have to do some detective-work to try and figure out how this works in Pandas, and whether we can use it for our purposes.\n",
    "\n",
    "Consider the first rows of the share-prices for ticker MSFT:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2007-01-03</td>\n",
       "      <td>29.86</td>\n",
       "      <td>22.25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-04</td>\n",
       "      <td>29.81</td>\n",
       "      <td>22.21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-05</td>\n",
       "      <td>29.64</td>\n",
       "      <td>22.08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-08</td>\n",
       "      <td>29.93</td>\n",
       "      <td>22.30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-09</td>\n",
       "      <td>29.96</td>\n",
       "      <td>22.32</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Close  Adj. Close\n",
       "Date                         \n",
       "2007-01-03  29.86       22.25\n",
       "2007-01-04  29.81       22.21\n",
       "2007-01-05  29.64       22.08\n",
       "2007-01-08  29.93       22.30\n",
       "2007-01-09  29.96       22.32"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df_prices.loc['MSFT']['2007-01-03':]\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The Pandas `shift` function moves the data forwards or backwards. When using the argument `periods` the data is moved in whole steps regardless of the index-dates in the DataFrame. For example, this moves the data one step forward:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2007-01-03</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-04</td>\n",
       "      <td>29.86</td>\n",
       "      <td>22.25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-05</td>\n",
       "      <td>29.81</td>\n",
       "      <td>22.21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-08</td>\n",
       "      <td>29.64</td>\n",
       "      <td>22.08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-09</td>\n",
       "      <td>29.93</td>\n",
       "      <td>22.30</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Close  Adj. Close\n",
       "Date                         \n",
       "2007-01-03    NaN         NaN\n",
       "2007-01-04  29.86       22.25\n",
       "2007-01-05  29.81       22.21\n",
       "2007-01-08  29.64       22.08\n",
       "2007-01-09  29.93       22.30"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shift(periods=1).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When we perform arithmetic operations with the original and shifted DataFrames, they are first aligned by the index-dates. This allows us to calculate relative changes over time.\n",
    "\n",
    "Let us explain what happens if we calculate `df/df.shift(periods=1)`. For the date 2007-01-04 the original Close value was 29.81, and for the preceding date 2007-01-03 the original Close value was 29.86. When we shift the DataFrame one step forward, the Close value from 2007-01-03 is moved to the following day 2007-01-04. We then divide the original Close value on 2007-01-04 which was 29.81, by the shifted Close value 29.86 that was originally at the preceding day 2007-01-03 but is now located at 2007-01-04 in the shifted DataFrame, so the result is a value of 29.81 / 29.86 = 0.998326 on 2007-01-04.\n",
    "\n",
    "This is perhaps a bit abstract and can be difficult to understand, so it is recommended that you play around with the `shift` operator, to understand how it moves the data forwards and backwards.\n",
    "\n",
    "Note that the first row in the DataFrame is NaN (short for \"Not-a-Number\"), because the arithemetic operation is trying to use data from the preceding time-step which does not exist in the DataFrame, so the result of the arithmetic operation for that row is NaN."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2007-01-03</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-04</td>\n",
       "      <td>0.998326</td>\n",
       "      <td>0.998202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-05</td>\n",
       "      <td>0.994297</td>\n",
       "      <td>0.994147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-08</td>\n",
       "      <td>1.009784</td>\n",
       "      <td>1.009964</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-09</td>\n",
       "      <td>1.001002</td>\n",
       "      <td>1.000897</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2007-01-03       NaN         NaN\n",
       "2007-01-04  0.998326    0.998202\n",
       "2007-01-05  0.994297    0.994147\n",
       "2007-01-08  1.009784    1.009964\n",
       "2007-01-09  1.001002    1.000897"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(df / df.shift(1)).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Pandas has a built-in function for calculating the relative change in a time-series, so we don't have to shift the DataFrame ourselves. The function's name is `pct_change` which is somewhat misleading, because it actually calculates the relative change instead of the percentage change. It calculates the relative change from the previous time-step $t-periods$ until the current time-step $t$ using this formula:\n",
    "\n",
    "$$\n",
    "    pct\\_change_t = \\frac{data_t}{data_{t-periods}} - 1\n",
    "$$\n",
    "\n",
    "For example, we can calculate the relative change of the share-prices:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2007-01-03</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-04</td>\n",
       "      <td>-0.001674</td>\n",
       "      <td>-0.001798</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-05</td>\n",
       "      <td>-0.005703</td>\n",
       "      <td>-0.005853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-08</td>\n",
       "      <td>0.009784</td>\n",
       "      <td>0.009964</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-09</td>\n",
       "      <td>0.001002</td>\n",
       "      <td>0.000897</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2007-01-03       NaN         NaN\n",
       "2007-01-04 -0.001674   -0.001798\n",
       "2007-01-05 -0.005703   -0.005853\n",
       "2007-01-08  0.009784    0.009964\n",
       "2007-01-09  0.001002    0.000897"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.pct_change(periods=1).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can check that the result of the `pct_change` function and our calculation using the manually shifted DataFrame, is exactly the same if we also subtract 1:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2007-01-03</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-04</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-05</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-08</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-09</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Close  Adj. Close\n",
       "Date                         \n",
       "2007-01-03    NaN         NaN\n",
       "2007-01-04    0.0         0.0\n",
       "2007-01-05    0.0         0.0\n",
       "2007-01-08    0.0         0.0\n",
       "2007-01-09    0.0         0.0"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Relative change using manually shifted data.\n",
    "df1 = df / df.shift(periods=1) - 1\n",
    "\n",
    "# Relative change using Pandas' pct_change function.\n",
    "df2 = df.pct_change(periods=1)\n",
    "\n",
    "# Difference.\n",
    "(df1 - df2).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What if we want to shift the share-prices one year to calculate the 1-year returns? Most years have 365 days, except for every 4th year which is a leap-year that has 366 days. So on average a year has 365.25 days. We cannot call `df.shift(periods=365.25)` because the number of periods must be an integer.\n",
    "\n",
    "Furthermore, there aren't 365 trading- or business-days in a normal year, because the stock-markets are not open during weekends and holidays.\n",
    "\n",
    "Using the share-prices for ticker MSFT again, we can count the average number of trading-days in the 12 years between 2007 and 2018 (both inclusive), which is about 251.67 trading-days per year."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "251.66666666666666"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df['2007':'2018'])/12"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can also plot a histogram of the number of trading-days per year between 2007 and 2018 (both inclusive), where we can see that about half of the years had 252 trading days:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f5656359320>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df2 = df['2007':'2018']\n",
    "df2.groupby(df2.index.year).apply(lambda x: len(x)).hist()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we use the average of 251.67 trading-days per year, then in order to shift the DataFrame with share-prices one year, we need to shift it 251.67 periods or steps. In practice we round this up to the nearest integer 252. But if we want to calculate the number of periods for multiple years, then it is best to do the rounding to the nearest integer at the end:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "755"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "int(round(251.67 * 3))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If instead we had used the rounded number 252 we would get a small error, which grows bigger for more years:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "756"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "252 * 3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "But even so, we can still get a small drift in the dates, because 251.67 is an average number and some years will have a different number of trading-days per year.\n",
    "\n",
    "For example, if we calculate the 5-year changes, we can see that it has drifted 3 time-steps. A few days of drift for a 5-year period is typically deemed acceptable when analyzing long-term stock returns, as it will most likely have a negligible impact on the statistics."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2011-12-29</td>\n",
       "      <td>-0.128600</td>\n",
       "      <td>-0.035506</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2011-12-30</td>\n",
       "      <td>-0.129151</td>\n",
       "      <td>-0.036020</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2012-01-03</td>\n",
       "      <td>-0.096997</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2012-01-04</td>\n",
       "      <td>-0.084531</td>\n",
       "      <td>0.013453</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2012-01-05</td>\n",
       "      <td>-0.076101</td>\n",
       "      <td>0.022849</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2011-12-29 -0.128600   -0.035506\n",
       "2011-12-30 -0.129151   -0.036020\n",
       "2012-01-03 -0.096997    0.000000\n",
       "2012-01-04 -0.084531    0.013453\n",
       "2012-01-05 -0.076101    0.022849"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "years = 5\n",
    "periods = int(round(251.67 * years))\n",
    "df.pct_change(periods=periods).dropna().head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Pandas DateOffset\n",
    "\n",
    "Pandas also provides more sophisticated ways of shifting DataFrames. One way is to use `pd.DateOffset`. Unfortunately, Pandas' [documentation](https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.tseries.offsets.DateOffset.html) is brief and unclear, so we have to partially guess how it is supposed to work. If we try and calculate 1-year returns on the DataFrame with share-prices for ticker MSFT, then we get an exception:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cannot reindex from a duplicate axis\n"
     ]
    }
   ],
   "source": [
    "try:\n",
    "    df.pct_change(freq=pd.DateOffset(years=1))\n",
    "except Exception as e:\n",
    "    print(e)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Pandas also allows a [string alias](https://pandas.pydata.org/pandas-docs/stable/user_guide/timeseries.html#timeseries-offset-aliases) as the date-offset. For example, we can try 'BY' for Business Year, but we still get an exception:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cannot reindex from a duplicate axis\n"
     ]
    }
   ],
   "source": [
    "try:\n",
    "    df.pct_change(freq='BY')\n",
    "except Exception as e:\n",
    "    print(e)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The problem in both cases, seems to be that leap-years cause duplicate entries in the index, because 29th February in a leap-year is shifted to 28th February in the following year, but there is already another row with that date:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2009-02-28</td>\n",
       "      <td>27.9300</td>\n",
       "      <td>21.17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-02-28</td>\n",
       "      <td>27.1999</td>\n",
       "      <td>20.62</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-02-28</td>\n",
       "      <td>31.8700</td>\n",
       "      <td>26.46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-02-28</td>\n",
       "      <td>31.7400</td>\n",
       "      <td>26.35</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Close  Adj. Close\n",
       "Date                           \n",
       "2009-02-28  27.9300       21.17\n",
       "2009-02-28  27.1999       20.62\n",
       "2013-02-28  31.8700       26.46\n",
       "2013-02-28  31.7400       26.35"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Shift the DataFrame.\n",
    "df_shifted = df.shift(freq=pd.DateOffset(years=1)).dropna()\n",
    "\n",
    "# Boolean mask for rows with duplicate index-values.\n",
    "mask = df_shifted.index.duplicated(keep=False)\n",
    "\n",
    "# Show the duplicate rows.\n",
    "df_shifted.loc[mask]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Another option is to specify a 365-day period instead of a 1-year period:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Relative change over 365-days.\n",
    "df_change = df.pct_change(freq=pd.DateOffset(days=365))\n",
    "\n",
    "# Remove leading NaN.\n",
    "start_idx = df_change.first_valid_index()\n",
    "df_change = df_change.loc[start_idx:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This does not raise an exception, but it has introduced a lot of NaN values in the data, for example around the problematic leap-year dates in 2009 and 2013:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2009-02-26</td>\n",
       "      <td>-0.418967</td>\n",
       "      <td>-0.406162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-02-27</td>\n",
       "      <td>-0.421769</td>\n",
       "      <td>-0.409069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-03-02</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-03-03</td>\n",
       "      <td>-0.411634</td>\n",
       "      <td>-0.398827</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-03-04</td>\n",
       "      <td>-0.415730</td>\n",
       "      <td>-0.402678</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2009-02-26 -0.418967   -0.406162\n",
       "2009-02-27 -0.421769   -0.409069\n",
       "2009-03-02       NaN         NaN\n",
       "2009-03-03 -0.411634   -0.398827\n",
       "2009-03-04 -0.415730   -0.402678"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_change['2009-02-26':'2009-03-04']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2013-02-28</td>\n",
       "      <td>-0.124134</td>\n",
       "      <td>-0.097533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-03-01</td>\n",
       "      <td>-0.134407</td>\n",
       "      <td>-0.108169</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-03-04</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-03-05</td>\n",
       "      <td>-0.108491</td>\n",
       "      <td>-0.081439</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-03-06</td>\n",
       "      <td>-0.109808</td>\n",
       "      <td>-0.082824</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2013-02-28 -0.124134   -0.097533\n",
       "2013-03-01 -0.134407   -0.108169\n",
       "2013-03-04       NaN         NaN\n",
       "2013-03-05 -0.108491   -0.081439\n",
       "2013-03-06 -0.109808   -0.082824"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_change['2013-02-28':'2013-03-06']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This has created hundreds of rows with NaN values, so this clearly did not work either:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2008-01-07</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2008-01-14</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2008-01-15</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2008-01-28</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2008-02-04</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-11-22</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-11-25</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-02</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-05</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-09</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>635 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Close  Adj. Close\n",
       "Date                         \n",
       "2008-01-07    NaN         NaN\n",
       "2008-01-14    NaN         NaN\n",
       "2008-01-15    NaN         NaN\n",
       "2008-01-28    NaN         NaN\n",
       "2008-02-04    NaN         NaN\n",
       "...           ...         ...\n",
       "2019-11-22    NaN         NaN\n",
       "2019-11-25    NaN         NaN\n",
       "2019-12-02    NaN         NaN\n",
       "2019-12-05    NaN         NaN\n",
       "2019-12-09    NaN         NaN\n",
       "\n",
       "[635 rows x 2 columns]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mask = df_change.isnull().any(axis=1)\n",
    "df_change[mask]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Pandas has several other ways of representing and manipulating dates, but none of them seem to work for our purposes. This is strange because Pandas was originally made for financial time-series. So I wonder if the Pandas developers never use investment periods of a year or more - or perhaps I have failed to understand how Pandas is supposed to do this. This is not easy to find out, because Pandas' documentation is brief and incomplete."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SimFin Helper-Function\n",
    "\n",
    "SimFin provides a helper-function named `sf.rel_change` that is specialized for calculating changes over time for this kind of financial data.\n",
    "\n",
    "The function assumes that the DataFrame is ordered ascendingly by the index-dates, and that the DataFrame is complete in the sense that data is present for all time-steps of the given frequency. The function does not check this, but the SimFin database ensures that the quarterly and annual financial data such as Income Statements and Balance Sheets are all complete without any missing reports. If you are using other data-sources, then you must ensure that yourself before using this function.\n",
    "\n",
    "Furthermore, the function supports both Pandas DataFrames and Series, as well as `DatetimeIndex` or `MultiIndex` with 2 indices, one of which is a `DatetimeIndex` and the other is e.g. TICKER or SIMFIN_ID.\n",
    "\n",
    "The function has a number of useful features, and all of this makes it easy to use on a wide variety of financial data, as we will demonstrate now."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Stock Returns\n",
    "\n",
    "Let us start with a basic example of calculating the daily changes in stock-prices for ticker MSFT. We set the argument `freq='bdays'` to indicate the DataFrame contains data for all business-days (roughly 252 data-points per year, as explained above). Pandas has functionality to detect the frequency of a DataFrame, but it doesn't seem to work properly, so we need to specify it manually.\n",
    "\n",
    "We then set the arguments for the time-interval we want, in this case we just set `bdays=1` which means we want the relative change between successive time-steps, because the DataFrame has the same frequency.\n",
    "\n",
    "Finally we set the arg `future=False` to indicate that we want the **past** change between the previous day $t-1$ and the current day $t$ so the function calculates `df_result[t] = df[t]/df[t-1] - 1`.  Note that this calculates the exact same as `df.pct_change(periods=1)` shown further above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2007-01-03</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-04</td>\n",
       "      <td>-0.001674</td>\n",
       "      <td>-0.001798</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-05</td>\n",
       "      <td>-0.005703</td>\n",
       "      <td>-0.005853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-08</td>\n",
       "      <td>0.009784</td>\n",
       "      <td>0.009964</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-09</td>\n",
       "      <td>0.001002</td>\n",
       "      <td>0.000897</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2007-01-03       NaN         NaN\n",
       "2007-01-04 -0.001674   -0.001798\n",
       "2007-01-05 -0.005703   -0.005853\n",
       "2007-01-08  0.009784    0.009964\n",
       "2007-01-09  0.001002    0.000897"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_prices.loc['MSFT'], freq='bdays',\n",
    "              bdays=1, future=False).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let us now try and change a single argument by setting `future=True` instead, to indicate that we want the change between the **future** time-step $t+1$ and the current time-step $t$, so the function calculates `df_result[t] = df[t+1]/df[t] - 1`. As you can see, this is effectively the same DataFrame as above, that has just been shifted one step backwards - and that is indeed how the helper-function implements it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2007-01-03</td>\n",
       "      <td>-0.001674</td>\n",
       "      <td>-0.001798</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-04</td>\n",
       "      <td>-0.005703</td>\n",
       "      <td>-0.005853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-05</td>\n",
       "      <td>0.009784</td>\n",
       "      <td>0.009964</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-08</td>\n",
       "      <td>0.001002</td>\n",
       "      <td>0.000897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-09</td>\n",
       "      <td>-0.010013</td>\n",
       "      <td>-0.009857</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2007-01-03 -0.001674   -0.001798\n",
       "2007-01-04 -0.005703   -0.005853\n",
       "2007-01-05  0.009784    0.009964\n",
       "2007-01-08  0.001002    0.000897\n",
       "2007-01-09 -0.010013   -0.009857"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_prices.loc['MSFT'], freq='bdays',\n",
    "              bdays=1, future=True).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can also use an arbitrary time-interval that is a multiple of the DataFrame's frequency. A simple example is to get the future change in stock-prices for 1-year intervals, so we effectively calculate `df_result[t] = df[t+1 year]/df[t] - 1`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2007-01-03</td>\n",
       "      <td>0.184528</td>\n",
       "      <td>0.200449</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-04</td>\n",
       "      <td>0.153304</td>\n",
       "      <td>0.168843</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-05</td>\n",
       "      <td>0.167679</td>\n",
       "      <td>0.183424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-08</td>\n",
       "      <td>0.117608</td>\n",
       "      <td>0.132735</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-09</td>\n",
       "      <td>0.149533</td>\n",
       "      <td>0.165323</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2007-01-03  0.184528    0.200449\n",
       "2007-01-04  0.153304    0.168843\n",
       "2007-01-05  0.167679    0.183424\n",
       "2007-01-08  0.117608    0.132735\n",
       "2007-01-09  0.149533    0.165323"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_prices.loc['MSFT'], freq='bdays',\n",
    "              years=1, future=True).dropna().head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can also combine several arguments for the time-interval. For example, let us say we want 3 business-days plus 1 year, so we effectively calculate `df_result[t] = df[t+1 year+3 bdays]/df[t] - 1`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2007-01-03</td>\n",
       "      <td>0.120228</td>\n",
       "      <td>0.135281</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-04</td>\n",
       "      <td>0.155317</td>\n",
       "      <td>0.171094</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-05</td>\n",
       "      <td>0.158232</td>\n",
       "      <td>0.173913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-08</td>\n",
       "      <td>0.132977</td>\n",
       "      <td>0.148430</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2007-01-09</td>\n",
       "      <td>0.147864</td>\n",
       "      <td>0.163530</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2007-01-03  0.120228    0.135281\n",
       "2007-01-04  0.155317    0.171094\n",
       "2007-01-05  0.158232    0.173913\n",
       "2007-01-08  0.132977    0.148430\n",
       "2007-01-09  0.147864    0.163530"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_prices.loc['MSFT'], freq='bdays',\n",
    "              bdays=3, years=1, future=True).dropna().head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For periods of several years, it is often useful to calculate the annualized returns. For example, if we want the past annualized returns for 3-year investment periods, we are effectively calculating `df_result[t] = (df[t]/df[t-3 years]) ** (1/3) - 1`. Note that there is a small drift of a few days, but that is deemed acceptable for periods of 3 years, as we discussed further above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>0.006874</td>\n",
       "      <td>0.025416</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-04</td>\n",
       "      <td>0.012588</td>\n",
       "      <td>0.031280</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-05</td>\n",
       "      <td>0.014630</td>\n",
       "      <td>0.033441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-06</td>\n",
       "      <td>0.009269</td>\n",
       "      <td>0.027914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-07</td>\n",
       "      <td>0.005444</td>\n",
       "      <td>0.024058</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2009-12-31  0.006874    0.025416\n",
       "2010-01-04  0.012588    0.031280\n",
       "2010-01-05  0.014630    0.033441\n",
       "2010-01-06  0.009269    0.027914\n",
       "2010-01-07  0.005444    0.024058"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_prices.loc['MSFT'], freq='bdays',\n",
    "              years=3, future=False, annualized=True).dropna().head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once again, we can combine different arguments for the time-interval, and the function automatically calculates the correct annualized return for that time-interval, but the formula is too long and complicated to write here, so we just demonstrate the function:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2010-04-27</td>\n",
       "      <td>0.009842</td>\n",
       "      <td>0.028099</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-04-28</td>\n",
       "      <td>0.010995</td>\n",
       "      <td>0.029293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-04-29</td>\n",
       "      <td>0.013659</td>\n",
       "      <td>0.032135</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-04-30</td>\n",
       "      <td>0.006057</td>\n",
       "      <td>0.024341</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-05-03</td>\n",
       "      <td>0.008971</td>\n",
       "      <td>0.027252</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close  Adj. Close\n",
       "Date                            \n",
       "2010-04-27  0.009842    0.028099\n",
       "2010-04-28  0.010995    0.029293\n",
       "2010-04-29  0.013659    0.032135\n",
       "2010-04-30  0.006057    0.024341\n",
       "2010-05-03  0.008971    0.027252"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_prices.loc['MSFT'], freq='bdays',\n",
    "              bdays=2, weeks=3, months=3, years=3,\n",
    "              future=False, annualized=True).dropna().head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the examples above we only used a DataFrame with a single stock, but all the demonstrated features also work seamlessly for DataFrames with multiple stocks:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj. Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ticker</th>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">AAPL</td>\n",
       "      <td>2010-03-11</td>\n",
       "      <td>0.364285</td>\n",
       "      <td>0.364139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-03-12</td>\n",
       "      <td>0.356989</td>\n",
       "      <td>0.356793</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-03-15</td>\n",
       "      <td>0.354814</td>\n",
       "      <td>0.354833</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-03-16</td>\n",
       "      <td>0.353878</td>\n",
       "      <td>0.354048</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-03-17</td>\n",
       "      <td>0.319785</td>\n",
       "      <td>0.319687</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">MSFT</td>\n",
       "      <td>2019-12-05</td>\n",
       "      <td>0.346979</td>\n",
       "      <td>0.372822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-06</td>\n",
       "      <td>0.356728</td>\n",
       "      <td>0.382734</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-09</td>\n",
       "      <td>0.354154</td>\n",
       "      <td>0.380097</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-10</td>\n",
       "      <td>0.354838</td>\n",
       "      <td>0.380798</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-11</td>\n",
       "      <td>0.357777</td>\n",
       "      <td>0.383798</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>7371 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                      Close  Adj. Close\n",
       "Ticker Date                            \n",
       "AAPL   2010-03-11  0.364285    0.364139\n",
       "       2010-03-12  0.356989    0.356793\n",
       "       2010-03-15  0.354814    0.354833\n",
       "       2010-03-16  0.353878    0.354048\n",
       "       2010-03-17  0.319785    0.319687\n",
       "...                     ...         ...\n",
       "MSFT   2019-12-05  0.346979    0.372822\n",
       "       2019-12-06  0.356728    0.382734\n",
       "       2019-12-09  0.354154    0.380097\n",
       "       2019-12-10  0.354838    0.380798\n",
       "       2019-12-11  0.357777    0.383798\n",
       "\n",
       "[7371 rows x 2 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_prices, freq='bdays',\n",
    "              weeks=1, months=2, years=3,\n",
    "              future=False, annualized=True).dropna()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Share-Price : Close vs. Adj. Close\n",
    "\n",
    "It is important to understand the difference between the 'Close' and 'Adj. Close' share-price, which was also explained in Tutorial 01.\n",
    "\n",
    "The Close share-price is only adjusted for stock-splits, while the Adj. Close share-price is adjusted for both stock-splits and dividends. This means the Close share-price should be used when calculating valuation ratios such as P/Sales or P/E (which is demonstrated in Tutorial 04), while the Adj. Close should be used for calculating stock-returns.\n",
    "\n",
    "For daily stock-returns the Close and Adj. Close are identical, except on the days where dividends have been paid on the stock. The Adj. Close is adjusted for the dividend by assuming it is immediately reinvested in the stock, without having to pay any dividend-taxes, trading commissions, etc. This is also known as the Total Return.\n",
    "\n",
    "For longer time-intervals of one or more years, it becomes increasingly important to use the Adj. Close when calculating stock-returns, so as to include the dividends. Although some stocks pay little or no dividends, for other stocks the dividend might be a significant part of the annual returns.\n",
    "\n",
    "SimFin's keywords for the data-columns provide aliases that reflect this relationship so that `SHARE_PRICE == CLOSE == 'Close'` and `TOTAL_RETURN == ADJ_CLOSE == 'Adj. Close'`. So you just need to remember always to use the keyword `SHARE_PRICE` when calculating valuation ratios such as P/E and P/Sales, and use the keyword `TOTAL_RETURN` when calculating stock-returns.\n",
    "\n",
    "For example, we can calculate the future 5-year annualized returns from each stock's Total Return as follows:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Ticker  Date      \n",
       "AAPL    2007-01-03    0.370454\n",
       "        2007-01-04    0.364427\n",
       "        2007-01-05    0.370643\n",
       "        2007-01-08    0.370793\n",
       "        2007-01-09    0.351933\n",
       "                        ...   \n",
       "MSFT    2014-12-05    0.281061\n",
       "        2014-12-08    0.288033\n",
       "        2014-12-09    0.287912\n",
       "        2014-12-10    0.291291\n",
       "        2014-12-11    0.290795\n",
       "Name: Adj. Close, Length: 6003, dtype: float64"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_prices[TOTAL_RETURN], freq='bdays',\n",
    "              years=5, future=True, annualized=True).dropna()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Growth\n",
    "\n",
    "The same helper-function can be used to calculate growth-rates for quarterly or annual data. Let us first consider the quarterly Income Statements for ticker MSFT:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2009-06-30</td>\n",
       "      <td>1.309900e+10</td>\n",
       "      <td>3045000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-09-30</td>\n",
       "      <td>1.292000e+10</td>\n",
       "      <td>3574000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>1.902200e+10</td>\n",
       "      <td>6662000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-03-31</td>\n",
       "      <td>1.450300e+10</td>\n",
       "      <td>4006000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-06-30</td>\n",
       "      <td>1.603900e+10</td>\n",
       "      <td>4518000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-09-30</td>\n",
       "      <td>1.619500e+10</td>\n",
       "      <td>5410000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-12-31</td>\n",
       "      <td>1.995300e+10</td>\n",
       "      <td>6634000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2011-03-31</td>\n",
       "      <td>1.642800e+10</td>\n",
       "      <td>5232000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  Revenue  Net Income\n",
       "Report Date                          \n",
       "2009-06-30   1.309900e+10  3045000000\n",
       "2009-09-30   1.292000e+10  3574000000\n",
       "2009-12-31   1.902200e+10  6662000000\n",
       "2010-03-31   1.450300e+10  4006000000\n",
       "2010-06-30   1.603900e+10  4518000000\n",
       "2010-09-30   1.619500e+10  5410000000\n",
       "2010-12-31   1.995300e+10  6634000000\n",
       "2011-03-31   1.642800e+10  5232000000"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_income_qrt.loc['MSFT'].head(8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can calculate the quarterly growth-rate using `sf.rel_change` with the argument `freq='q'` to indicate the DataFrame contains data with quarterly frequency, and `quarters=1` if we want to calculate the relative change over 1 quarter, and `future=False` to get the past growth from the previous to the current quarter:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2009-06-30</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-09-30</td>\n",
       "      <td>-0.013665</td>\n",
       "      <td>0.173727</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>0.472291</td>\n",
       "      <td>0.864018</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-03-31</td>\n",
       "      <td>-0.237567</td>\n",
       "      <td>-0.398679</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-06-30</td>\n",
       "      <td>0.105909</td>\n",
       "      <td>0.127808</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Revenue  Net Income\n",
       "Report Date                      \n",
       "2009-06-30        NaN         NaN\n",
       "2009-09-30  -0.013665    0.173727\n",
       "2009-12-31   0.472291    0.864018\n",
       "2010-03-31  -0.237567   -0.398679\n",
       "2010-06-30   0.105909    0.127808"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_income_qrt.loc['MSFT'], freq='q',\n",
    "              quarters=1, future=False).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can also use a time-interval of 1 year with the argument `years=1`, and the function automatically converts it into 4 quarters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2009-06-30</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-09-30</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-03-31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-06-30</td>\n",
       "      <td>0.224445</td>\n",
       "      <td>0.483744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-09-30</td>\n",
       "      <td>0.253483</td>\n",
       "      <td>0.513710</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-12-31</td>\n",
       "      <td>0.048943</td>\n",
       "      <td>-0.004203</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2011-03-31</td>\n",
       "      <td>0.132731</td>\n",
       "      <td>0.306041</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Revenue  Net Income\n",
       "Report Date                      \n",
       "2009-06-30        NaN         NaN\n",
       "2009-09-30        NaN         NaN\n",
       "2009-12-31        NaN         NaN\n",
       "2010-03-31        NaN         NaN\n",
       "2010-06-30   0.224445    0.483744\n",
       "2010-09-30   0.253483    0.513710\n",
       "2010-12-31   0.048943   -0.004203\n",
       "2011-03-31   0.132731    0.306041"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_income_qrt.loc['MSFT'], freq='q',\n",
    "              years=1, future=False).head(8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we set a time-interval that is not a multiple of the DataFrame's frequency, which is quarterly in this case, then it is automatically rounded to the nearest multiple of the frequency. For example, if we use a time-interval of 2 months and 1 year, then it is rounded to 5 quarters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2009-06-30</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-09-30</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-03-31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-06-30</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-09-30</td>\n",
       "      <td>0.236354</td>\n",
       "      <td>0.776683</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-12-31</td>\n",
       "      <td>0.544350</td>\n",
       "      <td>0.856184</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2011-03-31</td>\n",
       "      <td>-0.136368</td>\n",
       "      <td>-0.214650</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Revenue  Net Income\n",
       "Report Date                      \n",
       "2009-06-30        NaN         NaN\n",
       "2009-09-30        NaN         NaN\n",
       "2009-12-31        NaN         NaN\n",
       "2010-03-31        NaN         NaN\n",
       "2010-06-30        NaN         NaN\n",
       "2010-09-30   0.236354    0.776683\n",
       "2010-12-31   0.544350    0.856184\n",
       "2011-03-31  -0.136368   -0.214650"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_income_qrt.loc['MSFT'], freq='q',\n",
    "              months=2, years=1, future=False).head(8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When the time-interval is more than a year, it is often useful to consider the annualized growth-rate, which is done by setting the argument `annualized=True`. For example, let us calculate the past 3-year annualized growth-rates:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2012-09-30</td>\n",
       "      <td>0.074051</td>\n",
       "      <td>0.077097</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2012-12-31</td>\n",
       "      <td>0.040952</td>\n",
       "      <td>-0.014468</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-03-31</td>\n",
       "      <td>0.122073</td>\n",
       "      <td>0.147628</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-06-30</td>\n",
       "      <td>0.074475</td>\n",
       "      <td>0.031948</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-09-30</td>\n",
       "      <td>0.045900</td>\n",
       "      <td>-0.010334</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Revenue  Net Income\n",
       "Report Date                      \n",
       "2012-09-30   0.074051    0.077097\n",
       "2012-12-31   0.040952   -0.014468\n",
       "2013-03-31   0.122073    0.147628\n",
       "2013-06-30   0.074475    0.031948\n",
       "2013-09-30   0.045900   -0.010334"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_income_qrt.loc['MSFT'], freq='q',\n",
    "              years=3, future=False, annualized=True).dropna().head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can do the exact same calculations on DataFrames with multiple stocks:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ticker</th>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">AAPL</td>\n",
       "      <td>2012-06-30</td>\n",
       "      <td>0.532336</td>\n",
       "      <td>0.690038</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2012-09-30</td>\n",
       "      <td>0.433599</td>\n",
       "      <td>0.480888</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2012-12-31</td>\n",
       "      <td>0.514797</td>\n",
       "      <td>0.570220</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-03-31</td>\n",
       "      <td>0.478217</td>\n",
       "      <td>0.458997</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-06-30</td>\n",
       "      <td>0.310346</td>\n",
       "      <td>0.284858</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">MSFT</td>\n",
       "      <td>2018-09-30</td>\n",
       "      <td>0.125876</td>\n",
       "      <td>0.216459</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2018-12-31</td>\n",
       "      <td>0.109168</td>\n",
       "      <td>0.188303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-03-31</td>\n",
       "      <td>0.141913</td>\n",
       "      <td>0.328619</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-06-30</td>\n",
       "      <td>0.084306</td>\n",
       "      <td>0.243209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-09-30</td>\n",
       "      <td>0.146602</td>\n",
       "      <td>0.235129</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>75 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Revenue  Net Income\n",
       "Ticker Report Date                      \n",
       "AAPL   2012-06-30   0.532336    0.690038\n",
       "       2012-09-30   0.433599    0.480888\n",
       "       2012-12-31   0.514797    0.570220\n",
       "       2013-03-31   0.478217    0.458997\n",
       "       2013-06-30   0.310346    0.284858\n",
       "...                      ...         ...\n",
       "MSFT   2018-09-30   0.125876    0.216459\n",
       "       2018-12-31   0.109168    0.188303\n",
       "       2019-03-31   0.141913    0.328619\n",
       "       2019-06-30   0.084306    0.243209\n",
       "       2019-09-30   0.146602    0.235129\n",
       "\n",
       "[75 rows x 2 columns]"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_income_qrt, freq='q',\n",
    "              years=3, future=False, annualized=True).dropna()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Invalid Data\n",
    "\n",
    "Some data causes the calculations for the relative change to become mathematically invalid. One example is if some values are zero, so the relative change causes division by zero, which creates a value of infinity in Pandas.\n",
    "\n",
    "Another example is when one number is negative and the other is positive, so the fraction of the two numbers is negative. This causes a problem when calculating annualized changes for multiple years.\n",
    "\n",
    "For example, consider the following which calculates the annualized change from a value of 456 to -123 over a 3-year period. The result is a complex number here, but in the Pandas DataFrames returned by `sf.rel_change` this would be a NaN (\"Not-a-Number\"):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-0.6769398225681011+0.559556641214265j)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(-123 / 456) ** (1/3) - 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To demonstrate this, consider the annual Income Statements for ticker AMZN, and note how the Net Income is negative in several rows:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2008-12-31</td>\n",
       "      <td>1.916600e+10</td>\n",
       "      <td>645000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>2.450900e+10</td>\n",
       "      <td>902000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-12-31</td>\n",
       "      <td>3.420400e+10</td>\n",
       "      <td>1152000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2011-12-31</td>\n",
       "      <td>4.807700e+10</td>\n",
       "      <td>631000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2012-12-31</td>\n",
       "      <td>6.109300e+10</td>\n",
       "      <td>-39000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-12-31</td>\n",
       "      <td>7.445200e+10</td>\n",
       "      <td>274000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2014-12-31</td>\n",
       "      <td>8.898800e+10</td>\n",
       "      <td>-241000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2015-12-31</td>\n",
       "      <td>1.070060e+11</td>\n",
       "      <td>596000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2016-12-31</td>\n",
       "      <td>1.359870e+11</td>\n",
       "      <td>2371000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2017-12-31</td>\n",
       "      <td>1.778660e+11</td>\n",
       "      <td>3033000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2018-12-31</td>\n",
       "      <td>2.328870e+11</td>\n",
       "      <td>10073000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  Revenue   Net Income\n",
       "Report Date                           \n",
       "2008-12-31   1.916600e+10    645000000\n",
       "2009-12-31   2.450900e+10    902000000\n",
       "2010-12-31   3.420400e+10   1152000000\n",
       "2011-12-31   4.807700e+10    631000000\n",
       "2012-12-31   6.109300e+10    -39000000\n",
       "2013-12-31   7.445200e+10    274000000\n",
       "2014-12-31   8.898800e+10   -241000000\n",
       "2015-12-31   1.070060e+11    596000000\n",
       "2016-12-31   1.359870e+11   2371000000\n",
       "2017-12-31   1.778660e+11   3033000000\n",
       "2018-12-31   2.328870e+11  10073000000"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_income_ann.loc['AMZN']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can calculate the relative change without problems:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2008-12-31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-12-31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2011-12-31</td>\n",
       "      <td>1.508452</td>\n",
       "      <td>-0.021705</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2012-12-31</td>\n",
       "      <td>1.492676</td>\n",
       "      <td>-1.043237</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-12-31</td>\n",
       "      <td>1.176704</td>\n",
       "      <td>-0.762153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2014-12-31</td>\n",
       "      <td>0.850947</td>\n",
       "      <td>-1.381933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2015-12-31</td>\n",
       "      <td>0.751526</td>\n",
       "      <td>-16.282051</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2016-12-31</td>\n",
       "      <td>0.826506</td>\n",
       "      <td>7.653285</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2017-12-31</td>\n",
       "      <td>0.998764</td>\n",
       "      <td>-13.585062</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2018-12-31</td>\n",
       "      <td>1.176392</td>\n",
       "      <td>15.901007</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Revenue  Net Income\n",
       "Report Date                      \n",
       "2008-12-31        NaN         NaN\n",
       "2009-12-31        NaN         NaN\n",
       "2010-12-31        NaN         NaN\n",
       "2011-12-31   1.508452   -0.021705\n",
       "2012-12-31   1.492676   -1.043237\n",
       "2013-12-31   1.176704   -0.762153\n",
       "2014-12-31   0.850947   -1.381933\n",
       "2015-12-31   0.751526  -16.282051\n",
       "2016-12-31   0.826506    7.653285\n",
       "2017-12-31   0.998764  -13.585062\n",
       "2018-12-31   1.176392   15.901007"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_income_ann.loc['AMZN'], freq='y',\n",
    "              years=3, future=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "But if we try and calculate the annualized change, then the rows with negative Net Income become NaN (\"Not-a-Number\"):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Revenue</th>\n",
       "      <th>Net Income</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Report Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2008-12-31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-12-31</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2011-12-31</td>\n",
       "      <td>0.358737</td>\n",
       "      <td>-0.007288</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2012-12-31</td>\n",
       "      <td>0.355882</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2013-12-31</td>\n",
       "      <td>0.295985</td>\n",
       "      <td>-0.380417</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2014-12-31</td>\n",
       "      <td>0.227811</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2015-12-31</td>\n",
       "      <td>0.205421</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2016-12-31</td>\n",
       "      <td>0.222382</td>\n",
       "      <td>1.053022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2017-12-31</td>\n",
       "      <td>0.259661</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2018-12-31</td>\n",
       "      <td>0.295923</td>\n",
       "      <td>1.566281</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Revenue  Net Income\n",
       "Report Date                      \n",
       "2008-12-31        NaN         NaN\n",
       "2009-12-31        NaN         NaN\n",
       "2010-12-31        NaN         NaN\n",
       "2011-12-31   0.358737   -0.007288\n",
       "2012-12-31   0.355882         NaN\n",
       "2013-12-31   0.295985   -0.380417\n",
       "2014-12-31   0.227811         NaN\n",
       "2015-12-31   0.205421         NaN\n",
       "2016-12-31   0.222382    1.053022\n",
       "2017-12-31   0.259661         NaN\n",
       "2018-12-31   0.295923    1.566281"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sf.rel_change(df=df_income_ann.loc['AMZN'], freq='y',\n",
    "              years=3, future=False, annualized=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The function `sf.rel_change` currently does not do any additional processing to help with such invalid data, so you must do any pre- and post-processing of the data yourself."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Column Names\n",
    "\n",
    "In all the example above, the column-names were the same after calculating the relative changes. We can change the column-names by passing the argument `new_names` to the `sf.rel_change` function. This argument is passed directly to the Pandas `rename` function and therefore accepts the same arguments.\n",
    "\n",
    "For example, we can use a dict to map from the old to new names:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>Share Price 3-Year</th>\n",
       "      <th>Total Return 3-Year</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ticker</th>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">AAPL</td>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>0.359871</td>\n",
       "      <td>0.359641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-04</td>\n",
       "      <td>0.356921</td>\n",
       "      <td>0.356789</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-05</td>\n",
       "      <td>0.360938</td>\n",
       "      <td>0.360885</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-06</td>\n",
       "      <td>0.351463</td>\n",
       "      <td>0.351568</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-07</td>\n",
       "      <td>0.315176</td>\n",
       "      <td>0.314955</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">MSFT</td>\n",
       "      <td>2019-12-05</td>\n",
       "      <td>0.355349</td>\n",
       "      <td>0.379921</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-06</td>\n",
       "      <td>0.362851</td>\n",
       "      <td>0.387593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-09</td>\n",
       "      <td>0.351098</td>\n",
       "      <td>0.375590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-10</td>\n",
       "      <td>0.353064</td>\n",
       "      <td>0.377583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-11</td>\n",
       "      <td>0.347731</td>\n",
       "      <td>0.372150</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>7512 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                   Share Price 3-Year  Total Return 3-Year\n",
       "Ticker Date                                               \n",
       "AAPL   2009-12-31            0.359871             0.359641\n",
       "       2010-01-04            0.356921             0.356789\n",
       "       2010-01-05            0.360938             0.360885\n",
       "       2010-01-06            0.351463             0.351568\n",
       "       2010-01-07            0.315176             0.314955\n",
       "...                               ...                  ...\n",
       "MSFT   2019-12-05            0.355349             0.379921\n",
       "       2019-12-06            0.362851             0.387593\n",
       "       2019-12-09            0.351098             0.375590\n",
       "       2019-12-10            0.353064             0.377583\n",
       "       2019-12-11            0.347731             0.372150\n",
       "\n",
       "[7512 rows x 2 columns]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# New column-names.\n",
    "SHARE_PRICE_3Y = 'Share Price 3-Year'\n",
    "TOTAL_RETURN_3Y = 'Total Return 3-Year'\n",
    "\n",
    "# Dict mapping old to new column-names.\n",
    "new_names = {CLOSE: SHARE_PRICE_3Y,\n",
    "             ADJ_CLOSE: TOTAL_RETURN_3Y}\n",
    "\n",
    "# Calculate the annualized 3-year changes and rename columns.\n",
    "sf.rel_change(df=df_prices, freq='bdays',\n",
    "              years=3, future=False, annualized=True,\n",
    "              new_names=new_names).dropna()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Another way of changing the column-names is to use a lambda-function. This is useful if you want to quickly rename all the columns in the same way, and don't need to address them later, in which case you would have to use a string such as 'Close 3-Year Ann. Chg.' instead of keywords like SHARE_PRICE_3Y."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>Close 3-Year Ann. Chg.</th>\n",
       "      <th>Adj. Close 3-Year Ann. Chg.</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ticker</th>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">AAPL</td>\n",
       "      <td>2009-12-31</td>\n",
       "      <td>0.359871</td>\n",
       "      <td>0.359641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-04</td>\n",
       "      <td>0.356921</td>\n",
       "      <td>0.356789</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-05</td>\n",
       "      <td>0.360938</td>\n",
       "      <td>0.360885</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-06</td>\n",
       "      <td>0.351463</td>\n",
       "      <td>0.351568</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2010-01-07</td>\n",
       "      <td>0.315176</td>\n",
       "      <td>0.314955</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td rowspan=\"5\" valign=\"top\">MSFT</td>\n",
       "      <td>2019-12-05</td>\n",
       "      <td>0.355349</td>\n",
       "      <td>0.379921</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-06</td>\n",
       "      <td>0.362851</td>\n",
       "      <td>0.387593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-09</td>\n",
       "      <td>0.351098</td>\n",
       "      <td>0.375590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-10</td>\n",
       "      <td>0.353064</td>\n",
       "      <td>0.377583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2019-12-11</td>\n",
       "      <td>0.347731</td>\n",
       "      <td>0.372150</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>7512 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                   Close 3-Year Ann. Chg.  Adj. Close 3-Year Ann. Chg.\n",
       "Ticker Date                                                           \n",
       "AAPL   2009-12-31                0.359871                     0.359641\n",
       "       2010-01-04                0.356921                     0.356789\n",
       "       2010-01-05                0.360938                     0.360885\n",
       "       2010-01-06                0.351463                     0.351568\n",
       "       2010-01-07                0.315176                     0.314955\n",
       "...                                   ...                          ...\n",
       "MSFT   2019-12-05                0.355349                     0.379921\n",
       "       2019-12-06                0.362851                     0.387593\n",
       "       2019-12-09                0.351098                     0.375590\n",
       "       2019-12-10                0.353064                     0.377583\n",
       "       2019-12-11                0.347731                     0.372150\n",
       "\n",
       "[7512 rows x 2 columns]"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Lambda function for changing column-names.\n",
    "new_names = lambda old_name: old_name + ' 3-Year Ann. Chg.'\n",
    "\n",
    "# Calculate the annualized 3-year changes and rename columns.\n",
    "sf.rel_change(df=df_prices, freq='bdays',\n",
    "              years=3, future=False, annualized=True,\n",
    "              new_names=new_names).dropna()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting\n",
    "\n",
    "If we want to plot the relative changes that we have calculated for a single stock, then it is quite easy because the function `sf.rel_change` returns a Pandas DataFrame that provides a `plot` function. For example, we can plot the 3-year annualized growth-rate for the Revenue and Net Income for the stock with ticker MSFT:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f5653589898>"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# New column-names.\n",
    "SALES_GROWTH = 'Sales Growth'\n",
    "EARNINGS_GROWTH = 'Earnings Growth'\n",
    "\n",
    "# Dict mapping from old to new column-names.\n",
    "new_names = {REVENUE: SALES_GROWTH,\n",
    "             NET_INCOME: EARNINGS_GROWTH}\n",
    "\n",
    "# Calculate annualized 3-year growth-rates and rename columns.\n",
    "sf.rel_change(df=df_income_ann.loc['MSFT'], freq='y',\n",
    "              years=3, future=False, annualized=True,\n",
    "              new_names=new_names).dropna().plot(kind='bar')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we want to make a single plot for multiple stocks, then we can use the Seaborn plotting package."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, '3-Year Annualized Sales Growth')"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Calculate annualized 3-year growth-rates.\n",
    "df_growth = sf.rel_change(df=df_income_ann, freq='y',\n",
    "                          years=3, future=False, annualized=True,\n",
    "                          new_names=new_names)\n",
    "\n",
    "# Remove rows where ALL items are NaN.\n",
    "# This keeps the rows where e.g. the Revenue growth was valid,\n",
    "# but the Net Income growth was invalid due to negative numbers.\n",
    "df_growth = df_growth.dropna(how='all')\n",
    "\n",
    "# Title of the plot.\n",
    "title = '3-Year Annualized Sales Growth'\n",
    "\n",
    "# Plot with connected lines.\n",
    "sns.lineplot(x=REPORT_DATE, y=SALES_GROWTH, hue=TICKER,\n",
    "            data=df_growth.reset_index()).set_title(title)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We would like to plot this as a bar-chart, but the x-axis looks bad:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, '3-Year Annualized Sales Growth')"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(x=REPORT_DATE, y=SALES_GROWTH, hue=TICKER,\n",
    "            data=df_growth.reset_index()).set_title(title)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can use the Fiscal Year as the index instead of the Report Date, to better align the x-axis in the bar-chart, but we removed that from the DataFrame above, so we need to reload the data and use Fiscal Year as the index:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset \"us-income-annual\" on disk (9 days old).\n",
      "- Loading from disk ... Done!\n"
     ]
    }
   ],
   "source": [
    "# Load annual Income Statements and set the index.\n",
    "df = sf.load_income(variant='annual', market='us',\n",
    "                    index=[TICKER, FISCAL_YEAR])\n",
    "\n",
    "# Select the tickers and columns we are interested in.\n",
    "df = df.loc[tickers, [REVENUE, NET_INCOME]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can then calculate the growth-rates and plot them in a bar-chart using Seaborn:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, '3-Year Annualized Sales Growth')"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Calculate annualized 3-year growth-rates.\n",
    "df_growth = sf.rel_change(df=df, freq='y',\n",
    "                          years=3, future=False, annualized=True,\n",
    "                          new_names=new_names)\n",
    "\n",
    "# Remove rows where ALL items are NaN.\n",
    "# This keeps the rows where e.g. the Revenue growth was valid,\n",
    "# but the Net Income growth was invalid due to negative numbers.\n",
    "df_growth = df_growth.dropna(how='all')\n",
    "\n",
    "# Plot bar-lines.\n",
    "sns.barplot(x=FISCAL_YEAR, y=SALES_GROWTH, hue=TICKER,\n",
    "            data=df_growth.reset_index()).set_title(title)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Mean Log-Change\n",
    "\n",
    "The simfin package provides another function `sf.mean_log_change` which is used to calculate the average change over many periods. This is useful for calculating the mean annualized return on a stock e.g. over all 1-3 year periods, so as to smoothen out the short-term volatility and get a less noisy estimate of the future return. We will use this in other tutorials to study the relation between predictor variables such as P/E ratios and average stock-returns over the following 1-3 years.\n",
    "\n",
    "We will first show how the mathematical formulas are derived and then show how the function `sf.mean_log_change` is used on the data. The math gives the precise definition of what is being calculated, but if you understand the general idea then you can probably skip the math."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Mathematical Derivation\n",
    "\n",
    "Let us consider a time-series $x_t$ which is indexed by time-step $t$. This could e.g. be the Total Return time-series of a stock, which is the share-price with dividends reinvested immediately and without having to pay any taxes, so it is a tax-neutral measure of the total return to investors from dividends and changes in share-price.\n",
    "\n",
    "We are then interested in the average change between the current time-step $t$ and all the future time-steps between $t + N$ and $t + M$, e.g. we could choose $N$ to be 1 year and $M$ to be 3 years, so we are interested in the change between the current time-step $t$ and all future time-steps between $t + 1\\ Year$ and $t + 3\\ Years$.\n",
    "\n",
    "For each value $i$ between $N$ and $M$ we want to normalize the change between the current time-step $x_t$ and the future time-step $x_{t+i}$ so they are comparable for different choices of $i$, because there is a big difference in a 10% change that occurs over a single year, compared to a 10% change that occurs gradually over 3 years. An easy way of doing this normalization, is to use the Geometric Mean, which is the average cumulative change per step, in order to go from $x_t$ to $x_{t+i}$ in $i$ steps. In this case it is simply the $i$'th root of the overall change, which is calculated as follows:\n",
    "\n",
    "$$\n",
    "    Geometric\\ Mean\\ Change_{t, i} = \\left( \\frac{x_{t+i}}{x_t} \\right) ^ {1/i}\n",
    "$$\n",
    "\n",
    "We can easily check that the Geometric Mean takes us from $x_t$ to $x_{t+i}$ in $i$ steps, by taking it to the $i$'th power:\n",
    "\n",
    "$$\n",
    "\\left( Geometric\\ Mean\\ Change_{t, i} \\right) ^ i\n",
    "= \\left[ \\left( \\frac{x_{t+i}}{x_t} \\right) ^ {1/i} \\right] ^ i \n",
    "= \\frac{x_{t+i}}{x_t}\n",
    "$$\n",
    "\n",
    "If we wanted to calculate the Annualized Change instead, then we need to adjust for a factor $a$. For example, if we are using daily share-price data then $a$ would be roughly 252 because that's how many trading-days there typically are in a year. If we are using quarterly financial data, then $a$ would be 4 because that's how many quarters there are per year. The formula with this adjustment is only slightly different:\n",
    "\n",
    "$$\n",
    "    Annualized\\ Change_{t, i} = \\left( \\frac{x_{t+i}}{x_t} \\right) ^ {a/i}\n",
    "$$\n",
    "\n",
    "Now that we have normalized the changes from $x_t$ to $x_{t+i}$ for different choices of $i$, we want to take their average. Ideally we would like to calculate the following:\n",
    "\n",
    "$$\n",
    "    Mean\\ Annualized\\ Change_t = \\frac{1}{M-N} \\cdot \\sum_{i=N}^{M} \\left( \\frac{x_{t+i}}{x_t} \\right) ^ {a/i}\n",
    "$$\n",
    "\n",
    "This is easy enough to implement using for-loops, but it would be very slow in native Python code. Instead we want a formula that can be calculated efficiently using Pandas and Numpy's vectorized functions. So we use the logarithm to transform the formula into the following:\n",
    "\n",
    "$$\n",
    "Mean\\ Log\\ Change_t = \\frac{1}{M-N} \\cdot \\sum_{i=N}^{M} \\log \\left[ \\left( \\frac{x_{t+i}}{x_t} \\right) ^ {a/i} \\right] \\\\\n",
    "= \\frac{1}{M-N} \\cdot \\sum_{i=N}^{M} (a/i) \\cdot \\left( \\log(x_{t+i}) - \\log(x_t) \\right) \\\\\n",
    "= \\frac{1}{M-N} \\cdot \\left[ \\sum_{i=N}^{M} (a/i) \\cdot \\log(x_{t+i}) - \\log(x_t) \\cdot \\sum_{i=N}^{M} (a/i) \\right] \\\\\n",
    "$$\n",
    "\n",
    "This formula may look weird and intimidating, but it is reasonably simple if we break it down. The hardest part of the formula is the following, which is a so-called *windowing* function, that takes the dot-product between a slice of the time-series from $x_{t+N}$ to $x_{t+M}$ and a vector of values $(a/i)$ where $i$ goes from $N$ to $M$:\n",
    "\n",
    "$$\n",
    "     \\sum_{i=N}^{M} (a/i) \\cdot \\log(x_{t+i})\n",
    "$$\n",
    "\n",
    "If instead we are interested in the change from **past** time-steps $x_{t-i}$ to the current time-step $x_t$, then the formula becomes:\n",
    "\n",
    "$$\n",
    "Mean\\ Log\\ Change_t = \\frac{1}{M-N} \\cdot \\sum_{i=N}^{M} \\log \\left[ \\left( \\frac{x_t}{x_{t-i}} \\right) ^ {a/i} \\right] \\\\\n",
    "= \\frac{1}{M-N} \\cdot \\sum_{i=N}^{M} (a/i) \\cdot \\left( \\log(x_t) - \\log(x_{t-i}) \\right) \\\\\n",
    "= \\frac{1}{M-N} \\cdot \\left[ \\log(x_t) \\cdot \\sum_{i=N}^{M} (a/i) - \\sum_{i=N}^{M} (a/i) \\cdot \\log(x_{t-i}) \\right] \\\\\n",
    "$$\n",
    "\n",
    "The vector of values $(a/i)$ can be pre-calculated for efficiency, as can its sum. The windowing can be calculated using the Pandas `rolling` operator. And we can handle DataFrames with multiple stocks by using the `groupby` operator. So altogether we can implement these log-transformed formulas quite efficiently using Pandas and Numpy's vectorized operators, instead of slow for-loops. But the implementation is a bit complicated and may be hard to understand without the mathematical derivation above."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Future Mean-Log Annualized Stock Returns\n",
    "\n",
    "Let us now try and use the function `sf.mean_log_change` to calculate and plot the mean-log annualized stock-returns for all 1-3 year periods. That is, we want to calculate the mean-log annualized changes between the current time-step: `df_prices[t]` and all future time-steps in the slice: `df_prices[t + 1 Year:t + 3 Years]`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Future 1-3 Years Mean-Log Ann. Total Return')"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# New column-names.\n",
    "SHARE_PRICE_1_3Y = 'Mean-Log Share-Price Return 1-3 Years'\n",
    "TOTAL_RETURN_1_3Y = 'Mean-Log Total Return 1-3 Years'\n",
    "\n",
    "# Dict mapping old to new column-names.\n",
    "new_names_1_3y = {SHARE_PRICE: SHARE_PRICE_1_3Y,\n",
    "                  TOTAL_RETURN: TOTAL_RETURN_1_3Y}\n",
    "\n",
    "# Calculate the future mean-log annualized 1-3 year changes.\n",
    "df_mean_chg = sf.mean_log_change(df=df_prices, freq='bdays',\n",
    "                                 future=True, annualized=True,\n",
    "                                 min_years=1, max_years=3,\n",
    "                                 new_names=new_names_1_3y)\n",
    "\n",
    "# Title of the plot.\n",
    "title = 'Future 1-3 Years Mean-Log Ann. Total Return'\n",
    "\n",
    "# Plot the result.\n",
    "data = df_mean_chg.dropna().reset_index()\n",
    "sns.lineplot(x=DATE, y=TOTAL_RETURN_1_3Y, hue=TICKER,\n",
    "             data=data).set_title(title)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The simfin function actually uses the natural logarithm, which is almost linear for returns between $\\pm 20\\%$, but beyond that range it underestimates both the gains and losses compared to the non-log returns. But for practical purposes, you may often consider the mean-log changes to be roughly equivalent to the normal percentage changes.\n",
    "\n",
    "For example, consider the mean-log Total Return on 15th November 2016 for ticker MSFT:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Ticker  Date      \n",
       "MSFT    2016-11-15    0.337589\n",
       "Name: Mean-Log Total Return 1-3 Years, dtype: float64"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_mean_chg.loc[('MSFT', '2016-11-15'), TOTAL_RETURN_1_3Y]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This value of about 0.34 is the mean-log annualized return for *all* periods starting on 2016-11-15 and ending somewhere between 2017-11-15 and 2019-11-15. So if you had bought the MSFT stock on 2016-11-15 and held it somewhere between 1 and 3 years, then on average, you would have made a log-return of 0.34 per year.\n",
    "\n",
    "Because the log-transform underestimates the normal returns for values beyond $\\pm 20\\%$, the mean annualized return was probably around 40% for all the holding-periods starting on 2016-11-15 and ending somewhere between 2017-11-15 and 2019-11-15.\n",
    "\n",
    "As will be shown in other tutorials, the mean-log stock-returns are very useful for studying the relation between predictor variables such as P/E ratios and future returns."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Past Mean-Log Annualized Stock Returns\n",
    "\n",
    "We can also calculate mean-log returns for the **past** instead of the future as we did above. For example, we may calculate the mean-log annualized change between all the **past** time-steps in the slice: `df_prices[t - 3 Years:t - 1 Year]` and the current time-step: `df_prices[t]`. This is done simply by changing the function's argument to `future=False`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Past 1-3 Years Mean-Log Ann. Total Return')"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOydd3gU1dfHv1uy2YSEFNKARGoIEFogBCWEEloghC5FpCNFERBs+CoiioA/AQERURBFpClNQlMI0qsGAxggAdKAFNLL9rnvH7Ntdmd203YDyXyeJ09m7tyZe3d3Zs6955x7joAQQsDDw8PDU2cR1nQHeHh4eHhqFl4Q8PDw8NRxeEHAw8PDU8fhBQEPDw9PHYcXBDw8PDx1HF4Q8PDw8NRxeEHAw8PzzDFy5EgcOXKkprtRZ+AFQTUSGRmJDh06ICQkBN27d8fixYtRWlpa5WtevHiR87hSqcS8efMQGRmJoKAgXLlyxeL1kpOTMXLkSHTt2hVdu3bFlClTkJyczFp37dq1mDx5MqPs4cOH6Ny5M+7evVvxD1NNXLx4EUFBQZg/fz6j/ObNmwgKCsKUKVPs3qfU1FQEBQXZrb23334bwcHBePr0qV3amzZtGkJCQhASEoLg4GC0a9dOv79s2TKL55aWliIoKAg5OTnV0pdVq1YhODgYISEhCA0NxSuvvILbt2+X+/xu3brhxo0b1dKX2gIvCKqZb7/9FvHx8Thw4ABu3ryJTZs22bzNzp0744svvoC3t7fVuj4+Pli/fj2uXr2Ky5cvIzIyEm+99RZr3TfeeANPnz7F3r17AQCEEHz00UeYOnVqtb/01Gp1hep7eXnh2rVrKCoq0pcdOnQITZs2rdZ+PYuUlJTg5MmTcHFxQWxsrF3a/OGHHxAfH4/4+HgMGjQIs2bN0u8vWbLELn0wZtSoUYiPj8fFixcRHByMhQsX2q3tit6rzwO8ILARvr6+iIiIQFJSEgBg3759GDRoEEJCQtC3b1/s3r1bXzcvLw+zZs1CaGgowsLC8Morr4CiKLzzzjt4/PgxZs+ejZCQEHz//fdm7UgkEkyZMgWhoaEQCq3/nPXr14e/vz8EAgEIIRCJREhLS2OtK5FI8Pnnn2P16tXIysrCnj17UFhYiNmzZ+vr/Prrr4iKikLXrl0xY8YMPHnyRH9s2bJl6NmzJzp37oxRo0bhn3/+0R9bu3YtFixYgIULFyIkJAS///47bty4gREjRqBz587o3r07Vq1axfk5HB0d0adPHxw9ehQA/XAeP34cMTExjHrJycmYMmUKwsLCEBUVhRMnTuiPnTp1CsOGDUNISAh69+6NjRs36o/pRvgHDx5Ez5498eKLL+K7776z+v2yoVAo8Nlnn6FHjx6IiIjAihUroFQq9cc3b96M8PBwREREYO/evQgKCkJGRgbn9Y4fPw5PT0/MmjULBw4cYBxbu3YtFi5ciLfffhshISEYMmQIY7Tcs2dPbNu2DTExMejSpQsWLlzI6EtV2L59O/r27Ytu3bph3rx5yM3NBQC8+uqrAID+/fsjJCQEp0+fxtOnTzF9+nR069YNYWFh+kFHRZFIJIiJiUFKSgpkMpm+/MSJE4iJiUFoaCheffVV3L9/HwA9uCkoKMDkyZMREhKCX375BadPn0ZUVBTjusazhlWrVuGdd97BvHnzEBISghMnTmDVqlV49913sWDBAoSEhGDo0KE1OkuuMoSn2ujTpw+5cOECIYSQx48fk8GDB5O1a9cSQgg5ffo0SU1NJRRFkStXrpAOHTqQW7duEUII+fLLL8lHH31ElEolUSqV5Nq1a4SiKLNrWiMiIoJcvny5XHW7dOlC2rRpQ4KCgsjGjRst1l2xYgWZNGkSCQsLIwkJCfryY8eOkQEDBpD79+8TlUpF1q9fT8aPH68/fuDAAZKfn09UKhXZvHkz6dGjB1EoFIQQQtasWUOCg4PJqVOniEajITKZjIwcOZIcPnyYEEJIcXExuXHjBmt/Lly4QPr06UOuXr1Kxo4dSwgh5OTJk+S1114jO3fuJJMnT9Zfo0ePHuTAgQNEpVKRmzdvkrCwMHL//n1CCCEXL14kd+/eJRqNhiQmJpKwsDASFxdHCCEkJSWFtGrVinz00UdELpeTW7dukeDgYPLw4UPWPunqs7F69WoyduxYkpubS54+fUpGjx5NNmzYQAghJC4ujoSHh5Pk5GRSWlpK3nrrLdKqVSuSnp7O+XtMmDCBrF69mmRmZpLWrVuTxMRE/bE1a9aQ9u3bk7NnzxK1Wk1WrlxJxo0bpz8eERFBXn75ZZKdnU3y8vLIgAEDyN69eznbYmPRokVk/fr1jLKTJ0+S8PBwcvfuXSKTycjixYvJ9OnTCSGElJSUkFatWpHs7Gx9/aysLHLq1Ckil8tJYWEhmTlzJlm0aJH++IgRI0hsbCxr+ytXriQfffQRIYQQuVxOli1bRiIiIvTHr127RsLDw8mtW7eIWq0mO3bsIFFRUUStVhNCCAkLCyPx8fH6+nFxcWTgwIGMNozrrFy5krRr146cOXOGUBRFZDIZWblyJenYsSO5ePEiUavVZNmyZfr77nmEnxFUM2+88YZeb9m1a1f96Ll379544YUXIBAIEBYWhvDwcFy/fh0AIBaLkZOTg8ePH8PBwQGhoaEQCAQ27ef169dx/fp1fPTRR2jbtq3FugsWLEBaWhqGDRuG9u3b68v37NmD2bNno3nz5hCLxZgzZw4SEhKQlZUFABg+fDjc3d0hFovx2muvoaSkBKmpqfrzO3fujMjISAiFQkilUjg4OCAlJQX5+flwcXFBx44dLfYrNDQUOTk5SEtLw6FDhzB8+HDG8bi4ODRt2hTDhw+HWCxGu3bt0LdvX/2s4KWXXkKrVq0gFArRunVrREdH4+rVq4xrvPnmm3B0dERwcDACAwNx584d61+uCYcPH8bcuXPh6emJBg0a4I033sChQ4cAAMeOHcPo0aPRokULODs744033rB4rfT0dFy/fh1DhgyBr68vwsLCcPDgQUadrl27IiIiAiKRCMOGDTPr8+TJk+Ht7Q0PDw/07t0biYmJFf5MbJ9x3LhxaNWqFaRSKd555x2cO3cOeXl5rPV9fHwQGRkJR0dH1K9fH6+99hquXbtW7vb279+P0NBQdOzYEcePH8e6dev0x/bs2YNJkyYhODgYIpEIEyZMQElJSZU+54svvoiePXtCIBBAKpUCALp3746XXnpJ/z1Xx/dYU4hrugO1jY0bN6J79+5m5WfOnMHGjRuRkpICiqIgl8vRqlUrAMD06dPx9ddfY9q0aQCAsWPHYubMmVXuy+PHjxEdHa3fj4+PZxx3dnbG+PHj8dJLL+Ho0aNo0KAB63WkUin8/f0RGBhodv1ly5Zh+fLl+jKhUIjMzEz4+vri+++/x759+5CTkwOBQACZTIb8/Hx93YYNGzKu9/nnn2PDhg2IiopCQEAA3nzzTfTq1Yvz8wkEAgwdOhTbt2/H9evX8b///Q/79+9n9C8+Ph6hoaH6Mo1GgxEjRui/jzVr1iApKQkqlQpKpRJDhgxhtGFsd5FKpSgrK4NGo2Fc8/jx45x9BICcnBw0atRIv9+oUSO9sMzOzkaXLl04vxNTDh06hFatWunvnZiYGKxduxbvvPMORCKRWZ+dnJxQVlbGuIaXlxfjMxUWFlpsszxkZ2czfisPDw84OzsjOzsbAQEBZvVLSkqwfPlyXLx4EcXFxQAqpnsfOXIkli1bhpycHMyZMwd3795FSEgIAPp3//PPPxmqPJVKhaysLLRr165Sn4/tdzH9Hk2/5+cJXhDYAZ1nz6pVq9C3b184ODjg9ddfB9EGfnVxccH777+P999/H0lJSZg0aRLat2+Pl156qUrtNmrUyOzlbwpFUZDJZMjKyuIUBFz4+flh/vz5GDx4sNmxy5cv48cff8SPP/6Ili1bAgC6dOmi/8wAzGY9zZs3x9q1a0FRFI4fP44333wT165dg6OjI2cfhg8fjqioKIwaNcqsnp+fH1588UVs2bKF9dyFCxdi2rRp2LJlCxwdHbFs2bJyPcwikcjsezWe6Zji7e2Nx48fo3nz5gCAJ0+ewNfXV38sMzNTX9fYxmIKIQQHDx5EdnY2wsPDAdAvz4KCApw/f96i0LQ1Pj4+ePTokX6/oKAAZWVl8PHxYZ3dbtq0CTk5Odi/fz8aNGiA69evY+rUqRVu19vbG5988gkmT56MqKgouLu7w8/PDwMGDDDzeNNh2h8nJyfI5XL9vlKp1AunugKvGrIDSqUSSqUSnp6eEIvFOHPmDC5cuKA/fvr0aaSmpoIQAhcXF4hEIr3h18vLC+np6Vavr1AoANAjH4VCwXjhGnPhwgX8999/0Gg0KCkpwcqVK1G/fn20aNGiwp9r3Lhx2LRpk94QV1RUpB8dl5aWQiQSwcPDAyqVChs2bGAY89g4ePAg8vLyIBQK4eLiAoFAYFVF1qRJE/z888+YN2+e2bG+ffsiOTkZhw8fhkqlgkqlQkJCAh48eKDvo5ubGxwdHXHjxo1q8VtXKBSMP4qiEB0djY0bNyIvLw95eXn45ptvMHToUADAoEGD8Ntvv+HBgweQyWT45ptvOK/9999/IzMzE/v27cPBgwdx8OBBxMbGYtCgQWZGY3szZMgQ7NmzB0lJSZDL5fjyyy/Ro0cPeHp6wtnZGc7Ozoz7uLS0FE5OTnB1dUVeXh6+/fbbSrcdHByMjh074scffwRA35c//fQTbt++DUKI3stK94yYPlMtW7bE06dPceXKFSiVSoaaqa7ACwI74OLigg8//BALFixA165dERsbi8jISP3x1NRUTJ06FSEhIRg7dizGjx+Pbt26AQBmzpyJTZs2ITQ0FFu3bmW9flRUFDp06ICsrCxMnz4dHTp0YIzOjCkqKsLChQsRGhqKfv36ITU1VT8iriiDBg3C1KlTMX/+fHTu3BlDhw7F+fPnAQC9evVC9+7dMWDAAERGRsLFxcWqe+vZs2cxePBghISE4IsvvsDatWshkUis9iM0NBQ+Pj5m5a6urti6dSt+//139OjRAz169MDq1av1XjJLly7FmjVrEBISgm+//RaDBg2q8HdgSocOHRh/165dw9y5c9G6dWsMHToUQ4cORYcOHTBr1iwA9DqR8ePH49VXX8WAAQPQuXNnAGD93AcOHEC/fv0QGBgIb29v/d+kSZNw6tQphittZUhPT0dISIhebVUR+vXrh+nTp2PWrFno1asXCgoKGF5f8+fPx9y5cxEaGoq//voLM2bMQE5ODrp164YJEyagT58+Ver7jBkzsGPHDhQVFaFr16744IMP8OGHHyI0NBRRUVE4evSoflAxZ84c/O9//0PXrl2xc+dOeHl5YfHixXjrrbfQp08fNGrUCK6urlXqz/OGgHANHXl4eOzO3bt3MXLkSNy8ebNc7sA8PNUBf6fx8NQwf/75J5RKJfLz87F69Wr07duXFwI8doWfEfDw1DBTpkxBQkICxGIxunXrho8//pjhkcLDY2t4QcDDw8NTx+Hnnzw8PDx1nOduHcGNGzcq5eEC0K59lT23qtTFtvnPzLddW9utybYr265CoUCnTp1Yjz13gsDR0RFt2rSp1LmJiYmVPreq1MW2+c/Mt11b263JtivbrqUQGLxqiIeHh6eOwwsCHh4enjoOLwh4eHh46jjPnY2Ah4cNlUqFjIwMffAwlUpVI2GBa6pdXdsPHz6Ev78/HBwcaqQPPM8nvCDgqRVkZGTA1dUVTZs21Ye8dnJysns/aqpdACgrK0NZWRkyMjLQrFmzGukDz/MJrxriqRXI5XI0aNDA5gl9nmUEAgEaNGjACKnMw1MeeEHAU2uoy0JAB/8d8FQGXhDw8PDw6HjwF/A0uaZ7YXd4GwEPTzWQn5+PKVOmgKIofXIdT09PAHQGrN27d3OeO3HiRLz77ruMfNA8NcT2YfT/pVVP3/k8wQuCOgwhBGv+vIfWfvUR3cFyrlwey3h4eODQoUOQyWTYsmULnJ2dMX36dJu3q9Fo9LmKeaqIvGqJfZ5neNVQHeZ6aj42xCXj/X0JNd2VWo0uqToAfP/994iJicHQoUPx5ZdfMupRFIX33nsPa9euBQCcP38eY8eOxYgRIzBv3jyUlpYCoLOaff311xg/frw+NShPNVDAnXe6tsPPCOowi/b+CwAoVqhruCd1gzNnzuDUqVPYu3cvnJycUFBQoD+m0Wjw9ttvIzAwEHPmzEFeXh42bdqEbdu2wdnZGd999x22bduGuXPnAqBjbu3ataumPkrtpFCb3tW57uWC4AVBHSYtr6ymu1CnuHTpEkaOHKlfZ+Du7q4/tmTJEgwaNAhz5swBAPz7779ITk7G+PHjAdCLxYwjRw4ePNiOPa8jqGX0f2HdU7XZVBCcPXsWy5cvB0VRePnllzFz5kzG8f379+OLL76Ar68vAODVV1/Fyy+/bMsu8WihKEM+Ii+XmgnjW9cghHC6d4aEhODKlSuYNm0aHB0dQQhBeHg41qxZw1q/phat1Wo0Kvp/SRagKAbEToCoboyVbWYj0Gg0WLZsGbZs2YIjR44gNjYWycnmblmDBw/GoUOHcOjQIV4I2JH/O3hLv92gnqQGe1J3CA8Px759+yCT0SNPY9XQ6NGj0atXL8yfPx9qtRqdOnXCP//8g9RUWm8tk8nw8OHDGul3nUGtMGyv8Ad+m1pzfbEzNhMECQkJaNKkCQICAiCRSBAdHY1Tp07ZqjmeCrLralpNd6HO0bNnT0RGRmLUqFEYNmwYfvjhB8bxqVOnom3btnj33Xfh7u6OFStWYOHChYiJicGYMWPw4MGDGup5HUGjZO4n/l4z/agBbJaz+Pjx4zh37hyWL18OADh48CASEhKwZMkSfZ39+/djzZo18PDwQLNmzbB48WI0bGjZjbEqGcrkcjmkUmmlzq0qz1rbg34yvFQa13fAlhEBdmnXVqhUKgQGBur3LalhbElNtWvcdlJSkt2DztXU/V1d7UrzEuHz70bUy77OKE8ce9nmbVeUqrTLldDGZgowNvli+oD06dMHQ4YMgUQiwa5du/Dee+9h+/btFq/LZyirrraNRpdCsU36Zs/PnJiYyNCb18Wgc7q2HRwc7H6vPW/ZusxY+iJrsaVrP2+fuUYylPn5+SEzM1O/n5WVBR8fH0YdDw8PSCS0fnrMmDG4ffu2rbrDY4RaQzH2FWqKoyYPTx2g9GlN96DGsZkgaN++PVJSUpCeng6lUokjR44gMjKSUSc7O1u/HRcXhxYtWtiqO88Uu6+m4XxSzd18aoo5W3taouCoycNTy1nVFPhf3XjvWMJmqiGxWIwlS5ZgxowZ0Gg0GDVqFAIDA7Fu3Tq0a9cOffv2xc8//4y4uDiIRCK4ublhxYoVturOMwMhBO/vvwkASFkZXSN90FDmartdV9MwPuyFGugND081U5IDyAsBr5aW62nUgCzfch2KAoS1PwCDTZ1ke/XqhV69ejHK5s+fr99etGgRFi1aZMsuPHN8HVeNkQ0zbwJSN8C9Yi9wjdZ+My28GeLT8xGfVoDfbzzmBQFP7WBDZ0BRZD1wnKrUvKzDWCBhj2FfowSENePkYU9qv6h7RsgpVePWo0Ks/vNe9V302x7A+hDr9UzQaGhBEODpBGcJvYry0oNcxiIzHp7nFkU5g8c9vmFe5h3E3Dd1Ka2l8ILATkz6LQ1DNpxnlKVXJcSDSrscnqp4nCDdjEAsFEAqNiynL5KrKt8fHgC0rSsoKAj3799nlP/4449o3749iouL9WVXrlxBly5dMHz4cAwaNAhff/21vnzWrFl27XetRGUlU1tpDnN/xGag+zxmGS8IeGxBRKAhoFXEF6dZ3WzLhcywKhXyogqF0NXZCIRCARwdDLdASTmDzz0plGH8d5eRWcinRDTl2LFj6NKlC44ePcooj42NRfv27fHnn38yykNDQ3Hw4EHs27cPv//+O27dugWeamLvRPZyWT6wfyZQks0sDxoMiByAqccAr1Z0mbpuOFLwgsDOUCYv/iJ5JSN/Eo1he2UA/VdOdILAdEYgU2q4TmGw60oaLj3Ixffn+JWuxpSWluLGjRtYvnw5jhw5oi9PS0tDWVkZFixYwCg3xtnZGcHBwUhL41d8VxtJf7CXX9xA2wEub2KWO7rS/5t0B3ospLc1dUMQ1I2ISjWMTvdeHyVIecpcbJRXqoSbUyVWgZLK+/7rZwQCARwdDIKgrJyCQCe8HhfIKt0HW7Lv7wzsvpoKYTV6e4wJDcCoLv4W65w8eRLh4eFo1qwZ3N3dcfv2bQQHByM2NhbR0dEIDQ3F+++/j9zcXDRo0IBxbn5+Pv7991+8/vrryMvLq7Z+87BwbrV2w2Q2brzgVaR9JtW8asiM0tJSPvBVJSiWqxEtvIwE6Uy4F/7HOJZXWr4Rh1pDIafYqG41CAKxSABHseEWsCYINBTB0t9vIymb1nM3cOGD1Rlz5MgRDBw4EAAdTDE2NhYAcPToUURHR0MoFKJ///6MZDLXr1/H8OHDMX36dLz22muMMBk8lSCNOyQEAODSN4ZtoYVxsFgbxoafEdBMmzYN69evB0VRGDZsGOrVq4fevXvj7bfftkf/nl+UpUDufeDxP3BI+B0bJXTAPX/BU9wmzfTVckvKN+L4+Pfb2HklBbc+GYR6jmJ2QaAoARxdrF5LbTQjkBgJApnKsprq+K1M/HgxRb+fX/psGpdHdfHH4LYN7BrqIT8/H5cvX8a9e/fw2WefQaPRQCAQYNiwYUhJScG0adMAAEqlEgEBAZgwYQIA2kawefNmu/Wz1vPDQMvHz6wybFMWBj5i7b1jzeBcS7A6I8jLy4OLiwv++OMPDB06FIcPH8aZM2fs0bfnjoz8MshV2ptr13hgcwRweD6cUw1RVwsI80W98vidcl07N+EEHkpfhSr9H7qAzcgst55wmxCCb07TaxnEQiEcRIbpsLUZwRs7/2HsF8jqxrS5PJw4cQLDhw/HsWPHEBcXhzNnzsDf3x/Lly/Hm2++ibi4OMTFxeH8+fPIysrCo0eParrLdQPTF7nODgAAhen0Ohw2nD3p/2W5tunXM4ZVQaBWq1FYWIgTJ04gMjKyxiIrPuso1RR6rDqNt3+l0z/iIbuwFAooSI08dR7ksCxqYaEbRfs8i1O012WbEZQjs1J6ngz74+mXkEgIjAhprD9WXhuBDpWaX3eg48iRI+jXrx+jbMCAAXj06JFZef/+/TmNxjouXbqEnj176v/i4+Orvc+1DjZ9vsrERdu7tdEOAVz8gDYxQPh8Zr163vT/W79VaxefVayqhmbNmoUxY8agc+fO6NixI9LT09GoUSN79O25Qmc4vZacCfnJz8G1FlEMDepJxJCrKjCavrwJ7lQeIAJIkXYkySYIymE3kKkML3svF0e09DGMkBg2iHKgovhgdTp+/vlnANAnnQGASZMmYdKkSWZ1Fy9erN/u1q2b2fFu3bohISHBBr2s5chYjOy6gataQc+inTyA+v5AUQZdXpwJzL1qfp6LNkDmrX3A6B/Mj9cyLAoCjUaDoqIinDhxQl8WEBDA6zRZyCySY5PDWgyirgHnueuJoUH3+jn4s1QKOWiDlFpDQSzimJwRAhx/HyO0g/2Cx0lwBSotCMqUBjtAYw+mDv1/J+7ijT5W4rMYEZ9WYL0SD489ODwf8GlrXq5ToX6mfbE36wnUb2QQBAoOdarYERA5Al2mVHtXn0UsqoZEIhHDw4GHGw1FMEh0zWo9H0EBNuTPwVLxT/qy1h9Z+I7VTB3nvSytKqnSgsAwI3B3qrrXj43yGvHwlB+KAv7+ETj2Lr3fvI/hGKGA1EuG/eJMwNXXsO/RlPu6EmeYuZjWUqzaCLp06YKVK1ciISEBycnJ+j8eJhqKoIg4W63XUUgvwgoSZqCVL204VlMExVzhHZRMGwIF7VSXzeOhgoLA2HUUAPq39TWtzsBVKsaU7k0ZZWyRTHl47IqpHWDg54btxN+BbVGG/af3AFejLIj9PuG+rkBo2bOoFmHVRnDpEi1NjY1VAoEAu3fvtl2vnkMoQqAox/o8VwGtQ/Zv4Ioezb1xL6sEAO2K6Sp1oCOKUhqgUSf6BLMAWjpBwCI4Ln0DDP7CYvulRmEkhMKKGf4VaooRkgKghZjYuo2ah8d2mAoCjyaG7SMs0Y1d/QzbwcO5rysQAll1I+SH1TfXnj17rFXhAS0IVBa+TgVxgKNAhRjhRQCAt7OQ4cNfrNC+2L/tQf/XhdAtzmK2oxMEGhZBcHWzVUGQWcTtF206uperNFj75z3MjWwJF0cxlGoKjiIhLi2OxEsr4gAAlx/koneQD9vleHjsg6kgkNQzbLPNkgPY01KaUZpD/yWdBAL7Wa//HFOuEBOXLl1CUlISFAqDV8lrr71ms049jzgUpaORgD00QHq3j3FfHITeF14xFBICiZEPfwlXzKHiJ4xdgyCouA9/Rn4ZVh6j1y24O5uHtTDNXNZh6R9QaihIHUSYG0kbkR1EQjR0MxiZp2y7VmMJdnh4ADCDxwnLEa6laTgQvRrIK2eUhNykWi8IrNoIPv30U+zevRvff/89ioqKsG/fPrMQuzyAR9YlzmMBUgUae5jYD8RSiIxi4SQ+KcL8pZ+an1zCnBG4OtPXIbteMa9rhYlbDW5yN5YMMDt+9h4zLK9Sm9u4oEypd4/NLaUF0EdDWDw0eMzCUGdkZCAoKAhfffWVvk5eXh6Cg4OxbNkyAMD06dMxbNgw/V+PHj3w8ssvAwDef/99REREQKlU6s81Tfla58lPMWz7h9L/h5sElHMxsX91nQEMXG7Tbj1PWBUEV69exbp16+Du7o5Fixbht99+Q05OjrXT6hxlTg25D/p3NV/spVFAqTEYor46lYR1+NL83OJM5mnaGYFAWWxWVe1hOfcqV5C4SS8ZdKpJWebX/elSKr49Q7/YdDkLGrrV/qxNlYEtDHVAQAD++usv/f7x48fRsqXBTXfr1q04dOgQDh06hF27dsHFxQULFizQHxeJRPjttxpc2JT3EDj7Jftq9mcB45DsAu0rTaB93hy0aiJCaJVQ348r0UDtX0RrVRBIpfQD7+joiNzcXEilUmRkZNi8Y4/dtngAACAASURBVM8NhADHP4Bf6mHWwymv3QEC+4EITASBWgm5yqC/LCjj8Boych/NIW4QEAobT7N7bSn8OlvsqkJNtzejRzNG+bJh7fTbh248xtbz9JQ55AV3ffmuq+kAgI+HBAMAemjzKnRp4mGxzboEVxhqqVSKFi1a4OZNOlf1sWPHMGjQINZrLF++HL169UJ4eLi+bPLkyfjpp5+gVlcyZHlVWd8JiPsUyHtGw44bB4bTDbh0AkGXxrU0G5h+AohYaN++PSdYtRGEh4ejqKgIU6dOxbBhwyAWizF48GB79O35gFDA5Y1gy/abS1zRtDE9UyCmkQ7VMkNcIktQalBOnmie/zX+lLyDYpkcP5+4iDfYBuRVWOn7x1s9MWDtWXytFTJeLhLWBWNuWttCfakDGrs7oWmDemZ1apwbuyD5+6dyhdwoNyGvAp3GW6zCFobazY2OZTN48GAcPXoU3t7eEAqF8PHxQXY2MzHKn3/+iVu3buHXX39llDds2BCdO3fGoUOH0KdPH7BSkgUhHAHYMNDeta1A1OfW69kb4+QxOhuBTu0qqoZI+8ffA16cXfXrPMNY/ZZ0U9To6Gj07t0bcrncLJZ6rSLnLr3y0Dg4lSUs+O7nDdgA/TdlOiNQljFmBJxQalAC+mdSQwgRCH6SrGKtSsoZmjoj31xF5OLIvBXm72bJ52qCg0gANR9mQs+RI0cwbtw4AIYw1LoooxEREVi3bh0aNGjAOpDKysrCZ599hq1bt0IiMV/oN3v2bMyZMwe9e/dmb7zoMb1O3aXiOawtwuad9qxh7GKtG3DpZgTPqjrrGcOqIFAqldi+fTseP36MJUuWIDc3F4mJiejRo4c9+md/NoYBTcKBqUet1wUsCoJGPoa0lJTQ5OFWlkKuLseMQKPWq5WIQAQR0aCRwDwiooKIQcr5Ul4SY27orSexPnKa1bM5Y99RLCrfrMbedBoPZdDwZyIM9Suv0EZ9iUSC4OBgbNu2DbGxsTh9+rT+XEII3nvvPcycOZNhOzCmSZMmaNOmDY4dO2Z+0JYvO5XRoOFO7LM5I7iwzrCtFwTagVcV8nbUJazaCBYvXgy5XI6rV2mPEy8vL3z5JYtRszagW0WYeqHcp6gs6G2FEqMXkcjErU1ZglCtDt7b1RFmS9kv0onMUZYLtVjrcSQQQQQKTjAPDlcKKTN9JQsOIgH6tvZBI3fzF2Q9R3Y1ypzeBgO06UzCVSoud57j2g5XGOqsLIPX17Rp0/D222/Dw4NpV9m6dSscHR31swcuZs+ejR9+MAqAppZr71kbCgLjGUFBKlD02HZtVQc6VZDOvVqWX3N9eY6wKggePnyIuXPnQiymv2BnZ2dQtVUdUIlpcKmcO2Kn0MGCZw3RYEq3Rjj3bh/0aOkFoenDfHMv/V9RBLmEVjARrSAQC8y/fzkkVmcEHs4SrdAxhyvonbtRGk1j7yKAFgQXknNx+UHdiNluCa4w1N9++61+PzAwECNGjDA796uvvsL9+/cZLqQTJ5onXg8MDETbttrZnEoGZCcCmQmAphqE8d1jwOVvgcxbTE810/Uqa9pUva3q5oWXDNs6G8FNrZ3FZB0ODztW9QEODg5QKpX6PASPHj2Cg0Mlcuw+D1DmDxQhBL//+xhR7fzgyBJLoUSmgrtxwcSDwM/0snWJxLKLpUBVhgBPTyRnl0BgKgjkWr2nqgwqIR2TSKYBRGAf9auIGAorqiaKEIgqGFbCePVzK1+m3aSe1q4w7rvLdX5RWUXCUAPAyJEjMXLkSADArVvcYQxWrlzJ2P/6669pVdATIxtO9u2KdbYsjzaw6kItCATArnGG4/W8gXe0nmlsoZ2fNYyTy+hUQ7xtoEJYnRHMmTMH06dPR1ZWFj744ANMnDgRCxfWUhesTPMY8CduZ2H+7hsI+pA9QmiZ0mTE5GwwpAvEJnaB7m8y95V0nCGVhjKfEeQ/RFFeFqCSo1BN39warbGYLbidEg74O8XyQ6umKi4IdC6nACASMc+NTeBHWzVCVV9yX7YC1rQGPnEHtrF4AJbm0B5Cf/8EfNe7am3ZA+OZvE411Gcxe93KknzKep3nGM4ZAUVREAqF6N27N9q1a4e///4bhBC89dZb8Pb2tmcf7cc2g293XqkSnvUkjPj9bJSaRg01DhstMhEETiY+9zn3APcX0EpahDyYL+S6uX0RwgUy3HpKC5tWggx4CEtwVtMePUU30VS+EylS2hipti7ToamEINhyzrAMX8KVM4HHzlRREBgHLEy7yFyQpePUJ+VKfVppFMWAxMWQOKYqGKuvdDMCFz/2uhXBvyuQoQ0tv2Mk8METbWjq2gfnkz169Gh9liQvLy8MHDgQUVFRtVcImPBIaxitL7WsBitVmMwI3PwN22ZxT0xu+pP0Ksf1mRNwVfqG2bWL1CIQlQwyQgsUDwE9g+gpuol0yvx3EIBYNN6qNQSicjx41z/spw8hoQuVDQBSB6ZqbKaJFxGPnahuT5gsFtVS/cbmZdVFzj1ghT+dQ6CKSPPuACnnDAW6Z05sYgt7/UrFLz7jJHPfNLhdLYJTEHz44YdYunQpPv74YxQVmYZCLh9nz57FwIED0b9/f3z33Xec9Y4fP46goCD9ystngdV/3gUAqyPoMrmJIKhvlMbT1FNI54rn14H+326U2fUKRZ767VxBA0AlgwLmfuUKOODWJwONSgQQgCAtl/1mJYRAodbASWJ9kZWDUIgB2twEwzpxpyVtzOJ9xGMHitgT35NyzApZUbNEpM3+z7Dt7AWELzCvU1k2dqX/x1b9mt4J3zALdCN2sYl9zqc1qsz9OGCpG1BS+0LscN45nTt3xm+//YaWLVti1KhRWL58OVasWKH/s4ZGo8GyZcuwZcsWHDlyBLGxsawJbUpKSvDzzz+jY8eOVfsk1cxfd+kfWxd4DaBTSpqiKCvhvoipakg3otDFQGcxTv8Mg9FVTYSASgYZzD19PN1cGYvACAQQAPj6dBJrV5QaChQxH9Wz4SAWIMDTGXc+jcKY0ADOesazD4pPUGM/jFUhRgsVzVavl5cdI5n7pmGapfWBNkMrd21TqtmIK1JpEzeJnWgbXE9tljLjGcGEaorTdFbrNv/o7+q53jOExSFEcXEx7t27Bzc3NwQGBqJly5b6P2skJCSgSZMmCAgIgEQiQXR0NE6dMje4rFu3DjNmzICjI7tbY02he8mqjF7+KSyj7capBwAAsZoX8VfL9+lCXZ5T0+mpLuQBIfSIRVHMkgGJYLFqOgBAQskgoFSQE/MZASWir71OPRIXNMF6rfHRm5lmdQFArqQ/R7kEgdYWIHUQ6b3F2BjX1SAkHuaWctbjqWYc65uXlSf8cnkxTXqU9wDw7wK06Et7FFUF4wVqTSMM28pSQMWdK4MLykE7K63nDQz4DHDUqjKNw4uY2uYqi24gZzrTrwVwDiF+/fVXbN68GVOmTMEnn3wCobBi086srCz4+RkMNr6+vnqbg47//vsPmZmZ6NOnD3OhjAUUCgUSExMr1BcAEMnzIH18FVxnCpXFCDLaL1GoMf+ncziUaFCLpTx8AFUu8yYQZ9C6x4tUMGRUBHwTE4EWMyFoNhXkzh19PblcjqwSCr4AcjOS4CZyQnFWOp7Gn0eg0fU6aW5jtWYRVjhsxStlOwAATYSZ+KyPH3DeUE+uBhITE7FWPRoAcESyGDojoun3I5fLkfAf3ZeCp9lITLT8wCXdvcMqACx976kPH0D5lPndyOXySv1WlUGlUjFcNwkhjH170KlTJ0RHR2P5cjq8sVqtRv/+/dGuXTts2LABubm5WLp0KbKysqBWq9GwYUNs3LgRjx49wsiRI9GkiWGdxpgxY7B3L72W5MGDB2jatCmEQiHCw8OxaOYE6L5pAgIBAEogBAgFmUwGlUpl8XvnWgnwtO0UeP33I/KcmsITzFFvYmIifAXuqK9WI4nl2lZ/a0LQ6NJHcH18Xj/6LFZSyNCe02bPi1C4BOBB9K/c12DBT+SCegDS289FiUn7us+ZWOgIFFfuPjT+rtSyQogBpGY8QZkq0a73tzG2aJdTEFy4cAE7d+6Ej0/lsk+xJTU3frlQFFVuNZMxjo6OaNOmEotaLm8Cri0F+r1mGDUYkxir3ywi9CjDWAgAQNNmzdDSh+lLv7KsDTo4/IsDmnAMlrpy9i0xMRG+AxcCJf+hQfQS4MfL8HB2gIcgjVEvyFUBIhdCSUSQCOjZwijRefzXNpAhCMROurboiJA+UgqUnHYfNe1DYmIiHqncAKShU1BTtGnDlZuYvpZ+0ZJJOftn0+ZgbhVoZjNITEys3G9VCRITExkhJWQymV1DTAD0Ysvk5GQIBAJIpVKcOXMGvr6+EIlEcHJywnfffYeIiAhMnjwZAHDnzh04OTlBKpXihRdewOHDzAi2ukVlkZGR+Pnnn+HpqbUfGa3u1T1RQpEYlEoBJycnODg4cH/vbB5CWrxGrQY6DIBn0GBgtwy4awiz0qZNG+BhA+CRgPXaVn9reRGwl2l8dRWq0KaJL+BMfy7HkvQK3y9FFwSAdxsE9JvFWadNcPsKXZMLsTbKaZNmLYAmbex6fxtT2XYtCQ/OYf5XX31VaSEAAH5+fsjMNKgpsrKyGNcrLS3FvXv3MGnSJERGRuLGjRuYM2eO7QzGOl0qm2EMoPWgWuoLZKgPc1WHaQYvum4ZlESELi0aY37fQLPjDJw9gYn7AbfGdDKNm3uZ+VMBeDnTP4kSzNG16YpgjZC5761MR3thCpzB/vnm7qRzTrfwZhGCVjj/Xh9c+aCvxTq8jYAmPDxcn3vgyJEjiI422Hyys7MZs+TWrStpwGTTswvE5osS2TC2B4z8nnlMJAZaR9MuneN3GcpDp2vbEFbeY0lh4h5dzwfIuAp80axKge0ElMp+qhpduOvK2mKeYWz2idq3b4+UlBSkp6fD19cXR44cwerVq/XHXV1dceWKwaVr4sSJePfdd9G+ffVIbzN0+vor3wKRH7IcZ44eGwlyUUSYIZbVGoIt5x4gtKknOgXQ64mD3CiUlblgx2vlzINqiskSfgGlQUM3KVwUhhf6O6pZWFmPaSegTA3RWtZJNwMw90bSUd+J+6Hp4O+GhAxz33F/0+xqLFDP0ErO3+//jn1391VYnWmJEYEjMLSFdYNpVFQUtm7dij59+uDu3bsYNWoU/v6bVrNMmDABb731Fnbs2IHu3btj5MiR8PWlZ2dpaWkYNmwYANpR4+OPLSVQ0X7XElf9okRaJ16O30Bn6Ow0Aegwho7cyZbg3RidB45AWHljr6kgqN+QzhEA0MHsKolT7m3AgcO++NJcwLcd+7GqUJ3hzZ8RbLZCSCwWY8mSJZgxYwYGDx6MQYMGITAwEOvWrWM1GtscnX//2f+xHzcxkGlYvhoNRfDZkUQM32gISiehZFAIq6CCMB0NBfbHuXeZMecviruZu7FyjEr6w9xfukRpMEh71mMXIACwe+aLVkf+XBRz5VyuY7Rq1QoZGRmIjY1Fr169GMciIiJw8uRJjBkzBg8ePMCIESOQl0er81544QV9ljLLQgD0qFwoBryMnDaEooq9pHUj+9Yx3HU+zKHdRntrnSAqMyPIvkO7XP53kFn+5F/Dduxb5b9e1m3g0Fx9fCWxIh8oe8ped+ByqzkkrMLi4l0bsTojIIRAIBBArVYjOTkZjRs3hqtr+WL19+rVy+xhmD9/PmtdXawWm9G8t+XjB+cwdimW9HS7rhr0+bqVxxKqDApBFQSB0kQF1W8pxCYjWQHLCMTXnZ6tOIqFjDAQAGg9sJMhAlJGYfmm3s4SMZzLEY6ajV1X07B8hI1mcxVkaIuh6N+ov91tBDoiIyPxxRdfYPv27SgoYOrk3d3dERMTg5iYGMyaNQvXrl1DcHBwxRogBHrrgFcgHTtIIKJVQ+UVBi5aNa2l0a1YAvT/xLBfkAooi4HCR7R6szykXaT//2XBFqgbDHmWY4HirnFAQRotoBq0ABEIIehkOWprlWD1knp2Zr/VBeeMIC4uDuHh4YiIiMDp06fxyiuv4NNPP8XgwYMZ+VefG6xN57QJsP+maD0/24xg97V0/fbW87SRtJviEl5QPzSrW25Mw+Sy9ZNFByoR02V925jbcciF9Yx9Fwn9WfpxGomrTkSgl/VKdYTRo0fj9ddfR1BQEKP80qVLek+mkpISpKWloWFDC7muuZDlGWawknqAe4AhEUtxJv2itKZ3d9W261ABYZmoNWbfqoBfvqlgGvAZMGIzs0yv3iqHrr9AOxjTKAFFEQSEAhpYztVdJdgE642dtmuvhuAUBOvXr8eBAwewc+dOLFq0CCtWrMAvv/yCnTt3YsOGDfbso215cAbYPky/G0/R020RLE+BN56+X7V2o9fQ/8u0IZzbjwE6T2at+rhE25fZRnkStLMGJwfzEbzpravz1oruUA3xV0zYM5O2jbAZ0usqfn5+es8gY27fvo1Ro0YhJiYG48aNw8svv4wOHTpU7OJcIad1g4USrYMGl1OEDp2KR1KJVKMKC4sordF2ONBxHPsx05DXliAaoFib68G1EsK0vAhYXpHXttiuvRrCoh5A5+XTqFEjtGhBS92AgIDalY9gO9MAqJsJNBNk4j6xPP2VK9UQEhFuvvAqulS0XV0qTJ2L3sDPARfDNPTA4L8x4ih9VUonr/3a0TrLW/uAUlovuiSmLV7wdGa4llKEKeF1rrzC6gjwZUJDN3pEWa60m7Wc+Ph4s7UL3bp1Q7du3QAAM2bMwIwZM8zO8/f3R2wst8E0Li7OsKNLPmTqLGA6k7SmIjJdyMjhfMDK2S+AyP8rX13TcBhaV1GEzQSumoSdUXPn9jCjJMuwUKy8aWUrA9sMvToX7z0jcM4ICCH6F/6nn36qL6coCirVc5DH1BIWHhJdvJYo0TWLl4gI9MKj7KeQCDQQ1atEDmfdFDdHu+jMRP1DiThyGZRoPS20AsTNyQHz+zHdVotkhpHVnmtp+OIcfY6lVcKVRepAf1+yZzFlZW1Ed++ajYIr+NsaZ7ObnwC8VY6cBiO444UxyL0PFGbQ27r/OnQzkH6fMA2xjToD6gosAPxlDJCqtT842DAiKJsgqIg67TmBUxB88skn+hd+SIghIfbjx48xffp02/fMlliIIviHhh6F/0c14awD0B5E91Ppl7lj/Urox01vJhNB0N7fDZXl8L+GBUfv7buJpFxaMFQwAnW58KgngVAAZBdVPDwAT2XQzrxMVRamQt6a0G8catj2aGIwHlsiIMx6HQDY0BlYqzWAGztDRK0ybEucgWFGAePcAyoWYoJogD+0buCVUW+VFwGLILClcbqG4BQEnTp1Mov/c+fOHfj7+7Om23sekNfXeiVY0EXqIn0KLNgIOvq7QUMRfbx2b+9KGGHbDmPum0zNW/m64m8qEIc1JusTdC+ABtzxnkoU7DM2W6iGHERCNHJ3QlpezYfoZVvNXuvQqXTMRqr0b0t/Bxa8h/zDgEYhQLMI9uOWsOZwsT4EuGq0SE0lY8YW6mqiFjOOxVXPh54RWPoNU86zl1dBEDyVPcXHFz+GQsOhlmL7zHXJWMzG4sXVnPXHzgh00+Fb+zjrKLRmE7OMYUY4OohAEYK4W6kAAJFjJW5EFxPhwbIuYJTyE7ypmscc3OluzB7cvtdcr/s7mebJb6oDhZrCoRs1m9RcKpUiNze39gsDvSAwuV8EAhBCkFuqhrTwAThdHAlV+SBsxoOVpW6GdKoAHUk37wFw9G1D2XI/esGaVxAw6XdD9jCjPht1jP7HlWxeJQd+5EiHWgXV0Lp/1mF/0n4cfXCUUb7l5hZ89fdX7DMChQ0T9tQQFXIaf94fMrFcu/DEeDGLCbrQDpYEgUQkRJlSjdSsPEBiJUk9FyIH2ke5NAfwa29xKv96b2P3OG09Nm8GIz6N/Q/TezRjlN15Urm8EtZ4Fm4Lf39/ZGRkICeHDh+uUqlqJLe2zdtVltDrBvLFTGGgUQLF2ZAWPoD/P6uAXq9wXMBoDUJFMR28lGTpQ7MIubyUMq4BgQOB5r3Yj+uIpwMs4vpWoOc75sfjPjUv0yFlicZaTkTaF72KUqFEWYL/O/9/aOrWFD/cooNgzmvyKvtouTw3PSG0cLSle2s1USFBMGfOHOuVnmGISAqoSoF/tgNDN5h5TmSL/JBL6JsqTHgHmzXsqy6FQgFuPSpCbyGtgqlfzgV2ZgQOAG78Yh7/3YSQAKMRnF4AsD/MMiKBCBpsPf8QW88z1zd4OFfAM6QCxHRsiN+uZ1ivaEMcHBzQrJlB8D1vAcHKzfm1wMml5mkTH/0D/DbWsM+2AjjhVzrERMt+lWvbdLCSn0ovaAPgWGiea0RPeYyrIa/SbpmmxmUdl77mPlfqzn3MCg5aDyAVpcKiM4tw8fFFwLBcCKUCCqxPN6EgVBYDqZeAJi+xXzzxMLB3ItBtNjBoFXudZ4QKqYYGDqQzYqWkpNiiLzbncbclhh1CzIzGr6g+RM82tMtoX1E8XhKye1KcvZcDpYZCW9A3v6CyXgQ6HamVIFZdmhgJAp1qiGMGoYIIYg77hqODbSKKiAQCaJ6FaUFdQJYPiBzNX66mK9TZ2K/V0VuZTZab4if0f0UxmsbNZh57/bJh29LzMWQtMPhLoNd79D5b+kqutRM6qmD7ctA6adzIvoGOXiFmx5VctzWh0CRuDrAtSqsmY1EXpWrX/Vz5ttL9sxeVuiOmTJlSzd2wD6V+3Qw7cZ+ZeSnIVRoM6WhIzdgIufpt4yQsOhaI99Mbpmnxyovu5Wkl45FIZHSjW5kRQCiGCMyZzph29Iipe4tKuLmWA5FQgDLlM+I+mvE37VpolvCnliAvpFUhpi+/F0xGpZYEs7XFZhXpC8C0FQD0wMbHaFZk6f4OnQaEvWYI5VCPxXvJ2NV1pNFirma9IPM0DZleMU6knAAAHE85jk0JG82OF3OtmaLUkBrPgvIemNcxFgCVSLpjTziHolx5AgghKC62jdHRrpz7EujCXP0phhrNvQ36RoGAfpha+7li5agO+hATvvUdkVVk5GVQWUHwVJtWMuOqxWpiIYsg4BjVURDBwUQQTO3iiQ9Hh6G+1Da667g79DqFzWfuw1EsxJTwZlbOsCFbIgEAjs0mAI8UtH+6DbylagxKwz6DFIlR1qAdnHNv0fuWgsMpq+DhNXAFcELrNKITBKbu2M7aAUev94Azq4DOk6xfVyAAWg9hf6EaC7X2o+mZiJs/0G4kUhITORPtlIfssmyLx2VcAvWmaZgNK/eYohiojC3RTnAKgr179+Ltt9+GRGKuVxaLn+N43MEjgdvakbzOte3F15F6Nx4ZT7zRxMsQr18AgiPzesDfndbFxr7ZAxn5Zdj/zyP88V+W4ZqVVQ1ZWVI/rmsAdl9LZwah073UOF5uMkrImBEs6BcIQGMzIQAASdl0yIEVx+jFce393dCliafN2mNFJWd8nx73DwB/HACGbaT1z7UGYkG1w+KFw8aj65Vv3ni9C9eMQCcI+nwA9F5cfkGsUQLZ/7EEtdN+Fod69LXC51Wq62x0b9SdtgtwoOCaWRaZeMnpfhO1Aoj/GegyjXncwtqlZwHON3q7du3Qtm1bxmIyHV9/bcFw8zxRSnuYILA/1hWOhUfJUzg7Gm50F8gQ3MiwsKtdYze0a+wGiViIP/7LQhLVGL4SGeo7V/KlpwscVp89lMVnw9vhnYFBkIiNBYFlbZ4GQogFhtEgnYjGvu5unx+9g31zutu1TfwyGkg5p9/1uE/nkkZeFQICVgOhn/2Jnq28sWZMp+q5IOEWBALjUOqmM4LqsuEYz0aubgYadjCEdg8cACT9ATgZPQ8VmY0l/UH/P/sFELPOUK7rey8Wb6IqIFPLOIWAInsgHH1OIEGRA9ZfjjKxW+g+55lVwLnVdDwmiYshoJ7KvmlTKwrnW2XdunVm0RN1nDlzxmYdsj1GD4TOwCZ1w/74R8gpZi4qKQa7f3Jka1+09nOFGkJk1KtC4gtvbYaqSb+zHhaLhGjgYpp0g+PBmnQImHgQaiKCGIabVGyL5cQmfDeRGWnpno3WK1jESAgwsGUcmnLwtESJ/f8w4+2sP5WEYUY5LSoEocB1Dzjl3zGqZ/Liry67gKla6tAbgKNWnfri6/R/1ypGuTWLe6T7LNz38vcJ3+N2Lu3cIVPL8FTGnqNg7929aP9Te5SpyrDt1jbO61Eq2q72v1yud51ZaEf6ny4Q3smPDUIA0M8ICCFIL0rHswanIPD09ISzsw1jeNQUxiMlXZArEXuGoxfbcvv/+rlJ6QilbAtOysvgL4Gpx5gJRqzBNSNo3hto0YeeERh5DQntIAjaNGT6cfcPtl246wpz8uMqpUK0BWv+vId/0wtwLSUPHxy4WbH1OYQq5yjb5JrVNSJlSwvp5g+5eyB9Dw7dAAxYXrU2KrhSWKFRYH38eoyLHYfM0kyE/RKGsbFjWet+ce0LAECJqgROYm6VLiVvxCxwf4H+L9VqCEx/M91vYjJTOF7PGQ8cxPrvf9qJaRh8YDDSipi5ymsam2Uoe2Yx/gF1gkDsCIlIiNm9mC/+0Y1ywYVULKIFQVXSITq6AE0qqELRuY9yGANVYHoNiexgKDVtQq15xlxJTdMk2gm1xnJE1pe/vYSdV9LwdyrHalo2LKiGGJiqLmwpCO4eBRGI6Buh8yQ6DWVVEIiAHaOAb7SeULpnluNe3n1nt367/2/9AXAbgXWhJB4WPoSbI3c8L0pl5LK9tBBo0oPelmhnmPICkzPMBcF9BzHe8fHCMP9G+u//ehZtn3lQ+AAz/piBycfYQ8/bmzooCIweTu1NoYQDlBpKH0lTz5lVnKNJihAIQUFQlRlBZdC9BDiMWG71pIwZAcP11EaYptFUWXkB2p0a8hoqLadL7ehvL5X/ooQqnyAwXVfAlc6xonCGc6gG4e+uDfR47ksg+SRtOGZcm/13sa/PvAAAIABJREFUrMzoesYfM/Tqow5eHRAseFd/rPT+QoAY1FOJuYmGE3WqRtNQGLlaV1IjO01SoFHK2XvHGNXfjHsTV55cwT/Z/6BIaZsV/xWh7gkCYx2nVm/6sICW4s29Xczrf+oFqM29e/74L0s7I7C3INC2R3GoO4Riho2goZvtXdZMZx01nqSmo0l4hepaQFVBypSG36HawrPc+g3IK0dSJGNPnqInwOae1dM+h4ecSFkNs64ZHLnMrcwISCWFkM5GsKn/JjR0DIa6uA3KUmaDUjLXMvyb8y9ua7Sfz1H7jjAVBHsn0v+NFr+Jw2YBAFoqlRaT2eTKuDUP9sKqH2hqaiq2bduGR48eQaMxjHB++OEHm3bMZtQzChmtfcGnFdKfi/OlSakAmLvRCgUEUkfbhG3gpL5Wd1n4iPUwEdAqqxebeyK0iSda+9VHYj573erC1A5RI4uMdXGbAGDAp8C/xhEia2ZGkG201uTgjUcYEeJftQsWZ1o8XBjQD27pJ+kdY3VY7IKqtWsMVwKbiia1Z8OFLT8wYG1GoOIaFJlQZuLCWaKijbkOQgc0dq8HWTy7mmb5Fdrm8Q8AB4lBEKikXnCQm8y0jASEitBCIVnrgv+k5Anr9Tkjn9oRq4Jg3rx5GD16NIYOHQqRyM6jX1tgHDgri158M39fIgAHM68ha4hAob6rnZNUtImhp87+XVkPC8USiCFHz1beeL13BYzQVcA0vHWNBCc0XpPh3ABJQw4gMLZmw6UbR2T95vR9tPJ1ZbgjV5hi9heJDgeZkV7cOEJmdXpOcYRDEdgqobtGbXVGcDD5IGt5dlk2CCFIKkiCu6M7xh8Zz1pPIpTo1cLfvtoFs3ewr4QuFQrhrls8+jgexMVEsBdmMEJrGwuoG44SzD/CnqJTQ2p+FbxVQSAUCjFx4kR79MU+NOsJnNZ6Ndz4BQAgo2gBR3G9wExHOzd24mhfJaSXCFyd2D2ObEajTsBHT9mNdgB83OpBQxF0tuPqXlPVkN3FACHM1bICAdT1jA2WNaOqauxhGCQkZZcgev15PPh8MGtdpZqCRCzEzYxCxKfnY9JLTc0rWQmb4fw0wbCjVgClubRACBwA3PzVcMy7Cmtxje+7IV/Ri8f2TqyeGQEAdHoVuLHDsC/LM2qTXRB4OHogX2FucD+ddhrb/9uOtGKmDcHfxR8ZJXRwuxcbvgiRUASlmu7/QAseb8VCIf54cg4va3tCjFNW+rU3JOMBgFf2QqUxuI9ObOQHyPNYr6t5BsKhWFWe9u3bF3v27EFeXh5KSkr0f88tL5hH+tSlp/Stz6EaWtuOqXM9OAdtL8yHB5UPgb1tBACnEAAAoUiMxvXFkDrYr1+mjlOcAtVWyAtp9V3/ZcCHOebHaygg3pUHuagvZY61mn/AjHvfrw394rmeSr8kYr4+jyWHONJGWnGDzW8x0rBDaYBN3elkMcazpQ8eA7OqsA5IlxBp7A4gdKphBFxd33HTHsz90hyrM4IGTuwxtOpJ6pkJAQCY2WGmfru5G52sSqGhBbFAIMCpRb3wVr9WZueNbeSHT708cUOrDibGtqfMm4Zt/65Aq4E4/4gjkQ6AYIVB+zDl+BTOevbCqiD49ddf8c0332DEiBGIjo5GdHQ0hgwZYo++2Z2uTbUrIl+LYx6QFxhyDJv5Dz9j6jKRg7nroK2bNLER2N1WrEuQ7hYAiNl02DUjCEoUajRyt6w6DPCkj7/y/RVGOcX2JVrRhWd2eQeYd4PeIRqgRGtT0KV0BGgffXEVZrGOrrQ7ZRttiHbtQIhw2Q4qiunvV/wE2KILm20uCIpURUguYIbA7uDdAQCw+Bx7Iq3hLYfrt488PIL0vDJsPvNA/5238HYxywMOAMUi+nXpSCzbLHRhsU+mnWQUN5R6oblbc9x8+Qx2P87CK4W0HUdFqSzaCR4UPED4rnBkFNsu1LtFQUBRFNatW4czZ84w/v766y+bdcgutLYiyBp3MS/T+WEXmqwKrIkZgSWEYrsLghq3EegCgNXjMDbW0IxAqaas5oDgclBQsrngWvtdBQKDV4+xukFnwJx82PL5lUGb7SyvFbv+u8K07A+0fxkYtZXe3znO4CXFMiP45M4nZmWvd3yd8/LtvdpDYHQdsUCMuTv/AVB+bzfdEy8tTAZeZcl2qB2YhDcOZxQ/kT9Fk/pN9OuXREYDlNAdoZzPzc47O1GkLMLJ1JOsx6sDi4JAKBTi888/t1njNcad2Iqfo9CqhkpNvASetRmBUGT3EMzmgsCuzQPn19D/TX3cdSE8agiFml6b0qOlF2cd41wTTd8/ot9mfSlZi8sPGO5HNp19kx7mZVWlcRdgfgLyW42pnutJ6wOjtgAt6CiyzFmQuSBIlzEHZt0admO8gMUmxu01vdcw9ud1nlfhFK6MX8EkyY8SwF5ZGihCsY7gxYpi4C79O79czFSxL720FCVKZtmTkifYc3cPAGB9/HqbDbKsqoa6d++OkydtJ4lqhMr4levc8UxHZVVZWWwLhGJ6cUtuOXzNqwlT1VBl/borzIHZdFIQHaYqj1BtBMganBFIxEKUKrlf4FxRWhmrkgsfAdsGAw9O0/szLej4dTNU08FA8962u1c9mlT/Ncvh6WTqDgoAQpNXmtroeX0/7H341fNjHB8ZOBIvafN0HJ7LFJSqovas7VIW3JE3u7vhUy9PHHt4DOnF6RgVOIpxXPzgDHBkEQDAX8W8L/Yn7cdnVz5j9FkXPwmgVUj7krjzrVcFq3fGL7/8grlz56JDhw4ICwtD165dERYWVq6Lnz17FgMHDkT//v3x3XffmR3ftWsXYmJiMGzYMIwfPx7JyRbS3VUngQP0m7u95wMATi60klNVNyMwNdg9czMCB9o4uKGz/Zo0eS64cnlUO//uYu5XRfdtAxRqDRzFIsSnMcMRTOneFAAwuD39Uto0gflbjRCeg+vmUIM68sJXdLary9/Q+5bCl+tXnj9b8ZUqjMjB/Nkymnl+eP5DrPl7DSpCPQf2GEZSsQitfF3Q3p/p2ttUM5u1/lEX7hhsuVoX+3x5PihCwd+V6WLqYDQoYXP5OPLgCD6/YtDCmHpDtfIwN2JXB1bdRy9fvmytCisajQbLli3Dtm3b4Ovri9GjRyMyMhItWxp822NiYjB+PO3be+rUKaxYsQJbt26tVHsVYuT3dMjbjOs4lTkY7Rqr0NKHZVWxMfoZgckDlp1oXrcmKTDykiDELuEVBCZtXHpQQyslzVRiun7V7IzAlKVDg7F0qMHVMKodc5S6VrKJjhy+dzLwyh7zWail2Ek6VciZ/zHLn8dUohb86w/dP8Ra7iLhfo6lRgmkjo08pl9QRhFipt4EgOMLeqL9T+bX2eFWH+/lGQn37m8CFzcAXq0A0KrjVdfoHMWpRamMc0k5FjeeSjuFJS/RaXULTGIadfDugMSn1f/OsToj+Pfff1n/rJGQkIAmTZogICAAEokE0dHROHWKuYTcxcXwo8lkMrMXis2Q1qczHQ1aCZlGAEdxOUb1BWm0GiLhV2Z5eZb725N0I8FtZSVqddL5hconEK82TFUUuvvJhi9BtYZCQkYBq+5WqTEXBO9GmYd2FwgE6B1EG7odjDXQSSfoJCemgoBjMSEAg2pIUche/hxzvewxVBoVMkuZ9/UPAw1RDhpIudOxdvI2ZBbwd/VHa0/ahsQlCCzBmPQO+Iz2pPJqhVPOTC+xAFdmitsikXX1nO5eelLyBOvj10MAAU6POY2/xvxVoT5WBKszgm+++Ua/rVAocPv2bbRt2xY7duywcBaQlZUFPz/DSMfX1xcJCQlm9X755Rds27YNKpUKP/3EIn5NUCgUSEysnESUy+WMcwkhuP0oHy08JWbXNF1yU5Z8gc5OcIP5uVWyYiSXoz+mbdsK437fT4yHsn6BXdqWy5nRLRMTE23ebqMmUXBLPQ4AyA16BdnJKUb9kSMzMxN+AO4l3YVGWk1B10x478RjJGTK8V6ED3o3d2F85jKFCqVFhhHdT6NegHc9Jet3QilofXdbQQqjPPfORYiUhTAWs4kPHwEwDxsil8uReDeJNXVjSZkM6Tb8LWzxWxt/jtm+3riQ+huQapoiEsh7ZFioRZVQnP3IT8tHPswXnhUVl0CpUFeo/xrQo2jjcwIKnqLAiSlww8XhSPdLw5WMfXgiFkNgNGAo8Q0DYD5Yy1fkIzExEWOu0gZ4AoKcFHp9TDaybfJdWxUEW7YwgyVlZGRg7dq1Vi/MNkJiG/FPmDABEyZMwOHDh7Fp0yasWrXK4nUdHR3Rpk3lVkYmJiYyzlWqKeTLHiK8tT/atLGse3MWs48qHVwalKs/pm3bgxaPDgLdvrNL2y5nCgAYfKHvyF0Rfy8fn4zpZLuZ3j0vQDvzbuDthwZGnzExMRF+DenVxa1aBlY9WQoHCT/ROXYLBC5o06YN47tWkxT4+XjhYGR7nLuXg15dzX3TdTS5qwFSSiE2yTfdICAQyLnLKOP6LRMTE9EmiP0+dnF1s+k9YMt7TBU8EhfKuNNrdmnbBUGPgnA3/y5a+rdEm9ZtAKM04P2b9MeU4Clow7Gi2vlyMRRQsvefI524RiBA1ovLmOdccYCEIlBqjWY/Rv2IYN9gfNr2MxyLzcG7uReh0T0LHV+BS1kuoGaftZ9RMh0C2pjc25X5ri0Jjwq7Efj7++PevXtW6/n5+SEz0/Ahs7Ky4OPjw1k/Ojra7t5JuhWwjqbhpwHDghkd2rhEeqK0Aqu6FtJUF//P3nXHR1Wl7edOTa9AAiQEAgECBJAuRZpIUxRBWAEXC5ZdWcGCFVBZUcSK2D6XtYEoiEpXUUAQpbcECTWEhPTeJtPv98e5vcydmRQSluf3Q+f2M5OZ8563PU93QXdpY/LwS+b6x9edxJfHy5CW04AymULbXKPQUXwVcwQ0TcPudMOs16F3fAT+NVrdCAB8yEjG2eN289rA3oBNFrP/78MQqTW1ogZvEEOqdtwXPM8LLQJbIKUlOVev8D6f6PsE12SWnleJJ9efFFVludy0z4sVNwC3SSzIhEH/RAcHn0PsG8P3I5mY/ITbEEBYVu/4ENAb8ZzNjFaB8nnxgxMfcK+f7PukT2PzB5qG4NVXX8Vrr72G1157DUuXLsWsWbOQlOT5Sw0AKSkpyMzMRHZ2Nux2O7Zt24ZRo0aJzsnMzORe//bbb0hIaIAyNA9gDYFifHD6GuA5D6ydicOB9sOAW32rXGhwxAqkM32RJ8w5SgTg/USrUOWKnYbtMhbc/Phq+WFpjuDIp8DKfvU6ArYPIELSOPbzX0Sy0Or0roQqNMCIVqFmeSqxPBM49xO/3UKjaoSiiBFgJS0TmJp6D7QkTRZzfgGeyYRbZZLeNnkbvuz7JQDPTYxtQ3hN8PnfnMB3x67gYhGv10DT8hJoKcyUuKLIBcBYLRGw7zoBVVHtAQAvDxY3upkZNTQnRQFx/bjmvxlWFx7v9zgAoF1oO9lz+7Tqg3t73OtxbPUBzdCQcNI3GAwYM2aMV+WjBoMBixcvxpw5c+ByuTBlyhQkJSVhxYoV6NGjB0aPHo01a9Zg//79MBgMCAsL0wwL1TdczCylquLl6ccTGAXc60djWmPi4i7tcwDCV/+fUUCvu4HJH/v1qJcndcfgji2w7Md0VFr55GaDKmVqEp0JPAJ7DbD1cWazfqqpdl7eiXNB80HpF6FW0ivwwg+Ee6ak2kOppwQWuws6xrhd6TgDcRfXyunGZ8pj5DJQevLZGMyEWyuuPzB6sdfjaDIwBgLGQNAP/QZsuR3J4R2RXsEXZ7QLa4cavViAR2llT1EU/v7pIZzJq4SRSdbmV1rRJZb0KrjctOb3tLokBcYonjvITVFwhPBylrXOWrx+6HXk1uRiVvIs3Jl0p+h6I+sRCMdnDgWslZjYYSL6tOqDNiFt8NHJj/DhCT4v2z/WQ2FAPULTENhsNsyaNUu0b82aNbJ9Shg+fDiGDxfX58+bN497vXDhQukljQp2taqq66vzYAh81FVt0mB7JLJ8UMqSIDrEjBkD2+G3s4XYcbqA2+9rNYZPEK4CoxUot4UewTcz+f1OG2Csu2DP9kuEQM4QmoYTV9qLjpXUEANwWy/vZRvtLjcoiryn8y1GIy5zg7xiyJuVvU5Pypz1ZlJJNad5N4TSTH6ndViCyBCIzmEMKKVSnrn3HAkdsiv/2Z8eQuayiQA8Vw0tHLgQi7buByUxOHYKqI3qxm1vy9jGNXtFmOUVdHpGF12UAQoIB6wVoAC0YYzKrORZIkMQpKoIV7/QDA199528k23DBi9WJc0ALMmU6mrAUydmI/2BfIcfE6+NaWv3hsJAA4tv6ybabtjydcHN7/5G4bjgs2C7cgHfQmYesDub3JPS13ATjRQjuqjnxZTA5gj+7/dLpHEs83fxCZ4WJ9xNVEpImyncjOfXJ6YP1k5YKzrmctN4esNJlDGGlwKFCgsfp5eGaFySWOWBjBIcvFSKK2XKms7Tu06HvXgMKL34uJXSwQWaq/Mvt/HVYUpayHoDawgEzw8IJwZboCdtkuQcA/QNrzAIePAItm/fjm3btuHKlSuYO3cut7+mpgahofUodNHIqLI6oKMoBJsNcDGzlMf44B0fARv/Id/f1KgllJByl3fnsU0r9cApH2QSf6UcDdlmLLQy4Z7Uv2ggvB1QwTTb+WEIXG4XBq4diKf6PYW/df0b3LSbEx7RB2UC9dBDJ/wW0rSal+pF0ldFPKa5gjUEekrPJYVZlNa6sP7IFUS0KwSCmZ6MN3cDTLpRGqJhMbwz6dv4/hjhA8opVzYELAyhRD+5W3Q3nC45jVqKwifZq7HzxB4cu+cYiiz8QsCzRyD4zpqZZLOtEjCRhaWBEv/tKhrJmKt+Y1JSUhAREYH8/HzMnMm71cHBwY1eBlmfSHlpBwKNeqT/e5zAI/BgCHrPIPHl7U810gjrEQ7PX24ObHWRIoWzb5DpF7sa0iUQ3FupeosdyzvdgfhBdTIEp0pOweayYenBpfhb17/BLqB5MIScB2Wonx8s6xEktgoh3cVSeBUaagaLFB/AJoL/yq1EQRz/t1vw7Ul0DiV/B0tFR5iDD6GsrCXKLJUwFY/CPYPVyf7KLXZcLqnB+iPeUTu7HRHQm4twS8ItOF1yGtU6HQ6UHwZAJmu7m/8+hJnDZNfrmRyBLDQEALXlQCjpudJLDL2TbhwmYVVDEB8fj/j4eAwePBj5+fnIzMzEoEGDYLfb4XA4YDQ2wyoEBrUO8ueotpEPOcSssYJqsmEgDWgImXBgPYF68Aj0eqkhaEiPgLn3jXOVV8rCUsJqQb22H9VRs7aLc2JSndyQpNdgsTMNQMzEFRXsm2GlKN4Q2NSqjXwJDV0jYOP/G47m4OtfdyKUWYd+e5SfxO0VvXD0yX8hZfFesl10CxbdSHIAVocLOkpcwZaWU4Hhb/zm/RhcZA5gk9FLWkTCpDOjxmXByPUjManjJO7cIIN8vugaloBplVW4xymormPnFaf6gk3qITQUNJcOGzZswD/+8Q8usZuTk4N//lOd77sp4+19haJtVqNY8wfrD1vp1UKAID7pqy5BeVadGeMMkjCbrUENAU2SxGOXKh8XJoTLMvnX9ZAjUBISeW8nIU1kJ5zZSnKTHvDanSmICiITvcWh4kl5E/a5BugkhHBzCxTy3XJa2sNeNlB2XqhJHLJmPf7sUgvcNLDib72x+oEBeHh4os9lzdacmXik+9OIDSIr94smEyocvMtmFXynlD2CQCwqKUN7oUvAeuBOcWXZtsnb8GDKgwDU1dfqG5oz3OrVq7Fu3TqOF6hDhw4oKblKpGJ1gNXhwi8XxVzfbGVHS5UaeA6NxYFUH+gzmzS7tenjgyEQ/Cr2vaV9+qXfgWrl5Kg03/LJngwvx+APaHhMjhtUEm1OdTUob1BQUyCKCbNYdzgLx/Nquf4UL2hlRJh8QxxW/q0XAKDWoeYReDHJsx6BwbM6WnMBR2vO5E1qLz8CW/5k0TkxYfLfcGYJqfRhIwAhZgOGJbVEYgvfK/5oZxhGtrkdKS1IjqK93QG3gHFox+UdAIAvxn3ByV+KwIYuhYsTJm8AyaKiXVg7PNLrETw74FkZjXVDQfOrajabYTLxK2aX6+oLLfuD/Ar5KrCUMQTaLrxksjE14WS53gAMeoS4nd4K1AiTrpl/eD436wDwxa3Am50U7y/NESRfXkPI+vxpVnO7gRNr1UNcNO3ZW5MagijmB+rBFVd+jHj5ePOGm5FeKm/Xr3Bl4/kd2Vxlil/UGsyjqqwqRtybe7LGotd035/fBPHladI0Jq3cEcKhkIsqZDx+VuSHXaT8cFzcm9EqlJQ+a8HpdiM+LB6BbjfyVIgq+8So0L+zq3+hR8fSpjvlvSYmvQkzk2fKcgYNBU1D0KdPH/znP/+BzWbDgQMH8Pjjj2PEiBGNMLT6xa/pBaLtiloHyi3kDxARpBF3Ff74nr4EPH5K/dymAp3eP8lKCe2tDELair1vyh8r8QgeNjDyiLVysi+PqMwFlkSSiq3fVbyU0xuB4rPKxwBetpHFgIfJ/330CKoccqqOs6XkuW4bXx4anPguzLGbOLuq1a2qDJr5bx28UNY4BjQBRth6ACfRqFNeTNyc3AoOpxsOSRiSNQSsYTYwSfR5o8Xd2YdeuBlL7+gBNTw7nrCUsvc30IDN14Q8m9sRLlxYL8GDXnFjQfPdLFiwAMHBwUhMTMSXX36JQYMG4fHHH2+MsdUr2C8Fi+e/T0NpjQPBJr02DbXwjxcUBQQ2gx+YwBDoa0tkxGUiCFe8ucc931eoHCWtcVeAjWZ+AL7G5Q8KOpzLs9XP8wSpRxAU5fNY5uyYgyFfD5Ht359Hmu8MRY+gW1RP/pGhp2FnEr3+2YF6MASsBxUgr2dvbqBpGtlV5O9vL75Z8ZzW4YGwu9z444KYYbaw0oqf/8rHsh/PAOANc3Jr/jvMhok8eW89GcEau5P8bYwq3FVfTfjK0zsh/xMm8jmPQNsQuNw0V9zSENA0BHq9HjNmzMAHH3yADz/8EDNmzEBaWlqDDaihsGCsmAe+zGJHtc2BkIBrq+aag0DEvtPWycAH3qnKaULoZajU7p9eMpZ7bWd1mHyNy4cLeNyVQjm2avk+KaRuNSO0rhWmcrldeOvIWzicfxgH8w5y+98Yzou9XKq4BACotYTgkZTHuP0UaBy5TGiRyyx+qISd3+H7NVKwurfXgCG4UC5QLaSVPXe9joLN6ca9n5Fyzn+M6AgAKK624+HVR3H0MvFGDUxFW1gAf59WCrkFKVhaCidTSKGWZ+4UodDdzoINowq/k5xH4JmGhKZpPLn+BHq8+DMONpDok6ohcLvd+PHHH/H555/j4kXS1r13717MnDnzqlND+AOjXoeJXfhsvstNY/fZIgQYvUm+NaNkMQudgfvy6dga58IzKif7UEIhNAQq4vDCpjIn6xI7ahTPVYWwlFVp4s4+KN8nhdQjCGPIxzQ8giMFR/D5X5+LWv0BcBUjQthdbgSb+MmEdgXhgS8IZXLqFY0wmxIO/wcAEEMJQmlzj4pOefPns/jnV+J9IrDhu2ZiCLIqszD0m6HIqsySHSuqJUn5/oGPyY6x2HhCHPMPDTAg1GyAzSnOYbEVbcLw5VO3yIWCpGANARsaKtMrzxlmvQejEt0JCG0NjFkiGJC2R1Brd6HDc9ux8QQhuDudV6k5Xn+gagheeOEFrF27FgUFBXjppZewcOFCvPLKK7jnnnuwZcuWBhlMQ+OuHvwP4+ClUhRV2XC5RC6ALUNzKh9lodPLpf4+lJfcAeBDQ1GJpNrIE4Q0FL++CJz7WfG0Q8+PxphOIbDRjFGwe/E5i54jWE3nnRAfK8sEirWp0NGyCzBIUOocwsTzdyzyeBkbijhScAQdwzty+5XKAgHA5uK9E0rPv8+68CyZIVglhrUBOo3hNt/ffQHb0zypzzF/T5XxNjV8cOIDVNgqMPGHidw+N+1GRkUGDuQSxb3f/iK/waWTe+Cx0Ul4cFgH7txnx4kXJDRNVv/SZkaDIK5v1FMY2CEK/dpHaY7PyHgSSglpFosGLfKc2DUFAU+eATqN5vexhsOiLpr00195om27l2y2vkI1LpKamorNmzdDr9fDarVi0KBB2LFjh0dNgaaOmBB/m+CaoUeQzhjrEm+kNFn2PaO26LldkjhdO43I9EnQKiwALYIMxBBQ8L7LmYXQ86gUrPhoGljRi98eqED/IcS414C/NgJVubyHoNG27xQ82+K0IKVFCh7q+RASwxPx2A2P4b3j7zFD0ePUy2NRYuPHRxksCO64HDUXn/Y8Lg3ohWKIlA6YsR5Sz83tptUJEwF+xdnE4KbdmPD9BDzc82FMTpqMowVi72b+7vnYmSWWtXXbyYQ9cyBPVT9rUAIuXLgIc6S4gWvOsA544+ezWH1ArBcsTN6f/fd4rx19qUegBIM/tB7s32fnEmCYsubA4+vEssDd2oQBzvpX21Nd6ppMJugZFyggIAAdOnRo1kaARaA3oSApmmNoiMWmR70/V2/0XHJachH49l6vb9falYsBOiZJ7WvXsppB2r5AvJ04Qvtej58CFhbxVUQa3blCQ1Bpr0Svlr0wIp48R6hBa829CyFmAxLCxDoaOlMp/EYw+Y1tdd/I76N0hDZCp0dJNR9G0AwTNFFDcKXqCnKqc7BkPwmTVNrJ+wg2BuOnSz/JjECvlr0AWl7inRAdjDZhRoQF8pPwsKQWqsUfBkHXu05HeV3e68kQsKR2SRHaGi3yG/ve5zEsqaXvz/ECqmYsIyMDkyeTpg2apnHp0iVMnjwZNE3UfH744YcGGVBDw+lP52wdG5CuCtoNBrL+BCweJqWLu0g4iA0N6QyAQxDCyfiN9ExUZBE+FA2lKCluzVvJb1QwlT8VOUB4W+ULhJAaJJeT9EgwMXQO3tbVs257aGugk3L1CfcoQUhJP2rsAAAgAElEQVStxlGDMIESlU4YJqT9q/EutZbiYvlFZa55UxDOBo+HI0vw0xQ8M/UK781kltSgR1sPeYCmpp7H4GI58VJbBRGjFx8aj3Nl50DTNBbsXSA7v2tUV+yT7eUhTP56Ktn1r5zXc2hocqfJuLH1jWgd4j3dOD8gwYLkzHag6wS/xlcfUDUEzTUPoAVPcT5VXFHXS22yaNmFGAJPcpWrJ5PJ4rYVZNvtJLF3t5u8/vJ2kLiOf8Rx5lpB78b2p4C2fYH/jCTdz5Pe83yxtAeippDEyqXwMX9zwmTE/eV78ZOlkJuIpJDyCAlzA8I4cEK0+oouNPlZOB3vKx6bv3s+jhcex+GZhxEgTWi7XbBI3npmaS1sLqBLbCju+/wwt1+T0K+JGoIaJykcCGQ6n1l6BotTOY90tvQs2kYMxo0dlekWwgIFhsDDwkBKf+ItWI/gdK7cA6Moyj8jIMXRz66qIVD9FbVr187jv/8pNMfQ0CAmdj7wYc/nuey8R8DqMp/fAaxlKawVJpsuE+X7FNCWliQ0Wa/g2BfaF7scZCKbvoZsq/US+Pi3+SqAggM0DuWrqJJDHBoCxBw2Qk1cq8FDbwYAC3VJcf/lShK7Lq5ViPW6HOjcRpzAHPHWHox9dy97AoyR+2CO/R42pwNLDyzFhbIL8vsATTY05GK8PYfbgQpbBawuq0dytbNlZ1FjdyLYpOyBCZkBFPXHGRh85fxgwBqCz//M9Ot6r6BQNnwsi1SOtY1oeKqQZlgOU794e1ov7ZMaVl2lYcCuYn99Ufm4p1zAqQ0kLKSGOz/Rfr7SZybkvtH6TN1OEqoKYqiE2fLTthLNYR89AlYq0NPE45JUWwXoQnAim5SCCkNDZTWeG3zG9VCmQWZDTSVWhZpwtwPBAVKOJN7YmVruQEDsVpgiD+HPnMP45uw3mLx5Mo7kH+GpkNkxNlGPIK2Y9CFlVZGyUbvLjqhA9eqd1eNXw2JzIcgDSzA7WXpqDvXXIzDoG2khKCmoOMb0P7B9EQ2J/zlD0D6arzBoFxWEO/t4EjRhwMbyghsmUdMgUJsgLzCJOFEtvWRSTvvW873NIeJtpUldKa+yViCUwxoip12Z8bSmmOhCM4IdKGLKRU3BYuoEHw2BDWzXr/KPu8JWIdIaAIAPd+Xijg/+gMXuFBmQyV1HeHxW97bK5Zsso+Ss7bM47wAAKbG1lABHP8f6h2+UXXepuAaGEN4L+SmD16S+7+f78NZ5hoqDrWBpooZg3dl1ou1yWzlubifO21SdfYl7nRieBLvL7bHQg43/mw3q3wd/cwQmPz0JnyFkyAUfxp7SJw7bHxuGn+YPa7BH/88Zgg3/GMy9zir1srZ92JNAv/uBeakNNKoGgNoEueZOkngVTtS+eDy3M01WTwvCHkqkcFrEbm4nSVa/0hJYf4/8eCWTVGb7Fn56hrnOBQQIJlgfDAFN09jD2HRL7jFM2jgJh/JIiKi4thjHC49j6DdD8empT0XXnThPQkOVtU6Ydfyzh8QN0nyeEiLNkdzrd4++yx/glOJcGNBBvkIe+eZv0Afw4TZjlJgg8HgFQw/CGoImFhralrEN355TXmSEmPjFhSXrAcAdAEv2vZjXZx7PFeRhZc7O8Z4MQV1zBFIs7FIPjbVjX+VffzgIOMVLA7uYBZJBT6FbmzB0jW24vhBVX+uOO+5QLK9q7lVDLUL8+HEERQG3vlP/g2lIeCrXtFdL3FBmwpr+FbBupuIlHNjJJSgKuPllEnpyOwAIVp9uF1nRA/jBNQST9QqMpk4rSVYDwJmtzDBowjGUPIkYgta9yD8AiGBKNN1OQDBp+GIIhCvRReeJ9u2yw8vw/aTvMXL9SNG5oS43qrgJgPxMhry+Cy7UIpRpRo3QaNhyq/wNMip4am6RroEX78VR0QvGcFJbzgrdy8AagibUCGlxWPDs78+qHhdq87pqSCmmq7orpiXdgu8YOUlPieBMpjHU7IXX4CvUrusZ3lNxv0/oOR34+Xl++9eXgR6Eepr1CPw1YL5A1RC8955GVcc1gP/O7qd9UnOFtKtYdMytTLMQ0137vsKSN/a11CNYOx248AsAAemcFHYFyomSi8BPz5J/ANBlAikZTRxJzi/PBrIPALGCH6APalwci6UAF8svyhqaAKBKr8OOrBzsHP8LFqaTVThZmfITlkmw+txyxxaU28rxzPZVyKX2AICIr14IlyD0VGARVFaxn+PwZ9TfBKVu4HuH9yYv/r4ROPZlk6KYYKkiWIxvPx4/Zv7Ibet1etRkzIcuQCwdeSqnAi/8cIo5R3tClFZS3TekPT77IxOAuLO4sbDxeA6Kq22YM0xBowCQ5+rK+VDhip3nAfhJZ+4jVA3B/0JlUOeYJqwrUFeoibIA5Msn9AhYDiJv9HCF92Ubs6SlnowRAAAbVO7pUAjLSVvtQ5myPGMQ8TC2PUG2hfQSJu9lRMd3GI+D+WKOIjftxr0/3at4fmuXC98fUuclEhqC9uHtAQDzu0/G06eJIVCruqXLLgGM1Ou5snNw026ShGbLVqOUJ41OrULgSWHXyP492vYl/xoZq9JWoXNkZ9wUd5PsmFTIp19sPwQaA/H9+e8BAOGmKLhtRrhthM9pTLcY/HK6AJW1/CLDG8oO9m/y2OgklNbYkNya99oaLekrwPx1hB5F1RCEtAJCYoBqwYKAphu9UlHTRKampmLatGno27cvUlJS0KNHD/Tpo8FH00zgq6Zss0JQFBCbonyMdok9ghK2/FDy5RstqTi6c5WI8wZ6Zh3hQRvZBpXP+H0Fb6xWQtIWx5xjDCA5B5avSDj2IHWBcilqvCC+WzZsGQDggfIKIDAKrT2U7pkU4tHtQ9rgw1Gk6U3NI5Dah3eOMmFHNm/DhN8m217GY3a+MzwyyOixia3M7qPmQz1jxbEVeHSncid7bk2uaLttSFsMjOW5r/Ly+IXnW3f1wjPjSPxNyODqzURuYs55YkxnvHJHiiic5Cm01BBweaOHSVHiPAEAOGobjFNIDZqG4OWXX8ayZcsQHx+PY8eOYfHixZg9e3ZjjK3B8MbUntj+2DAEa4nWN3e07q283+0UT6YRDG2C9IdilKy2e95FqA5YsCvQKvGPXIioTj7QX0vDVSxbqDGQeDAGBaPiTZcygzeOvKF5jlFnxI6sHPyjvAKIbM/p3iqeq5JEbB1CjJNastgtMbhbM7aS6im2+Y8psz1OJ2Gzm9dCqLa5EGJSz3FdqLmAKikXVAPC5Xbh45Mfo0KDuwkAyqxiI5UQliBS86qu5T/LKX3jYGLobYTKgt54BHpJ+CexJS9L6ZGXSQP/HNGRi9ULheo94cXNvICVR6NgkzSqZR9EldUPCvM6QNMQ0DSNxMREuFwuGI1GTJs2Dfv372+MsTUY7uoXT8ibrnWoVQO5XWJqZy60I/ihdJkIdL/D8/3ZUNJ/RvH7ysVUwiWB7eDwhooh8w953oAt1zUEkuRx2WX5dfWMEFMIWrdIhpkGcuyBnN6tEG5HBFy2VqplhWxMV80jkKK4tphUT/2X8bZUOGjS8ypRZXMhgOIrimoy5ovOKddSmKtHbM3Yig9OfICX97+savRYVNgqRM14Mz8+i4uF/HcwpS0xnqHM4oz1tlb9zifWs8u0q/ykXkNcpPehQ8/31cHppkHTNJYOXYq02dqaLF8f4psgPYrKSH4z+P4h1NgaVxJY0xAEBgbCbrejS5cuePvtt7F69WrU1PjILX8dVwdq8XPaBVwWVPIoSVre+E/t8kMlxsV3xeEouz4Et2KFxkABfD4B2PRP8T5W6JudGEsZJtV7NmrfTwHD44YDAHQqk9ZzA57Dja1v5FbmFwqqcKFQLoBTc+FZWDKeUC1VZA2B2uSoGTDwQEZGgQZNA7bCsajNmQ63LRaW7L/j3u73AgB0DZgQXXN6DXZe5gnhFv5Byid/ufwL38wG4MU/X8TZ0rNcBzFAegVCjGGwZD6M2pxpuFxixcz/+4s7HhlI/tarmAIO1hDU2Pl7ZBZrzztSr8FTp7EvYL0Bb6I9LITJbak2gggdmYXUTQzP0g0zUdnUPIJly5aBpmksXrwYOp0Oly9fxsqVK7Uuu46mgJHPA3EKxGZuN7DvbX6brdWnKF5sRm/WbkiS5gZK5ZQKDn0wKmnBxOZPSaPQU+g9E+g4Uv1cD4gPjUewIRDzyuSr5j/u/gMzkmeQSZyp6qFA40qZej+EUo6AXKdhCLQiFAqGoDvnwdKodbhhLxkJZ+UNAABXdTd0jiQ6vGdKz+CRXx+BRSEZX2WvQqnVf2bU1w+/jvm/zVc8VuPi/0bfn/8eU7dMRe/VvbHxwka8sO8FzhC4ajvAWcmHhDbevhHfTfoOM1aRJL7OQ2PYir/doDlGaallgJYMrZdgJ/WOz29Hbrl3lOrCOL/HmH+HmwiV+wimjNQQgE/3KdOTNBQ0f5V79uyB2WxGWFgY5s+fj4ULF2LfPk9cgDz27t2LsWPHYsyYMfjkEzktwWeffYYJEybgtttuw+zZs5GTk6Nwl+vwG4GRwJxfce72H8X73U6g2+38dj7bKEcBk1aS8szYHrxwhhqkUpUKTKd2QxDK6RBgwpukIe/B3b6/D2G4g+0hCAgHhvLa2alFqUj5IgVH8tUJAt20GzrocF9FFXZm5aCTnUz4n479VMQw6i1ldohKjomlodBKFk+srkFXOzHCtcKVrEFuCP4SEJ4JQ1I3dW6JlqFmjg//3aPv4o+cP7AvR/4bvWvLXRi+brj6G/ISbtotqwI6Uqb8uS/6YxE2X9yMXy7/ggCduErPbNChY0RHzogB/IpeamQzl030Sk1QmgdQM9a+QmhgzuZr52GkiwCvkr86HfGyXXYUMXTj3z4i7zBvCGh+St99951s34YNGzRv7HK5sGTJEqxatQrbtm3D1q1bceGCmBwrOTkZ3333HbZs2YKxY8fijTe0k3nX4TtcAZHASEEX5M4lwOlN/DabpKUoIH4A8MjvZFXqSXEJADoMA2JS+FJFhRATrTPD5nSh65Y4bLps4AXkfcEQwSqUrY1/Ngu4+SVu9/5ckre67+f7FG+x9MBSrD2zFnpKBwpAK5cLYUznZnRANPE6WPoNxhBQzJT97HhlSU41EjMd87OiaRpbLm7B8cLjouM0gDuqqrGsqAR3VJEJfkD7eHD+lUeeehrhgSb0aReBaf3i0CE6CHanm4u/tw4mJbc51eJFVV51nmyfLxA2xz2/73mM+naU6Piqy6s073GlRDxJs93TwoQ8O98KJ94HhnaAFtgVu0mhsuj7fw7G5rlDZPt9gTDM441x6fDcdtG23YOojQiBUaguzsHv50kpdb+ESI0L6geqZTPbt2/Htm3bcOXKFcydO5fbX11djdBQ7fr71NRUJCQkID6eVKRMnDgRO3fuRKdOvMDzoEF8i37v3r2xefNmv97EdXiB4QvIJPrjAuDsNrLPFEISx5V5ytd4U24X0Y5PdinRSugouGnA6nBjxa/ncftjCqEqxWcLjFDrnoAxmBDPqTRJKRK4MbA6rfjm7DcAgHIHv5pbXliCH8e/iA7hHYAN9wN/fQ888CtQQ1a7rCFIiApC+pJx2HA0G5/+kYlLWrFq5mNzup1Y/OdiABAlF2n+FIQKJsG1YaGYXVklMwS94sKRUVSDKpsToIAAox7f/5NMbK9sPU0MAVPBxeYIhMR5xbXFuOW7WzyPWQPVDj5Xsi1jm383MWeKNtnJPi2HrzpiPQJhE9WiW7tp3npqnzisO5KtaJz7tKv7ZCqsEPNEY6EGr8tBW3YR9ck0RjMZ4MEQpKSkICIiAvn5+Zg5k6cdCA4ORnJysuaNCwoKEBvLi33HxMQgNVWdq2fDhg246SZ5I4oUNpsN6enpmucpwWq1+n1tXdEUnt0q4wSEjO4utxuugGiYqkmb0rnzF+AKEE+o5lu+QOIOUi6sNP62FivMlkpkpKcjJOcc4iXHS0v4+4UaXEi/mAWtb8/lkR/CERQDh+B5yUwPQFHOJRQrjOPrM19zr9PT00Wf997ivbLzASDG5cIQqjfOnDmDxKyjMAPAf3nyM/YnaCnNR+bFcvSLAPpNjMH4LzK450hhtVpReLEQALAmdQ23/2jaUQQZ+OQ9e+9oFz9hW3QUjptN+GzXQjySyC++Fg2LBEVF4lKZHT+VByGz1sk9u6qiDDanC7lXiPGqriITdn5BPtIN5Jzz1edFYzycdhghBglxIIOc2hw8nvY4EoMSMS1uGvpEkHh+qd233MInN3yC18+9jos1vFSqjeaNcJcWZlRWVSM9PR0PbuCrZi5fzoSpWhyS9PS7Yf/ON8a4sA5AW11Fg/zOigr50FxOdhaCawt8+k2fvXAJxioPTZ4M4uxuOGv5z0ntO1bf71HVEMTHxyM+Ph6DBw9GaWkpTp0iNbFxcXEwGrU7UJUSZWrWbdOmTTh16hTWrFmjeFwIs9nslSFSQnp6ut/X1hVN4tlZ4pWRHjT0QWEAs9jr3K2nnFkUyUDeXUDbvsrjP9MKqDpPjrlOk32jFgG7/g0AaNmiBQCy4osIC5XfIywOqBT0y874Fgmd1VevLYP1aKkwDtMRE1e5kpycLPq8p30xTfV+SUlJRPBmnbw0Vcd4BANSuiA+lBKs1DO450iRnp6OmPgY4CRwrlqwsmtFITmGnO/+kwJLExTv4MNpn4WH4ZOIcDhLfofV5ASlHw3aFYI+Kd2g01HoC+Dw3nDkukzcs+PzzsOdVo42bdsB54EK5rOObhmN5ORkbLm4BS+cfkH2niOEDK4CvPnzm+QdWjKw7NwyzpPJq84DTsjP/2fvf2JQ60H4+49/BwCMiB+B+X3mo2NER3zW+TMcKzgGg86Aubvmok/ITDA91wgOCkSQ2YDk5GSM7u7gSi07JSYimUuMq3/Ows87OTkZyQDuGKadTPYXJ6uyAJBwTWKHDkiOC1f8TZ/OrcStK3+XXd8mLh7JnbxoftxVCjCGe+bAdqrfMX/mEk/GQ9PH2bFjB+68805s2rQJmzZtwtSpU/HLL79oXYbY2Fjk5/NMiQUFBYqax3/++Sc+/vhjfPTRRzCZruFO36aAAQ+Kt521QOFpflstNj1lFS90I4XBzHfEZjIJSoY0CxBHl7jX5nCil3Dzy2IjAKhXBLFJ4jCxGtTWjK1I+SJFVL7IIr8mH28c1sg7eah/Z0ndWtVeBJbGAuneqfbpFCqjagVhMxIaIvdu7eQNgYEGnMyHtD9vP4wRhBlVmAClQXNVSQCQ2JJ8LgUVJMOQWZkJgI/pbzgnz+fRHgpY2RwDC7YvQarRwCI5Khk3tLoBL3V9CY/2fhQrR61ExwjCnx8ZEInRCaMxPH440manoYPxNgDAgrFdoKMorsnqYhEfahOWe/757CjsWTBCdayNCWGOwOXhO7P20GWuxPTlSd2x6VESwvNYPspgZ3oBUMRP1pNv8L5Zsq7QbK19//33sWHDBrRoQaxZcXEx7r//fowZM8bjdSkpKcjMzER2djZiYmKwbds2vPXWW6JzTp8+jcWLF2PVqlWIjlaWobuOekRke8/HtZLDitcYCE9K7gkgi+HxEeg2UFLaCgB4TtBAIxTOmXtUne9ofhrw03NA/zniW/3+nGi7Yzgv4vHSny/hj1y+XyImKAaLb1wMrLpNcIUHQ8AcM/9nKNlx6jsg+TbV85UQaY5Ema1MbAh0Ou5TMYbz1ApJ5kgcc/LxcnOrHXDVSoJttNjQtAolYRSLXRyDZg2BiN2Uu4Xye7a77Dicf1i07+YNN8Pmssn0Alj8mfsnRsSPQLewbpqrVIvdidbhAXh0ZCfsPVcEJzNjFlfZkNgyGJ1ahiBe0ADWphGUubyFsFHN7cEQCPsY2kQEcollm0M7R/DAF0eQyUWPaL8V1fyBV53FrBEAgKioKM0uQgAwGAxYvHgx5syZgwkTJmD8+PFISkrCihUrsHMnqc5Yvnw5LBYL5s2bh9tvvx2PPPJIHd7KdVwVMHTTWDeLcBuFtyPhpQd34+L4ryGs5vPYjDNkPtCik/rxoCjgzv8DzJ4LFaod1fjx0o+wuqyiSpekyCRsu3ObnBDNw3c5wKDD2jk8Hw6qCDHY8UVjcGKx+kJIOFGzescH88RkdxQAjPk38HgaVkeTMTkj5ESPxghxWaa0JJVNYpbXiksa2feuxK+kRo/90C8PIbcmF2a9Gb9MZdhjGUPya5acuRUg4u3eosbmQhAjNxlsNsDCls46XOjbLhKf/L1fvZV71jeEyWJP859w2RMeaOQSyzaFZHF2qQU9XvwZGUXypsUg2BCoa7ymMk2PYMiQIXjwwQdx6623AiDVREOHDvXq5sOHD8fw4eK65Xnz5nGvP//8cx+Geh31AkrPU1THpgD52q3yHsGKxFRkA6nf8A1sbfvAXpnOpgcAwDN/iiCcVBcUWArw9N6nEWOOQZeWXbj974x4B2a2LyJ+EKGzBuDJI6BpGoM7CGLptYQvJ1KDrFBoCGKDYnG58jLKbXwvBA2AatEFGPIYAKD3rR9g7J6ncLZUroHsdog9ZVYPhEVpDQmJrdkdCCQIrqPdoGmaCxUBQLvQdsiqktAZCMDScdtcNsQGx6JDeAdcqvDc2NQ1Srm0VglWhwuBAkNQU+zi9nvTI3A1IZTA9FQJKvzbBJn0nD6CUmho88lcVNuc+PboFTwzjnyOm103YpJ+Pw6Y5yLsPxbSaNYI0DS/zzzzDCZPnoy0tDSkpqbijjvuwNNPP90YY7uOhsCjjGg7pQPu2VR38RJp93FIjGhT2Agl5YoXQWEc3nieaqh0VvK0zABCjIIkeDWfu2I9giu0PJGXEBUIFAgMpR89ECzt9U+ZP6HSTj4LGpARvYaZwkTGghueU0wTIs0RsGGKcov4s/r01Kfo+aVYOOWuzncx13gOU9zekTQbahmBce3H+VTeuPNMIU7lkM9g95lCXCquQftnt6HGzhuIpgphyWiNXZ03SHheoEkPIxNSsit8949nydliT7pJaDOMYjrDX40jOh0NDNVZ4P777wdALNyECROwcOFCLFq0COPHj2+02tbraAC06EQ6fBcWAsHRwICH6nY/afdxkHgFK0z+eUpSgqJA0zTya8gkbXFY0PPLnvjiry88Pr5PK56uYFz7caJjO7N4XhyhFKK4A5qMSafQBRxo1AGfjOB3tNKuZweUk8UAMOTrIXC4HKI+AhahplBFQwBKvpIUGoJBieTzfmJMZ9l5Qvww6QcEmwgTp5aBXTJkieqx6ADyvL4xffHGcP8bQIUkbHanG+GBXmhhXEUIQ0NCjQQpwgTvo2WoGUamr8MpcCMKK604mV2OX9NJmXFpNV/okEhJenrsVcDB/6vT2L2BqiEoLfWfk+Q6mjgiE/ikLDtxe1LF8gS9JLooqTwSzjlFVTYPdLwU+q3phzEbxuBw/mHOIChVvQjRIpBfyQtpFWpd4uY2s9BgtRO07TMD1MONr50jsdrJJ0UDpPFqT6pvonfCT9SfjBFTq9Q4ahhDIDYFIooLAZ4a21G0LQ0NsUlMvY5CQmAC1NApshPf8axikAP0AZjUcZLMkE3oMIF7fU83oi/dMrAlvAFN0ziTXwmrhMW1RYjYk0xu3bRFomoEhqvSqu4RvPEzH94LCzByfx+hN/yPr47h9g/4IoZ1R7LR/lnSpPe+U4Hx1xQEXPrdN21xH6GaI6iqqsKOHTtUL7zllrp1Kl5HE8Hgf5HyTwFvj0+QhoYkVT8OwUqooNKG7FIL2rfgOeIxaiGw6xUgJIYrAf0t+zeMiB8BAAg3e5ZbZMMcZr0ZvVv1VuTYkWHqp8Af7wJ734DQI3BDh5ecs9E5NgwDqXT5xC+VFVSBaMXeehB6RPfAqRLSh+OknXBQgEky2YaalCdCJy2edKSTOL/ipNEltAsu16pTdWuxogYaAhEo4Dn6YtwXWLBnAV4Y9AImJk6EjtJhaNuhaBfWDoNaD1K8hxS3rtyHv3Ir8foUwkr77zt6AABemtQdc9fy1BtNXS0wLor/XDxpVEjBehION/87OHpZWUAoNMCAPKtC9eT+D4B97wBT/gukTPX62b5A1RBUV1dj9251grDrhuAagTEQGL3I/+tlhkC8LeVhl3GuDHsKGPgP/N+Zr7hdVqcVP2f+DEAS0lGAk3aiS2QXbJi0AQU1Bbh5g3KZowjmECCaqVBy1AKVeWhJVcIFHVzQY1/nZzGw4FmgupC/JiDCa49AOFlTFIVu0d04Q3C29CxqKQqHHeIwkHQV3qtlL5wuOQ2ngL9p0R+LROEugO8xKKq2AhocgRwrqopH4IZbNI4+MX2wcxp5nrDaakyC59JxIdgc0Rd/EgMVFmAQjYVFUw8NdY0Nw84nh2P0W3u8Ux5jwBkCJ7nmh+PqYqNVap4Gq3FdckH5eD1A1RC0adMGr732WoM9+DquEUjr/rUMgbSMjqLgNAbg/RPvc7tctAsGinw1Y4LEyWcpXG4X9Ez/Q3SgL70ozET00WBuj4mhfQsPNBIGUDaXMPY14OfngONrgNs/0LyzqPlqeSIMsXxj0J+5fwIATrvE5Z4ltTwVx667diHMHIZh3wwTGYKNF9R1GNYcyMKtNyrX3d+SQBZtnGCOSrKY005uAJzOIwbBrEILHdjEq4YAEvMHPPcRtI8OQmaJBW9MJYl6vY4CRQFOxiN4fN1J7x/Y/0Hg8H/4bae8J6S+oPpXr0vFxnX8D0GaLJYYhmrJKsehUHsnTZJ+d/47rspHq8LFTbs5o2HQGfD8wOe9GrYSoV40RSbnYLMBKM3gJTiFAj2F2hwvojCPpQSuCl6piq3rfz5YXHY5q9ss7nXLoJYw682gQPn0O9xxQbnUMCGM5A6kHsGFsgtwCDQlaJpWbgD0E0cvy/OMrHRkZDD5ntw7uD32LBjRqM1T/oLVPPbkEdQ6XJjeLx539eMbAWka2HJSXc5ViNUPDIAljMkL3fQU0Eno4TbcnLx6IDUAACAASURBVKz66S9fvrzBHnod1xBkHoHYMMRHicsflST4HFKBG5B+AEB7QeKknaJV7N1d78bfuvzN4zUESoagghmjU9TqD4MZMDAtn/8dq3lns+QzOC2gTvnuPKF172kQ5z6CjcE4Ousofp/O89ToKB3XQKbUISyFvWSE4n62q5n1CEqtpcirzsPkzZPx9lEiUJRdlY1qRzUXkqsPTPlILmnL5gIGd2yBT+7pi+cnJCMhOlh2XlMESzPx2o9nVM+xqJTCZpZYcCJbvOBJaSvPfwWZDAh6YDMw8W0gNBZIEnzfGnBxrmoIOnf2XI52HdcBQG4IJPKW79/dR7S9aNMpSKE0ybErZ48lp1AOZwhX5A/0eACfjv1UfqGCRxBvqsHILi0Jx8toAfWFIQBowxCaeSHULkWJXj4xBFDyfSa9SUQGJ/QIhALxS4cuVX6QOwBHZx1FqFGceL2hFRk7K9rzztF3kF1FvJS0YtIn8e7RdwEARbViwRl/UFRl46pgPOGW7rFNtpNYCcKvjNoCpdbOd09LseBbPiw0oEMU1j8sF50x6XVE8Kn/A2SHkARSQXWuvtB8/grX0TQhTRYHihlOw4OMSH3pFjw8PBEAFLn8lQjjdmdrK5n9cvkXHM4/zNE4sIgL5ZXT5vedj/6xChoICpxGLagKfHbfAESHmIG4fvwBgxnIkq9uvUV/q1W2T+9FLJ6iKC40xhqCmKAYUTmnFCa9CVWM5sKKkSuwd/pe3NKe5AiGtiWMAJ0iOiG9lHg8rNFkieLqA9K/8VCGdVNtgmwuMAj0oJXCQ3anG043rfo+hftv792GazYTQmYYhd5yrUKfST3huiG4jrpBZgjk9MZhAUb8Yzg/0bjcNEqqeS9AGBrqFCHmG2JXXmdLz3K9BSye+O0JAJDp8N7RSaEWWwqFaiTqpqf4DZ2gjiLQP2GTXlbyHhcXl2JRiLgZLTRP7hlJoaN0nCFg8yivDH2Fk6Vksf2xYTDpdRjTUfyews3hiAzgxz4ynjC7tgxsiTePELrpc2WEKrtWSVTITwgrgB4ensgJ0j8/4erQsNcXhAykTgVDUGsnYc9Ak3INTo2dD4ve3b+d6H4sDFLjYBSEVq0NRzehaQhuu+022b8ZM2bg1VdfRVmZcj1scwVN05i2ZRp2ZKr3T1wreGjHQ9hy0TtaZY/oPA7oNIbX2Q1QnjQjgniD8bdP9qPf0l+5SZ71CBbfuBhP91emL5m6ZSrGbBCXLbYPaw8AMOnExsiryheJIfgw9mVg2JOCmwg8hrb9fNZaPtRhNj7LI3kOU1gcbqzgq4K+ys1HdJ66SBM3BErHhcYqbaTqJtwkjyt3axOGNhEBcEgmJxGtBkgyXUfpRNKZhZZCWJ1WVNmJF/H2iLe9eXsekV9JPKD5NyfhufHJCDDqkblsImYNUm94ay5gZUuFlUNpVypQUeuAhdGWUPMIohmOqpOLb4FORykyNMg8DSEHV7Gci6q+oPmLGTZsGEaMGIE333wTb775JkaNGoX+/fujRYsWeO6557Qub1Yos5UhvTQdT+55UvvkZgyby4b9efvx/D4vK2w8wRwCzNoAhDMlkqYgz+cDOJxZBprmGRl/OP8DAKBDWAfZatdTjqBPDMk/+EV1IOmIDgyUeAhCSm5jANBWkOtwykNZUgQ6LDACRFu5TW/EW/n4borNDozw7rNnjSUrFSmd3Lkh6nX47VINaJrmDGGwUZyEpSgKAfoAETU3QDyq3OpcdI3q6lOPgBpmf0r4rHrHK4vfNGdIK4domsZt7+/D3Z8cQAlDFaFmCA5eKsXgjtEID1LvmUhqJf0e6oBb3yGvyzKBqnzZNfUBTUNw7NgxPPnkk+jSpQu6dOmCxx9/HAcPHsRDDz2EnBz/xbCbIlgRDoCU1l2LoGmaSwwCkAmr+w0u+alefji+R6xou9buQoWtAj9cIIbApDeJiOIAdUNgc9nw/fnvAYhpJlg8k/QMvp/0vfp4JQYnKCjY43ER1vHSrSg6C7wUDpwRi5Xjd0Z7Q2ck4bPCv7hDFAC06a1+fwZl1jKsP7cedped85pM0lAcO35m8rG73FzVkpLRCDDI5RLHfjcW+/P2Izmq7qEbYdftwA7XnsYI28BXbiHhTLZB8nReJW5dSbrapT0RgxJ5skKjSpnsnKEdcO4VFR63fvfzr3+vu8emBE1DYLFYcPIkn+1OTU2FxUJWN3qFaojmDGE9+4K9C67iSBoO6aXpWJPOS4KyEoN1BvsF9tB9Ky2rc7jdIoZLk94k8ggMlEG1OiOtyDN9dt/IvkiKTFI/oVV30abeLDUECt/tuYw2wPkdgI1pCPtgAPn/ya/l5wPEQ/pLwSCpqcEJwBrBBXsWYHsGMTRqhmBcD6Is5nYTyu2BrQcq0lZI8ylCCOkl/EV2GZkbXrszpckzivqDnDKSS5n7NVlAWe3yPpcgSY5AaBgMCnkBAFh4azfPFVQPMHoQbfuon1MHaOoRvPLKK3jhhRdQU0MqAYKDg7F06VJYLBY89FAdmSubGISVKixl8LUGpYn1dMlptApqpbiy9hpsUstDrbO0q9TlprmQByCP9bcJaaPqEVypVm/V9wp6A3DbCmAL0cfQB0omTaUuzhCB1KqlRCySY5eIiySOIPQVAkS5XChlF0/SRjwP2JW9i3utZgjYCcbpdmNI2yEY0naI1/dnIfXG/EFuOckPtIvSDhE2RwSbyd/vZHY5gCjUOuQLH6kBFBoGpQSxV4jvDzyTSQoX6lm4HvDCEPTs2RNbtmxBVVUVaJpGWBhfqjdhgnoZW3PE5399zr3u1bLX1RtIA0JKYgYA07dOR7AxGAdmHFC4wktM+wI48ikQ0131lACjeMXjdNF479h73LZRb+S8sl4te5HXKnYlp7oewpKCaqWYDiniYyy1Q5KAU0s0UVJk+c0iKlF8va2aF+1h8GN2LpzsPGDwT59bzRCwIQu350Zsjyi0FGqfpIH8SmL8YsPlIahrAWwTWGvm/SkZAmmO4LZerbEtjdBLy6qCfIGf1WveQNMQVFVV4f3338fhw0TLdMCAAXj00UcRGtq02QLrCo9hhWYMtpFIihpHjYzi2CdEtANufsnjKVIVKofLzdWzA8QjaBdG5Bpnd5+N9469p+oR5FTV3RA4HDYYAXzqHIcJLSXx9Lb9gJEviOOzwnAO7QYOfsRvh/O9CwAAew0QJhaCD6Jp3rD54BEIwdJpyPYLPAJ/0S+2n/ZJGsivIJ5UbNi1aQiGMD0Rswe3B2DnSkaFMEtCPIM78Z62Xtc0K/Y1R/X8888jODgYK1aswIoVKxASEnLNVQuxGNp2KHpE94CO0onIvq4V2Fw2kdi7NIZc53CLBqQ/EGktdrg5HC0CWyBtdhpXvSI1BBW2CmRXZmNLRt1LXzcdyQQAuCgjJwLPQacDhj8NBAvCZRQFpBCVL7idQJGAasAhaRqz1/AlqkEKITcN7WU1qBlqNuTgUgnNXSquQVGVWIS+Z0tlBbO6IL+iFqEBBsLXdA2C/ZzZj5n1CAZ3jMYX9w/AkE7RiIsUh8VMggSxUhNZU4CmIcjKysJjjz2G+Ph4xMfHY+7cucjOVl5VNmfQNI204jSEmEJgoAzXpCFgE44AMCB2ABYOXCg6vv7s+gZ9vrRiwumi0SG8A7ctrWihKDnp2lN7nsKEH+onJJlZSMJQNlrHhVY00ZVod8NlB4QdzU6JIXA7+YTzw3vE6mbT1wAR8ahPcIZAhRBt5Ju/of/SX3F/50X4ecrPSJudhtXjV4vOqQ/lwfxK6zXrDQCAjisfJZ4XK7gzb3QShnduia/mDJIlfUWGQOIRLBjbBcuniA3y1YCmIQgICMCRI0e47aNHjyIg4Nr7Q5fbylFhq0CoKRQG3bVpCFia5smdJuO/Y/8rq9lXlEqsR+w6I45BO91urpN4fp/5svMpUDKP4EBeHfIYAtTYnEijSVz/iNsHXi02Rv/RYKCmmN8vNQQAuFLa8Dig60Tyus0NQPJtvg9Ya1geDIFFoLE77+tUtAlpA8DLxjsfkV9pu2bzA4DwcybbxxiRGaOHih+djuI8AWmO4NGRnTCtf/0uCvyBpv/28ssv4+mnn0Z1NamKCAsLw7Jlyxp8YI0NlvhsSJshOJB3AKdLTl/lEdU/XIzC1vSu0wEAJ4pOiI6z3aWNBYeLRq2zFt2iu+GBlAdkx1lKZCFXf32hpNqOPe5e6G/9AEXwIQknTNZWCEJpMkMgmZCHzCdehJeNZFJEB0SjxKr+OXiiSH57xzm/nukP8itq0bmVdzKWzRGs48iG4FgD0CLYc87HpNfB4XKp9hFcbWiOqmvXrti8eTP3b+PGjbh8WV0Or7mC5bsx6o2oslfhSMERHC04epVHVb/4PYdQHOuZ5i+d5M/v8kKKkaZpZFRk4L9p/0WRxTemSukkZXO4UGApUE2AUhSF3OpcjFg/QvWe/lY6sdKBRYjEkE4+ND4JO6crmYS1KUSeI6BpMV2lOQQYs4R0KfuBrZO3Yu/0varH2ZXmiexylFvEnc81Ao9AuiL9YLS20I63cLrcKKq6tj0CiqKg11Fc4xzbI8CWlaqhHUO1rdZHcLXhtXkKCQlBSAhJfl1rymVOtxP7cklXoLCWWq3Cprni23PfAuBDAlM6TxEdVyotleKnzJ9w+8bb8e6xd/H0XmVeIDVI2+crrU6cLzuP1OJU5FcohVY8N0ABchoFb2FzEEMwvkcs/jtbgZ1UDYF8lyhXYmoKUfEI6u9HH2IKERHIScHGrud9cwJTPxYzpQ7txK/QpVUrg9sMRn2hqNoGN33tlo6y0FMU5xGwixut/oB0RqHNolBu2hTgl59yramXfXTyI7x68FUA4qamCj+455VgcViw/sp6ONxyAZarAdYjECZqAWBfzj7NMbJslYA8tKSFSIZ0K4LhWqm08s8a9NpOxWukIi9C6BU4/b2FzUl+kNP6x8vKWj0iSGgImB+1WcEQSD2COmB0u9Ga5whXmhcK+ea2wkorHl17jNuWRiakeaK6gDXm13KyGCAFZawBYMnntIoN2Aa7pjp3+mUI6qO6oCnhk9RPuNdGAU/9rqxdSqf7df8NuRvw4h8vihqoGhN51Xnca0+J8PePE+3gYd8Mw8ztM0XHtlzcgsuVfFjQ14T63FGd8PcbE7BjPhFCr6wlhsBRfoPi+RRFeTRMayasUT2mBSvjEUhLWjUhbOphJ39TcIN6BG8Of1PzHLWJaOFGMd11Q6qBrT9CciYtQvzrkWguoGng+2M5KK5x8h6BxpzI6nE0VaguB267Tb2yobi4WPVYc4dwhRRullP++gOWU56tfX+sz2P1cl9fIKy7F9I6SJFRkQGAVBCVF5VjVdoqzEmZAwAytlJfmSrDAoxYcnsPuN00KAqosjqhp/SwOZU/5/Nl51XvNTxuOHq06MFt0zSNTSdyMSGltVeqV6xH4JM3AIgFbWqZsJUpVJ4jAOrNI/Bm1a4We24h6I+ICDLC4axD67EGLjKeSPtmIj3pL2xON2zVNjy7Iw8zB5MJ3m/qiCYC1W/Yxx9/XOeb7927F0uXLoXb7cZdd90l4yY6fPgwXn31VZw9exZvv/02xo0bV+dn+oNu0d24KiHh5K8lnO4NjhYcxWd/fVbn+9QFF6ovYOXplQBI9YkSy2SEOQLltnKZ67ri2AqEmcIUm4385WPS6SjQNHC+sAou2gXQZDL2pbP5pribRNsvbDyFtQezMH/dCXx2X3/EqlzHwm+PQAmmIKBaQs/QyCEAJY+g0urA+QK+Eiw62CQKx9U3hia1wKHMUs3E6bWCnEoHlzTWaXxv2f4Bu7NphoZUDUHbtm1l+3bv3o2RI0d6dWOXy4UlS5bgs88+Q0xMDKZOnYpRo0ahUydegap169Z47bXX8OmnCpqyjYisyizuddfIrtxri7PuGqG+JlQbAl9lf8W93jVtl2L9eLAxGOW2csVwz78P/FuWTwgxhqDWUTdVq+1pOQhNBkCT8RRV2dDKi/jyN7d+g+7RYk4jIf3xfZ8dxo+zPbviLH2wX4YgbgBwhXDuI6wt0TRu4GSxFpQmoifWncThTF48KjrEjMpa+d/39WGve0xEewsuXn6NhY7V0DbMyPUTaHkEbH6stEaBzLAJwKdfwXvveR/fTk1NRUJCAuLj42EymTBx4kTs3ClOCMbFxaFr167QXUX+DYfLgWpHNSLMEZjfZz70AvphmxIDpQ+wuWyKRF6NnTC6YuXr3aVG4IEeD+DmdjfzOrgq3+dfL/8q2o4LjUOtq47yhhRbQUE+czV6BClaBsrr1BNbisMRbo17ebuSU8RkgbdcmaNsCOoxWewNqq3yCf5CIe8NvHRbN8RFBCp6BBMSJ+DGNnIhdV/BfuT/I3YArYIN3HdWKzKUEE2Sxa1Cm2Yi3aeSAV8msIKCAsTG8g56TEwMUlO15fm0YLPZkO4nDavVasWkbychJiAGT3Ri9G7tJM47tfVUDDYMFt272lLt97MAoMyuLOXZ88ueWD+gYekchBDq0Urfz9igsUAQCYOtN6yHrUb58117Zi33+qbom+CGG+drznv8fKxWq+pxHQVQgcQTM0YchL1kBM6eO4/yEG0q5MwLmSgxipur3vr5kmi7qMLicWxXcsgkeSkjA7Zi3+iXjVXZECorl9VYEVJbjQvp6dx7TnI5UVlWjoI6fH+WdV+G05WnvfoOpmeIw3Tp6ekwCMqBB0ZZcex8FcpV/r51hdVqRWGRBRSAM2fOaJ5fn89tiPfjDY7n1eJ4HsljefOeX7k5Ft1a6es83oZ4zz4ZgiVLlnh9rpLRqI9qI7PZjORk/5SUUv9KxSXLJVyyXOLusfniZgDA4C6DkRxL9lGHCLWB0Wz0+1kASIWNSoVl165dG636anjmcPxS+AsAeHw/CZkJMBlNMLVWp0j+ZuI36N6iO1458Ar2lexDm8Q2qkn19PR01eeFmLNQayKTuc5EDGaHxI6iqpZ5znlYcWwF3EWTQUVtA6UnjVK9uveSlZXaXBmi7aOFLjx5o/p7PVWTDaAInZM6Id5X7vxS8bMjIyKBjGIkd2yH9ItZ5D1v1iMqKgpRdfj+JCMZEzHRq3PNLarx/oE9/LXJyYjaUw6UlnLb7bPPouZsJbp06eo9t5KXSE9PR3R0EHS6ijr9Zvx5bmM+jyBDtsebMdTXMP19z56Mh1cxmWPHjmHLli3IyMjAxo0bsXHjRs1rYmNjkZ/P62sWFBSgVatWHq5oeJysOCnbl19DxijUH/jz7j8xqPUgksj0EXN+noPZP84GQPoHACLK/mXfLxEVwNegs9KDjQE9E3p5boBn1thAfSBSi1KRWZmpeHxI2yHo3oLE5tnwzN3b7vZrTJVWJ3QmUn1WmzsVAJBRXCOi9Z2TMgdps9Mwuu1kVJ/jFyGeegtYrNzvubKtbmEMyUUnvyH//0MQOm3k8F9iyxC8PJr3wGvtLplASligETQt7jSuT7hpWjNEch1NE5qGYMGCBVi+fDmOHj2KtLQ0pKWl4dSpU1qXISUlBZmZmcjOzobdbse2bdswatSoehm0P6iwVWD5+eWy/dX2apj1ZpHgR4gpBBHmCL+I5w7mH8SxQtLAsy+HdCvHh8YjQB+AIW141SirIklZw8ANNyLNkZiRPMPjeQfzD6LaUY2PT5IYuFSxLNzEr/zZfovsqmw43A4sP7zcJ06gAcnFMEUTygu3jSwQ7vvsMKZ89KfsXDVGTSF6xsm9kqIq9RwPG9v1q+xPaj1Y1lSrsAGxcZPFAJDckjeQN7+9B3vOEQqQ16cQ0Z2wAPI3q6htmMohN33t9Rj9r0AzNHTq1Cls377d5z+wwWDA4sWLMWfOHLhcLkyZMgVJSUlYsWIFevTogdGjRyM1NRVz585FZWUldu/ejZUrV2Lbtm1+vxlPWHd2neL+KkeVosh3fTCQvnecrBBZ8rS7u97N1fPbXb55BBaHBRfKL8g45L2Bm3b7xDR5ppTEO5/s96RIv6B9eHvutTD0tztrN1afXo1yazleHfaqV8+YMTgKi9g5n+Zj9Kfz5CWprCFwVPSCzlSMcosdEUHi8JVRr0PX2FCcyecTpP2X/orUl27hJkAh6lbhIrnGHArYKngdY4CxA407KYYKyjZzykleqEtMKKb3J2I/LPVDdmmtjDO/PkDTdCObvuuoL2gagqSkJBQVFfkV1hk+fDiGDx8u2jdv3jzudc+ePbF3rzqRVn0itzpXtq/GUYMN5zagfVh72TE9pfcrNMRCuDruG9MX58vPI6VlCpIik3C+7DysLt88gif3PIl9OftwcMZBBBl9+xHToP2iHA42iCtxogN4crYIc4RobIBvtMZCFkba7TlZywrYWHNJGIp1EGauOoBTOZU4+eItsDldHHWFEE+tP4lP/i5X3mLv4ZchEL5PYzDQfgiQug6wCmm8G98jUILRwI+hI8P1dKm4Bjd29IFoz0vYnO7/mYqhaw2ahqCsrAwTJ05Ez549YTTyP7T6aDhrTChR+P77wL8BQDEmbtAZvGLjVMPsn2aL7sXioZ4PYcGeBT57BH/k/AHAO2I4KfKseX7xHJkN4li8sIFsQuIEvLT/JdFxv+U9ac+GQFoK6mSKt/+4QP6mVocLdqcb4RHy+/x+XjlXwJeP+jxa8Up/zi9Enzh1HXBmK1rbACR/1ejlo2owCEqzwwPJ51Nja5gcwed/ZjbIfa+j4aFpCP71r381xjgaHLOSZ+G37N+4bYfbwdFMxwbL+1ANOgMKawsxaeMk/DDpB1F/gRqE4ZKogCgRLw+LAD1xzx/59RH8MvUXr8fPCrRU2ioRZgrTOJvHQzsewukq77QVvprwlYhfSFoNVOOo4V4HGgIxJmEMfrnMvwc1YXVN0OKvobTDWCppKd0+lVMBm9ONAKMenVqFiEjXah0ubDmZi4ToIPSMi0BRlQ2hAQbOuPhHDSC4Jqa7KDEccWkrT0bXBDwCYZ4kgGmeUxJcV8PFomq0jw7Gp/suYen2dNw9IB6v3SkPTzpcTbNjtjGw9V9Dr/YQ6gxNX37AgAGK/5obBrYeiHvb3cttT986nasY6h8jpyFmmS0vVVzChfILXj2DFbcB+En0nRHviM6JDyVqROyzfcV357/jXjvdTtA0jR2ZO/DQjocUKTH25+2X7VODlHqic4RYuUtK+VxpE8fzWU0Hb0AJJkmaFn8NpZOKW2oIJMdX7rqAyyUWmA06/PqEOBQJAP/6+jiXhO6/9Ff0enkHl3fwK7kpvYaigAS+EIAolzWNiZHNFQDacpZSZJdaMPqtPXjj57NYup2UHn59SJmavZah7JiY0rouw22W6NG2fjjJriY0DcGJEycwZcoU3HDDDejRoweSk5PRp0+fxhhbvWN8zHhuNS0kNOsY0VF2rjAsMnXLVK4U1OV2ocyq3CgmXDGzr6WrZKVnaUFIh82GlGiaRp/VffDCvhfw5J4nsT9vPwatHaR6D28I4ox6Ix67gSfEk5Kdze42W7QdZhZ7Jv7TbIu/hlfKxNQeTrfb4zZbHeOJbE5ILmdzurHjrwIAfnoEbFgvtA2/L/sQ/7oqr8mEhoSgKAoUpd11zYLtQv7trLw7Xginy43p64j3ey1Mir5g7ZyBV3sI9QJNQ7BkyRK8/fbbSEhIwMmTJ/HKK69g1qxZjTG2egdFUYoJzdndZ8v2ScMvbOPZO0ffwU3rbkK1Xc7gyRoLgDcESupb07sQqUh/Jk42Tl9iLQENWsQqWuusxckiea8EAPRp5Zvx/nu3v8tWy0KKbgBYNGiRaPt44XGv7y+6Ny1+zqYTfGLf6nDhQEYpesdHoEML4pHQABwuufdjNpDJ/vbebWTHIoKMIgK2jGLy9/ErMhQWB/ScDsz8lt8nrDCrLsTV8gh6xUeItmcObCfa1lOU1x4BazyFlVhK+P54Dvf6zj5yjrJrDdse40NBfRLqztHUFOBVmUdCQgJcLhf0ej2mTJmCQ4cOaV/URMHSLQgFaJRofqUlpWyy+YvTXwBQZt6scfIeAcv/r3TvSxWEDsEfbYJQUygAMW2EELO2i400G6K6M+lOr+7Peh8XKy6K9n8x7gvZuZEBkRgVz/eG+BJmEZbm3tQ5BgeeG42vmNVVdAj/t3l5C8lvZJVa8MQYEqqiaVqRW4clkLtvSAfZsVq7C2Pe4SvUWDFxv6qG9Abgzk+A2B6CnYLJtbYUcFiuikew5oEBmNiTD8+wxpOFTqCupQVpSA6QczoB4JoAx3WPRcw1LkoDAN3b8F6PzzTmTRSahiAwMBB2ux3JyclYvnw5Pv/8c1gsdWflvFpg4/hsPF1KZ8wixCQ2BJHmSFGlj1L5J+sRhJnCUGYj4SOlJHO7MLJK25qx1asxC8tY2TF4qjoSehqR5kjcGHWj1yWnwSbyQx8QS/JAq8evxm/TfkOfGGWPQuglCGU+1ZBWlAY37RZpGwQZjIgND8CgxGjoKKCwks+1sBz31TYnN6+6abItRTnTKKXEzV9cLf688hg1rQZhytz0KPm/vfF/J6EBRjw7jmfQ7StZsep03jc9Sw1GbFgAJ/EpBKvt8PoU33tcrqNpQNMQLF++HDRNY/HixQgKCkJeXh5WrlzZGGNrULBlmO+NVF6VS0NDVfYqXKnmWTyVEqNvHiFKUsJJWskjWNBvAQAgpUWKV2Mdvo5PgNpddlgcFhzIUxdtn7tzLj9OtwNGyntStTk95uCHST/g/h73AwB6t+qN6ED1mnOh51RlVw8h2Fw2vH7odczYPgNvHH5DdMzEhHT0OgpuGnh/tzw5b9LruEl7/8USRRbNRGb1275FMGJDDIhU6CuQokEERVhvp+KK5/MaCPFRQchcNhFpL92CG9qJDYEvoSFpUr5Di2DYJMI2y386g1e3n4FJTyEssP5kL6+jcaH5l2vbti1hFiwsxNy5c7VOb/J4fdjreOb3ZwAAocZQ1bJQKZ9Npb0Sa07z0ojSFXmVOsNP5QAAG1BJREFUvQppxWkAxN6CkiEIMgahe3R3r/iGpOR9drcdz/3+HHZlq8to/pnL0zTYXXafdGmNeiM6RXbSPpGB0NNQypvUOGrwwM8PwO62cwn6NeliiUmjXt29NjAhnFqHi4vnv7j5Ly72PaBDFA5dIsRqI7oQ/qMQswGfTWmHD09YseWkvJFQiAblxqkuaMCbayNUoaNap/PeEEjPa98iGKdy+cKFoiobPvyNhBDNBuo6vUQzhqZHsGvXLtx+++2YM4fIFaanp+ORRx5p8IE1FCYkTuBeu6GuQCaVc6yyV6FzJF9OuecKYXqkaRr5Nfmq5aBKyWIASC9N55rEWOzK2oWL5eLYvNRY2F12nC6V9wX0iObj1fd0u0d0vS8ega8QxvqFORIWX5/5Gn+V/KUoO9k78H5Upb8Gk0F9AmFXpTpKnIM4kEFyNi/d1h2Hnh+NhROT0bGlOJyn92JeqrfJq98DQGwK3MIFRB31LBoCxOvybAiOZZVh/8USWWgoLNAAm8PNLU5YAwwAVbaGk8C8joaHpiF4//33sWHDBoSFkVBJcnIycnJyNK5qHhCWe0ohjXfbXDasSlvFbf9f6v8BAF7880WM2TAGxwqOcceWDOaZMtXi5ko1//N2z8Mdm+4Q7ZOS06nV6gvLXYUegN1l9yp27y+E3seliktIL0nHhnMbkPL/7Z17WJR12se/c2CGwwACclIJxcDIA8hmHgIJDMQGRBHMcs21rLY3pcQ0DXN3dbVLX9e97Np112pdfVty84yFmyaWuB4y80AWGCokqKAiIMgwx+f9Y3jmwMwAM8w8Msz9uS6vi5nncP+emfG5n9/9u+/vvX0k/l75dzQayS4YEyocC4CHi9VNJttYYbQzVdqbjZtBaAgArt7RfndBPmIE+bhjfkKEyU29Y/zfw02Af72sT/ebYE+ZhfSNwG//C75BLQlsTqd1HF2Fhn6pf4CszSfx/Een0dZeeBYzyBcbZ8ZgcIAXFGoNfrypTZT46LipHDPhnHTpCAQCAby9vbkYS6+C16EqVKlRoq7VdKq/78o+AEBtq35GEBOkl7S2FJZ5tJ8+/HJXdhfVzeYLdQwdwfCA4VBqlGZ7PSQMTNC/aN/MMAzkarlRhpS9YRvbs8z8Yib+cOoPAICjd45aTCld8sQSNLRox2UoNJfUHt6530EhM0AiMqnT9fN0Q4CX5Wvr6BieGxOG+Ei9omrHCmW7Y0WBHVfwu5gR7Dmnf8j78pL2N/3OlMeQFTdIl4HErs+wv8PlUx7DpzMfAeG8dOkIIiMj8fnnn0OtVqOqqgqrV6/G6NGjuRibw/ETdz8HuOPs4fGAx41eG9YQGD6BMxbyyUcHaT/DgrICJO9MxrN7nzW7H5vltDZ+LcQCMRQahVkxPMM1DdamTCWDhtHAU2B/pUmWzKGZ2JS0CZlDM81uN9f3+fyc83hx+IsQtN+oV2fq+w9nxmrz0DvWCRS8PM4kVOEpEnYa2hF0+HWzT8IrpNoK6vER9hdeM6IbsiRc09WMoKlVH4r8v1PaIjFWr4hNuWUrvy/WaGdyryUOhZ+LLRSXLEnC1qywhz0Mu9GlI3jvvfdw5coViEQi5OXlQSKRID8/n4uxOYxhfsMAGMsqd0WjXB/iCPcJh7ebt8XtQZ56pVZDxU5D2Cf9LRe3WHQWgN4BuQvdIRKIoFArzIrhhUr0ueNs3J5d53CkI+DxeEh+JBkeQg+z283Jc7CzJPaJ3MdD7zjZ6mCFWgO1RtvoJDf5UTwS4Imi0ltG5zGUTzAHGxqaPDwYOb8ahLee0YrizU+IwK7fjsebk2wUyesuL3DXjrS7CPg8qDXAvQcKfHvNVIjRnI9gHSqrGNvcptQJ/7kqjwR4ItTbcSFXrunSjXt4eGDRokVYtGgRF+PhhE+ln2L16dVYENv9LChDmYcA9wB8W/st7sr0ypasI/i39N8QC8R4PeZ1JAxMMKlHYHlj9Bv4/NrnunoDS/yn8j8AtLLPIoEITfIms44jKSwJf530V7xT8g6UGiUUagWO12gbv3gIzN+k7Ul3MpOWjllqNJNStN9MDIty2JuNQqVB7f02aBgg0FtstK27sLMFfy+RiVDamMH+5g7pMW39IuHeWKGVn+jvYEdjA3y+VmJi8c4L+PryHVxYmWLU20GmVGNgPw8o1BqdYJ2gfUbg2d7xbMGn55EQ2d/05ITTYvF/b1eZQc4mQ22ISCDC6qdWW3WM4RM/qyr67nF9URS7aDrMXzvb+J/Y/+n0fIZtKztyV3ZX1x3sgfIBBDwBngh+Ap+WfQqlRokA9wCj8QDaXgATB02Ep9ATSo0SMw7M0Mlrc+EIOluQjh8Yj8xHM5E2OM3o/eb2WLOvmRmBUq3BO7tLAQD3Hmj3eyNpKPac635uPpsaymX1p4oNNz7k1FFLsKGhhtb2xfjKe0gdrlff3f296efLFugFGVQNs/LeK9MfN9mfcD4sOoILFy4gNDQUUqkUMTExZhco+zKGjVcAfYhmZtRMFF8vBmAqMyHii7qds9/ZjfNa4zWdI2hTtyHYMxg8Hg9uAjco1AqTFpJG5xW4YW/FXqP3GpWWM3fsRWfXvfqp1WbHzMpEGDmC9qd+uUqDW03a0A9bqGS4HwC8lhjR6ZjYYjFPEXeOQFLXLr/Sg6ZGjoTP0y4Wh/l74kJ1I2oazIfXJGKhrnqbDbF5mHGofUViwdWxONc+ceIEFi1ahIqKCqxZswYnTpyAn5+f08pQW8vIwJH42zN/M3k/fWi6rgitYwpoR+fRGUK+EEN8TTVxAGDrj1t1f8tUMl38Xa1R43rzddxouWExJm/OwYzwGWFmT/vS2aKtJcclFLCLkPqfITsjuNXYpksRnTJCu/4hNrjpVL7/rJGUgjnYNQhzNzBHofQ07W3Rm2Czhthvq8Kgd4PhAv2IgfrKeraoz1xozktMjqAvYNERCAQCTJw4EevWrcPOnTsRHh6OOXPm4JNPPuFyfA+V+IHx+GTKJxgbqs899xTqF147FkmNHzDeqvMbFqgZEu4drvu7VdUK9/bm6Id/OQwAqGutQ/xA880wzD2ZB4mtbzNqLZ3VZFhi8+w4vJIwRCcNAehnBIt36VVU2SZbYgOZaa2kcufFYKwuDpdPrTeffK/rnR4ibGiITSHdceY6Lt3Qrn+drdKvV72RpE9v7kyGQyJ2rWyhvkqnq28KhQKHDx/G22+/jYKCAsyZMwepqalcja1XEBsUi4ESvbSul5uX7km+Y9vIjvr8XcHKVKyfuF73nrebt1FVc5uqzezTv1ggNpHBAPTqpCx7pu4x2ccRdJSX+FPin7o8Zkh/L+RLHwff4EZjrqcAG5owJybXGWxLRk8RdzcrjZnvpDfBb88aMtQMqqrXOnE2FDTA1x0JkYG67QIDhxv/qPHsjsvPlnAcFr/Fd955BxUVFUhISMCCBQsQFWX+6dUVMHQEHkIPrJ+43kgIbpjfMFxuuGx1H2K2RsBQuC1UEmok3iZTycymoAp4Avx+wu+x/Phyo/cNw1M/zNVqH5XVllk1Llt4ZdQrqLxfidI72gXe1MGp2Ny4GYMjBlt1HnEnjsBaOYi2dlXMqGDzmVuOgG9GlbY3cbdFjrJb942yftzbRf+83bW3gw05MUbHuBl8J/+9YtwDmlUeJZwbi46gsLAQHh4eqKysNAoHsf1kz507Z+nQPkegh/7pyEfsAze+G2ZHz0ZBWQEAfVjEsFVld2AX4A2f7CVuEqMwi0wl04WGDDlTewZ/jP+jiSNgs5GmDp1q1Vh6SrhPOAqeLcDJGyd1M6P+4v4I87Gu6Mbcwq7ARj2gdTNG4eTVu3jCQami5mDaW5zCy/HhOFtgU0INm82wlcJs/wF+h5kX6yAAbU3GoR/r4OvhhqhgCaefLeE4LDqC8vJyLsfRqzGMu7OLsYaZOaw8tbUzgldGvYJGeSOG+Q/DyvErwTAMjtUcw82Wm1h1ahVeHfWq0WKxIbce3DJzRn0jmgNXD2BN/BqrxmMPJgyc0KPjPcw4AjNN5bpFsI87po8e1KPxWIssMBZIWQXEvcipXWu50yxHYlQgjv18RyfnwVZud1wTkBiEf+aMG4xDP9ZhxEAfFMy33BqVcC5s/C/mWphbgH3MX5+xwvY0mPXYLKvOOyZkDHZm7ISv2Bc5UTmYOWwmvNy8cKXxCnb9vAubL2y2uEZgCXZ2MTbEOXupmos5O5W4MY8HPPUm4NH7WxiyIbN97a0mWemJjmJ95tZwzDWoIZwXcgTdQNA+3Tfsd7z8SW1IZoDXACQ9koQf5v6g0w/qCYbSFW2qNjTKGyESWBZWiwuKM2qiw85Kngx1zhRfcxkqhr3qj+RNxLfvTuJwRH2LX4/Ti8Ptb+8NzWoG/dze05n9Dgb2M30AMSz4I/oOtOTfDdi6AcNYPtu1y5wGf0+ovF+p+/s/VVp5icv3Lpvsx3Y525a2zWgRlW1TaU0zmt7K/2aPQnHZbaMY9aNBrqeEa09yfhWGf52+DkAbHnJ308p7N7UqsfagNhzMrskcfDNBl3nFMsBXu1719LDeuQZC2Ibz3y04gG0uY/i0ziqXSodI7Wqr7oGpNIG5bBkvNy+z21hnJe+FTVGsJXFYIHKe6DsKj72BYSH63/C6GSNx6Mc63GmW40iZ/nfH1m34eriZVHMH+bjj7Ipn4O/pOGlzgnsoNNQN2BkB29gd0Eo5nHr+FJaOWWpXW4O8TRc3WUf0Ztybuvf4FlZQpw6dCgFPgMmDJ9t1XA8DLiuCXQXD4rpnooPB5/FQXntf1/EN6LqPc3+J2CSziHBuHOoISkpKMHnyZKSkpODDDz802a5QKPDWW28hJSUFOTk5qKl5OM2+u4K96Rrm+wOARCSx2PPYVtYlrDN5L9JPq2JpWCNgricBoJXWvvDiBUT061yHxxkgHRvH4u4mwJGyOijVDHYZiM3Zmq5LOC8OcwRqtRqrVq3Cxx9/jKKiInzxxRe4csVYm37Xrl3w8fHBV199hd/85jfYsGGDo4bTI1hNoY6OwBH0czfVK8odnQsASBykL2IzrG3oq1grO01YhyVHS03oXQ+H/U8rLS1FeHg4wsLCIBKJIJVKUVxcbLTP0aNHMX36dADA5MmTcerUqV6pcspKPljqLWBvlj25zOi1m0Abpw30DMQPc3/AwayDmDhoIidjIfouAj4P/37VtBagq9AQ0fdw2GJxXV0dQkL0SozBwcEoLS012Sc0VKssKRQK4e3tjYaGBvj7W65WlMvlKCuzTTKhra3NpmOv3NbOZFQPVJzYjkOc0Wtzx5Wj+wV/tl53T7HVbpivG6qblD0as7Nd88OwXVZWBnmLyuT9ymtX0XbHuu5b9Hk7t12HOQJzT/Ydp5zd2acjYrEY0dHRNo2prKzMpmODhwSjuLEYixMWI9wnvOsD7GH7jP5PW6/XZtt2wla7h4ZGQqHSGHXO4sp2T3lYdq2x/fyTShy4cBPR0dHwa2oDcN1oe1j4EEQGW5emS59377fbmfNwWGgoJCQEtbW1utd1dXUICgoy2efWLa1UgkqlQnNzM/r1676mP1f4u/tjX+Y+m52ALcwfOZ8zW70NT5GwR06A6Jw/ThuJcytTABiHgabFDgAASNwpq9zVcJgjGDlyJKqqqlBdXQ2FQoGioiIkJycb7ZOcnIx9+/YBAA4dOoRx48bRQlU7r8e8jmVPLsP3v/7+YQ+F6GMI+DyI2xVHDaW9N86MxYllyQj1dXxrU6J34TDXLxQKsXLlSsyfPx9qtRozZsxAZGQkNm3ahBEjRmDSpEnIzs7GkiVLkJKSAl9fX/z5z3921HCcDpFAhNnRsx/2MIg+juHTP5/PMysrQfR9HDoHTExMRGJiotF7b76pL4oSi8X44IMPHDkEgiA6gVJ0CYAqiwmCIFweWhUiCBdn06xY+HvR4rwrQ46AIFyczNiBXe9E9GkoNEQQBOHikCMgCIJwccgREARBuDjkCAiCIFwccgQEQRAuDjkCgiAIF4ccAUEQhItDjoAgCMLF4TG9sSVYJ1y4cAFisfhhD4MgCMKpkMvliI2NNbvN6RwBQRAEYV8oNEQQBOHikCMgCIJwccgREARBuDjkCAiCIFwccgQEQRAuDjkCgiAIF8fpHcGtW7cwZ84cTJkyBVKpFNu3bwcANDY2Yt68eUhNTcW8efPQ1NQEALh69Sqee+45jBgxAv/4xz+MzlVSUoLJkycjJSUFH374IWd2ly9fjvHjxyM9PZ3Ta7Z0HkfblcvlyM7OxtSpUyGVSrvVt9qenzcAqNVqTJs2Da+99hqntpOTk5GRkYHMzExkZWVxZvf+/fvIzc1FWloapkyZgvPnz3Ni+9q1a8jMzNT9i4uLw7Zt2zi55m3btkEqlSI9PR15eXmQy+Wcfd7bt29Heno6pFJpp9drq+0DBw4gIyMDGRkZmDVrFsrLy3XnsuY+poNxcurq6phLly4xDMMwzc3NTGpqKlNRUcGsW7eO2bJlC8MwDLNlyxZm/fr1DMMwzN27d5mLFy8yGzduZD7++GPdeVQqFTNp0iTm+vXrjFwuZzIyMpiKigqH22UYhjlz5gxz6dIlRiqVcnrNls7jaLsajYZpaWlhGIZhFAoFk52dzZw/f56Ta2bZunUrk5eXx7z66qud2rW37aSkJKa+vr5Lm/a2u3TpUmbnzp0MwzCMXC5nmpqaOLPNolKpmAkTJjA1NTUOt1tbW8skJSUxMpmMYRiGyc3NZfbs2cPJNV++fJmRSqVMa2sro1Qqmblz5zKVlZV2tf39998zjY2NDMMwzDfffMNkZ2frPmNr7mMsTj8jCAoKwvDhwwEAEokEERERqKurQ3FxMaZNmwYAmDZtGo4cOQIACAgIwKhRoyAUGnfpLC0tRXh4OMLCwiASiSCVSlFcXOxwuwAwZswY+Pr6cn7Nls7jaLs8Hg9eXl4AAJVKBZVKBR6Px8k1A0BtbS2++eYbZGdnd2rTEbatwV52W1pa8N133+muVyQSwcfHh/NrPnXqFMLCwjBwoOXWmPa0q1ar0dbWBpVKhba2NgQFBXFyzVevXkVMTAw8PDwgFAoxZswYfPXVV3a1HRcXp7tnxMbGora2FoD19zEWp3cEhtTU1KCsrAwxMTGor6/XffFBQUG4d+9ep8fW1dUhJCRE9zo4OLjTm6K97PYUe9k2PA8XdtVqNTIzMzFhwgRMmDCh23btYXvt2rVYsmQJ+Hzrf/72+LxffvllZGVl4bPPPuPEbnV1Nfz9/bF8+XJMmzYN+fn5aG1t5cS2IUVFRd0Of/bUbnBwMF566SUkJSUhPj4eEokE8fHxnNiOiorC2bNn0dDQAJlMhpKSEt2N2hG2d+/ejYkTJwKw/T7WZxzBgwcPkJubi3fffRcSicTq4xkzShtdPaXaw25PsJdta89jD7sCgQCFhYU4duwYSktL8fPPPztkrB35+uuv4e/vjxEjRlh9rD2ue8eOHdi3bx8++ugjFBQU4LvvvnO4XZVKhZ9++gnPP/889u/fDw8Pj27Hju31G1MoFDh69CjS0tI4sdvU1ITi4mIUFxfj+PHjkMlkKCws5MT20KFDMX/+fLz00kuYP38+hg0bBoFA4BDbp0+fxu7du/H2228DsP0+1iccgVKpRG5uLjIyMpCamgpAO227ffs2AOD27dvw9/fv9BwhISFGXruurq7LqaQ97NqKvWybOw8Xdll8fHwwduxYHD9+nBPb586dw9GjR5GcnIy8vDycPn1a95/I0bYB7RMae2xKSgpKS0sdbjckJAQhISG6WVdaWhp++umnLsdqz++6pKQEw4cPR//+/Tmxe/LkSQwaNAj+/v5wc3NDampqlwvk9rINADk5Odi3bx8KCgrQr18/hIeH2912eXk5VqxYgc2bN8PPzw+AbfcxoA84AoZhkJ+fj4iICMybN0/3fnJyMvbv3w8A2L9/PyZNmtTpeUaOHImqqipUV1dDoVCgqKgIycnJDrdrC/aybek8jrZ779493L9/HwDQ1taGkydPIiIighPbixcvRklJCY4ePYqNGzdi3Lhx2LBhAye2W1tb0dLSovv7xIkTiIyMdLjdwMBAhISE4Nq1awC0sfqhQ4d2eoy9f99FRUWQSqVd7mcvuwMGDMDFixchk8nAMAzn11xfXw8AuHnzJg4fPtxlSMxa2zdv3sTChQuxfv16DBkyRLe/tfcxFqdXHz179ixmz56NqKgoXcw3Ly8Po0aNwltvvYVbt24hNDQUmzZtQr9+/XDnzh3MmDEDLS0t4PP58PT0xMGDByGRSHDs2DGsXbsWarUaM2bMwOuvv86J3by8PJw5cwYNDQ0ICAjAwoULkZOT43Db5eXlZs+TmJjoULs1NTVYtmwZ1Go1GIZBWloaFixYwNn3zPLtt99i69at2LJlCye2Gxoa8MYbbwDQrpGkp6dz9hsrKytDfn4+lEolwsLC8P7773eaoGBP2zKZDE8//TSOHDkCb29vzr7nDz74AAcPHoRQKER0dDTWrFkDkUjEie0XXngBjY2NEAqFuvRwe153fn4+Dh8+jAEDBgDQhlr37t0LAFbdx1ic3hEQBEEQPcPpQ0MEQRBEzyBHQBAE4eKQIyAIgnBxyBEQBEG4OOQICIIgXJyeiaEQhAsQHR2NqKgoqFQqCAQCTJ8+HXPnzu1UoqKmpgbnz59HRkYGhyMlCNsgR0AQXeDu7q6TJ6ivr8fixYvR3NyM3Nxci8fcuHEDX3zxBTkCwimgOgKC6ILRo0cbyRNUV1cjOzsbp0+fxo0bN7B06VLIZDIAwHvvvYe4uDjMnDkTV69exaBBgzB9+nTMmTMHGzZswJkzZ6BQKDB79mzMmjXrYV0SQRhBMwKCsJKwsDBoNBrU19cjICAA//znPyEWi1FVVYW8vDzs3bsXixcvNqpa/uyzz+Dt7Y09e/ZAoVBg1qxZeOqppxAWFvaQr4YgyBEQhE2wE2mVSoVVq1ahvLwcfD4fVVVVZvc/ceIELl++jEOHDgEAmpub8csvv5AjIHoF5AgIwkqqq6shEAgQEBCAv/zlL+jfvz8KCwuh0WgwatQos8cwDIMVK1YgISGB49ESRNdQ+ihBWMG9e/fwu9/9DrNnzwaPx0NzczMCAwPB5/NRWFgItVoNAPDy8sKDBw90x8XHx2PHjh1QKpUAgMrKSquawxCEI6EZAUF0QVtbGzIzM3Xpo5mZmTqp4BdeeAELFy7El19+ibFjx8LT0xMAdM1Ipk6diqysLLz44ou4ceMGsrKywDAM/Pz8sHnz5od5WQShg7KGCIIgXBwKDREEQbg45AgIgiBcHHIEBEEQLg45AoIgCBeHHAFBEISLQ46AIAjCxSFHQBAE4eL8P+t9tjNkT+YGAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Calculate the past annualized 1-3 year changes.\n",
    "df_mean_chg = sf.mean_log_change(df=df_prices, freq='bdays',\n",
    "                                 future=False, annualized=True,\n",
    "                                 min_years=1, max_years=3,\n",
    "                                 new_names=new_names_1_3y)\n",
    "\n",
    "# Title of the plot.\n",
    "title = 'Past 1-3 Years Mean-Log Ann. Total Return'\n",
    "\n",
    "# Plot the result.\n",
    "data = df_mean_chg.dropna().reset_index()\n",
    "sns.lineplot(x=DATE, y=TOTAL_RETURN_1_3Y, hue=TICKER,\n",
    "             data=data).set_title(title)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For example, consider the mean-log Total Return on 15th November 2019 for ticker MSFT:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Ticker  Date      \n",
       "MSFT    2019-11-15    0.312211\n",
       "Name: Mean-Log Total Return 1-3 Years, dtype: float64"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_mean_chg.loc[('MSFT', '2019-11-15'), TOTAL_RETURN_1_3Y]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This value of about 0.31 is the mean-log annualized return for *all* 1-3 year periods ending on 2019-11-15 that started somewhere between 2016-11-15 and 2018-11-15. So if you had bought the MSFT stock somewhere between 2016-11-15 and 2018-11-15 and sold it on 2019-11-15, then on average, you would have made a log-return of 0.31 per year. Because the log-transform underestimates the normal returns for values beyond $\\pm 20\\%$, the mean annualized return was probably around 36%."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Geometric Mean Daily Stock-Returns\n",
    "\n",
    "The annualized returns are particularly useful for longer periods of several years, but for shorter periods the annualization formula can produce extreme values. So it is sometimes more useful to calculate the geometric mean, based on the frequency of the original data.\n",
    "\n",
    "For example, when we are using daily share-price data, the geometric mean tells us how much the daily return should be, in order to compound into the average return that was observed over an entire period.\n",
    "\n",
    "The following example shows the mean-log geometric-mean for all future stock-returns between 1-3 months. We switch from Annualization to Geometric Mean simply by setting the function's argument `annualized=False`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Future 1-3 Months Mean-Log Total Return')"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# New column-names.\n",
    "SHARE_PRICE_1_3M = 'Mean-Log Share-Price Return 1-3 Months'\n",
    "TOTAL_RETURN_1_3M = 'Mean-Log Total Return 1-3 Months'\n",
    "\n",
    "# Dict mapping old to new column-names.\n",
    "new_names_1_3m = {SHARE_PRICE: SHARE_PRICE_1_3M,\n",
    "                  TOTAL_RETURN: TOTAL_RETURN_1_3M}\n",
    "\n",
    "# Calculate the future mean-log 1-3 month changes.\n",
    "df_mean_chg = sf.mean_log_change(df=df_prices, freq='bdays',\n",
    "                                 future=True, annualized=False,\n",
    "                                 min_months=1, max_months=3,\n",
    "                                 new_names=new_names_1_3m)\n",
    "\n",
    "# Title of the plot.\n",
    "title = 'Future 1-3 Months Mean-Log Total Return'\n",
    "\n",
    "# Plot the result.\n",
    "data = df_mean_chg.dropna().reset_index()\n",
    "sns.lineplot(x=DATE, y=TOTAL_RETURN_1_3M, hue=TICKER,\n",
    "             data=data).set_title(title)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Mean-Log Sales & Earnings Growth\n",
    "\n",
    "The function `sf.mean_log_change` can also be used on other time-series data than share-prices. For example, we can use it to calculate the mean-log annualized growth-rates for Sales and Earnings for all the past 1-3 year periods.\n",
    "\n",
    "The function first calculates the growth-rates between all the time-steps in the slice: `df[t - 3 Years:t - 1 Year]` and the current time-step: `df[t]`, and then it calculates the annualized growth-rate, takes the logarithm, and finally it calculates the average of all those.\n",
    "\n",
    "This is not how you would typically calculate average growth-rates for Sales and Earnings, and it is merely shown for demonstration purposes:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset \"us-income-annual\" on disk (9 days old).\n",
      "- Loading from disk ... Done!\n"
     ]
    }
   ],
   "source": [
    "# Load annual Income Statements and set the index.\n",
    "df = sf.load_income(variant='annual', market='us',\n",
    "                    index=[TICKER, FISCAL_YEAR])\n",
    "\n",
    "# Select the tickers and columns we are interested in.\n",
    "df = df.loc[tickers, [REVENUE, NET_INCOME]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/magnus/development/simfin/simfin/rel_change.py:440: RuntimeWarning: invalid value encountered in log\n",
      "  df_log = np.log(df)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Past 1-3 Years Mean-Log Ann. Sales Growth')"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# New column-names.\n",
    "SALES_GROWTH_1_3Y = 'Mean-Log Sales Growth 1-3 Years'\n",
    "EARNINGS_GROWTH_1_3Y = 'Mean-Log Earnings Growth 1-3 Years'\n",
    "\n",
    "# Dict mapping from old to new column-names.\n",
    "new_names = {REVENUE: SALES_GROWTH_1_3Y,\n",
    "             NET_INCOME: EARNINGS_GROWTH_1_3Y}\n",
    "\n",
    "# Calculate mean-log ann. growth-rates for all 1-3 year periods.\n",
    "df_growth = sf.mean_log_change(df=df, freq='y',\n",
    "                               future=False, annualized=True,\n",
    "                               min_years=1, max_years=3,\n",
    "                               new_names=new_names)\n",
    "\n",
    "# Remove rows where ALL items are NaN.\n",
    "# This keeps the rows where e.g. the Revenue growth was valid,\n",
    "# but the Net Income growth was invalid due to negative numbers.\n",
    "df_growth = df_growth.dropna(how='all')\n",
    "\n",
    "# Title of the plot.\n",
    "title = 'Past 1-3 Years Mean-Log Ann. Sales Growth'\n",
    "\n",
    "# Plot bar-lines.\n",
    "sns.barplot(x=FISCAL_YEAR, y=SALES_GROWTH_1_3Y, hue=TICKER,\n",
    "            data=df_growth.reset_index()).set_title(title)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Disk Cache\n",
    "\n",
    "We only used 3 stocks in the examples above, so the stock-returns were calculated very quickly. But for DataFrames with thousands of stocks and share-prices for a decade or more, the returns can take a while to compute, especially for `sf.mean_log_change`.\n",
    "\n",
    "Both functions `sf_rel_change` and `sf.mean_log_change` can automatically save the results in a cache-file, so the next time you call the function, it first checks if the cache-file exists and is recent enough, and then it loads the DataFrame from the cache-file, otherwise it computes the results and saves them in the cache-file for later use.\n",
    "\n",
    "This is implemented by using a wrapper-function ` @sf.cache` on the signal-functions. You can also use this wrapper on your own functions (see Tutorial 06).\n",
    "\n",
    "A few things should be noted:\n",
    "\n",
    "1. The wrapper adds three more arguments to the original function: `cache_name` which allows you to distinguish cache-files from each other. `cache_refresh` which determines if the signal function should be called and the results saved to the cache-file on disk. `cache_format` which is the format of the cache-file. See Tutorial 06 for details.\n",
    "\n",
    "2. Because of the new function arguments, you **MUST** use keyword arguments when calling the functions `sf.rel_change` and `sf.mean_log_change`, otherwise you will get an exception (see Tutorial 06 for an explanation).\n",
    "\n",
    "Here is how we can use the caching-feature with `sf.mean_log_change`. First we need to set a name for the cache, and determine how often it should be refreshed:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Refresh the cache once a day.\n",
    "cache_refresh_days = 1\n",
    "\n",
    "# Name for the cache. Use the concatenated tickers.\n",
    "cache_name = '-'.join(tickers) + '-1-3y'\n",
    "\n",
    "# Dict with the cache-arguments. This makes it easier\n",
    "# to pass the arguments to multiple functions.\n",
    "cache_args = {'cache_name': cache_name,\n",
    "              'cache_refresh' : cache_refresh_days}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we may call `sf.mean_log_change(..., **cache_args)` which unpacks the arguments in the dict as if they were normal arguments passed to the function. Because the function gets the argument `cache_refresh` it automatically enables the caching feature."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cache-file 'mean_log_change-AAPL-AMZN-MSFT-1-3y.pickle' not on disk.\n",
      "- Running function mean_log_change() ... Done!\n",
      "- Saving cache-file to disk ... Done!\n",
      "CPU times: user 125 ms, sys: 4.26 ms, total: 129 ms\n",
      "Wall time: 121 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "df_mean_chg = sf.mean_log_change(df=df_prices, freq='bdays',\n",
    "                                 future=True, annualized=True,\n",
    "                                 min_years=1, max_years=3,\n",
    "                                 new_names=new_names_1_3y,\n",
    "                                 **cache_args)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that the cache's filename is `mean_log_change-AAPL-AMZN-MSFT.pickle` which is constructed from the function's name `mean_log_change`, the cache-name we supplied `AAPL-AMZN-MSFT-1-3y`, and the file-extension `.pickle`. This keeps the cache-files neatly organized on disk, while still allowing us to designate different caches for different calls of the same function.\n",
    "\n",
    "Because these DataFrames are so small, it is very fast to compute the function and there is no real need for the caching-feature. But for DataFrames with thousands of stocks and a decade of share-price data, it might take a while to calculate the results, while it takes less than a second to load it from the cache-file. So the caching-feature may save you a lot of time."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## License (MIT)\n",
    "\n",
    "This is published under the\n",
    "[MIT License](https://github.com/simfin/simfin-tutorials/blob/master/LICENSE.txt)\n",
    "which allows very broad use for both academic and commercial purposes.\n",
    "\n",
    "You are very welcome to modify and use this source-code in your own project. Please keep a link to the [original repository](https://github.com/simfin/simfin-tutorials).\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}