Day3/OOP_QUAIL.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# QUAIL Pipeline\n",
    "**Quick-RAIL, A Simplified RAIL Degradation Pipeline**\n",
    "\n",
    "This pipeline heavily references the pipeline created in RAIL's Degradation Demo notebook, which can be found [rendered on ReadTheDocs](https://rail-hub.readthedocs.io/projects/rail-notebooks/en/latest/rendered/creation_examples/degradation-demo.html) and in [notebook form on GitHub](https://github.com/LSSTDESC/rail/blob/main/examples/creation_examples/degradation-demo.ipynb)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Welcome!\n",
    "\n",
    "The recommended order of action is as follows: \n",
    "\n",
    "_(However, because the pipeline is already able to run as-is, you are free to skip any steps and come back to them later.)_\n",
    "\n",
    "1. Read through the whole notebook.\n",
    "1. Implement an Incorrectness stage.\n",
    "    1. Each row will have some change of its redshift value being changed to a random value.\n",
    "    1. This kind of degrader is especially useful in cases where you need a pipeline to test for the impact of incorrect values in your training set, and how it will impact your estimations on your test set. In essense, how sensitive is your estimator to bad values in your training set? If even a tenth of a percent of values being bad will ruin things, that's good to know up front!\n",
    "1. Implement InvRedshiftIncompleteness.\n",
    "    1. This degrader applies a selection function, which keeps galaxies with probability $p_{\\text{keep}}(z) = \\min(1, \\frac{z_p}{z})$, where $z_p$ is the ''pivot'' redshift. We'll use $z_p = 0.8$.\n",
    "    1. If this speaks to you, consider other incompleteness degrader stages you could implement. Different survival functions can be applied to different bands (see [Malmquist bias](https://en.wikipedia.org/wiki/Malmquist_bias)), and this will result in different redshift distributions. \n",
    "    1. We recommend using scipy's [continuous distributions](https://docs.scipy.org/doc/scipy/reference/stats.html#continuous-distributions), which have survival functions as a method called `sf` (eg, see [alpha.sf docs](https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.alpha.html#scipy.stats.alpha:~:text=sf(x%2C%20a%2C%20loc%3D0%2C%20scale%3D1))).\n",
    "1. Once you have at least a couple degraders to choose from, take some time to play around with the pipeline and see how different adding and removing degraders in different combinations will affect the results.\n",
    "1. Implement LSSTErrorModel to replace the BadErrorModel.\n",
    "    1. More info on how the error model is generated is in the LSST docs.\n",
    "    2. If you'd like a hint, we recommend looking at John Franklin's code in his implementation of the LSSTErrorModel ([repo here](https://github.com/jfcrenshaw/photerr/tree/main/photerr)).\n",
    "1. Implement better plotting functionality (as described at the end).\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setting things up"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Start by making a conda environment (or translate these instructions for your preferred environment manager):\n",
    "\n",
    "`conda create -n quail python=3.10`\n",
    "\n",
    "Install the following packages:\n",
    "```\n",
    "pip install pzflow\n",
    "pip install pz-rail-pzflow\n",
    "pip install matplotlib\n",
    "pip install ipykernel\n",
    "```\n",
    "\n",
    "Make your conda environment available as a kernel for your notebook:\n",
    "\n",
    "`python -m ipykernel install --user --name quail`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Then our imports:\n",
    "from numbers import Number\n",
    "import os\n",
    "\n",
    "import pzflow\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generate our truth data from PZFlow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "flow_file = os.path.join(\n",
    "    os.path.dirname(pzflow.__file__), \"example_files\", \"example-flow.pzflow.pkl\"\n",
    ")\n",
    "flowCreator_truth = pzflow.Flow(file=flow_file)\n",
    "\n",
    "n_samples = int(1e5)\n",
    "samples_truth = flowCreator_truth.sample(n_samples, seed=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>redshift</th>\n",
       "      <th>u</th>\n",
       "      <th>g</th>\n",
       "      <th>r</th>\n",
       "      <th>i</th>\n",
       "      <th>z</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.890625</td>\n",
       "      <td>27.370831</td>\n",
       "      <td>26.712662</td>\n",
       "      <td>26.025227</td>\n",
       "      <td>25.327188</td>\n",
       "      <td>25.016504</td>\n",
       "      <td>24.926823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.978239</td>\n",
       "      <td>29.557047</td>\n",
       "      <td>28.361183</td>\n",
       "      <td>27.587229</td>\n",
       "      <td>27.238544</td>\n",
       "      <td>26.628109</td>\n",
       "      <td>26.248562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.974287</td>\n",
       "      <td>26.566011</td>\n",
       "      <td>25.937712</td>\n",
       "      <td>24.787411</td>\n",
       "      <td>23.872452</td>\n",
       "      <td>23.139565</td>\n",
       "      <td>22.832047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.317978</td>\n",
       "      <td>29.042734</td>\n",
       "      <td>28.274597</td>\n",
       "      <td>27.501108</td>\n",
       "      <td>26.648792</td>\n",
       "      <td>26.091450</td>\n",
       "      <td>25.346502</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.386366</td>\n",
       "      <td>26.292625</td>\n",
       "      <td>25.774780</td>\n",
       "      <td>25.429962</td>\n",
       "      <td>24.806532</td>\n",
       "      <td>24.367950</td>\n",
       "      <td>23.700010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99995</th>\n",
       "      <td>2.147171</td>\n",
       "      <td>26.550978</td>\n",
       "      <td>26.349936</td>\n",
       "      <td>26.135284</td>\n",
       "      <td>26.082018</td>\n",
       "      <td>25.911028</td>\n",
       "      <td>25.558134</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99996</th>\n",
       "      <td>1.457508</td>\n",
       "      <td>27.362207</td>\n",
       "      <td>27.036274</td>\n",
       "      <td>26.823139</td>\n",
       "      <td>26.420128</td>\n",
       "      <td>26.110035</td>\n",
       "      <td>25.524902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99997</th>\n",
       "      <td>1.372993</td>\n",
       "      <td>27.736044</td>\n",
       "      <td>27.271955</td>\n",
       "      <td>26.887581</td>\n",
       "      <td>26.416138</td>\n",
       "      <td>26.043434</td>\n",
       "      <td>25.456161</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99998</th>\n",
       "      <td>0.855022</td>\n",
       "      <td>28.044552</td>\n",
       "      <td>27.327118</td>\n",
       "      <td>26.599014</td>\n",
       "      <td>25.862333</td>\n",
       "      <td>25.592169</td>\n",
       "      <td>25.506388</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99999</th>\n",
       "      <td>1.723768</td>\n",
       "      <td>27.049068</td>\n",
       "      <td>26.526745</td>\n",
       "      <td>26.094595</td>\n",
       "      <td>25.642973</td>\n",
       "      <td>25.197958</td>\n",
       "      <td>24.900501</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>100000 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       redshift          u          g          r          i          z  \\\n",
       "0      0.890625  27.370831  26.712662  26.025227  25.327188  25.016504   \n",
       "1      1.978239  29.557047  28.361183  27.587229  27.238544  26.628109   \n",
       "2      0.974287  26.566011  25.937712  24.787411  23.872452  23.139565   \n",
       "3      1.317978  29.042734  28.274597  27.501108  26.648792  26.091450   \n",
       "4      1.386366  26.292625  25.774780  25.429962  24.806532  24.367950   \n",
       "...         ...        ...        ...        ...        ...        ...   \n",
       "99995  2.147171  26.550978  26.349936  26.135284  26.082018  25.911028   \n",
       "99996  1.457508  27.362207  27.036274  26.823139  26.420128  26.110035   \n",
       "99997  1.372993  27.736044  27.271955  26.887581  26.416138  26.043434   \n",
       "99998  0.855022  28.044552  27.327118  26.599014  25.862333  25.592169   \n",
       "99999  1.723768  27.049068  26.526745  26.094595  25.642973  25.197958   \n",
       "\n",
       "               y  \n",
       "0      24.926823  \n",
       "1      26.248562  \n",
       "2      22.832047  \n",
       "3      25.346502  \n",
       "4      23.700010  \n",
       "...          ...  \n",
       "99995  25.558134  \n",
       "99996  25.524902  \n",
       "99997  25.456161  \n",
       "99998  25.506388  \n",
       "99999  24.900501  \n",
       "\n",
       "[100000 rows x 7 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "samples_truth"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## QUAIL Base Classes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We'll be using a few highly simplified versions of RAIL classes, namely:\n",
    "- QuailStage (accompanied by a NothingStage, to demonstrate how we'll inherit for our other degrader stages)\n",
    "- DataStore\n",
    "- Pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from abc import ABC, abstractmethod\n",
    "\n",
    "class QuailStage(ABC):\n",
    "    \"\"\"A class for the QuailStage stage.\"\"\"\n",
    "\n",
    "    def __init__(self, name):\n",
    "        \"\"\"Constructor.\n",
    "        \n",
    "        Parameters\n",
    "        ----------\n",
    "        name : str\n",
    "            The (human-readable) name of the stage (this will be used in the Pipeline's __repr__).\n",
    "        \"\"\"\n",
    "        self.name = name\n",
    "        self.data_in = None\n",
    "        self.data_out = None\n",
    "\n",
    "    @abstractmethod\n",
    "    def run(self):\n",
    "        \"\"\"To be implemented in subclasses. This should set the data_out attribute.\"\"\"\n",
    "        pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "class NothingStage(QuailStage):\n",
    "    \"\"\"A stage that does nothing.\"\"\"\n",
    "    def __init__(self, name):\n",
    "        \"\"\"Constructor.\n",
    "        \n",
    "        Parameters\n",
    "        ----------\n",
    "        name : str\n",
    "            The name of the stage.\n",
    "        \"\"\"\n",
    "        super().__init__(name)\n",
    "\n",
    "    def run(self):\n",
    "        \"\"\"Run the stage.\"\"\"\n",
    "        self.data_out = self.data_in"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Degrader 1: BadErrorModel\n",
    "\n",
    "We start by creating an incredibly naive error model as a stand-in for the LSSTErrorModel.\n",
    "\n",
    "This will generate u_err, g_err, etc. columns for us, but there's no scientific basis for the values it generates."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "class BadErrorModel(QuailStage):\n",
    "    \"\"\"A stage that applies bad errors.\"\"\"\n",
    "    def __init__(self, name):\n",
    "        \"\"\"Constructor.\n",
    "        \n",
    "        Parameters\n",
    "        ----------\n",
    "        name : str\n",
    "            The name of the stage.\n",
    "        \"\"\"\n",
    "        super().__init__(name)\n",
    "\n",
    "    def apply_errors(self):\n",
    "        \"\"\"Apply the errors to the input data.\n",
    "        \n",
    "        Returns\n",
    "        -------\n",
    "        errors : pandas.DataFrame\n",
    "            The input data with errors applied.\n",
    "        \"\"\"\n",
    "        # create a copy of the input data\n",
    "        errors = self.data_in.copy()\n",
    "        for col in [\"u\", \"g\", \"r\", \"i\", \"z\", \"y\"]:\n",
    "            # create a new column with the error\n",
    "            errors[f\"{col}_err\"] = 0.01 * self.data_in[col]\n",
    "        # reorder the columns\n",
    "        errors = errors[[\"redshift\", \"u\", \"u_err\", \"g\", \"g_err\", \"r\", \"r_err\", \"i\", \"i_err\", \"z\", \"z_err\", \"y\", \"y_err\"]]\n",
    "        return errors\n",
    "\n",
    "    def run(self):\n",
    "        \"\"\"Run the stage.\"\"\"\n",
    "        self.data_out = self.apply_errors()\n",
    "\n",
    "    def plot(self):\n",
    "        \"\"\"Plot the truth data and the errors.\"\"\"\n",
    "        if self.data_out is None:\n",
    "            raise ValueError(\"You must run the stage first.\")\n",
    "        \n",
    "        fig, ax = plt.subplots(figsize=(5, 4), dpi=100)\n",
    "\n",
    "        for band in \"ugrizy\":\n",
    "            # pull out the magnitudes and errors\n",
    "            mags = self.data_out[band].to_numpy()\n",
    "            errs = self.data_out[band + \"_err\"].to_numpy()\n",
    "\n",
    "            # sort them by magnitude\n",
    "            mags, errs = mags[mags.argsort()], errs[mags.argsort()]\n",
    "\n",
    "            # plot errs vs mags\n",
    "            ax.plot(mags, errs, label=band)\n",
    "\n",
    "        ax.legend()\n",
    "        ax.set(xlabel=\"Magnitude (AB)\", ylabel=\"Error (mags)\")\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Degrader 2: QuantityCut\n",
    "\n",
    "This degrader is based on our QuantityCut example in the degradation demo notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "class QuantityCut(QuailStage):\n",
    "    \"\"\"Applies a cut to the input data based on a quantity.\n",
    "    \n",
    "    Parameters\n",
    "    ----------\n",
    "    name : str\n",
    "        The name of the stage.\n",
    "    cuts : dict\n",
    "        A dictionary of cuts to apply to the data. \n",
    "\n",
    "    Notes\n",
    "    -----\n",
    "    The cut keys should be the names of columns you wish to make cuts on.\n",
    "\n",
    "    The cut values should be either:\n",
    "    - a number, which is the maximum value. E.g. if the dictionary\n",
    "    contains \"i\": 25, then values of i > 25 are cut from the sample.\n",
    "    - an iterable, which is the range of acceptable values. E.g. if the \n",
    "    dictionary contains \"redshift\": (1.5, 2.3), then redshifts outside that\n",
    "    range are cut from the sample.\n",
    "    \"\"\"\n",
    "    def __init__(self, name=\"QuantityCut\", cuts=None):\n",
    "        \"\"\"Constructor.\n",
    "        \"\"\"\n",
    "        super().__init__(name=name)\n",
    "        self.cuts = cuts \n",
    "        self.data_out = None\n",
    "\n",
    "    def apply_cut(self):\n",
    "        \"\"\"Apply the cuts to the input data.\n",
    "        \n",
    "        Returns\n",
    "        -------\n",
    "        data_out : pandas.DataFrame\n",
    "            The input data with the cut applied.\n",
    "        \"\"\"\n",
    "        # create a copy of the input data\n",
    "        data_out = self.data_in.copy()\n",
    "        \n",
    "        # Apply the cuts\n",
    "        for col, cut in self.cuts.items():\n",
    "            if isinstance(cut, Number):\n",
    "                # Apply cut as maximum value\n",
    "                data_out = data_out[data_out[col] <= cut]\n",
    "            elif hasattr(cut, \"__iter__\"):\n",
    "                # Apply cut as range of acceptable values\n",
    "                data_out = data_out[(data_out[col] >= cut[0]) & (data_out[col] <= cut[1])]\n",
    "            else:\n",
    "                raise TypeError(\"Cut must be a number or an iterable of (min, max)\")\n",
    "\n",
    "        return data_out\n",
    "\n",
    "    def run(self):\n",
    "        \"\"\"Run the stage.\"\"\"\n",
    "        data_out = self.apply_cut()\n",
    "        self.data_out = data_out"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Degrader 3: InvRedshiftIncompleteness\n",
    "\n",
    "This one is up to you! Feel free to reference "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Potentially, write an InvRedshiftIncompleteness for them here"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Making the Pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "class DataStore(dict):\n",
    "    def __init__(self):\n",
    "        dict.__init__(self)\n",
    "    \n",
    "    def __getattr__(self, key): # Code copied from RAIL's DataStore \n",
    "        \"\"\"Allow attribute-like parameter access\"\"\"\n",
    "        try:\n",
    "            return self.__getitem__(key)\n",
    "        except KeyError as msg:\n",
    "            # Kludge to get docstrings to work\n",
    "            if key in [\"__objclass__\"]:  # pragma: no cover\n",
    "                return None\n",
    "            raise KeyError from msg\n",
    "    \n",
    "    def __repr__(self): # Feel free to change this output into whatever you like!\n",
    "        \"\"\"Prints the DataStore keys and shapes in a human-readable format.\"\"\"\n",
    "        s = \"DataStore\\n\"\n",
    "        for key in self.keys():\n",
    "            s += f\"  {key}:\\n\\t{self[key].shape}\\n\"\n",
    "        return s\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Pipeline:\n",
    "    def __init__(self, data_store, stages):\n",
    "        self.data_store = data_store\n",
    "        self.stages = stages\n",
    "\n",
    "    def run(self):\n",
    "        current_data = self.data_store[\"input_data\"]\n",
    "        for stage in self.stages:\n",
    "            print(f\"Running stage: {stage.name}\")\n",
    "            stage.data_in = current_data\n",
    "            stage.run()\n",
    "            self.data_store[f\"{stage.name}_data\"] = stage.data_out\n",
    "            current_data = stage.data_out\n",
    "        self.data_store[\"output_data\"] = current_data\n",
    "\n",
    "    def __repr__(self):\n",
    "        s = \"Pipeline\\n\"\n",
    "        for stage in self.stages:\n",
    "            s += f\"  {stage.name}\\n\"\n",
    "        return s\n",
    "    \n",
    "    def get_stage(self, stage_name):\n",
    "        for stage in self.stages:\n",
    "            if stage.name == stage_name:\n",
    "                return stage\n",
    "        print(f\"Stage {stage_name} not found in pipeline.\")\n",
    "        return None\n",
    "\n",
    "    def plot(self):\n",
    "        pass\n",
    "\n",
    "    def save(self):\n",
    "        pass\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "DS = DataStore()\n",
    "DS[\"input_data\"] = samples_truth\n",
    "\n",
    "stages = [\n",
    "    NothingStage(\"do_nothing\"),\n",
    "    BadErrorModel(\"bad_error_model\"),\n",
    "    QuantityCut(\"quality_cut\", cuts={\"i\": 25.3}),\n",
    "]\n",
    "\n",
    "pipeline = Pipeline(DS, stages)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DataStore\n",
       "  input_data:\n",
       "\t(100000, 7)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Our DataStore before running the pipeline\n",
    "\n",
    "DS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Running stage: do_nothing\n",
      "Running stage: bad_error_model\n",
      "Running stage: quality_cut\n"
     ]
    }
   ],
   "source": [
    "pipeline.run()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DataStore\n",
       "  input_data:\n",
       "\t(100000, 7)\n",
       "  do_nothing_data:\n",
       "\t(100000, 7)\n",
       "  bad_error_model_data:\n",
       "\t(100000, 13)\n",
       "  quality_cut_data:\n",
       "\t(51834, 13)\n",
       "  output_data:\n",
       "\t(51834, 13)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Our DataStore after running the pipeline\n",
    "\n",
    "DS"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Looking at the data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Poke around and see the changes you've made! Note that overloading the `__getattr__` in our DataStore class lets us access data via `DS.key` syntax."
   ]
  },
  {
   "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>redshift</th>\n",
       "      <th>u</th>\n",
       "      <th>u_err</th>\n",
       "      <th>g</th>\n",
       "      <th>g_err</th>\n",
       "      <th>r</th>\n",
       "      <th>r_err</th>\n",
       "      <th>i</th>\n",
       "      <th>i_err</th>\n",
       "      <th>z</th>\n",
       "      <th>z_err</th>\n",
       "      <th>y</th>\n",
       "      <th>y_err</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.974287</td>\n",
       "      <td>26.566011</td>\n",
       "      <td>0.265660</td>\n",
       "      <td>25.937712</td>\n",
       "      <td>0.259377</td>\n",
       "      <td>24.787411</td>\n",
       "      <td>0.247874</td>\n",
       "      <td>23.872452</td>\n",
       "      <td>0.238725</td>\n",
       "      <td>23.139565</td>\n",
       "      <td>0.231396</td>\n",
       "      <td>22.832047</td>\n",
       "      <td>0.228320</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.386366</td>\n",
       "      <td>26.292625</td>\n",
       "      <td>0.262926</td>\n",
       "      <td>25.774780</td>\n",
       "      <td>0.257748</td>\n",
       "      <td>25.429962</td>\n",
       "      <td>0.254300</td>\n",
       "      <td>24.806532</td>\n",
       "      <td>0.248065</td>\n",
       "      <td>24.367950</td>\n",
       "      <td>0.243679</td>\n",
       "      <td>23.700010</td>\n",
       "      <td>0.237000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.781594</td>\n",
       "      <td>25.260241</td>\n",
       "      <td>0.252602</td>\n",
       "      <td>24.760660</td>\n",
       "      <td>0.247607</td>\n",
       "      <td>24.190182</td>\n",
       "      <td>0.241902</td>\n",
       "      <td>23.490292</td>\n",
       "      <td>0.234903</td>\n",
       "      <td>23.254755</td>\n",
       "      <td>0.232548</td>\n",
       "      <td>23.142416</td>\n",
       "      <td>0.231424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.241211</td>\n",
       "      <td>25.242661</td>\n",
       "      <td>0.252427</td>\n",
       "      <td>24.297731</td>\n",
       "      <td>0.242977</td>\n",
       "      <td>23.660130</td>\n",
       "      <td>0.236601</td>\n",
       "      <td>23.434042</td>\n",
       "      <td>0.234340</td>\n",
       "      <td>23.290945</td>\n",
       "      <td>0.232909</td>\n",
       "      <td>23.230438</td>\n",
       "      <td>0.232304</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.679428</td>\n",
       "      <td>25.597195</td>\n",
       "      <td>0.255972</td>\n",
       "      <td>24.576441</td>\n",
       "      <td>0.245764</td>\n",
       "      <td>23.358215</td>\n",
       "      <td>0.233582</td>\n",
       "      <td>22.386457</td>\n",
       "      <td>0.223865</td>\n",
       "      <td>22.045208</td>\n",
       "      <td>0.220452</td>\n",
       "      <td>21.766764</td>\n",
       "      <td>0.217668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99985</th>\n",
       "      <td>0.655446</td>\n",
       "      <td>27.938507</td>\n",
       "      <td>0.279385</td>\n",
       "      <td>26.373728</td>\n",
       "      <td>0.263737</td>\n",
       "      <td>24.935673</td>\n",
       "      <td>0.249357</td>\n",
       "      <td>23.836836</td>\n",
       "      <td>0.238368</td>\n",
       "      <td>23.449917</td>\n",
       "      <td>0.234499</td>\n",
       "      <td>23.138634</td>\n",
       "      <td>0.231386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99987</th>\n",
       "      <td>0.533671</td>\n",
       "      <td>25.913958</td>\n",
       "      <td>0.259140</td>\n",
       "      <td>23.062689</td>\n",
       "      <td>0.230627</td>\n",
       "      <td>21.498064</td>\n",
       "      <td>0.214981</td>\n",
       "      <td>20.530350</td>\n",
       "      <td>0.205303</td>\n",
       "      <td>20.118843</td>\n",
       "      <td>0.201188</td>\n",
       "      <td>19.873358</td>\n",
       "      <td>0.198734</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99990</th>\n",
       "      <td>0.890144</td>\n",
       "      <td>25.153248</td>\n",
       "      <td>0.251532</td>\n",
       "      <td>24.985104</td>\n",
       "      <td>0.249851</td>\n",
       "      <td>24.572035</td>\n",
       "      <td>0.245720</td>\n",
       "      <td>23.884954</td>\n",
       "      <td>0.238850</td>\n",
       "      <td>23.539755</td>\n",
       "      <td>0.235398</td>\n",
       "      <td>23.409140</td>\n",
       "      <td>0.234091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99992</th>\n",
       "      <td>0.857895</td>\n",
       "      <td>26.644789</td>\n",
       "      <td>0.266448</td>\n",
       "      <td>26.049950</td>\n",
       "      <td>0.260499</td>\n",
       "      <td>24.878296</td>\n",
       "      <td>0.248783</td>\n",
       "      <td>23.759335</td>\n",
       "      <td>0.237593</td>\n",
       "      <td>23.270386</td>\n",
       "      <td>0.232704</td>\n",
       "      <td>23.129335</td>\n",
       "      <td>0.231293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99994</th>\n",
       "      <td>0.767410</td>\n",
       "      <td>25.276413</td>\n",
       "      <td>0.252764</td>\n",
       "      <td>24.816841</td>\n",
       "      <td>0.248168</td>\n",
       "      <td>24.226608</td>\n",
       "      <td>0.242266</td>\n",
       "      <td>23.505165</td>\n",
       "      <td>0.235052</td>\n",
       "      <td>23.275276</td>\n",
       "      <td>0.232753</td>\n",
       "      <td>23.138519</td>\n",
       "      <td>0.231385</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>51834 rows × 13 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       redshift          u     u_err          g     g_err          r  \\\n",
       "2      0.974287  26.566011  0.265660  25.937712  0.259377  24.787411   \n",
       "4      1.386366  26.292625  0.262926  25.774780  0.257748  25.429962   \n",
       "5      0.781594  25.260241  0.252602  24.760660  0.247607  24.190182   \n",
       "7      0.241211  25.242661  0.252427  24.297731  0.242977  23.660130   \n",
       "8      0.679428  25.597195  0.255972  24.576441  0.245764  23.358215   \n",
       "...         ...        ...       ...        ...       ...        ...   \n",
       "99985  0.655446  27.938507  0.279385  26.373728  0.263737  24.935673   \n",
       "99987  0.533671  25.913958  0.259140  23.062689  0.230627  21.498064   \n",
       "99990  0.890144  25.153248  0.251532  24.985104  0.249851  24.572035   \n",
       "99992  0.857895  26.644789  0.266448  26.049950  0.260499  24.878296   \n",
       "99994  0.767410  25.276413  0.252764  24.816841  0.248168  24.226608   \n",
       "\n",
       "          r_err          i     i_err          z     z_err          y     y_err  \n",
       "2      0.247874  23.872452  0.238725  23.139565  0.231396  22.832047  0.228320  \n",
       "4      0.254300  24.806532  0.248065  24.367950  0.243679  23.700010  0.237000  \n",
       "5      0.241902  23.490292  0.234903  23.254755  0.232548  23.142416  0.231424  \n",
       "7      0.236601  23.434042  0.234340  23.290945  0.232909  23.230438  0.232304  \n",
       "8      0.233582  22.386457  0.223865  22.045208  0.220452  21.766764  0.217668  \n",
       "...         ...        ...       ...        ...       ...        ...       ...  \n",
       "99985  0.249357  23.836836  0.238368  23.449917  0.234499  23.138634  0.231386  \n",
       "99987  0.214981  20.530350  0.205303  20.118843  0.201188  19.873358  0.198734  \n",
       "99990  0.245720  23.884954  0.238850  23.539755  0.235398  23.409140  0.234091  \n",
       "99992  0.248783  23.759335  0.237593  23.270386  0.232704  23.129335  0.231293  \n",
       "99994  0.242266  23.505165  0.235052  23.275276  0.232753  23.138519  0.231385  \n",
       "\n",
       "[51834 rows x 13 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DS.quality_cut_data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plotting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can access the plot method of our error model like so:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pipeline.get_stage(\"bad_error_model\").plot()  # Compare this to the plot in the RAIL degrader notebook,\n",
    "                                              # where we see very different errors!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And here, we grab a snippet of plotting code from the degrader demo notebook. \n",
    "\n",
    "As it is now, we can compare our DataStore's input data and its output data. \n",
    "\n",
    "What else would you want to show? And where would you fit the code in?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeIAAAF2CAYAAACyHpv2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy80BEi2AAAACXBIWXMAAA9hAAAPYQGoP6dpAABXz0lEQVR4nO3de1wU9foH8M9yv99UboqI4g1E8e5284auSqbJKS1TvGUSmIhXPGqhKelR8ZpUFnhKMy31eMCjIoqmYhZK4iVSQ0VjIU0gEReE+f3Bj8mV2y4s7MJ+3q/Xvo4z852ZZ6Y9+zAz33m+EkEQBBAREZFWGGg7ACIiIn3GRExERKRFTMRERERaxERMRESkRUzEREREWsRETEREpEVMxERERFrERExERKRFTMRERERaxERMRESkRTqTiD/66CNIJBKEhoaK8x4/fozg4GA0a9YMVlZWCAgIQHZ2ttJ6t2/fhr+/PywsLODo6Ih58+bhyZMnSm2SkpLQo0cPmJqawtPTE7GxsQ1wRERERDUz0nYAAPDjjz/ik08+QdeuXZXmz549G/Hx8dizZw9sbW0REhKCMWPG4PTp0wCAkpIS+Pv7w9nZGWfOnEFWVhYmTpwIY2NjrFy5EgCQkZEBf39/zJgxAzt27EBiYiKmTZsGFxcXyGQyleIrLS3F77//Dmtra0gkEs0ePBERNRqCIOCvv/6Cq6srDAw0dC0raNlff/0ltG/fXkhISBD69+8vzJo1SxAEQcjNzRWMjY2FPXv2iG2vXr0qABCSk5MFQRCEgwcPCgYGBoJcLhfbbN26VbCxsREUCoUgCIIwf/58wdvbW2mfY8eOFWQymcoxZmZmCgD44YcffvjhRwAgZGZm1jbtVaD1K+Lg4GD4+/vDz88PH374oTg/JSUFxcXF8PPzE+d16tQJrVu3RnJyMvr164fk5GT4+PjAyclJbCOTyRAUFITLly+je/fuSE5OVtpGeZunb4E/S6FQQKFQiNPC/w9QlZmZCRsbm7oeMhERNVL5+flwc3ODtbW1xrap1US8a9cunD9/Hj/++GOFZXK5HCYmJrCzs1Oa7+TkBLlcLrZ5OgmXLy9fVl2b/Px8FBYWwtzcvMK+IyMjERERUWG+jY0NEzEREWn0MaXWOmtlZmZi1qxZ2LFjB8zMzLQVRqXCw8ORl5cnfjIzM7UdEhERNVFaS8QpKSnIyclBjx49YGRkBCMjI5w4cQIbN26EkZERnJycUFRUhNzcXKX1srOz4ezsDABwdnau0Iu6fLqmNjY2NpVeDQOAqampePXLq2AiIqpPWkvEgwcPRlpaGlJTU8VPr169MH78ePHfxsbGSExMFNdJT0/H7du3IZVKAQBSqRRpaWnIyckR2yQkJMDGxgZeXl5im6e3Ud6mfBtERETapLVnxNbW1ujSpYvSPEtLSzRr1kycP3XqVISFhcHBwQE2NjaYOXMmpFIp+vXrBwAYOnQovLy8MGHCBKxevRpyuRyLFy9GcHAwTE1NAQAzZszA5s2bMX/+fEyZMgXHjh3D7t27ER8f37AHTERqEwQBT548QUlJibZDIT1haGgIIyOjBn1VVeu9pqsTFRUFAwMDBAQEQKFQQCaT4eOPPxaXGxoaIi4uDkFBQZBKpbC0tERgYCCWLVsmtvHw8EB8fDxmz56NDRs2oFWrVti2bZvK7xATkXYUFRUhKysLjx490nYopGcsLCzg4uICExOTBtmfRCh/N4eqlJ+fD1tbW+Tl5fF5MVEDKC0txbVr12BoaIgWLVrAxMSExXSo3gmCgKKiIvzxxx8oKSlB+/btKxTtqI98oNNXxESkn4qKilBaWgo3NzdYWFhoOxzSI+bm5jA2NsatW7dQVFTUIG/16EytaSKiZ2mshCCRGhr6e8dvORERkRYxERMRkRKJRIL9+/drOwy9wWfEVK27uYV4UFBU5XJ7SxO0tKu8MAoRVe2PP/7A0qVLER8fj+zsbNjb26Nbt25YunQpnn/+eW2HRw2IiZiqdDe3EH5rT6CwuOp3OM2NDXF0Tn8mYyI1BQQEoKioCNu3b0fbtm2RnZ2NxMRE3L9/X9uhUQPjrWmq0oOCIhQWl2D9WF/EzXyhwmf9WF8UFpdUe8VMRBXl5ubi+++/x6pVqzBw4EC4u7ujT58+CA8PxyuvvAIAWLduHXx8fGBpaQk3Nze8++67ePjwobiN2NhY2NnZIS4uDh07doSFhQX+8Y9/4NGjR9i+fTvatGkDe3t7vPfee0oFUdq0aYPly5fjjTfegKWlJVq2bIktW7ZUG29mZiZef/112NnZwcHBAaNGjcLNmzfr5dzoIyZiqpGnoxW6tLSt8PF0tNJ2aESNkpWVFaysrLB//36lIVefZmBggI0bN+Ly5cvYvn07jh07hvnz5yu1efToETZu3Ihdu3bh0KFDSEpKwquvvoqDBw/i4MGD+PLLL/HJJ5/g22+/VVrvX//6F7p164YLFy5g4cKFmDVrFhISEiqNo7i4GDKZDNbW1vj+++9x+vRpWFlZYdiwYSgq4h/hGqGxkY2bsLy8PAGAkJeXp+1QGlTanVzBfUGckHYnt1bLiWqrsLBQuHLlilBYWKjtUOrNt99+K9jb2wtmZmbCc889J4SHhws///xzle337NkjNGvWTJyOiYkRAAjXr18X573zzjuChYWF8Ndff4nzZDKZ8M4774jT7u7uwrBhw5S2PXbsWGH48OHiNABh3759giAIwpdffil07NhRKC0tFZcrFArB3NxcOHz4sPoH3ghU9/2rj3zAK2IiIi0ICAjA77//jgMHDmDYsGFISkpCjx49EBsbCwA4evQoBg8ejJYtW8La2hoTJkzA/fv3lUp+WlhYoF27duK0k5MT2rRpAysrK6V5Tw+MA6DCoDdSqRRXr16tNM6ff/4Z169fh7W1tXgl7+DggMePH+PGjRt1PQ0EdtYiItIaMzMzDBkyBEOGDMGSJUswbdo0vP/++xgwYABefvllBAUFYcWKFXBwcMCpU6cwdepUFBUVidXGjI2NlbYnkUgqnVdaWlrrGB8+fIiePXtix44dFZa1aNGi1tulvzERExHpCC8vL+zfvx8pKSkoLS3F2rVrxSpPu3fv1th+zp49W2G6c+fOlbbt0aMHvvnmGzg6OrLWfj3hrWkiogZ2//59DBo0CF999RUuXryIjIwM7NmzB6tXr8aoUaPg6emJ4uJibNq0Cb/99hu+/PJLREdHa2z/p0+fxurVq/Hrr79iy5Yt2LNnD2bNmlVp2/Hjx6N58+YYNWoUvv/+e2RkZCApKQnvvfce7ty5o7GY9BmviImIGpiVlRX69u2LqKgo3LhxA8XFxXBzc8Pbb7+NRYsWwdzcHOvWrcOqVasQHh6Ol156CZGRkZg4caJG9j9nzhz89NNPiIiIgI2NDdatW1fl0LAWFhY4efIkFixYgDFjxuCvv/5Cy5YtMXjwYF4hawiHQVSBvg6DeOluHl7edApxM19Al5a2ai8nqq3Hjx8jIyMDHh4eDTL6jT5p06YNQkNDERoaqu1QdFZ137/6yAe8NU1ERKRFTMRERERaxGfERER6hKUpdQ+viImIiLSIiZiIiEiLmIiJiIi0iImYiIhIi5iIiYiItIiJmIiISIv4+hIRNSp3cwvxoKBhBqS3tzRBSzvzBtkX6S8mYiJqNO7mFsJv7QkUFpc0yP7MjQ1xdE5/lZPxpEmTsH37dgCAkZERHBwc0LVrV7zxxhuYNGmSOJISUDbO75IlS3D27Fnk5+fD2dkZffv2xaZNm+Do6Fjp9gcMGABfX1+sX7++zsemjg8++AD79+9Hampqje0iIiIAAIaGhrCzs4OXlxfGjBmDoKAgmJqaqrzPpKQkDBw4EA8ePICdnV0dotd9TMRE1Gg8KChCYXEJ1o/1haejVb3u63rOQ4R+k4oHBUVqXRUPGzYMMTExKCkpQXZ2Ng4dOoRZs2bh22+/xYEDB2BkZIQ//vgDgwcPxssvv4zDhw/Dzs4ON2/exIEDB1BQUFCPR1X/vL29cfToUZSWluL+/ftISkrChx9+iC+//BJJSUmwtrbWdoi6R9Cijz/+WPDx8RGsra0Fa2troV+/fsLBgwfF5f379xcAKH3eeecdpW3cunVLGDFihGBubi60aNFCmDt3rlBcXKzU5vjx40L37t0FExMToV27dkJMTIxacebl5QkAhLy8vFofa2OUdidXcF8QJ6Tdya3VcqLaKiwsFK5cuSIUFhYqzW/I71xt9hUYGCiMGjWqwvzExEQBgPDZZ58JgiAI+/btE4yMjCr8VtWkf//+wqxZs8Rpd3d3YcWKFcLkyZMFKysrwc3NTfjkk0/E5RkZGQIA4euvvxakUqlgamoqeHt7C0lJSWKbmJgYwdbWVmk/+/btE8rTQ0xMTIXf4ap+Q99//32hW7duFeZfvXpVMDExEf75z3+K8/79738LPXv2FKysrAQnJyfhjTfeELKzs5XifvoTGBgoCIIg/O9//xOef/55wdbWVnBwcBD8/f2F69evq3EWa1bV908Q6icfaLWzVqtWrfDRRx8hJSUFP/30EwYNGoRRo0bh8uXLYpu3334bWVlZ4mf16tXispKSEvj7+6OoqAhnzpzB9u3bERsbi6VLl4ptMjIy4O/vj4EDByI1NRWhoaGYNm0aDh8+3KDHSkT6a9CgQejWrRv27t0LAHB2dsaTJ0+wb98+CHUcAG/t2rXo1asXLly4gHfffRdBQUFIT09XajNv3jzMmTMHFy5cgFQqxciRI3H//n2Vtj927FjMmTMH3t7e4u/w2LFj1YqxU6dOGD58uHj8AFBcXIzly5fj559/xv79+3Hz5k1MmjQJAODm5obvvvsOAJCeno6srCxs2LABAFBQUICwsDD89NNPSExMhIGBAV599VWUlpaqFZMu0eqt6ZEjRypNr1ixAlu3bsXZs2fh7e0NoGwsTGdn50rXP3LkCK5cuYKjR4/CyckJvr6+WL58ORYsWIAPPvgAJiYmiI6OhoeHB9auXQsA6Ny5M06dOoWoqKgqx98kItK0Tp064eLFiwCAfv36YdGiRXjzzTcxY8YM9OnTB4MGDcLEiRPh5OSk1nZHjBiBd999FwCwYMECREVF4fjx4+jYsaPYJiQkBAEBAQCArVu34tChQ/j8888xf/78Grdvbm4OKysrGBkZVflbrIpOnTrhyJEj4vSUKVPEf7dt2xYbN25E79698fDhQ1hZWcHBwQEA4OjoqPSMuPw4yn3xxRdo0aIFrly5gi5dutQ6Pm3SmdeXSkpKsGvXLhQUFEAqlYrzd+zYgebNm6NLly4IDw/Ho0ePxGXJycnw8fFR+uLKZDLk5+eLV9XJycnw8/NT2pdMJkNycnI9HxER0d8EQYBEIhGnV6xYAblcjujoaHh7eyM6OhqdOnVCWlqaWtvt2rWr+G+JRAJnZ2fk5OQotXn6N9XIyAi9evXC1atXa3kktfPs8aekpGDkyJFo3bo1rK2t0b9/fwDA7du3q93OtWvX8MYbb6Bt27awsbFBmzZtVFpPl2m9s1ZaWhqkUikeP34MKysr7Nu3D15eXgCAN998E+7u7nB1dcXFixexYMECpKeni7c35HJ5hb8ey6flcnm1bfLz81FYWAhz84qdMBQKBRQKhTidn5+vuQMmIr109epVeHh4KM1r1qwZXnvtNbz22mtYuXIlunfvjjVr1og9r1VhbGysNC2RSNS6TWtgYFDh9nhxcbHK66vq6eMvKCiATCaDTCbDjh070KJFC9y+fRsymQxFRdW/mjZy5Ei4u7vjs88+g6urK0pLS9GlS5ca19NlWk/EHTt2RGpqKvLy8vDtt98iMDAQJ06cgJeXF6ZPny628/HxgYuLCwYPHowbN26gXbt29RZTZGSk2AWfiKiujh07hrS0NMyePbvKNiYmJmjXrl299Jo+e/YsXnrpJQDAkydPkJKSgpCQEABAixYt8Ndff6GgoACWlpYAUOE1JRMTE5SU1P6VsV9++QWHDh1CeHi4OH3//n189NFHcHNzAwD89NNPFfYJQGm/9+/fR3p6Oj777DO8+OKLAIBTp07VOi5dofVb0yYmJvD09ETPnj0RGRmJbt26iQ/ln9W3b18AwPXr1wGUdXjIzs5WalM+Xf4so6o2NjY2lV4NA0B4eDjy8vLET2ZmZu0PkIj0ikKhgFwux927d3H+/HmsXLkSo0aNwssvv4yJEycCAOLi4vDWW28hLi4Ov/76K9LT07FmzRocPHgQo0aN0nhMW7Zswb59+/DLL78gODgYDx48EJ/R9u3bFxYWFli0aBFu3LiBnTt3IjY2Vmn9Nm3aICMjA6mpqbh3757SHcNnPXnyBHK5HL///jvS0tKwadMm9O/fH76+vpg3bx4AoHXr1jAxMcGmTZvw22+/4cCBA1i+fLnSdtzd3SGRSBAXF4c//vgDDx8+hL29PZo1a4ZPP/0U169fx7FjxxAWFqbZk6UFWr8iflZpaWmV/5HL/0pzcXEBUPbcY8WKFcjJyRFfgE9ISICNjY14e1sqleLgwYNK20lISFB6ZvIsU1NTtV48p9qrrkoSqxpRVa7nPNTZfRw6dAguLi4wMjKCvb09unXrho0bNyIwMFAs6OHl5QULCwvMmTMHmZmZMDU1Rfv27bFt2zZMmDBBk4cBAPjoo4/w0UcfITU1FZ6enjhw4ACaN28OAHBwcMBXX32FefPm4bPPPsPgwYPxwQcfKN2RDAgIwN69ezFw4EDk5uYiJiZG7OH8rMuXL8PFxQWGhoawtbWFl5cXwsPDlQp6tGjRArGxsVi0aBE2btyIHj16YM2aNXjllVfE7bRs2RIRERFYuHAhJk+ejIkTJyI2Nha7du3Ce++9hy5duqBjx47YuHEjBgwYoPFz1pAkQl37ztdBeHg4hg8fjtatW+Ovv/7Czp07sWrVKhw+fBht27bFzp07MWLECDRr1gwXL17E7Nmz0apVK5w4cQJA2S0LX19fuLq6YvXq1ZDL5ZgwYQKmTZuGlStXAih7falLly4IDg7GlClTcOzYMbz33nuIj49Xudd0fn4+bG1tkZeXBxsbm3o7H7rm0t08vLzpFOJmvoAuLW3VXl6TmqokqVvViJqOx48fIyMjAx4eHjAzMxPn63plLV1z8+ZNeHh44MKFC/D19dV2OI1GVd8/oH7ygVaviHNycjBx4kRkZWXB1tYWXbt2xeHDhzFkyBBkZmbi6NGjWL9+PQoKCuDm5oaAgAAsXrxYXN/Q0BBxcXEICgqCVCqFpaUlAgMDsWzZMrGNh4cH4uPjMXv2bGzYsAGtWrXCtm3b+OqSDqiuSlJtqxpR09bSzhxH5/RnrWlqUrSaiD///PMql7m5uYlXvtVxd3evcOv5WQMGDMCFCxfUjo8ahqejVa2uqEk/tbQzZ3KkJkXnnhETEZFmtGnTps6Vu6j+ab3XNBERkT7jFbEeqGn81ro+B6uud6kub5uISBcwETdxqvQyrW3PUHtLE5gbGyL0m9RGtW0iIl3CRNzE1TR+a116J9fUg1VXt01EpEuYiPVEffVMrs8erOwdS0T6gJ21iIiItIhXxETUuORmAo9UG9S+ziyaAXZuDbMv0ltMxNSosVe1nsnNBLb0AYof1dxWE4wtgOBzKifjSZMmiUMYGhkZwcHBAV27dsUbb7yBSZMmibWmAeDnn3/GkiVLcPbsWeTn58PZ2Rl9+/bFpk2bxNr5lbl8+TIiIiJw/Phx5Ofnw93dHePGjcPChQthYWGh8qElJSVh4MCBePDgAezs7FReTxWqltYsb1fOysoKrVu3xoABAxAaGor27durtd82bdogNDQUoaGhtYxcO5iIqVFir2o99eh+WRIe8xnQvEP97uver8Det8v2qcZV8bBhwxATE4OSkhJkZ2fj0KFDmDVrFr799lscOHAARkZG+OOPPzB48GC8/PLLOHz4MOzs7HDz5k0cOHCg2mEQz549Cz8/P/j5+SE+Ph5OTk44d+4c5syZg8TERBw/flwcPrAxOXr0KLy9vfHo0SOkpaVhw4YN6NatG/773/9i8ODB2g6v/glUo7y8PAGAkJeXp+1Q1JZ2J1dwXxAnpN3JVXt5TevW575VcefBIyHtTm6ln33n79Rp26RdhYWFwpUrV4TCwkLlBXcvCML7NmX/W99qsa/AwEBh1KhRFeYnJiYKAITPPvtMEARB2Ldvn2BkZCQUFxervO3S0lLBy8tL6NWrl1BSUqK0LDU1VZBIJMJHH30kCIIgZGRkCACECxf+jv3BgwcCAOH48ePi8qc/gYGBgiAIQv/+/YXg4GAhODhYsLGxEZo1ayYsXrxYKC0tFbcFQNi3b59SDLa2tkJMTIy4/OlP//79Kz2myuIUBEEoKSkRBgwYILi7uwtPnjwRBEEQrl+/LrzyyiuCo6OjYGlpKfTq1UtISEgQ1+nfv3+F/QqCINy7d08YN26c4OrqKpibmwtdunQRdu7cWe25rvL7J9RPPmBnLWq0WtqZo0tL20o/lb2qRaQtgwYNQrdu3bB3714AZeOkP3nyBPv27VO5BGVqaiquXLmCsLAwpVvcANCtWzf4+fnh66+/Vmlbbm5u+O677wAA6enpyMrKUhoHfvv27TAyMsK5c+ewYcMGrFu3Dtu2bVNp2wBw7tw5AGVXullZWeJxq8rAwACzZs3CrVu3kJKSAgB4+PAhRowYgcTERFy4cAHDhg3DyJEjcfv2bQDA3r170apVKyxbtgxZWVnIysoCUDaSUs+ePREfH49Lly5h+vTpmDBhghijLmAiJiJqAJ06dcLNmzcBAP369cOiRYvw5ptvonnz5hg+fDj+9a9/ITs7u8r1f/31VwBA586dK13euXNnsU1NDA0N4eDgAABwdHSEs7MzbG3/fr3Rzc0NUVFR6NixI8aPH4+ZM2ciKipKpW0DZeMNA0CzZs3g7Ows7ksdnTp1AgDxnHXr1g3vvPMOunTpgvbt22P58uVo164dDhw4AKBsXGVDQ0NYW1vD2dkZzs7OAMrGNZ47dy58fX3Rtm1bzJw5E8OGDcPu3bvVjqm+MBE3AXdzC3Hpbl6ln4YYQJ2IaiYIAiQSiTi9YsUKyOVyREdHw9vbG9HR0ejUqRPS0tJq3E5969evn1KsUqkU165dQ0lJw4wDDfx9nOVxPHz4EHPnzkXnzp1hZ2cHKysrXL16VbwirkpJSQmWL18OHx8fODg4wMrKCocPH65xvYbEzlqNnKolLO0tG18HDqKm5OrVq0o9hIGyK8bXXnsNr732GlauXInu3btjzZo1Ys/rp3Xo0EHcTvfu3Svdfnmb8lvXTyft4uJijR2LRCKp8AeBJrcPlB0PAPGczZ07FwkJCVizZg08PT1hbm6Of/zjHygqqn5s6n/961/YsGED1q9fDx8fH1haWiI0NLTG9RoSE3EjV1MJS4Cv8RBp27Fjx5CWlobZs2dX2cbExATt2rWrste0r68vOnXqhKioKIwbN67Cq1BHjx5FZGQkgL9vDWdlZYlJOzU1tcL+AFR6lfvDDz8oTZ89exbt27eHoaGhuP3yZ7AAcO3aNTx69PcrZdVtWxWlpaXYuHEjPDw8xPhPnz6NSZMm4dVXXwVQdoVcftv66f0+u8/Tp09j1KhReOutt8Rt//rrr/Dy8qpVbPWBibiJqK8SlkSkHoVCAblcrvT6UmRkJF5++WVMnDgRABAXF4ddu3Zh3Lhx6NChAwRBwH//+18cPHgQMTExlW5XIpHg888/x5AhQxAQEIDw8HA4Ozvjhx9+wJw5cyCVSsX3Z83NzdGvXz989NFH8PDwQE5ODhYvXqy0PXd3d0gkEsTFxWHEiBEwNzeHlVXZH/O3b99GWFgY3nnnHZw/fx6bNm3C2rVrxXUHDRqEzZs3QyqVoqSkBAsWLICxsbG43NHREebm5jh06BBatWoFMzMzpWfQz7p//z7kcjkePXqES5cuYf369Th37hzi4+PF5N++fXvs3bsXI0eOhEQiwZIlS1BaWqq0nTZt2uDkyZMYN24cTE1N0bx5c7Rv3x7ffvstzpw5A3t7e6xbtw7Z2dlMxEREdXJPtU5J2tjHoUOH4OLiAiMjI9jb26Nbt27YuHEjAgMDxatYLy8vWFhYYM6cOcjMzISpqSnat2+Pbdu2YcKECVVu+7nnnsPZs2cRERGB4cOH46+//kLr1q0RGBiI8PBwmJqaim2/+OILTJ06FT179kTHjh2xevVqDB06VFzesmVLREREYOHChZg8eTImTpyI2NhYAMDEiRNRWFiIPn36wNDQELNmzcL06dPFddeuXYvJkyfjxRdfhKurKzZs2CD2bgbKipls3LgRy5Ytw9KlS/Hiiy8iKSmpyuPy8/MDAFhYWMDd3R0DBw7Ep59+Ck9PT7HNunXrMGXKFDz33HNo3rw5FixYgPz8fKXtLFu2DO+88w7atWsHhUIBQRCwePFi/Pbbb5DJZLCwsMD06dMxevRo5OXlVfNfsWExERNR42HRrKza1d63G2Z/xhZl+1RRbGysmMyq07ZtW3z66ae1CsnHxwfffvttje06d+6MM2fOKM179rnukiVLsGTJkgrrGhsbY/369di6dWul23Z1dcXhw4eV5uXm5ipNT5s2DdOmTas2xjZt2qjc+axNmzY4duyY0rzg4GCl6X79+uHnn39Wmufg4ID9+/ertA9tYSImosbDzq2s5CRrTVMTwkRMACqv2cxXn0gn2bkxOVKTwkSs52qq2cxXn4j0S3XPcql+MBHruZZ25jg6pz8eFFT+Th1ffSIiql9MxISWduZMtkREWsISl0SksxqinCPRsxr6e8dETEQ6p7w4xNPVmogaSvn37ukiJfWJt6apSauu5zeff+suQ0ND2NnZIScnB0BZoYenByEgqg+CIODRo0fIycmBnZ2dWNWrvmk1EW/duhVbt24V64V6e3tj6dKlGD58OICycSTnzJmDXbt2QaFQQCaT4eOPP4aTk5O4jdu3byMoKAjHjx+HlZUVAgMDERkZCSOjvw8tKSkJYWFhuHz5Mtzc3LB48WJMmjSpIQ+VGlhNvcGBsh7hR+f0ZzLWUeXD2JUnY6KGYmdnJ37/GoJWE3GrVq3w0UcfoX379hAEAdu3b8eoUaNw4cIFeHt7Y/bs2YiPj8eePXtga2uLkJAQjBkzBqdPnwZQVlDc398fzs7OOHPmDLKysjBx4kQYGxtj5cqVAICMjAz4+/tjxowZ2LFjBxITEzFt2jS4uLhAJpNp8/CpHtXUG/x6zkOEfpOKBwVFTMQ6SiKRwMXFBY6Ojhof2YeoKsbGxg12JSwSdIy9vb2wbds2ITc3VzA2Nhb27NkjLrt69aoAQEhOThYEQRAOHjwoGBgYCHK5XGyzdetWwcbGRlAoFIIgCML8+fMFb29vpX2MHTtWkMlkKseUl5cnABDy8vLqcmj1Iu1OruC+IE5Iu5Or7VAqqCk2bcauy+eNiHRXfeQDnemsVVJSgl27dqGgoABSqRQpKSkoLi4Wi4EDQKdOndC6dWskJycDAJKTk+Hj46N0q1omkyE/Px+XL18W2zy9jfI25dsgIiLSJq131kpLS4NUKsXjx49hZWWFffv2wcvLC6mpqTAxMYGdnZ1SeycnJ8jlcgCAXC5XSsLly8uXVdcmPz8fhYWFMDeveFtSoVBAoVCI08+O8EHqqarDFEtoEhHpQCLu2LEjUlNTkZeXh2+//RaBgYE4ceKEVmOKjIxERESEVmNoClTtMKXNEppV/THAHtVE1FC0nohNTEzEMSd79uyJH3/8ERs2bMDYsWNRVFSE3Nxcpavi7OxssTebs7Mzzp07p7S97OxscVn5/5bPe7qNjY1NpVfDABAeHo6wsDBxOj8/H25uLDKvrpo6TAHaS3iq1Nhmj2oiaghaT8TPKi0thUKhQM+ePWFsbIzExEQEBAQAANLT03H79m1IpVIAgFQqxYoVK5CTkwNHR0cAQEJCAmxsbODl5SW2OXjwoNI+EhISxG1UxtTUVGmAbao9XS2fWd0fCexRTUQNSauJODw8HMOHD0fr1q3x119/YefOnUhKSsLhw4dha2uLqVOnIiwsDA4ODrCxscHMmTMhlUrRr18/AMDQoUPh5eWFCRMmYPXq1ZDL5Vi8eDGCg4PFRDpjxgxs3rwZ8+fPx5QpU3Ds2DHs3r0b8fHx2jx00gG6+kcCEekXrSbinJwcTJw4EVlZWbC1tUXXrl1x+PBhDBkyBAAQFRUFAwMDBAQEKBX0KGdoaIi4uDgEBQVBKpXC0tISgYGBWLZsmdjGw8MD8fHxmD17NjZs2IBWrVph27ZtfIeYiIh0glYT8eeff17tcjMzM2zZsgVbtmypso27u3uFW8/PGjBgAC5cuFCrGImIiOqTzrxHTEREpI+YiImIiLSIiZiIiEiLmIiJiIi0iImYiIhIi3SuoAeRrqiuFjZLYBKRpjAREz1D1RrZLIFJRJrAREz0jJpqZLMEJhFpEhMxUSVY/pKIGgoTMVEt8RkyEWkCEzGRmvgMmYg0iYmYSE18hkxEmsRETFQLfIZMRJrCgh5ERERaxCti0q7cTODR/cqXWTQD7NwaNh4iogbGREzak5sJbOkDFD+qfLmxBRB8jsmYiJo0JmLSnkf3y5LwmM+A5h2Ul937Fdj7dlkbJmIiasKYiEn7mncAXH21HQURkVao3VmroKCgPuIgIiLSS2onYicnJ0yZMgWnTp2qj3iIiIj0itq3pr/66ivExsZi0KBBaNOmDaZMmYKJEyfC1dW1PuIjbauuVzPAns1ERHWkdiIePXo0Ro8ejT/++ANffvklYmNjsWTJEshkMkyZMgWvvPIKjIz46LlJqKlXM8CezUREdVTrjNmiRQuEhYUhLCwMmzZtwrx583Dw4EE0b94cM2bMwMKFC2FhYaHJWKmhVderGWDPZiIiDah1Is7Ozsb27dsRGxuLW7du4R//+AemTp2KO3fuYNWqVTh79iyOHDmiyVhJW9iruUHdzS2sso41UPPITnVdn4galtqJeO/evYiJicHhw4fh5eWFd999F2+99Rbs7OzENs899xw6d+6syTiJmpSqkuX9giLM+DIFhcUlVa5b3chOd3ML4bf2RK3XJ6KGp3Yinjx5MsaNG4fTp0+jd+/elbZxdXXFP//5zzoHR9QU1ZQszY0NsX1KHzSzNKmwrKaRnR4UFKGwuATrx/rC09FK7fWJqOGpnYizsrJqfPZrbm6O999/v9ZBETVlNSVLTdw69nS0QpeWtnXaBhE1DLXfI7a2tkZOTk6F+ffv34ehoaFa24qMjETv3r1hbW0NR0dHjB49Gunp6UptBgwYAIlEovSZMWOGUpvbt2/D398fFhYWcHR0xLx58/DkyROlNklJSejRowdMTU3h6emJ2NhYtWIlLbn3K/B7auWf3ExtRlZn5cny2Q+vVIn0i9pXxIIgVDpfoVDAxKTirbTqnDhxAsHBwejduzeePHmCRYsWYejQobhy5QosLS3Fdm+//TaWLVsmTj99RV5SUgJ/f384OzvjzJkzyMrKwsSJE2FsbIyVK1cCADIyMuDv748ZM2Zgx44dSExMxLRp0+Di4gKZTKZWzNRALJqVvRq19+2q2/DVKSJqAlROxBs3bgQASCQSbNu2DVZWf99SKykpwcmTJ9GpUye1dn7o0CGl6djYWDg6OiIlJQUvvfSSON/CwgLOzs6VbuPIkSO4cuUKjh49CicnJ/j6+mL58uVYsGABPvjgA5iYmCA6OhoeHh5Yu3YtAKBz5844deoUoqKiGkUirq4X7PWchw0cTQOxcytLslUVEyl/dep2ModRJKJGTeVEHBUVBaDsijg6OlrpNrSJiQnatGmD6OjoOgWTl5cHAHBwcFCav2PHDnz11VdwdnbGyJEjsWTJEvGqODk5GT4+PnBychLby2QyBAUF4fLly+jevTuSk5Ph5+entE2ZTIbQ0NA6xdsQVO0Fa19Jx55Gz86t6kTaCK6Yq/ojSZf/eOKrT0QNT+VEnJGRAQAYOHAg9u7dC3t7e40GUlpaitDQUDz//PPo0qWLOP/NN9+Eu7s7XF1dcfHiRSxYsADp6enYu3cvAEAulyslYQDitFwur7ZNfn4+CgsLYW6u/MOiUCigUCjE6fz8fM0dqJpq6tgD6OmPo6pXzFooNmJvaQJzY0OEfpNaZRtd/OOJrz4RaYfaz4iPHz9eH3EgODgYly5dqjCYxPTp08V/+/j4wMXFBYMHD8aNGzfQrl27eoklMjISERER9bLt2mIv2EpUd8WsRS3tzHF0Tv9Gd2XJV5+ItEOlRBwWFobly5fD0tISYWFh1bZdt26d2kGEhIQgLi4OJ0+eRKtWrapt27dvXwDA9evX0a5dOzg7O+PcuXNKbbKzswFAfK7s7Owsznu6jY2NTYWrYQAIDw9XOs78/Hy4ueneDz7prpZ25vWarOrztjf/6CNqWCol4gsXLqC4uFj8d1UkEolaOxcEATNnzsS+ffuQlJQEDw+PGtdJTU0FALi4uAAApFIpVqxYgZycHDg6OgIAEhISYGNjAy8vL7HNwYMHlbaTkJAAqVRa6T5MTU1hamqq1rEQNYTGetubiKqmUiJ++na0Jm9NBwcHY+fOnfjPf/4Da2tr8Zmura0tzM3NcePGDezcuRMjRoxAs2bNcPHiRcyePRsvvfQSunbtCgAYOnQovLy8MGHCBKxevRpyuRyLFy9GcHCwmExnzJiBzZs3Y/78+ZgyZQqOHTuG3bt3Iz4+XmPHotfu/Vr1MvZc1qjGetubiKpW5/EK8/PzcezYMXTq1Ent15e2bt0KoKxox9NiYmIwadIkmJiY4OjRo1i/fj0KCgrg5uaGgIAALF68WGxraGiIuLg4BAUFQSqVwtLSEoGBgUrvHXt4eCA+Ph6zZ8/Ghg0b0KpVK2zbtq1RvLqk0xpBz+WmSBO3vSu7ha3LvbmJmjK1E/Hrr7+Ol156CSEhISgsLESvXr1w8+ZNCIKAXbt2ISAgQOVtVVUcpJybmxtOnDhR43bc3d0r3Hp+1oABA6q9rU61oMM9l6lyNd3a5m1tooandiI+efKkOKDDvn37IAgCcnNzsX37dnz44YdqJWJqAnS05zJVrqZb27ytTdTw1E7EeXl5YsGNQ4cOISAgABYWFvD398e8efM0HiARaVZ99+gmIvWoPeiDm5sbkpOTUVBQgEOHDmHo0KEAgAcPHsDMzEzjARIRETVlal8Rh4aGYvz48bCysoK7u7vY0erkyZPw8fHRdHxERERNmtqJ+N1330WfPn2QmZmJIUOGwMCg7KK6bdu2+PDDDzUeIBERUVNWq9eXevXqhV69einN8/f310hARERE+kTtRFxSUoLY2FgkJiYiJycHpaWlSsuPHTumseCIiIiaOrUT8axZsxAbGwt/f3906dJF7bKWRERE9De1E/GuXbuwe/dujBgxoj7iIdKs6spvVoelOatUXQUuvodMpD61E7GJiQk8PT3rIxYizVGl/GZ1WJqzAlUHnOB4xUTqUTsRz5kzBxs2bMDmzZt5W5p0V03lN6vD0pyVqqkqF8crJqodtRPxqVOncPz4cfzvf/+Dt7c3jI2NlZbv3btXY8ER1QnLb2ocq3IRaZ7aidjOzg6vvvpqfcRCTVVVz2lr+/yWiKgJUTsRx8TE1Ecc1BSpOkyiRbOGi4mISMfUqqDHkydPkJSUhBs3buDNN9+EtbU1fv/9d9jY2MDKykrTMVJjpcpzWvZOJiI9p3YivnXrFoYNG4bbt29DoVBgyJAhsLa2xqpVq6BQKBAdHV0fcVJjxee09JS7uYUcgpHoGbUq6NGrVy/8/PPPaNbs71uKr776Kt5+u5avihBRk3c3txB+a0+gsLik0uV89Yn0ldqJ+Pvvv8eZM2dgYmKiNL9Nmza4e/euxgIjoqblQUERCotLsH6sLzwdlR9h8dUn0mdqJ+LS0lKUlFT8i/bOnTuwtrbWSFBE1HhVVXmrfL6noxW6tLRtyJCIdJraiXjo0KFYv349Pv30UwCARCLBw4cP8f7777PsJZEeU7Xylr2lSZXLifSR2ol47dq1kMlk8PLywuPHj/Hmm2/i2rVraN68Ob7++uv6iJGIGoGaKm8B7JBFVBm1E3GrVq3w888/Y9euXbh48SIePnyIqVOnYvz48TA35//BiPQZK28Rqa9W7xEbGRnhrbfe0nQsREREekelRHzgwAGVN/jKK6/UOhiqJ7mZLKpBRKSjVErEo0ePVpqWSCQQBKHCPACV9qgmLcrNBLb0AYofVd2GQ/4REWmNgSqNSktLxc+RI0fg6+uL//3vf8jNzUVubi7+97//oUePHjh06FB9x0vqenS/LAmP+QyYfqLiZ8xnZctrM1wgERHVmdrPiENDQxEdHY0XXnhBnCeTyWBhYYHp06fj6tWrGg2QNKR5B8DVV9tRNC7VjQ7F2/lEpCFqJ+IbN27Azs6uwnxbW1vcvHlTAyERaZmqo0bxdj4RaYBKt6af1rt3b4SFhSE7O1ucl52djXnz5qFPnz5qbSsyMhK9e/eGtbU1HB0dMXr0aKSnpyu1efz4MYKDg9GsWTNYWVkhICBAad8AcPv2bfj7+8PCwgKOjo6YN28enjx5otQmKSkJPXr0gKmpKTw9PREbG6vegZP+KB81qrJb+bydT0QapvYV8RdffIFXX30VrVu3hptb2dVAZmYm2rdvj/3796u1rRMnTiA4OBi9e/fGkydPsGjRIgwdOhRXrlyBpaUlAGD27NmIj4/Hnj17YGtri5CQEIwZMwanT58GUNY5zN/fH87Ozjhz5gyysrIwceJEGBsbY+XKlQCAjIwM+Pv7Y8aMGdixYwcSExMxbdo0uLi4QCaTqXsKSB9w1KgmhyM/ka5SOxF7enri4sWLSEhIwC+//AIA6Ny5M/z8/MSe06p6tnNXbGwsHB0dkZKSgpdeegl5eXn4/PPPsXPnTgwaNAgAEBMTg86dO+Ps2bPo168fjhw5gitXruDo0aNwcnKCr68vli9fjgULFuCDDz6AiYkJoqOj4eHhgbVr14rxnjp1ClFRUUzERDqkqjrVQN2SJUd+Il1Wq4IeEokEQ4cOxdChQzUaTF5eHgDAwcEBAJCSkoLi4mL4+fmJbTp16oTWrVsjOTkZ/fr1Q3JyMnx8fODk5CS2kclkCAoKwuXLl9G9e3ckJycrbaO8TWhoaKVxKBQKKBQKcTo/P19Th0hNCTtzaYyqdaprmyw58hPpslol4vpQWlqK0NBQPP/88+jSpQsAQC6Xw8TEpELnMCcnJ8jlcrHN00m4fHn5sura5Ofno7CwsEJpzsjISERERGjs2KiJYWcujaupTrWmkiVHfiJdpDOJODg4GJcuXcKpU6e0HQrCw8MRFhYmTufn54vPw5usyq7uqrvi02flnbmq6qx179eyJP3oPhOxGlinmvSVTiTikJAQxMXF4eTJk2jVqpU439nZGUVFRcjNzVW6Ks7Ozoazs7PY5ty5c0rbK+9V/XSbZ3taZ2dnw8bGptKBKkxNTWFqaqqRY9N5NV3dGVuUtSFl7MxFRBqi1UQsCAJmzpyJffv2ISkpCR4eHkrLe/bsCWNjYyQmJiIgIAAAkJ6ejtu3b0MqlQIApFIpVqxYgZycHDg6OgIAEhISYGNjAy8vL7HNwYMHlbadkJAgbqMhVNdjE9Bir82aru74rJOIqF6pnYj79++PqVOn4rXXXqvzsIfBwcHYuXMn/vOf/8Da2lp8pmtrawtzc3PY2tpi6tSpCAsLg4ODA2xsbDBz5kxIpVL069cPADB06FB4eXlhwoQJWL16NeRyORYvXozg4GDxqnbGjBnYvHkz5s+fjylTpuDYsWPYvXs34uPj6xS/qmrqsQloudcmr+6IiLRG7UTcvXt3zJ07FzNnzsTrr7+OqVOniklRXVu3bgUADBgwQGl+TEwMJk2aBACIioqCgYEBAgICoFAoIJPJ8PHHH4ttDQ0NERcXh6CgIEilUlhaWiIwMBDLli0T23h4eCA+Ph6zZ8/Ghg0b0KpVK2zbtq3BXl2qrscmwF6bRET6TO1EvH79eqxZswYHDhzA9u3b8dJLL8HT0xNTpkzBhAkTKvROrs6zIzhVxszMDFu2bMGWLVuqbOPu7l7h1vOzBgwYgAsXLqgcW32otx6b1Q1zyA5XREQ6rVbPiI2MjDBmzBiMGTMGOTk5+PTTT7FkyRIsWrQII0aMwHvvvScW4KB6puowh+xwRUSkk+rUWevcuXOIiYnBrl274OjoiEmTJuHu3bt4+eWX8e6772LNmjWailMvVFZVqLpKQwCUhzls3qHyNuxwRUSks9ROxDk5Ofjyyy8RExODa9euYeTIkfj6668hk8nEEpeTJk3CsGHDmIhVVFNVIXNjQ9hbmlS/EQ5zSETUKKmdiFu1aoV27dphypQpmDRpElq0aFGhTdeuXdG7d2+NBKgPaqoqxIL0RERNl9qJODExES+++GK1bWxsbHD8+PFaB6WPWFWIiEg/qT0ecWVXwOUOHz5cp2CIiIj0jdqJuEePHhVeJVIoFAgJCcGoUaM0FhgREZE+UDsRx8bGYunSpRgxYgSys7ORmpqK7t274+jRo/j+++/rI0YiIqImS+1E/Prrr+Pnn39GcXExvL29IZVK0b9/f5w/f54dtIiIiNSkdiIuV1RUhJKSEpSUlMDFxQVmZmaajIuIiEgvqJ2Id+3aBR8fH9ja2uLXX39FfHw8Pv30U7z44ov47bff6iNGIiKiJkvtRDx16lSsXLkSBw4cQIsWLTBkyBCkpaWhZcuW8PX1rYcQiYiImi613yM+f/48OnbsqDTP3t4eu3fvxpdffqmxwIiI1FHdmN81loqtoQ2L6lB9UjsRP5uEnzZhwoQ6BUNEVBuqjvldWanYmkrMlq+rtfHCqcmr1aAPd+7cwYEDB3D79m0UFSn/Bbpu3TqNBEZEpKqaxvwGqr6qranELMcLp/pWqxKXr7zyCtq2bYtffvkFXbp0wc2bNyEIAnr06FEfMRI1TtWNBc0RsWqlqtvH5fNrO+Y3S8ySNqmdiMPDwzF37lxERETA2toa3333HRwdHTF+/HgMGzasPmIkalwsmpWNAb337arbGFsAweeYjFWk6u3jGkcpI9JBaifiq1ev4uuvvy5b2cgIhYWFsLKywrJlyzBq1CgEBQVpPEiiRsXOrSzJPrpf+fJ7v5Yl6Uf3mYhVVNPtY4AdqqjxUjsRW1pais+FXVxccOPGDXh7ewMA7t27p9noiBorOzcmWQ3j7WNqqtROxP369cOpU6fQuXNnjBgxAnPmzEFaWhr27t2Lfv361UeMRERETZbaiXjdunV4+LCsY0RERAQePnyIb775Bu3bt2ePaSJ1VNWZix25SA3VvT8N8JZ9Y6B2Im7btq34b0tLS0RHR2s0IKImr6bOXOzIRSpS9f3p6Ak90ayKd6iZpLWvVu8RE1EdVNeZix25SA01vT99v6AIM75MQeAX5ypdn4VKdINKidje3h4SiUSlDf755591CohIL7AzF2lQde9PV9XbnIVKdIdKiXj9+vX1HAYREdUH9jbXfSol4sDAwPqOg4iISC/V6Rnx48ePK9SatrGxqVNARERE+kTt8YgLCgoQEhICR0dHWFpawt7eXumjjpMnT2LkyJFwdXWFRCLB/v37lZZPmjQJEolE6fNsGc0///wT48ePh42NDezs7DB16lTx9apyFy9exIsvvggzMzO4ublh9erV6h42ERGp6W5uIS7dzav0cze3UNvh6Qy1r4jnz5+P48ePY+vWrZgwYQK2bNmCu3fv4pNPPsFHH32k1rYKCgrQrVs3TJkyBWPGjKm0zbBhwxATEyNOm5qaKi0fP348srKykJCQgOLiYkyePBnTp0/Hzp07AQD5+fkYOnQo/Pz8EB0djbS0NEyZMgV2dnaYPn26mkdPRESqqOnVKvbY/pvaifi///0v/v3vf2PAgAGYPHkyXnzxRXh6esLd3R07duzA+PHjVd7W8OHDMXz48GrbmJqawtnZudJlV69exaFDh/Djjz+iV69eAIBNmzZhxIgRWLNmDVxdXbFjxw4UFRXhiy++gImJCby9vZGamop169YxEZPu4shN9P+qK9hR1WhUuqC6V6vYY1uZ2on4zz//FIt62NjYiK8rvfDCC/Uy4ENSUhIcHR1hb2+PQYMG4cMPP0SzZs0AAMnJybCzsxOTMAD4+fnBwMAAP/zwA1599VUkJyfjpZdegonJ3y+zy2QyrFq1Cg8ePKj0drpCoYBCoRCn8/PzNX5cRJXiyE1NUnXJtLqiGqoW7NDlUadqOzSlPqlVZa2MjAy0bt0anTp1wu7du9GnTx/897//hZ2dnUaDGzZsGMaMGQMPDw/cuHEDixYtwvDhw5GcnAxDQ0PI5XI4OjoqrWNkZAQHBwfI5XIAgFwuh4eHh1IbJycncVlliTgyMhIREREaPRYilXDkpianLrdoayrYAbA6VlOgdiKePHkyfv75Z/Tv3x8LFy7EyJEjsXnzZhQXF2u81vS4cePEf/v4+KBr165o164dkpKSMHjwYI3u62nh4eEICwsTp/Pz8+Hmxh89aiAs9tGkaOIWLa8qmza1E/Hs2bPFf/v5+eGXX35BSkoKPD090bVrV40G96y2bduiefPmuH79OgYPHgxnZ2fk5OQotXny5An+/PNP8bmys7MzsrOzldqUT1f17NnU1LRCpzAiorpgMqWqqP360rPc3d0xZsyYek/CAHDnzh3cv38fLi4uAACpVIrc3FykpKSIbY4dO4bS0lL07dtXbHPy5EkUFxeLbRISEtCxY0e1X7ciIiLSNJWviAsLC5GYmIiXX34ZQNnt26c7NBkaGmL58uUwMzNTeecPHz7E9evXxemMjAykpqbCwcEBDg4OiIiIQEBAAJydnXHjxg3Mnz8fnp6ekMlkAIDOnTtj2LBhePvttxEdHY3i4mKEhIRg3LhxcHV1BQC8+eabiIiIwNSpU7FgwQJcunQJGzZsQFRUlMpxEhFVp7H2bCbdoHIi3r59O+Lj48VEvHnzZnh7e8PcvOy5xi+//AJXV1elW9c1+emnnzBw4EBxuvy5bGBgILZu3YqLFy9i+/btyM3NhaurK4YOHYrly5cr3TbesWMHQkJCMHjwYBgYGCAgIAAbN24Ul9va2uLIkSMIDg5Gz5490bx5cyxdupSvLhGRRmiiZ3NVyZpJXD+onIh37NiB+fPnK83buXOn+CrTV199hS1btqiViAcMGABBEKpcfvjw4Rq34eDgIBbvqErXrl3x/fffqxwXEZGq6tKz2d7SBObGhgj9JrXK7ev660lUdyon4uvXr8PHx0ecNjMzg4HB34+Y+/Tpg+DgYM1GR0TUSNSmM1ZLO/Mqhyksx9eTmj6VE3Fubq7SM+E//vhDaXlpaanSciIiqhmHKSSVE3GrVq1w6dIldOzYsdLlFy9eRKtWrTQWGBFVo7oSmDVhicxa4XNcqi8qJ+IRI0Zg6dKl8Pf3r9AzurCwEBEREfD399d4gET0FFVKYNaEJTLVwue4VN9UTsSLFi3C7t270bFjR4SEhKBDhw4AgPT0dGzevBlPnjzBokWL6i1QvZabWX3JQ9IfNZXArAlLZKqNz3GpvqmciJ2cnHDmzBkEBQVh4cKFYm9niUSCIUOG4OOPPxZrOJMG5WYCW/oAxY+qbmNsUXalRPqBJTAbHJ/jUn1Sq8Slh4cHDh06hD///FMsxOHp6QkHB4d6CY5QduVS/AgY8xnQvEPlbfjMj4io0VK71jRQ9u5unz59NB0LVad5B8DVV9tREBGRhtW51jQRERHVXq2uiImIiFhjWzOYiImISG2aqLFNZZiIiYhIbXWpsU3KmIiJ9FFV75+zBz6pqTY1tstVd/tan5I4EzGRPqmpMherblEDULVa2dE5/fUiGTMRE+mT6ipzseoWNZCaqpVdz3mI0G9S8aCgiImYiJogVuYiHcBqZX9jIm4grrgHs3tpgKTyTg18Nkc6o7r65fyeEmkcE3EDMH54F0dN58FiXzXjNfPZHGmbKiM78XtKpHFMxA3A8PGfsJAokDlwA9za+1ZswGdzpAtqGtmJ31OiesFE3IAUdp6sF026jc+PiRocEzERqYfPkIk0iomYiFTDZ8hE9YKJmIhUw2fIRPWCiZiIVMdnyEQax/GIiYiItIhXxEREeowDL2gfEzERkR7iwAu6Q6u3pk+ePImRI0fC1dUVEokE+/fvV1ouCAKWLl0KFxcXmJubw8/PD9euXVNq8+eff2L8+PGwsbGBnZ0dpk6diocPlf/Cu3jxIl588UWYmZnBzc0Nq1evru9DIyLSaeUDL8TNfKHSz/qxvigsLqlyYAbSHK1eERcUFKBbt26YMmUKxowZU2H56tWrsXHjRmzfvh0eHh5YsmQJZDIZrly5AjMzMwDA+PHjkZWVhYSEBBQXF2Py5MmYPn06du7cCQDIz8/H0KFD4efnh+joaKSlpWHKlCmws7PD9OnTG/R4a1TZ+5nVvbNJRFQHHHhBN2g1EQ8fPhzDhw+vdJkgCFi/fj0WL16MUaNGAQD+/e9/w8nJCfv378e4ceNw9epVHDp0CD/++CN69eoFANi0aRNGjBiBNWvWwNXVFTt27EBRURG++OILmJiYwNvbG6mpqVi3bp3uJGJVxoi1aNawMRERUYPQ2WfEGRkZkMvl8PPzE+fZ2tqib9++SE5Oxrhx45CcnAw7OzsxCQOAn58fDAwM8MMPP+DVV19FcnIyXnrpJZiYmIhtZDIZVq1ahQcPHsDe3r7CvhUKBRSKvwdoyM/Pr6ej/H81vZ/JakVEpCVVdeaqrpMXqUdnE7FcLgcAODk5Kc13cnISl8nlcjg6OiotNzIygoODg1IbDw+PCtsoX1ZZIo6MjERERIRmDkRVfD+TiHSIqp257C1NqlxOqtHZRKxN4eHhCAsLE6fz8/Ph5sYkSUT6o7wzV3Wdtfh6k2bobCJ2dnYGAGRnZ8PFxUWcn52dDV9fX7FNTk6O0npPnjzBn3/+Ka7v7OyM7OxspTbl0+VtnmVqagpTU1ONHAcRUWOl7c5c+vKOs84mYg8PDzg7OyMxMVFMvPn5+fjhhx8QFBQEAJBKpcjNzUVKSgp69uwJADh27BhKS0vRt29fsc0///lPFBcXw9jYGACQkJCAjh07VnpbmoiItEvf3nHWaiJ++PAhrl+/Lk5nZGQgNTUVDg4OaN26NUJDQ/Hhhx+iffv24utLrq6uGD16NACgc+fOGDZsGN5++21ER0ejuLgYISEhGDduHFxdXQEAb775JiIiIjB16lQsWLAAly5dwoYNGxAVFaWNQyYiohrUdFv8es5DhH6TigcFRUzEdfXTTz9h4MCB4nT5c9nAwEDExsZi/vz5KCgowPTp05Gbm4sXXngBhw4dEt8hBoAdO3YgJCQEgwcPhoGBAQICArBx40Zxua2tLY4cOYLg4GD07NkTzZs3x9KlS3Xn1SUiIqpA27fFG5JWE/GAAQMgCEKVyyUSCZYtW4Zly5ZV2cbBwUEs3lGVrl274vvvv691nESkhuqK0PBVPKIKdPYZMRE1MjUVpgHKlgefYzImegoTMRFpRk2Fae79WpakH91nIiZ6ChMxEWkOC9MQqY2JmIgaVlXPkPn8mPQUEzERNQxVBjfh82PSQ0zERNQwqnuGzOfHpMeYiImo4fAZMlEFBtoOgIiISJ8xERMREWkREzEREZEWMRETERFpETtrEZHuYJ1q0kNMxESkfaxTTXqMiZiItI91qkmPMRETkW7gO8akp9hZi4iISIuYiImIiLSIiZiIiEiLmIiJiIi0iImYiIhIi5iIiYiItIiJmIiISIuYiImIiLSIBT2ISD/kZlZduYt1rEmLmIg15G5uIR4UFFW67I8/C+HZwPEQNUm1HRQiNxPY0gcoflT5ctaxJi1iItaAu7mF8Ft7AoXFJZUu95ZkYKApYGNu3MCRETURdR0U4tH9siQ85jOgeQflZaxjTVrGRKwBDwqKUFhcgvVjfeHpaFVhudk9W2Af4GhlqoXoiJoATQ0K0bwD4OpbLyES1ZZOd9b64IMPIJFIlD6dOnUSlz9+/BjBwcFo1qwZrKysEBAQgOzsbKVt3L59G/7+/rCwsICjoyPmzZuHJ0+e1Eu8no5W6NLStsLHs0XF5ExEarJzK0uilX2evcolakR0/orY29sbR48eFaeNjP4Oefbs2YiPj8eePXtga2uLkJAQjBkzBqdPnwYAlJSUwN/fH87Ozjhz5gyysrIwceJEGBsbY+XKlQ1+LERUz6p6hlzds2UiLdP5RGxkZARnZ+cK8/Py8vD5559j586dGDRoEAAgJiYGnTt3xtmzZ9GvXz8cOXIEV65cwdGjR+Hk5ARfX18sX74cCxYswAcffAATE5OGPhwiqg+qPkO2aNZwMRGpSOcT8bVr1+Dq6gozMzNIpVJERkaidevWSElJQXFxMfz8/MS2nTp1QuvWrZGcnIx+/fohOTkZPj4+cHJyEtvIZDIEBQXh8uXL6N69uzYOiYg0raZnyABfUSKdpdOJuG/fvoiNjUXHjh2RlZWFiIgIvPjii7h06RLkcjlMTExgZ2entI6TkxPkcjkAQC6XKyXh8uXly6qiUCigUCjE6fz8fA0dERHVGzs3JlpqlHQ6EQ8fPlz8d9euXdG3b1+4u7tj9+7dMDc3r7f9RkZGIiIiot62T0REVE6ne00/y87ODh06dMD169fh7OyMoqIi5ObmKrXJzs4Wnyk7OztX6EVdPl3Zc+dy4eHhyMvLEz+ZmZk1xuaKezC7lwb8nlrxw44iRERUBZ2+In7Ww4cPcePGDUyYMAE9e/aEsbExEhMTERAQAABIT0/H7du3IZVKAQBSqRQrVqxATk4OHB0dAQAJCQmwsbGBl5dXlfsxNTWFqanq7/waP7yLo6bzYLFPUU0jdhQhIqKKdDoRz507FyNHjoS7uzt+//13vP/++zA0NMQbb7wBW1tbTJ06FWFhYXBwcICNjQ1mzpwJqVSKfv36AQCGDh0KLy8vTJgwAatXr4ZcLsfixYsRHBysVqKtieHjP2EhUSBz4Aa4tfetvBE7ihDpttqWzySqI51OxHfu3MEbb7yB+/fvo0WLFnjhhRdw9uxZtGjRAgAQFRUFAwMDBAQEQKFQQCaT4eOPPxbXNzQ0RFxcHIKCgiCVSmFpaYnAwEAsW7asXuJV2Hmyag9RY1PX8plEdaTTiXjXrl3VLjczM8OWLVuwZcuWKtu4u7vj4MGDmg6NiJoKTZXPJKolnU7EREQNgq8+kRY1ql7TRERETQ0TMRERkRYxERMREWkREzEREZEWMRETERFpERMxERGRFjERExERaRHfIyYiUgVLYFI9YSImIqoOS2BSPWMiJiKqjqolMG8nV92GV8xUDSZiIqKaVFcCk1fMVEdMxEREdcFBI6iOmIiJiOpKlUEjqursVdNt69zMqpO8KuuTzmMiJiKqTzXduq7utnVuJrClD1D8qOrt87Z3o8dETERUn6q7dV1TR697v5Yl4TGfAc07VL0+b3s3akzERET1rapb16p29GotZaJtwpiIiYi0paaOXgCfAesBJmIiIm1SpaNXTVj1q1FjIiYiaqxUvbU99kvAonnl6zNJax0TMRFRY1XTre1H94BvJgBfBVS+nD2udQITMRFRY1bTre2aemyzx7XWMRETETVlNSVqPl/WOiZiIiJ91BA1squrCsYkL2IiJiLSR/U9qlRNVcH4fFrERExEpK80MapUVT2yq6sKxqEjlTARExFRRXXtkQ1UXRWMQ0cq0atEvGXLFvzrX/+CXC5Ht27dsGnTJvTp00fl9W+kJcPayrLC/NxblzQZJhGRbqhtj+xyVV3V1vG2uNm9h3DFvRqCbzz0JhF/8803CAsLQ3R0NPr27Yv169dDJpMhPT0djo6OKm2jXdxrsDGVVLrskWAKK3snTYZMRKTb6lIVrA63xT0BHDU1RebD3gBsa7d/HSIRBEHQdhANoW/fvujduzc2b94MACgtLYWbmxtmzpyJhQsXVrtufn4+bG1tcfzrdWjVuXelbazsneDcur3G4yYi0kvV9LjOvJYKt+OzcP3VeHh2e6FBwyrPB3l5ebCxsdHINvXiirioqAgpKSkIDw8X5xkYGMDPzw/Jyckqb8fWzavB/6MTEemlaq6YFX88BABk/lmIx3fzKm1jb2mClnbm9RaeJulFIr537x5KSkrg5KR869jJyQm//PJLhfYKhQIKhUKczssr+w/9sKAA+fn59RssERFVS1JShHyFgG/i/offhNRK25gZGWCOrCNszIw1uu+HBQUAAE3eTNaLRKyuyMhIREREVJj/kqya3oFERNTAoqpd+t9/1d+e79+/D1tbzTyf1otE3Lx5cxgaGiI7O1tpfnZ2NpydnSu0Dw8PR1hYmDidm5sLd3d33L59W2MnXh/k5+fDzc0NmZmZGnuW0tTxnNUOz5v6eM5qJy8vD61bt4aDg4PGtqkXidjExAQ9e/ZEYmIiRo8eDaCss1ZiYiJCQkIqtDc1NYWpqWmF+ba2tvzC1oKNjQ3Pm5p4zmqH5019PGe1Y2BgoLFt6UUiBoCwsDAEBgaiV69e6NOnD9avX4+CggJMnjxZ26EREZEe05tEPHbsWPzxxx9YunQp5HI5fH19cejQoQoduIiIiBqS3iRiAAgJCan0VnRNTE1N8f7771d6u5qqxvOmPp6z2uF5Ux/PWe3Ux3nTm4IeREREukhzT5uJiIhIbUzEREREWsRETEREpEVMxERERFrERPz/tmzZgjZt2sDMzAx9+/bFuXPnqm2/Z88edOrUCWZmZvDx8cHBgwcbKFLdos55i42NhUQiUfqYmZk1YLTad/LkSYwcORKurq6QSCTYv39/jeskJSWhR48eMDU1haenJ2JjY+s9Tl2i7jlLSkqq8D2TSCSQy+UNE7AOiIyMRO/evWFtbQ1HR0eMHj0a6enpNa6n779rtTlvmvhdYyLG32MVv//++zh//jy6desGmUyGnJycStufOXMGb7zxBqZOnYoLFy5g9OjRGD16NC5dutTAkWuXuucNKKvik5WVJX5u3brVgBFrX0FBAbp164YtW7ao1D4jIwP+/v4YOHAgUlNTERoaimnTpuHw4cP1HKnuUPeclUtPT1f6rqk67nhTcOLECQQHB+Ps2bNISEhAcXExhg4dioL/H7CgMvxdq915AzTwuyaQ0KdPHyE4OFicLikpEVxdXYXIyMhK27/++uuCv7+/0ry+ffsK77zzTr3GqWvUPW8xMTGCra1tA0Wn+wAI+/btq7bN/PnzBW9vb6V5Y8eOFWQyWT1GprtUOWfHjx8XAAgPHjxokJgag5ycHAGAcOLEiSrb8HetIlXOmyZ+1/T+irh8rGI/Pz9xXk1jFScnJyu1BwCZTKbW2MaNXW3OGwA8fPgQ7u7ucHNzw6hRo3D58uWGCLfR4net9nx9feHi4oIhQ4bg9OnT2g5Hq8qHcq1uoAJ+1ypS5bwBdf9d0/tEXN1YxVU9U5LL5Wq1b4pqc946duyIL774Av/5z3/w1VdfobS0FM899xzu3LnTECE3SlV91/Lz81FYWKilqHSbi4sLoqOj8d133+G7776Dm5sbBgwYgPPnz2s7NK0oLS1FaGgonn/+eXTp0qXKdvxdU6bqedPE75pelbgk7ZJKpZBKpeL0c889h86dO+OTTz7B8uXLtRgZNSUdO3ZEx44dxennnnsON27cQFRUFL788kstRqYdwcHBuHTpEk6dOqXtUBoVVc+bJn7X9P6KWN2xigHA2dlZrfZNUW3O27OMjY3RvXt3XL9+vT5CbBKq+q7Z2NjA3NxcS1E1Pn369NHL71lISAji4uJw/PhxtGrVqtq2/F37mzrn7Vm1+V3T+0T89FjF5crHKn76r5ynSaVSpfYAkJCQUGX7pqg25+1ZJSUlSEtLg4uLS32F2ejxu6YZqampevU9EwQBISEh2LdvH44dOwYPD48a1+F3rXbn7Vm1+l2rU1evJmLXrl2CqampEBsbK1y5ckWYPn26YGdnJ8jlckEQBGHChAnCwoULxfanT58WjIyMhDVr1ghXr14V3n//fcHY2FhIS0vT1iFohbrnLSIiQjh8+LBw48YNISUlRRg3bpxgZmYmXL58WVuH0OD++usv4cKFC8KFCxcEAMK6deuECxcuCLdu3RIEQRAWLlwoTJgwQWz/22+/CRYWFsK8efOEq1evClu2bBEMDQ2FQ4cOaesQGpy65ywqKkrYv3+/cO3aNSEtLU2YNWuWYGBgIBw9elRbh9DggoKCBFtbWyEpKUnIysoSP48ePRLb8HetotqcN038rjER/79NmzYJrVu3FkxMTIQ+ffoIZ8+eFZf1799fCAwMVGq/e/duoUOHDoKJiYng7e0txMfHN3DEukGd8xYaGiq2dXJyEkaMGCGcP39eC1FrT/mrNc9+ys9TYGCg0L9//wrr+Pr6CiYmJkLbtm2FmJiYBo9bm9Q9Z6tWrRLatWsnmJmZCQ4ODsKAAQOEY8eOaSd4LansfAFQ+u7wd62i2pw3TfyucRhEIiIiLdL7Z8RERETaxERMRESkRUzEREREWsRETEREpEVMxERERFrERExERKRFTMRERERaxERMRBUkJSVBIpEgNzdXI20/+OAD+Pr6Vpjn5OQEiUSC/fv31yleosaMiZiokZs0aRIkEgkkEgmMjY3h4eGB+fPn4/Hjx9oOTTR37lylOsZXr15FREQEPvnkE2RlZWH48OFo06YN1q9fr70gibSEwyASNQHDhg1DTEwMiouLkZKSgsDAQEgkEqxatUrboQEArKysYGVlJU7fuHEDADBq1ChIJBJthUWkE3hFTNQEmJqawtnZGW5ubhg9ejT8/PyQkJAAoGxUrMjISHh4eMDc3BzdunXDt99+q7T+wYMH0aFDB5ibm2PgwIG4efOm0vJbt25h5MiRsLe3h6WlJby9vXHw4EGlNikpKejVqxcsLCzw3HPPIT09XVz29K3pDz74ACNHjgQAGBgYQCKRYMCAAbh16xZmz54tXt0T6QsmYqIm5tKlSzhz5gxMTEwAAJGRkfj3v/+N6OhoXL58GbNnz8Zbb72FEydOAAAyMzMxZswYjBw5EqmpqZg2bRoWLlyotM3g4GAoFAqcPHkSaWlpWLVqldIVLgD885//xNq1a/HTTz/ByMgIU6ZMqTS+uXPnIiYmBgCQlZWFrKws7N27F61atcKyZcvEeUT6gremiZqAuLg4WFlZ4cmTJ1AoFDAwMMDmzZuhUCiwcuVKHD16VBxXtm3btjh16hQ++eQT9O/fH1u3bkW7du2wdu1aAEDHjh3FZFvu9u3bCAgIgI+Pj7iNZ61YsQL9+/cHACxcuBD+/v54/PgxzMzMlNpZWVnBzs4OAJQGnTc0NIS1tbVeDkRP+o2JmKgJGDhwILZu3YqCggJERUXByMgIAQEBuHz5Mh49eoQhQ4YotS8qKkL37t0BlHWc6tu3r9LyZweDf++99xAUFIQjR47Az88PAQEB6Nq1q1Kbp6fLB0XPyclB69atNXacRE0REzFRE2BpaQlPT08AwBdffIFu3brh888/R5cuXQAA8fHxaNmypdI6pqamKm9/2rRpkMlkiI+Px5EjRxAZGYm1a9di5syZYhtjY2Px3+XPeEtLS2t9TET6gs+IiZoYAwMDLFq0CIsXL4aXlxdMTU1x+/ZteHp6Kn3c3NwAAJ07d8a5c+eUtnH27NkK23Vzc8OMGTOwd+9ezJkzB5999plG4zYxMUFJSYlGt0nUGDAREzVBr732GgwNDfHJJ59g7ty5mD17NrZv344bN27g/Pnz2LRpE7Zv3w4AmDFjBq5du4Z58+YhPT0dO3fuRGxsrNL2QkNDcfjwYWRkZOD8+fM4fvw4OnfurNGY27Rpg5MnT+Lu3bu4d++eRrdNpMt4a5qoCTIyMkJISAhWr16NjIwMtGjRApGRkfjtt99gZ2eHHj16YNGiRQCA1q1b47vvvsPs2bOxadMm9OnTBytXrlTq9VxSUoLg4GDcuXMHNjY2GDZsGKKiojQa87Jly/DOO++gXbt2UCgUEARBo9sn0lUSgd92IiIireGtaSIiIi1iIiYiItIiJmIiIiItYiImIiLSIiZiIiIiLWIiJiIi0iImYiIiIi1iIiYiItIiJmIiIiItYiImIiLSIiZiIiIiLWIiJiIi0qL/A+ucGPyeWVmfAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 500x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(5, 4), dpi=100)\n",
    "\n",
    "zmin = 0\n",
    "zmax = 2.5\n",
    "\n",
    "hist_settings = {\n",
    "    \"bins\": 50,\n",
    "    \"range\": (zmin, zmax),\n",
    "    \"density\": False, # both versions are useful!\n",
    "    \"histtype\": \"step\",\n",
    "}\n",
    "\n",
    "ax.hist(DS[\"input_data\"][\"redshift\"], label=\"DS Input Data\", **hist_settings)\n",
    "ax.hist(DS[\"output_data\"][\"redshift\"], label=\"DS Output Data\", **hist_settings)\n",
    "# One could extend this to plot the intermediate data, too...\n",
    "\n",
    "ax.legend(title=\"Sample\")\n",
    "ax.set(xlim=(zmin, zmax), xlabel=\"Redshift\", ylabel=\"Galaxy density\")\n",
    "plt.show()\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "scratch",
   "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": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}