notebooks/benchmark-small-file-h5repack.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "dd29c436-f9b0-489c-b9e3-26749bd07d68",
   "metadata": {
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "# Benchmarking h5repack Format on smaller files\n",
    "\n",
    "This is a HDF5 format that has been rechunked from the original format from NSIDC with `h5repack`\n",
    "via the following libraries: h5py, h5coro and xarray"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7750e965-d0c3-4fca-97d5-3f81f0b558e6",
   "metadata": {
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "## Workflow 1 - Read full group\n",
    "\n",
    "### Setup Steps\n",
    "\n",
    "NOTE, below we use the `h5repack` directory, this should be replaced if you are working with a different dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ef607f86-b21f-46fd-95a2-36203d82ccef",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-08-14 21:12:27  560823635 ATL03_20220201060852_06261401_005_01.h5\n",
      "2023-08-14 21:12:40  568000000 ATL03_20220201060852_06261401_005_01_repacked.h5\n"
     ]
    }
   ],
   "source": [
    "# Checkout the files\n",
    "# !mamba install -c conda-forge awscli -y\n",
    "!aws s3 ls s3://nasa-cryo-scratch/h5cloud/h5repack_sm/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "8e42a3b7-ed22-411c-9863-c93569c18f20",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Create a list of s3 objects\n",
    "import h5py\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import s3fs\n",
    "import tqdm\n",
    "import xarray as xr\n",
    "import earthaccess\n",
    "import time\n",
    "\n",
    "try: \n",
    "    from h5coro import h5coro, s3driver, filedriver\n",
    "except:\n",
    "    !mamba install -c conda-forge h5coro --yes\n",
    "    from h5coro import h5coro, s3driver, filedriver\n",
    "\n",
    "# show full columns\n",
    "pd.set_option('display.max_colwidth', None)\n",
    "\n",
    "s3 = s3fs.S3FileSystem(anon=False)\n",
    "\n",
    "group = '/gt2l/heights'\n",
    "variable = 'h_ph'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a41ae3f0-c17b-4085-a7d6-e7da342bea77",
   "metadata": {
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "We could use earthaccess to directly open a granule if we know the file name. For now we are just using one previously pre processed with h5repack"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "d76f5b2c-8146-4042-b95a-5b0096a3b33a",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# earthaccess.login()\n",
    "# granule = earthaccess.search_data(\n",
    "#     short_name=\"ATL03\",\n",
    "#     granule_name=\"ATL03_20220201060852_06261401_005_01.h5\",\n",
    "#     cloud_hosted=True\n",
    "# )\n",
    "# display(granule[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "4c329e67-3553-434a-b8d0-acc29b7c58b9",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "path = \"s3://nasa-cryo-scratch/h5cloud/h5repack_rnd/\"\n",
    "\n",
    "repacked =  [\"s3://nasa-cryo-scratch/h5cloud/h5repack_rnd/ATL03_20220201060852_06261401_005_01_repacked.h5\"]\n",
    "original =  [\"s3://nasa-cryo-scratch/h5cloud/h5repack_rnd/ATL03_20220201060852_06261401_005_01.h5\"]\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e85dc339-f339-482c-8f29-70b728e67a11",
   "metadata": {
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "### Option 1: Read the original file with `h5py` "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "e54d588c-3910-4a53-96f0-7d46977c1187",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1/1 [00:05<00:00,  5.55s/it]\n",
      "100%|██████████| 1/1 [00:07<00:00,  7.71s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 3.86 s, sys: 680 ms, total: 4.54 s\n",
      "Wall time: 13.3 s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "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>tool</th>\n",
       "      <th>format</th>\n",
       "      <th>run</th>\n",
       "      <th>time</th>\n",
       "      <th>mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>h5py</td>\n",
       "      <td>original</td>\n",
       "      <td>0</td>\n",
       "      <td>2.196899</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>h5py</td>\n",
       "      <td>original</td>\n",
       "      <td>1</td>\n",
       "      <td>0.948688</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>h5py</td>\n",
       "      <td>original</td>\n",
       "      <td>2</td>\n",
       "      <td>0.773618</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>h5py</td>\n",
       "      <td>original</td>\n",
       "      <td>3</td>\n",
       "      <td>0.787938</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>h5py</td>\n",
       "      <td>original</td>\n",
       "      <td>4</td>\n",
       "      <td>0.836782</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>h5py</td>\n",
       "      <td>repacked</td>\n",
       "      <td>0</td>\n",
       "      <td>2.450255</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>h5py</td>\n",
       "      <td>repacked</td>\n",
       "      <td>1</td>\n",
       "      <td>1.312406</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>h5py</td>\n",
       "      <td>repacked</td>\n",
       "      <td>2</td>\n",
       "      <td>1.352802</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>h5py</td>\n",
       "      <td>repacked</td>\n",
       "      <td>3</td>\n",
       "      <td>1.224425</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>h5py</td>\n",
       "      <td>repacked</td>\n",
       "      <td>4</td>\n",
       "      <td>1.372676</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   tool    format  run      time        mean\n",
       "0  h5py  original    0  2.196899  314.330933\n",
       "1  h5py  original    1  0.948688  314.330933\n",
       "2  h5py  original    2  0.773618  314.330933\n",
       "3  h5py  original    3  0.787938  314.330933\n",
       "4  h5py  original    4  0.836782  314.330933\n",
       "5  h5py  repacked    0  2.450255  314.330933\n",
       "6  h5py  repacked    1  1.312406  314.330933\n",
       "7  h5py  repacked    2  1.352802  314.330933\n",
       "8  h5py  repacked    3  1.224425  314.330933\n",
       "9  h5py  repacked    4  1.372676  314.330933"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    " \n",
    "benchmarks = []\n",
    "\n",
    "tool = \"h5py\"\n",
    "\n",
    "# first let's to the originals:\n",
    "for file in tqdm.tqdm(original):\n",
    "    # we open the file 5 times and track the results\n",
    "    for run in range(5):\n",
    "        s3.invalidate_cache(file)\n",
    "        start = time.time()\n",
    "        with h5py.File(s3.open(file)) as f:\n",
    "            data = f[f'{group}/{variable}'][:]\n",
    "            data_mean = data.mean()\n",
    "            elapsed = time.time() - start\n",
    "            benchmarks.append({\"tool\": tool,\n",
    "                               \"format\": \"original\",\n",
    "                               \"run\": run,\n",
    "                               \"time\": elapsed,\n",
    "                               \"mean\": data_mean})\n",
    "\n",
    "for file in tqdm.tqdm(repacked):\n",
    "    # we open the file 5 times and track the results\n",
    "    for run in range(5):\n",
    "        s3.invalidate_cache(file)\n",
    "        start = time.time()\n",
    "        with h5py.File(s3.open(file)) as f:\n",
    "            data = f[f'{group}/{variable}'][:]\n",
    "            data_mean = data.mean()\n",
    "            elapsed = time.time() - start\n",
    "            benchmarks.append({\"tool\": tool,\n",
    "                               \"format\": \"repacked\",\n",
    "                               \"run\": run,\n",
    "                               \"time\": elapsed,\n",
    "                               \"mean\": data_mean})\n",
    "\n",
    "df_h5 = pd.DataFrame.from_dict(benchmarks)\n",
    "df_h5"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e353e811-35e6-4a51-9773-712c280f959c",
   "metadata": {
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "### Option 2: Read the group with xarray"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "6625b8a9-0528-47f4-950b-b43901637f37",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1/1 [00:17<00:00, 17.98s/it]\n",
      "100%|██████████| 1/1 [05:28<00:00, 328.80s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1min 33s, sys: 22.1 s, total: 1min 55s\n",
      "Wall time: 5min 46s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "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>tool</th>\n",
       "      <th>format</th>\n",
       "      <th>run</th>\n",
       "      <th>time</th>\n",
       "      <th>mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>xarray</td>\n",
       "      <td>repacked</td>\n",
       "      <td>0</td>\n",
       "      <td>4.621870</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>xarray</td>\n",
       "      <td>repacked</td>\n",
       "      <td>1</td>\n",
       "      <td>3.387357</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>xarray</td>\n",
       "      <td>repacked</td>\n",
       "      <td>2</td>\n",
       "      <td>3.373886</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>xarray</td>\n",
       "      <td>repacked</td>\n",
       "      <td>3</td>\n",
       "      <td>3.331894</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>xarray</td>\n",
       "      <td>repacked</td>\n",
       "      <td>4</td>\n",
       "      <td>3.265961</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>xarray</td>\n",
       "      <td>original</td>\n",
       "      <td>0</td>\n",
       "      <td>71.940291</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>xarray</td>\n",
       "      <td>original</td>\n",
       "      <td>1</td>\n",
       "      <td>64.167644</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>xarray</td>\n",
       "      <td>original</td>\n",
       "      <td>2</td>\n",
       "      <td>63.205375</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>xarray</td>\n",
       "      <td>original</td>\n",
       "      <td>3</td>\n",
       "      <td>65.394934</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>xarray</td>\n",
       "      <td>original</td>\n",
       "      <td>4</td>\n",
       "      <td>64.086429</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     tool    format  run       time  \\\n",
       "0  xarray  repacked    0   4.621870   \n",
       "1  xarray  repacked    1   3.387357   \n",
       "2  xarray  repacked    2   3.373886   \n",
       "3  xarray  repacked    3   3.331894   \n",
       "4  xarray  repacked    4   3.265961   \n",
       "5  xarray  original    0  71.940291   \n",
       "6  xarray  original    1  64.167644   \n",
       "7  xarray  original    2  63.205375   \n",
       "8  xarray  original    3  65.394934   \n",
       "9  xarray  original    4  64.086429   \n",
       "\n",
       "                                                            mean  \n",
       "0  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "1  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "2  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "3  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "4  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "5  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "6  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "7  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "8  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "9  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "xr_benchmarks = []\n",
    "\n",
    "tool = \"xarray\"\n",
    "\n",
    "for file in tqdm.tqdm(repacked):\n",
    "    # we open the file 5 times and track the results\n",
    "    for run in range(5):\n",
    "        start = time.time()\n",
    "        with s3.open(file, 'rb') as file_stream:\n",
    "            ds = xr.open_dataset(file_stream, group=group, engine='h5netcdf')\n",
    "            variable_mean = ds[variable].mean()\n",
    "            elapsed =  time.time() - start\n",
    "            xr_benchmarks.append({\"tool\": tool,\n",
    "                                  \"format\": \"repacked\",\n",
    "                                  \"run\": run,\n",
    "                                  \"time\": elapsed,\n",
    "                                  \"mean\": variable_mean})       \n",
    "\n",
    "\n",
    "# this is interesting, it takes forever with xarray on this granule... why?\n",
    "for file in tqdm.tqdm(original):\n",
    "    # we open the file 5 times and track the results\n",
    "    for run in range(5):\n",
    "        start = time.time()\n",
    "        with s3.open(file, 'rb') as file_stream:\n",
    "            ds = xr.open_dataset(file_stream, group=group, engine='h5netcdf')\n",
    "            variable_mean = ds[variable].mean()\n",
    "            elapsed =  time.time() - start\n",
    "            xr_benchmarks.append({\"tool\": tool,\n",
    "                                  \"format\": \"original\",\n",
    "                                  \"run\": run,\n",
    "                                  \"time\": elapsed,\n",
    "                                  \"mean\": variable_mean})\n",
    "            \n",
    "df_xr = pd.DataFrame.from_dict(xr_benchmarks)\n",
    "df_xr"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92005cb2-1807-4742-92f2-9361c6378d11",
   "metadata": {
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "### Option 3: Read the group with `h5coro`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "90c7790c",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1/1 [00:04<00:00,  4.98s/it]\n",
      "100%|██████████| 1/1 [00:06<00:00,  6.14s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 3.23 s, sys: 343 ms, total: 3.57 s\n",
      "Wall time: 11.1 s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "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>tool</th>\n",
       "      <th>format</th>\n",
       "      <th>run</th>\n",
       "      <th>time</th>\n",
       "      <th>mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>repacked</td>\n",
       "      <td>0</td>\n",
       "      <td>1.485010</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>repacked</td>\n",
       "      <td>1</td>\n",
       "      <td>0.929521</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>repacked</td>\n",
       "      <td>2</td>\n",
       "      <td>1.131845</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>repacked</td>\n",
       "      <td>3</td>\n",
       "      <td>0.783148</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>repacked</td>\n",
       "      <td>4</td>\n",
       "      <td>0.647872</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>original</td>\n",
       "      <td>0</td>\n",
       "      <td>2.180916</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>original</td>\n",
       "      <td>1</td>\n",
       "      <td>1.208768</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>original</td>\n",
       "      <td>2</td>\n",
       "      <td>1.027675</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>original</td>\n",
       "      <td>3</td>\n",
       "      <td>0.863832</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>original</td>\n",
       "      <td>4</td>\n",
       "      <td>0.858000</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     tool    format  run      time        mean\n",
       "0  h5coro  repacked    0  1.485010  314.330933\n",
       "1  h5coro  repacked    1  0.929521  314.330933\n",
       "2  h5coro  repacked    2  1.131845  314.330933\n",
       "3  h5coro  repacked    3  0.783148  314.330933\n",
       "4  h5coro  repacked    4  0.647872  314.330933\n",
       "5  h5coro  original    0  2.180916  314.330933\n",
       "6  h5coro  original    1  1.208768  314.330933\n",
       "7  h5coro  original    2  1.027675  314.330933\n",
       "8  h5coro  original    3  0.863832  314.330933\n",
       "9  h5coro  original    4  0.858000  314.330933"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "h5coro_benchmarks = []\n",
    "\n",
    "tool = \"h5coro\"\n",
    "\n",
    "repacked =  [\"nasa-cryo-scratch/h5cloud/h5repack_sm/ATL03_20220201060852_06261401_005_01_repacked.h5\"]\n",
    "original =  [\"nasa-cryo-scratch/h5cloud/h5repack_sm/ATL03_20220201060852_06261401_005_01.h5\"]\n",
    "\n",
    "for file in tqdm.tqdm(repacked):\n",
    "    for run in range(5):\n",
    "        start = time.time()\n",
    "        h5obj = h5coro.H5Coro(file, s3driver.S3Driver)\n",
    "        output = h5obj.readDatasets(datasets=[f'{group}/{variable}'], block=True)\n",
    "        variable_mean = h5obj[f'{group}/{variable}'].values.mean()\n",
    "        elapsed =  time.time() - start\n",
    "        h5coro_benchmarks.append({\"tool\": tool,\n",
    "                                  \"format\": \"repacked\",\n",
    "                                  \"run\": run,\n",
    "                                  \"time\": elapsed,\n",
    "                                  \"mean\": variable_mean})\n",
    "    \n",
    "    \n",
    "for file in tqdm.tqdm(original):\n",
    "    for run in range(5):\n",
    "        start = time.time()\n",
    "        h5obj = h5coro.H5Coro(file, s3driver.S3Driver)\n",
    "        output = h5obj.readDatasets(datasets=[f'{group}/{variable}'], block=True)\n",
    "        variable_mean = h5obj[f'{group}/{variable}'].values.mean()\n",
    "        elapsed =  time.time() - start\n",
    "        h5coro_benchmarks.append({\"tool\": tool,\n",
    "                                  \"format\": \"original\",\n",
    "                                  \"run\": run,\n",
    "                                  \"time\": elapsed,\n",
    "                                  \"mean\": variable_mean})\n",
    "\n",
    "df_h5coro = pd.DataFrame.from_dict(h5coro_benchmarks)\n",
    "df_h5coro"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "59f3bf8d-0888-499d-9105-21f92fa9066a",
   "metadata": {
    "tags": []
   },
   "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>tool</th>\n",
       "      <th>format</th>\n",
       "      <th>run</th>\n",
       "      <th>time</th>\n",
       "      <th>mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>h5py</td>\n",
       "      <td>original</td>\n",
       "      <td>0</td>\n",
       "      <td>2.196899</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>h5py</td>\n",
       "      <td>original</td>\n",
       "      <td>1</td>\n",
       "      <td>0.948688</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>h5py</td>\n",
       "      <td>original</td>\n",
       "      <td>2</td>\n",
       "      <td>0.773618</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>h5py</td>\n",
       "      <td>original</td>\n",
       "      <td>3</td>\n",
       "      <td>0.787938</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>h5py</td>\n",
       "      <td>original</td>\n",
       "      <td>4</td>\n",
       "      <td>0.836782</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>h5py</td>\n",
       "      <td>repacked</td>\n",
       "      <td>0</td>\n",
       "      <td>2.450255</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>h5py</td>\n",
       "      <td>repacked</td>\n",
       "      <td>1</td>\n",
       "      <td>1.312406</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>h5py</td>\n",
       "      <td>repacked</td>\n",
       "      <td>2</td>\n",
       "      <td>1.352802</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>h5py</td>\n",
       "      <td>repacked</td>\n",
       "      <td>3</td>\n",
       "      <td>1.224425</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>h5py</td>\n",
       "      <td>repacked</td>\n",
       "      <td>4</td>\n",
       "      <td>1.372676</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>xarray</td>\n",
       "      <td>repacked</td>\n",
       "      <td>0</td>\n",
       "      <td>4.621870</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>xarray</td>\n",
       "      <td>repacked</td>\n",
       "      <td>1</td>\n",
       "      <td>3.387357</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>xarray</td>\n",
       "      <td>repacked</td>\n",
       "      <td>2</td>\n",
       "      <td>3.373886</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>xarray</td>\n",
       "      <td>repacked</td>\n",
       "      <td>3</td>\n",
       "      <td>3.331894</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>xarray</td>\n",
       "      <td>repacked</td>\n",
       "      <td>4</td>\n",
       "      <td>3.265961</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>xarray</td>\n",
       "      <td>original</td>\n",
       "      <td>0</td>\n",
       "      <td>71.940291</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>xarray</td>\n",
       "      <td>original</td>\n",
       "      <td>1</td>\n",
       "      <td>64.167644</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>xarray</td>\n",
       "      <td>original</td>\n",
       "      <td>2</td>\n",
       "      <td>63.205375</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>xarray</td>\n",
       "      <td>original</td>\n",
       "      <td>3</td>\n",
       "      <td>65.394934</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>xarray</td>\n",
       "      <td>original</td>\n",
       "      <td>4</td>\n",
       "      <td>64.086429</td>\n",
       "      <td>&lt;xarray.DataArray 'h_ph' ()&gt;\\narray(314.33093, dtype=float32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>repacked</td>\n",
       "      <td>0</td>\n",
       "      <td>1.485010</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>repacked</td>\n",
       "      <td>1</td>\n",
       "      <td>0.929521</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>repacked</td>\n",
       "      <td>2</td>\n",
       "      <td>1.131845</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>repacked</td>\n",
       "      <td>3</td>\n",
       "      <td>0.783148</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>repacked</td>\n",
       "      <td>4</td>\n",
       "      <td>0.647872</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>original</td>\n",
       "      <td>0</td>\n",
       "      <td>2.180916</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>original</td>\n",
       "      <td>1</td>\n",
       "      <td>1.208768</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>original</td>\n",
       "      <td>2</td>\n",
       "      <td>1.027675</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>original</td>\n",
       "      <td>3</td>\n",
       "      <td>0.863832</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>h5coro</td>\n",
       "      <td>original</td>\n",
       "      <td>4</td>\n",
       "      <td>0.858000</td>\n",
       "      <td>314.330933</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     tool    format  run       time  \\\n",
       "0    h5py  original    0   2.196899   \n",
       "1    h5py  original    1   0.948688   \n",
       "2    h5py  original    2   0.773618   \n",
       "3    h5py  original    3   0.787938   \n",
       "4    h5py  original    4   0.836782   \n",
       "5    h5py  repacked    0   2.450255   \n",
       "6    h5py  repacked    1   1.312406   \n",
       "7    h5py  repacked    2   1.352802   \n",
       "8    h5py  repacked    3   1.224425   \n",
       "9    h5py  repacked    4   1.372676   \n",
       "0  xarray  repacked    0   4.621870   \n",
       "1  xarray  repacked    1   3.387357   \n",
       "2  xarray  repacked    2   3.373886   \n",
       "3  xarray  repacked    3   3.331894   \n",
       "4  xarray  repacked    4   3.265961   \n",
       "5  xarray  original    0  71.940291   \n",
       "6  xarray  original    1  64.167644   \n",
       "7  xarray  original    2  63.205375   \n",
       "8  xarray  original    3  65.394934   \n",
       "9  xarray  original    4  64.086429   \n",
       "0  h5coro  repacked    0   1.485010   \n",
       "1  h5coro  repacked    1   0.929521   \n",
       "2  h5coro  repacked    2   1.131845   \n",
       "3  h5coro  repacked    3   0.783148   \n",
       "4  h5coro  repacked    4   0.647872   \n",
       "5  h5coro  original    0   2.180916   \n",
       "6  h5coro  original    1   1.208768   \n",
       "7  h5coro  original    2   1.027675   \n",
       "8  h5coro  original    3   0.863832   \n",
       "9  h5coro  original    4   0.858000   \n",
       "\n",
       "                                                            mean  \n",
       "0                                                     314.330933  \n",
       "1                                                     314.330933  \n",
       "2                                                     314.330933  \n",
       "3                                                     314.330933  \n",
       "4                                                     314.330933  \n",
       "5                                                     314.330933  \n",
       "6                                                     314.330933  \n",
       "7                                                     314.330933  \n",
       "8                                                     314.330933  \n",
       "9                                                     314.330933  \n",
       "0  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "1  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "2  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "3  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "4  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "5  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "6  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "7  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "8  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "9  <xarray.DataArray 'h_ph' ()>\\narray(314.33093, dtype=float32)  \n",
       "0                                                     314.330933  \n",
       "1                                                     314.330933  \n",
       "2                                                     314.330933  \n",
       "3                                                     314.330933  \n",
       "4                                                     314.330933  \n",
       "5                                                     314.330933  \n",
       "6                                                     314.330933  \n",
       "7                                                     314.330933  \n",
       "8                                                     314.330933  \n",
       "9                                                     314.330933  "
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.concat([df_h5, df_xr, df_h5coro])\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "16078a45-2758-4a38-bc93-667880a31f74",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiMAAAHACAYAAABwEmgAAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAAA02UlEQVR4nO3de1xU1f7/8fcAwwApmqgohkrlrQdd/EIXTfNSYGh2L7+Hykw934wuR8l6aHZBs/RYx6Od8tK3FK00u18p5VRe0uoIqVleKtMwBUkrMdFhZPbvj37M1wlQBt2sAV7Px4OHzZq111p7fxh9t/eeGYdlWZYAAAAMCTG9AAAA0LgRRgAAgFGEEQAAYBRhBAAAGEUYAQAARhFGAACAUYQRAABgFGEEAAAYRRgBAABGEUaAOpaVlSWHw6G9e/ces1/fvn3lcDgq/Vx++eV1tFJUyM7OlsPhUF5enq3zOBwOZWVl2ToHEIzCTC8AQPVOP/10vfTSS35tzZs3N7MYALAJYQQIYpGRkbrooouMrqG8vFxHjhyRy+WqV2MDqD+4TAMYsmfPHv3lL39Rs2bNFBsbq+HDh2v//v21GmvXrl36n//5H8XHxys8PFxxcXG6/vrrtWfPHl+fgoIC3XzzzWrdurVcLpe6deumf/zjH/J6vb4+O3bskMPh0LRp0zR58mQlJCTI5XLpk08+kSS988476tGjh6KiotS0aVOlpKTos88+q9Eajzd2Xl6errzySrVo0UIRERHq3r27XnnlFb8xKi6X5Obm6rbbblOLFi10yimnaPDgwfrhhx/8+ubm5uqqq67SaaedpoiICJ155pm6/fbbq7w8tmXLFv3lL39RbGysXC6X2rdvr6FDh8rtdle7P4WFhUpKSlKnTp303XffSZJKSko0duxYJSQkKDw8XO3atdPo0aN18OBBv21LSkr017/+VTExMWrSpIkuv/xyffvttzU6jkBDxJkRwJDrrrtOQ4YM0YgRI7Rx40aNHz9ekjRv3jxfn23btqlFixYqKSlRhw4d9N///d968MEHFRkZ6euza9cunX/++fJ4PHrggQd0zjnnaN++fVq6dKl+/fVXxcbG6ueff1bPnj1VVlamRx99VB07dtR7772nsWPHatu2bZo1a5bf2p566il17txZTz75pKKjo9WpUyctWrRIN910k1JTU7V48WK53W5NmzZNffv21UcffaRevXrVaL+rGvuTTz7R5ZdfrgsvvFBz5sxRs2bN9PLLL2vIkCEqLS3VsGHD/MYYMWKEUlJStGjRIu3cuVMPPvig+vbtq6+++sp3GWvbtm3q0aOHRo4cqWbNmmnHjh2aPn26evXqpY0bN8rpdEqSNmzYoF69eqlly5aaNGmSOnXqpMLCQr3zzjsqKyur8qzN119/rYEDB+q0007TZ599ppYtW6q0tFR9+vTRTz/95KvDN998o4cfflgbN27Uv//9bzkcDlmWpauvvlpr1qzRww8/rPPPP1+rV69WWlpajY4f0CBZAOrUI488Ykmypk2b5teekZFhRUREWF6v17Isy5owYYI1a9Ys6+OPP7bef/9966677rLCwsKsSy65xCovL/dtN3z4cMvpdFqbNm2qds5x48ZZkqwvvvjCr/2OO+6wHA6HtXXrVsuyLGv79u2WJOuMM86wysrKfP3Ky8utuLg46+yzz/ab+8CBA1br1q2tnj17Hne/qxvbsiyra9euVvfu3S2Px+PXfsUVV1ht27b1zTl//nxLknXNNdf49Vu9erUlyZo8eXKVc3u9Xsvj8Vg//vijJcl6++23fc/179/fat68uVVcXFzt2ivmXbt2rZWbm2tFR0db119/vXXo0CFfnylTplghISHW2rVr/bZ97bXXLElWTk6OZVmW9cEHH1iSrJkzZ/r1e+yxxyxJ1iOPPFLtOoCGiss0gCFXXnml3+NzzjlHhw8fVnFxsSRp8uTJuuOOO9SvXz8NHDhQ//rXvzR16lStXLlSb7/9tm+7Dz74QP369VO3bt2qnevjjz/WWWedpQsuuMCvfdiwYbIsSx9//HGltVWcOZCkrVu3avfu3brlllsUEvJ/f200adJE1113nT7//HOVlpZK+r/7QCp+jr4MVNXY33//vbZs2aKbbrpJkvy2HThwoAoLC7V161a/MSr6VujZs6c6dOjgu+QjScXFxRo1apTi4+MVFhYmp9OpDh06SJI2b94sSSotLdWKFSt04403qlWrVtUevwoLFizQwIEDNXLkSL3yyiuKiIjwPffee+8pMTFR5513nt8+DBgwQA6HQ8uXL5ck3xr/vA/p6enHnR9oqAgjgCExMTF+jysuBxw6dKjabW6++WZJ0ueff+5r+/nnn3Xaaacdc659+/apbdu2ldrj4uJ8zx/tz30rnq9uDK/Xq19//VWSdMYZZ8jpdPp+Jk2adMyxK+5rGTt2rN92TqdTGRkZklTpPo82bdpUWkebNm186/R6vUpNTdUbb7yh+++/Xx999JH+85//+I5bxTH+9ddfVV5eftzjV+Hll19WZGSkRo4cKYfDUWk/vvrqq0r70LRpU1mW5duHffv2KSwsrFL9q9onoLHgnhGgHjr67ESrVq30008/HbN/TEyMCgsLK7Xv3r1bktSyZUu/9j//Q1vxD2d1Y4SEhOjUU0+VJL377rt+N35WBJ7qxq6Ye/z48br22murXH+XLl38HhcVFVXqU1RUpDPPPFPSH/d0bNiwQdnZ2br11lt9fb7//nu/bVq0aKHQ0NDjHr8KL730kh566CH16dNHy5Yt03nnnee3H5GRkX73/BytYj9jYmJ05MgR7du3zy+QVLVPQGPBmRGgHlmwYIEk+b3dNy0tTZ988kmlSxlHu/TSS7Vp0yZ9+eWXfu0LFy6Uw+FQv379jjlvly5d1K5dOy1atEiWZfnaDx48qNdff933DhtJOvvss5WcnOz7+XMYqWrsTp06acOGDX7bHf3TtGlTv23+/Nkra9as0Y8//qi+fftK+r/A8+ebT+fOnev3ODIyUn369NGrr7563A+hk/4IL//+97/VrVs39evXz+8M1RVXXKFt27YpJiamyn3o2LGjJPmO9Z/3YdGiRcedH2ioODMCBKFVq1bpscce0zXXXKPTTz9dhw8f1gcffKBnn31W/fv31+DBg319J02apA8++ECXXHKJHnjgAZ199tn67bff9OGHHyozM1Ndu3bVmDFjtHDhQg0aNEiTJk1Shw4d9P7772vWrFm644471Llz52OuJyQkRNOmTdNNN92kK664QrfffrvcbreeeOIJ/fbbb5o6deoJ7e/cuXOVlpamAQMGaNiwYWrXrp1++eUXbd68WV9++aVeffVVv/55eXkaOXKkbrjhBu3cuVMTJkxQu3btfJd1unbtqjPOOEPjxo2TZVlq0aKF3n33XeXm5laau+IdNhdeeKHGjRunM888U3v27NE777yjuXPnVgpCTZs21Ycffqhrr71WKSkpeuedd9SvXz+NHj1ar7/+ui655BKNGTNG55xzjrxerwoKCrRs2TLde++9uvDCC5WamqpLLrlE999/vw4ePKjk5GStXr1aL7zwwgkdQ6BeM3v/LND4VLyb5ueff/Zrr3jHxvbt263vvvvOGjhwoNWuXTvL5XJZERER1tlnn2099thj1uHDhyuNuXPnTmv48OFWmzZtLKfTacXFxVk33nijtWfPHl+fH3/80UpPT7diYmIsp9NpdenSxXriiSf83h1T8Y6XJ554osq1v/XWW9aFF15oRUREWKeccop16aWXWqtXr67Rfh9v7A0bNlg33nij1bp1a8vpdFpt2rSx+vfvb82ZM6fSMVq2bJl1yy23WM2bN7ciIyOtgQMHWt99953feJs2bbJSUlKspk2bWqeeeqp1ww03WAUFBVW+Y2XTpk3WDTfcYMXExFjh4eFW+/btrWHDhvmO9dHvpqngdrut6667zoqIiLDef/99y7Is6/fff7cefPBBq0uXLlZ4eLjVrFkz6+yzz7bGjBljFRUV+bb97bffrOHDh1vNmze3oqKirJSUFGvLli28mwaNlsOyjjrnCgBBLDs7W7fddpvWrl2r5ORk08sBcJJwzwgAADCKMAIAAIziMg0AADCKMyMAAMAowggAADCKMAIAAIyqFx965vV6tXv3bjVt2rTSR0kDAIDgZFmWDhw4oLi4OL+vsfizehFGdu/erfj4eNPLAAAAtbBz585jfiFlvQgjFR/HvHPnTkVHRxteTe14PB4tW7ZMqampfl+fDjOoR/CgFsGDWgSPhlKLkpISxcfHV/pahT+rF2Gk4tJMdHR0vQ4jUVFRio6Orte/WA0F9Qge1CJ4UIvg0dBqcbxbLLiBFQAAGEUYAQAARhFGAACAUfXinhEAAOzk9XpVVlZmehk+Ho9HYWFhOnz4sMrLy00vp1pOp1OhoaEnPA5hBADQqJWVlWn79u3yer2ml+JjWZbatGmjnTt3Bv3nazVv3lxt2rQ5oXUSRgAAjZZlWSosLFRoaKji4+OP+cFcdcnr9er3339XkyZNgmZNf2ZZlkpLS1VcXCxJatu2ba3HIowAABqtI0eOqLS0VHFxcYqKijK9HJ+Ky0YRERFBG0YkKTIyUpJUXFys1q1b1/qSTfDuIQAANqu4HyM8PNzwSuqvihDn8XhqPQZhBADQ6AX7fRnB7GQcO8IIAAAwijACAEA9t3z5cjkcDv3222+ml1Ir3MAKAMCfdBz3fp3Ot2PqoID69+3bV+edd55mzJghSerZs6cKCwvVrFkzG1ZnP8IIAAD1XHh4uNq0aWN6GbXGZRoAAOqRYcOGacWKFZo5c6YcDoccDoeys7P9LtNkZ2erefPmeu+999SlSxdFRUXp+uuv18GDB7VgwQJ17NhRp556qu6++26/T3gtKyvT/fffr3bt2umUU07RhRdeqOXLl9u+T5wZAQCgHpk5c6a+/fZbJSYmatKkSZKkb775plK/0tJSPfXUU3r55Zd14MABXXvttbr22mvVvHlz5eTk6IcfftB1112nXr16aciQIZKk2267TTt27NDLL7+suLg4vfnmm7r88su1ceNGderUybZ9Iowcxc5rhK5QS9MukBKzlspdbt9byAK97ggAqF+aNWum8PBwRUVF+S7NbNmypVI/j8ej2bNn64wzzpAkXX/99XrhhRe0Z88eNWnSRGeddZb69eunTz75REOGDNG2bdu0ePFi/fTTT4qLi5MkjR07Vh9++KHmz5+vxx9/3LZ9IowAANAARUVF+YKIJMXGxqpjx45q0qSJX1vFx7l/+eWXsixLnTt39hvH7XYrJibG1rUSRgAAaICcTqffY4fDUWVbxRcEer1ehYaGKj8/v9LHuh8dYOxAGAEAoJ4JDw/3u/H0ZOjevbvKy8tVXFys3r17n9Sxj4d30wAAUM907NhRX3zxhXbs2KG9e/f6zm6ciM6dO+umm27S0KFD9cYbb2j79u1au3at/v73vysnJ+ckrLp6hBEAAOqZsWPHKjQ0VGeddZZatWqlgoKCkzLu/PnzNXToUN17773q0qWLrrzySn3xxReKj48/KeNXx2FZlmXrDCdBSUmJmjVrpv379ys6Otq2eex/N0257v9PKO+mCQIej0c5OTkaOHBgpWuoqFvUIng0xlocPnxY27dvV0JCgiIiIkwvx8fr9aqkpETR0dEKCQnu8wbHOoY1/fc7uPcQAAA0eIQRAABgFGEEAAAYRRgBAABGEUYAAIBRhBEAQKNXD95YGrROxmec8AmsAIBGy+l0yuFw6Oeff1arVq3kcNj30QuB8Hq9Kisr0+HDh4P2rb2WZamsrEw///yzQkJCFB4eXuuxCCMAgEYrNDRUp512mn766Sft2LHD9HJ8LMvSoUOHFBkZGTQBqTpRUVFq3779CYUmwggAoFFr0qSJOnXqJI/HY3opPh6PRytXrtQll1wS1B9AFxoaqrCwsBMOTAGHkZUrV+qJJ55Qfn6+CgsL9eabb+rqq6+u0barV69Wnz59lJiYqPXr1wc6NQAAtggNDa30TbUmhYaG6siRI4qIiAjqMHKyBHxO5eDBgzr33HP19NNPB7Td/v37NXToUF166aWBTgkAABqwgM+MpKWlKS0tLeCJbr/9dqWnpys0NFRvvfVWwNsDAICGqU7uGZk/f762bdumF198UZMnTz5uf7fbLbfb7XtcUlIi6Y9raHZe03OF2vfWLleI5fenXYLpmmcwqzhOHC/zqEXwoBbBo6HUoqbrtz2MfPfddxo3bpxWrVqlsLCaTTdlyhRNnDixUvuyZcsUFRV1spfoM+0C24b2eTT5xN+PfSw5OTm2jt/Q5Obmml4C/j9qETyoRfCo77UoLS2tUT9bw0h5ebnS09M1ceJEde7cucbbjR8/XpmZmb7HJSUlio+PV2pq6jG/gvhEJWYttW1sV4ilR5O9eigvRG6vfW/T+jprgG1jNyQej0e5ublKSUlpFDeHBTNqETyoRfBoKLWouLJxPLaGkQMHDigvL0/r1q3TXXfdJemPD3KxLEthYWFatmyZ+vfvX2k7l8sll8tVqd3pdNpaFHe5/e/ldnsdts5Tn39pTbD7dwo1Ry2CB7UIHvW9FjVdu61hJDo6Whs3bvRrmzVrlj7++GO99tprSkhIsHN6AABQDwQcRn7//Xd9//33vsfbt2/X+vXr1aJFC7Vv317jx4/Xrl27tHDhQoWEhCgxMdFv+9atWysiIqJSOwAAaJwCDiN5eXnq16+f73HFvR233nqrsrOzVVhYqIKCgpO3QgAA0KAFHEb69u17zG83zM7OPub2WVlZysrKCnRaAADQQAXnVwECAIBGgzACAACMIowAAACjCCMAAMAowggAADCKMAIAAIwijAAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAKMIIAAAwijACAACMIowAAACjCCMAAMAowggAADCKMAIAAIwijAAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAKMIIAAAwijACAACMIowAAACjCCMAAMAowggAADCKMAIAAIwijAAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAqIDDyMqVKzV48GDFxcXJ4XDorbfeOmb/N954QykpKWrVqpWio6PVo0cPLV26tLbrBQAADUzAYeTgwYM699xz9fTTT9eo/8qVK5WSkqKcnBzl5+erX79+Gjx4sNatWxfwYgEAQMMTFugGaWlpSktLq3H/GTNm+D1+/PHH9fbbb+vdd99V9+7dA50eAAA0MAGHkRPl9Xp14MABtWjRoto+brdbbrfb97ikpESS5PF45PF4bFubK9Syb+wQy+9Pu9h5fBqSiuPE8TKPWgQPahE8Gkotarp+h2VZtf7X0eFw6M0339TVV19d422eeOIJTZ06VZs3b1br1q2r7JOVlaWJEydWal+0aJGioqJqu1wAAFCHSktLlZ6erv379ys6OrrafnUaRhYvXqyRI0fq7bff1mWXXVZtv6rOjMTHx2vv3r3H3JkTlZhl3421rhBLjyZ79VBeiNxeh23zfJ01wLaxGxKPx6Pc3FylpKTI6XSaXk6jRi2CB7UIHg2lFiUlJWrZsuVxw0idXaZZsmSJRowYoVdfffWYQUSSXC6XXC5XpXan02lrUdzl9oUE3xxeh63z1OdfWhPs/p1CzVGL4EEtgkd9r0VN114nnzOyePFiDRs2TIsWLdKgQYPqYkoAAFBPBHxm5Pfff9f333/ve7x9+3atX79eLVq0UPv27TV+/Hjt2rVLCxculPRHEBk6dKhmzpypiy66SEVFRZKkyMhINWvW7CTtBgAAqK8CPjOSl5en7t27+96Wm5mZqe7du+vhhx+WJBUWFqqgoMDXf+7cuTpy5IjuvPNOtW3b1vfzt7/97STtAgAAqM8CPjPSt29fHeue1+zsbL/Hy5cvD3QKAADQiPDdNAAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAKMIIAAAwijACAACMIowAAACjCCMAAMAowggAADAq4I+DB9C4dBz3vq3ju0ItTbtASsxaKne5w7Z5dkzlG8OBYMWZEQAAYBRhBAAAGEUYAQAARhFGAACAUYQRAABgFGEEAAAYRRgBAABGEUYAAIBRhBEAAGAUYQQAABhFGAEAAEYRRgAAgFGEEQAAYBRhBAAAGEUYAQAARhFGAACAUYQRAABgFGEEAAAYRRgBAABGEUYAAIBRhBEAAGAUYQQAABhFGAEAAEYRRgAAgFGEEQAAYFTAYWTlypUaPHiw4uLi5HA49NZbbx13mxUrVigpKUkRERE6/fTTNWfOnNqsFQAANEABh5GDBw/q3HPP1dNPP12j/tu3b9fAgQPVu3dvrVu3Tg888IDuuecevf766wEvFgAANDxhgW6QlpamtLS0GvefM2eO2rdvrxkzZkiSunXrpry8PD355JO67rrrAp0eAAA0MLbfM/LZZ58pNTXVr23AgAHKy8uTx+Oxe3oAABDkAj4zEqiioiLFxsb6tcXGxurIkSPau3ev2rZtW2kbt9stt9vte1xSUiJJ8ng8tgYYV6hl39ghlt+fdiHg1UzFceJ4HZ+drwuJ10Yw4XURPBpKLWq6ftvDiCQ5HA6/x5ZlVdleYcqUKZo4cWKl9mXLlikqKurkL/D/m3aBbUP7PJrstXX8nJwcW8dvaHJzc00vIejVxetC4rURTHhdBI/6XovS0tIa9bM9jLRp00ZFRUV+bcXFxQoLC1NMTEyV24wfP16ZmZm+xyUlJYqPj1dqaqqio6NtW2ti1lLbxnaFWHo02auH8kLk9lYdwk6Gr7MG2DZ2Q+LxeJSbm6uUlBQ5nU7Tywlqdr4uJF4bwYTXRfBoKLWouLJxPLaHkR49eujdd9/1a1u2bJmSk5OrPcAul0sul6tSu9PptLUo7nL7/iL0zeF12DpPff6lNcHu36mGoC5eFxKvjWDC6yJ41Pda1HTtAd/A+vvvv2v9+vVav369pD/eurt+/XoVFBRI+uOsxtChQ339R40apR9//FGZmZnavHmz5s2bp+eff15jx44NdGoAANAABXxmJC8vT/369fM9rriccuuttyo7O1uFhYW+YCJJCQkJysnJ0ZgxY/TMM88oLi5OTz31FG/rBQAAkmoRRvr27eu7AbUq2dnZldr69OmjL7/8MtCpAABAI8B30wAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAKMIIAAAwijACAACMIowAAACjCCMAAMAowggAADAq4O+mAepCx3Hv2zq+K9TStAukxKyltn1t/Y6pg2wZFwAaGs6MAAAAowgjAADAKMIIAAAwijACAACMIowAAACjCCMAAMAowggAADCKMAIAAIwijAAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAKMIIAAAwijACAACMIowAAACjCCMAAMAowggAADCKMAIAAIwijAAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAo2oVRmbNmqWEhARFREQoKSlJq1atOmb/l156Seeee66ioqLUtm1b3Xbbbdq3b1+tFgwAABqWgMPIkiVLNHr0aE2YMEHr1q1T7969lZaWpoKCgir7f/rppxo6dKhGjBihb775Rq+++qrWrl2rkSNHnvDiAQBA/RdwGJk+fbpGjBihkSNHqlu3bpoxY4bi4+M1e/bsKvt//vnn6tixo+655x4lJCSoV69euv3225WXl3fCiwcAAPVfWCCdy8rKlJ+fr3Hjxvm1p6amas2aNVVu07NnT02YMEE5OTlKS0tTcXGxXnvtNQ0aNKjaedxut9xut+9xSUmJJMnj8cjj8QSy5IC4Qi37xg6x/P60i53Hpy7ZWQupbupBLWo4Pq+NoFFxjDhW5jWUWtR0/Q7Lsmr8N8Du3bvVrl07rV69Wj179vS1P/7441qwYIG2bt1a5XavvfaabrvtNh0+fFhHjhzRlVdeqddee01Op7PK/llZWZo4cWKl9kWLFikqKqqmywUAAAaVlpYqPT1d+/fvV3R0dLX9AjozUsHhcPg9tiyrUluFTZs26Z577tHDDz+sAQMGqLCwUPfdd59GjRql559/vsptxo8fr8zMTN/jkpISxcfHKzU19Zg7c6ISs5baNrYrxNKjyV49lBcit7fqY3UyfJ01wLax65KdtZDqph7UomZ4bQQPj8ej3NxcpaSkVPs/i6gbDaUWFVc2jiegMNKyZUuFhoaqqKjIr724uFixsbFVbjNlyhRdfPHFuu+++yRJ55xzjk455RT17t1bkydPVtu2bStt43K55HK5KrU7nU5bi+Iut+8vQt8cXoet89TnX9qj1UUtJHvrQS0CnIfXRtCw++9a1Fx9r0VN1x7QDazh4eFKSkpSbm6uX3tubq7fZZujlZaWKiTEf5rQ0FBJf5xRAQAAjVvAl2kyMzN1yy23KDk5WT169NCzzz6rgoICjRo1StIfl1h27dqlhQsXSpIGDx6sv/71r5o9e7bvMs3o0aN1wQUXKC4u7uTuDQA0YB3HvW/r+K5QS9Mu+OPSnJ1nqXZMrf4NDGicAg4jQ4YM0b59+zRp0iQVFhYqMTFROTk56tChgySpsLDQ7zNHhg0bpgMHDujpp5/Wvffeq+bNm6t///76+9//fvL2AgAA1Fu1uoE1IyNDGRkZVT6XnZ1dqe3uu+/W3XffXZupAABAA8d30wAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAKMIIAAAwijACAACMIowAAACjCCMAAMAowggAADCKMAIAAIwijAAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAKMIIAAAwijACAACMIowAAACjCCMAAMAowggAADCKMAIAAIwijAAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAo8JMLwAAgPqm47j3bR3fFWpp2gVSYtZSucsdts2zY+og28YOBGdGAACAUYQRAABgFGEEAAAYRRgBAABGEUYAAIBRhBEAAGBUrcLIrFmzlJCQoIiICCUlJWnVqlXH7O92uzVhwgR16NBBLpdLZ5xxhubNm1erBQMAgIYl4M8ZWbJkiUaPHq1Zs2bp4osv1ty5c5WWlqZNmzapffv2VW5z4403as+ePXr++ed15plnqri4WEeOHDnhxQMAgPov4DAyffp0jRgxQiNHjpQkzZgxQ0uXLtXs2bM1ZcqUSv0//PBDrVixQj/88INatGghSerYseOJrRoAADQYAV2mKSsrU35+vlJTU/3aU1NTtWbNmiq3eeedd5ScnKxp06apXbt26ty5s8aOHatDhw7VftUAAKDBCOjMyN69e1VeXq7Y2Fi/9tjYWBUVFVW5zQ8//KBPP/1UERERevPNN7V3715lZGTol19+qfa+EbfbLbfb7XtcUlIiSfJ4PPJ4PIEsOSCuUMu+sUMsvz/tYufxqUt21kKqm3pQixqOz2ujxqhF8KAWJ3d8h2VZNd7T3bt3q127dlqzZo169Ojha3/sscf0wgsvaMuWLZW2SU1N1apVq1RUVKRmzZpJkt544w1df/31OnjwoCIjIyttk5WVpYkTJ1ZqX7RokaKiomq6XAAAYFBpaanS09O1f/9+RUdHV9svoDMjLVu2VGhoaKWzIMXFxZXOllRo27at2rVr5wsiktStWzdZlqWffvpJnTp1qrTN+PHjlZmZ6XtcUlKi+Ph4paamHnNnTlRi1lLbxnaFWHo02auH8kLk9tr3pUdfZw2wbey6ZGctpLqpB7WoGV4bNUctgge1qJmKKxvHE1AYCQ8PV1JSknJzc3XNNdf42nNzc3XVVVdVuc3FF1+sV199Vb///ruaNGkiSfr2228VEhKi0047rcptXC6XXC5XpXan0ymn0xnIkgNi5zcj+ubwOmydx87jU5fqohaSvfWgFgHOw2vjuKhF8KAWJ3f8gD9nJDMzU88995zmzZunzZs3a8yYMSooKNCoUaMk/XFWY+jQob7+6enpiomJ0W233aZNmzZp5cqVuu+++zR8+PAqL9EAAIDGJeC39g4ZMkT79u3TpEmTVFhYqMTEROXk5KhDhw6SpMLCQhUUFPj6N2nSRLm5ubr77ruVnJysmJgY3XjjjZo8efLJ2wsAAFBvBRxGJCkjI0MZGRlVPpednV2prWvXrsrNza3NVAAAoIHju2kAAIBRhBEAAGAUYQQAABhFGAEAAEYRRgAAgFGEEQAAYBRhBAAAGEUYAQAARhFGAACAUYQRAABgFGEEAAAYRRgBAABGEUYAAIBRhBEAAGAUYQQAABhFGAEAAEYRRgAAgFGEEQAAYBRhBAAAGEUYAQAARhFGAACAUYQRAABgFGEEAAAYRRgBAABGEUYAAIBRhBEAAGAUYQQAABhFGAEAAEYRRgAAgFGEEQAAYBRhBAAAGEUYAQAARhFGAACAUYQRAABgFGEEAAAYRRgBAABGEUYAAIBRtQojs2bNUkJCgiIiIpSUlKRVq1bVaLvVq1crLCxM5513Xm2mBQAADVDAYWTJkiUaPXq0JkyYoHXr1ql3795KS0tTQUHBMbfbv3+/hg4dqksvvbTWiwUAAA1PwGFk+vTpGjFihEaOHKlu3bppxowZio+P1+zZs4+53e2336709HT16NGj1osFAAANT1ggncvKypSfn69x48b5taempmrNmjXVbjd//nxt27ZNL774oiZPnnzcedxut9xut+9xSUmJJMnj8cjj8QSy5IC4Qi37xg6x/P60i53Hpy7ZWQupbupBLWo4Pq+NGqMWwYNanNzxHZZl1XhPd+/erXbt2mn16tXq2bOnr/3xxx/XggULtHXr1krbfPfdd+rVq5dWrVqlzp07KysrS2+99ZbWr19f7TxZWVmaOHFipfZFixYpKiqqpssFAAAGlZaWKj09Xfv371d0dHS1/QI6M1LB4XD4PbYsq1KbJJWXlys9PV0TJ05U586dazz++PHjlZmZ6XtcUlKi+Ph4paamHnNnTlRi1lLbxnaFWHo02auH8kLk9lY+VifL11kDbBu7LtlZC6lu6kEtaobXRs1Ri+BBLWqm4srG8QQURlq2bKnQ0FAVFRX5tRcXFys2NrZS/wMHDigvL0/r1q3TXXfdJUnyer2yLEthYWFatmyZ+vfvX2k7l8sll8tVqd3pdMrpdAay5IC4y+0ruG8Or8PWeew8PnWpLmoh2VsPahHgPLw2jotaBA9qcXLHD+gG1vDwcCUlJSk3N9evPTc31++yTYXo6Ght3LhR69ev9/2MGjVKXbp00fr163XhhRcGMj0AAGiAAr5Mk5mZqVtuuUXJycnq0aOHnn32WRUUFGjUqFGS/rjEsmvXLi1cuFAhISFKTEz0275169aKiIio1A4AABqngMPIkCFDtG/fPk2aNEmFhYVKTExUTk6OOnToIEkqLCw87meOAAAAVKjVDawZGRnKyMio8rns7OxjbpuVlaWsrKzaTAsAABogvpsGAAAYRRgBAABGEUYAAIBRhBEAAGAUYQQAABhFGAEAAEYRRgAAgFGEEQAAYBRhBAAAGEUYAQAARhFGAACAUYQRAABgFGEEAAAYRRgBAABGEUYAAIBRhBEAAGAUYQQAABhFGAEAAEYRRgAAgFGEEQAAYBRhBAAAGEUYAQAARhFGAACAUYQRAABgFGEEAAAYRRgBAABGEUYAAIBRhBEAAGAUYQQAABhFGAEAAEYRRgAAgFGEEQAAYBRhBAAAGEUYAQAARhFGAACAUYQRAABgVK3CyKxZs5SQkKCIiAglJSVp1apV1fZ94403lJKSolatWik6Olo9evTQ0qVLa71gAADQsAQcRpYsWaLRo0drwoQJWrdunXr37q20tDQVFBRU2X/lypVKSUlRTk6O8vPz1a9fPw0ePFjr1q074cUDAID6L+AwMn36dI0YMUIjR45Ut27dNGPGDMXHx2v27NlV9p8xY4buv/9+nX/++erUqZMef/xxderUSe++++4JLx4AANR/YYF0LisrU35+vsaNG+fXnpqaqjVr1tRoDK/XqwMHDqhFixbV9nG73XK73b7HJSUlkiSPxyOPxxPIkgPiCrXsGzvE8vvTLnYen7pkZy2kuqkHtajh+Lw2aoxaBA9qcXLHd1iWVeM93b17t9q1a6fVq1erZ8+evvbHH39cCxYs0NatW487xhNPPKGpU6dq8+bNat26dZV9srKyNHHixErtixYtUlRUVE2XCwAADCotLVV6err279+v6OjoavsFdGakgsPh8HtsWValtqosXrxYWVlZevvtt6sNIpI0fvx4ZWZm+h6XlJQoPj5eqampx9yZE5WYZd+Nta4QS48me/VQXojc3uMfq9r6OmuAbWPXJTtrIdVNPahFzfDaqDlqETyoRc1UXNk4noDCSMuWLRUaGqqioiK/9uLiYsXGxh5z2yVLlmjEiBF69dVXddlllx2zr8vlksvlqtTudDrldDoDWXJA3OX2Fdw3h9dh6zx2Hp+6VBe1kOytB7UIcB5eG8dFLYIHtTi54wd0A2t4eLiSkpKUm5vr156bm+t32ebPFi9erGHDhmnRokUaNGhQIFMCAIAGLuDLNJmZmbrllluUnJysHj166Nlnn1VBQYFGjRol6Y9LLLt27dLChQsl/RFEhg4dqpkzZ+qiiy7ynVWJjIxUs2bNTuKuAACA+ijgMDJkyBDt27dPkyZNUmFhoRITE5WTk6MOHTpIkgoLC/0+c2Tu3Lk6cuSI7rzzTt15552+9ltvvVXZ2dknvgcAAKBeq9UNrBkZGcrIyKjyuT8HjOXLl9dmCgAA0Ejw3TQAAMAowggAADCKMAIAAIwijAAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAKMIIAAAwijACAACMIowAAACjCCMAAMAowggAADCKMAIAAIwijAAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAKMIIAAAwijACAACMIowAAACjCCMAAMAowggAADCKMAIAAIwijAAAAKMIIwAAwCjCCAAAMIowAgAAjCKMAAAAowgjAADAKMIIAAAwijACAACMIowAAACjahVGZs2apYSEBEVERCgpKUmrVq06Zv8VK1YoKSlJEREROv300zVnzpxaLRYAADQ8AYeRJUuWaPTo0ZowYYLWrVun3r17Ky0tTQUFBVX23759uwYOHKjevXtr3bp1euCBB3TPPffo9ddfP+HFAwCA+i/gMDJ9+nSNGDFCI0eOVLdu3TRjxgzFx8dr9uzZVfafM2eO2rdvrxkzZqhbt24aOXKkhg8frieffPKEFw8AAOq/gMJIWVmZ8vPzlZqa6teempqqNWvWVLnNZ599Vqn/gAEDlJeXJ4/HE+ByAQBAQxMWSOe9e/eqvLxcsbGxfu2xsbEqKiqqcpuioqIq+x85ckR79+5V27ZtK23jdrvldrt9j/fv3y9J+uWXX2wNMGFHDto3ttdSaalXYZ4QlXsdts2zb98+28auS3bWQqqbelCLGo7Pa6PGqEXwoBY1c+DAAUmSZVnH7BdQGKngcPgfGMuyKrUdr39V7RWmTJmiiRMnVmpPSEgIdKlBJb0O5mj5jzqYpIGwux7UouZ4bQQPahE8GlItDhw4oGbNmlX7fEBhpGXLlgoNDa10FqS4uLjS2Y8Kbdq0qbJ/WFiYYmJiqtxm/PjxyszM9D32er365ZdfFBMTc8zQE8xKSkoUHx+vnTt3Kjo62vRyGj3qETyoRfCgFsGjodTCsiwdOHBAcXFxx+wXUBgJDw9XUlKScnNzdc011/jac3NzddVVV1W5TY8ePfTuu+/6tS1btkzJyclyOp1VbuNyueRyufzamjdvHshSg1Z0dHS9/sVqaKhH8KAWwYNaBI+GUItjnRGpEPC7aTIzM/Xcc89p3rx52rx5s8aMGaOCggKNGjVK0h9nNYYOHerrP2rUKP3444/KzMzU5s2bNW/ePD3//PMaO3ZsoFMDAIAGKOB7RoYMGaJ9+/Zp0qRJKiwsVGJionJyctShQwdJUmFhod9njiQkJCgnJ0djxozRM888o7i4OD311FO67rrrTt5eAACAeqtWN7BmZGQoIyOjyueys7MrtfXp00dffvllbaZqMFwulx555JFKl59gBvUIHtQieFCL4NHYauGwjvd+GwAAABvxRXkAAMAowggAADCKMAIAAIwijAAAAKMIIwAABKHG9P6SWr21F8f3008/afbs2VqzZo2KiorkcDgUGxurnj17atSoUYqPjze9RABAEHO5XNqwYYO6detmeim24629Nvj000+Vlpam+Ph4paamKjY2VpZlqbi4WLm5udq5c6c++OADXXzxxaaXCkk7d+7UI488onnz5pleSqNw6NAh5efnq0WLFjrrrLP8njt8+LBeeeUVv09xhn02b96szz//XD169FDXrl21ZcsWzZw5U263WzfffLP69+9veomNwtHfxXa0mTNn6uabb/Z9j9v06dPrcll1ijBig/PPP1+9evXSP//5zyqfHzNmjD799FOtXbu2jleGqmzYsEH/9V//pfLyctNLafC+/fZbpaamqqCgQA6HQ71799bixYvVtm1bSdKePXsUFxdHLerAhx9+qKuuukpNmjRRaWmp3nzzTQ0dOlTnnnuuLMvSihUrtHTpUgJJHQgJCdG5555b6TvYVqxYoeTkZJ1yyilyOBz6+OOPzSywDhBGbBAZGan169erS5cuVT6/ZcsWde/eXYcOHarjlTVO77zzzjGf/+GHH3TvvffyD2AduOaaa3TkyBHNnz9fv/32mzIzM/X1119r+fLlat++PWGkDvXs2VP9+/fX5MmT9fLLLysjI0N33HGHHnvsMUnShAkTtHbtWi1btszwShu+KVOm6H//93/13HPP+YU/p9OpDRs2VDqD2CBZOOkSEhKsefPmVfv8vHnzrISEhDpcUePmcDiskJAQy+FwVPsTEhJiepmNQuvWra2vvvrKry0jI8Nq3769tW3bNquoqIha1JHo6Gjru+++syzLssrLy62wsDArPz/f9/zGjRut2NhYU8trdP7zn/9YnTt3tu69916rrKzMsizLCgsLs7755hvDK6sb3MBqg7Fjx2rUqFHKz89XSkqKYmNj5XA4VFRUpNzcXD333HOaMWOG6WU2Gm3bttUzzzyjq6++usrn169fr6SkpLpdVCN16NAhhYX5/7XzzDPPKCQkRH369NGiRYsMraxxCwkJUUREhN9lgqZNm2r//v3mFtXInH/++crPz9edd96p5ORkvfjii3I4HKaXVWcIIzbIyMhQTEyM/vnPf2ru3Lm+U86hoaFKSkrSwoULdeONNxpeZeORlJSkL7/8stow4nA4GtVb6Ezq2rWr8vLyKr074F//+pcsy9KVV15paGWNT8eOHfX999/rzDPPlCR99tlnat++ve/5nTt3+u7lQd1o0qSJFixYoJdfflkpKSmN6nIl94zYzOPxaO/evZKkli1byul0Gl5R47Nq1SodPHhQl19+eZXPHzx4UHl5eerTp08dr6zxmTJlilatWqWcnJwqn8/IyNCcOXPk9XrreGWNz5w5cxQfH69BgwZV+fyECRO0Z88ePffcc3W8Mkh/fDxEfn6+LrvsMp1yyimml2M7wggAADCKT2AFAABGEUYAAIBRhBEAAGAUYQQAABhFGAEAAEYRRgDUibKyMtNLABCkCCMAbNG3b1/dddddyszMVMuWLdWpUyc5HA6tX7/e1+e3336Tw+HQ8uXLJUnLly+Xw+HQRx99pOTkZEVFRalnz57aunWrmZ0AUCcIIwBss2DBAoWFhWn16tVaunRpjbebMGGC/vGPfygvL09hYWEaPny4jasEYBofBw/ANmeeeaamTZsmSdqxY0eNt3vsscd8n4g7btw4DRo0SIcPH1ZERIQdywRgGGdGANgmOTm5Vtudc845vv+u+H6U4uLik7ImAMGHMALANkd/p0ZIyB9/3Rz9DRQej6fK7Y7+DqeKby7l+2qAhoswAqBOtGrVSpJUWFjoazv6ZlYAjRf3jACoE5GRkbrooos0depUdezYUXv37tWDDz5oelkAggBnRgDUmXnz5snj8Sg5OVl/+9vfNHnyZNNLAhAEHNbRF3ABAADqGGdGAACAUYQRAABgFGEEAAAYRRgBAABGEUYAAIBRhBEAAGAUYQQAABhFGAEAAEYRRgAAgFGEEQAAYBRhBAAAGEUYAQAARv0/bqcsjHV4w0UAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# df.set_index([\"tool\",\"file\"], inplace=True)\n",
    "\n",
    "df_filtered = df.loc[:, df.columns!='mean']\n",
    "\n",
    "for name, group in df_filtered.groupby(['tool', 'format']):\n",
    "    # print(name)\n",
    "    group.plot.bar(x=\"run\", y=\"time\", legend=True, title='-'.join(name), grid=True, subplots=False, sharex=True, sharey=True)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}