notebooks/sample_weights.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "7a00879e",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No normalization for SPS. Feature removed!\n",
      "No normalization for AvgIpc. Feature removed!\n",
      "Skipped loading some Tensorflow models, missing a dependency. No module named 'tensorflow'\n",
      "Skipped loading modules with pytorch-geometric dependency, missing a dependency. No module named 'torch_geometric'\n",
      "Skipped loading modules with pytorch-geometric dependency, missing a dependency. cannot import name 'DMPNN' from 'deepchem.models.torch_models' (/opt/conda/lib/python3.10/site-packages/deepchem/models/torch_models/__init__.py)\n",
      "Skipped loading some Jax models, missing a dependency. No module named 'jax'\n",
      "Skipped loading some PyTorch models, missing a dependency. No module named 'tensorflow'\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import sys\n",
    "import polars as pl\n",
    "import json\n",
    "from joblib import Parallel, delayed\n",
    "import deepchem\n",
    "from tqdm import tqdm\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "# Configure Polars \n",
    "cfg = pl.Config()\n",
    "cfg.set_tbl_rows(20)\n",
    "cfg.set_tbl_cols(50)\n",
    "from sklearn.model_selection import StratifiedGroupKFold, GroupKFold, KFold, GroupShuffleSplit, ShuffleSplit, StratifiedKFold\n",
    "import gc\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "dbdd8e37",
   "metadata": {},
   "outputs": [],
   "source": [
    "PERCENTILES = [0.01, 0.025, 0.05, 0.1, 0.25, 0.5, 0.75, 0.9, 0.95, 0.975, 0.99]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "7183f148",
   "metadata": {},
   "outputs": [],
   "source": [
    "def scatter_plot(x, y, show = True, xlabel = 'X', ylabel = 'Y', title = 'plot'):\n",
    "    # Create the scatter plot\n",
    "    plt.scatter(x, y)\n",
    "\n",
    "    # Add labels and title for the axes\n",
    "    plt.xlabel(xlabel)\n",
    "    plt.ylabel(ylabel)\n",
    "    plt.title(title)\n",
    "\n",
    "    # Show the plot\n",
    "    if show:\n",
    "        plt.show()\n",
    "        \n",
    "        \n",
    "def analyze_sample_weight(df):\n",
    "    assert df.shape[0] == 98415610\n",
    "    w_sum = df['sample_weight'].sum()\n",
    "    for protein in ['BRD4', 'HSA', 'sEH']:\n",
    "        print(protein, df.filter(pl.col(protein) == 1)['sample_weight'].sum() / w_sum)\n",
    "    print('Union:', df.filter((pl.col('BRD4') + pl.col('HSA') + pl.col('sEH')) != 0)['sample_weight'].sum() / w_sum)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "91df62d7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7b45f165",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e9926b84",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train_bbs --> 1145\n",
      "train_bb1s --> 271\n",
      "train_bb2s --> 693\n",
      "train_bb3s --> 872\n",
      "test_bb1s --> 341\n",
      "test_bb2s --> 1140\n",
      "test_bb3s --> 1389\n",
      "test_bbs --> 2110\n",
      "all_bbs --> 2110\n"
     ]
    }
   ],
   "source": [
    "with open('/home/dangnh36/datasets/competitions/leash_belka/processed/meta/building_blocks.json', 'r') as f:\n",
    "    bbs_meta = json.load(f)\n",
    "    \n",
    "for k, v in bbs_meta.items():\n",
    "    print(k, '-->', len(v)) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "88f00150",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(98415610, 8) 8601.91998577118\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (98_415_610, 8)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>molecule</th><th>bb1</th><th>bb2</th><th>bb3</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>scaffold_idx</th></tr><tr><td>str</td><td>u16</td><td>u16</td><td>u16</td><td>u8</td><td>u8</td><td>u8</td><td>i64</td></tr></thead><tbody><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>765</td><td>0</td><td>0</td><td>0</td><td>4283326</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>205</td><td>0</td><td>0</td><td>0</td><td>4486646</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1653</td><td>0</td><td>0</td><td>0</td><td>1015728</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>146</td><td>0</td><td>0</td><td>0</td><td>5301385</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>439</td><td>0</td><td>0</td><td>0</td><td>5301385</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>196</td><td>0</td><td>0</td><td>0</td><td>5301385</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>253</td><td>0</td><td>0</td><td>0</td><td>5301385</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1219</td><td>0</td><td>0</td><td>0</td><td>5301385</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>604</td><td>0</td><td>0</td><td>0</td><td>543172</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>121</td><td>0</td><td>0</td><td>0</td><td>2571428</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>536</td><td>0</td><td>0</td><td>0</td><td>4068980</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1367</td><td>0</td><td>0</td><td>0</td><td>2088996</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>329</td><td>0</td><td>0</td><td>0</td><td>3192765</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>757</td><td>0</td><td>0</td><td>0</td><td>4085046</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>335</td><td>0</td><td>0</td><td>0</td><td>4087757</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1792</td><td>0</td><td>0</td><td>0</td><td>4674025</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1178</td><td>0</td><td>0</td><td>0</td><td>5240137</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1699</td><td>0</td><td>0</td><td>0</td><td>5240137</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1308</td><td>0</td><td>0</td><td>0</td><td>3617024</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1415</td><td>0</td><td>0</td><td>0</td><td>268351</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (98_415_610, 8)\n",
       "┌───────────────────────────────────┬──────┬──────┬──────┬──────┬─────┬─────┬──────────────┐\n",
       "│ molecule                          ┆ bb1  ┆ bb2  ┆ bb3  ┆ BRD4 ┆ HSA ┆ sEH ┆ scaffold_idx │\n",
       "│ ---                               ┆ ---  ┆ ---  ┆ ---  ┆ ---  ┆ --- ┆ --- ┆ ---          │\n",
       "│ str                               ┆ u16  ┆ u16  ┆ u16  ┆ u8   ┆ u8  ┆ u8  ┆ i64          │\n",
       "╞═══════════════════════════════════╪══════╪══════╪══════╪══════╪═════╪═════╪══════════════╡\n",
       "│ C#CCOc1ccc(CNc2nc(NCC3CCCN3c3ccc… ┆ 1640 ┆ 1653 ┆ 765  ┆ 0    ┆ 0   ┆ 0   ┆ 4283326      │\n",
       "│ C#CCOc1ccc(CNc2nc(NCc3cccc(Br)n3… ┆ 1640 ┆ 1653 ┆ 205  ┆ 0    ┆ 0   ┆ 0   ┆ 4486646      │\n",
       "│ C#CCOc1ccc(CNc2nc(NCc3ccc(OCC#C)… ┆ 1640 ┆ 1653 ┆ 1653 ┆ 0    ┆ 0   ┆ 0   ┆ 1015728      │\n",
       "│ C#CCOc1ccc(CNc2nc(NCCNC(=O)C(=C)… ┆ 1640 ┆ 1653 ┆ 146  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      │\n",
       "│ C#CCOc1ccc(CNc2nc(NCC(=O)NCC=C)n… ┆ 1640 ┆ 1653 ┆ 439  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      │\n",
       "│ C#CCOc1ccc(CNc2nc(NCC(C)OCC=C)nc… ┆ 1640 ┆ 1653 ┆ 196  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      │\n",
       "│ C#CCOc1ccc(CNc2nc(NCCCOCC=C)nc(N… ┆ 1640 ┆ 1653 ┆ 253  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      │\n",
       "│ C#CCOc1ccc(CNc2nc(NCCOCC=C)nc(N[… ┆ 1640 ┆ 1653 ┆ 1219 ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      │\n",
       "│ C#CCOc1ccc(CNc2nc(Nc3cnn(C4CCN(C… ┆ 1640 ┆ 1653 ┆ 604  ┆ 0    ┆ 0   ┆ 0   ┆ 543172       │\n",
       "│ C#CCOc1ccc(CNc2nc(Nc3n[nH]c4c3CN… ┆ 1640 ┆ 1653 ┆ 121  ┆ 0    ┆ 0   ┆ 0   ┆ 2571428      │\n",
       "│ …                                 ┆ …    ┆ …    ┆ …    ┆ …    ┆ …   ┆ …   ┆ …            │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc(Nc2ncc… ┆ 141  ┆ 1415 ┆ 536  ┆ 0    ┆ 0   ┆ 0   ┆ 4068980      │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc(Nc2ncn… ┆ 141  ┆ 1415 ┆ 1367 ┆ 0    ┆ 0   ┆ 0   ┆ 2088996      │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc(Nc2ncn… ┆ 141  ┆ 1415 ┆ 329  ┆ 0    ┆ 0   ┆ 0   ┆ 3192765      │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc(Nc2noc… ┆ 141  ┆ 1415 ┆ 757  ┆ 0    ┆ 0   ┆ 0   ┆ 4085046      │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc(Nc2ncn… ┆ 141  ┆ 1415 ┆ 335  ┆ 0    ┆ 0   ┆ 0   ┆ 4087757      │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc(Nc2noc… ┆ 141  ┆ 1415 ┆ 1792 ┆ 0    ┆ 0   ┆ 0   ┆ 4674025      │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc(Nc2nnc… ┆ 141  ┆ 1415 ┆ 1178 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc(Nc2nnc… ┆ 141  ┆ 1415 ┆ 1699 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc(Nc2nnn… ┆ 141  ┆ 1415 ┆ 1308 ┆ 0    ┆ 0   ┆ 0   ┆ 3617024      │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc(Nc2noc… ┆ 141  ┆ 1415 ┆ 1415 ┆ 0    ┆ 0   ┆ 0   ┆ 268351       │\n",
       "└───────────────────────────────────┴──────┴──────┴──────┴──────┴─────┴─────┴──────────────┘"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_scaffolds = pl.scan_csv('/home/dangnh36/datasets/competitions/leash_belka/processed/train_scaffold.csv').collect()\n",
    "train_df = pl.scan_csv('/home/dangnh36/datasets/competitions/leash_belka/processed/train_v2.csv')\\\n",
    "    .select(\n",
    "        pl.col('molecule'),\n",
    "        pl.col('bb1', 'bb2', 'bb3').cast(pl.UInt16),\n",
    "        pl.col('BRD4', 'HSA', 'sEH').cast(pl.UInt8),\n",
    "        scaffold_idx = train_scaffolds['scaffold_idx'],\n",
    "    )\\\n",
    "    .collect()\n",
    "print(train_df.shape, train_df.estimated_size('mb'))\n",
    "train_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "1a696d3a",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df = train_df.with_columns(\n",
    "                pl.col('molecule').count().over('scaffold_idx').alias('mol_per_scaf'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c474c5ed",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "6188f6c2",
   "metadata": {},
   "source": [
    "# Sample weights by mol_per_scaffold v0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "af59ac61",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (98_415_610, 10)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>molecule</th><th>bb1</th><th>bb2</th><th>bb3</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>scaffold_idx</th><th>mol_per_scaf</th><th>sample_weight</th></tr><tr><td>str</td><td>u16</td><td>u16</td><td>u16</td><td>u8</td><td>u8</td><td>u8</td><td>i64</td><td>u32</td><td>f64</td></tr></thead><tbody><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>765</td><td>0</td><td>0</td><td>0</td><td>4283326</td><td>843</td><td>0.001186</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>205</td><td>0</td><td>0</td><td>0</td><td>4486646</td><td>9415</td><td>0.000106</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1653</td><td>0</td><td>0</td><td>0</td><td>1015728</td><td>9006</td><td>0.000111</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>146</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0.000037</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>439</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0.000037</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>196</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0.000037</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>253</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0.000037</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1219</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0.000037</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>604</td><td>0</td><td>0</td><td>0</td><td>543172</td><td>842</td><td>0.001188</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>121</td><td>0</td><td>0</td><td>0</td><td>2571428</td><td>864</td><td>0.001157</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>536</td><td>0</td><td>0</td><td>0</td><td>4068980</td><td>274</td><td>0.00365</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1367</td><td>0</td><td>0</td><td>0</td><td>2088996</td><td>20</td><td>0.05</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>329</td><td>0</td><td>0</td><td>0</td><td>3192765</td><td>260</td><td>0.003846</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>757</td><td>0</td><td>0</td><td>0</td><td>4085046</td><td>20</td><td>0.05</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>335</td><td>0</td><td>0</td><td>0</td><td>4087757</td><td>20</td><td>0.05</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1792</td><td>0</td><td>0</td><td>0</td><td>4674025</td><td>20</td><td>0.05</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1178</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td><td>0.016667</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1699</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td><td>0.016667</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1308</td><td>0</td><td>0</td><td>0</td><td>3617024</td><td>20</td><td>0.05</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1415</td><td>0</td><td>0</td><td>0</td><td>268351</td><td>20</td><td>0.05</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (98_415_610, 10)\n",
       "┌──────────────┬──────┬──────┬──────┬──────┬─────┬─────┬──────────────┬──────────────┬─────────────┐\n",
       "│ molecule     ┆ bb1  ┆ bb2  ┆ bb3  ┆ BRD4 ┆ HSA ┆ sEH ┆ scaffold_idx ┆ mol_per_scaf ┆ sample_weig │\n",
       "│ ---          ┆ ---  ┆ ---  ┆ ---  ┆ ---  ┆ --- ┆ --- ┆ ---          ┆ ---          ┆ ht          │\n",
       "│ str          ┆ u16  ┆ u16  ┆ u16  ┆ u8   ┆ u8  ┆ u8  ┆ i64          ┆ u32          ┆ ---         │\n",
       "│              ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆ f64         │\n",
       "╞══════════════╪══════╪══════╪══════╪══════╪═════╪═════╪══════════════╪══════════════╪═════════════╡\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 765  ┆ 0    ┆ 0   ┆ 0   ┆ 4283326      ┆ 843          ┆ 0.001186    │\n",
       "│ Nc2nc(NCC3CC ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ CN3c3ccc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 205  ┆ 0    ┆ 0   ┆ 0   ┆ 4486646      ┆ 9415         ┆ 0.000106    │\n",
       "│ Nc2nc(NCc3cc ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ cc(Br)n3…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 1653 ┆ 0    ┆ 0   ┆ 0   ┆ 1015728      ┆ 9006         ┆ 0.000111    │\n",
       "│ Nc2nc(NCc3cc ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(OCC#C)…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 146  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 0.000037    │\n",
       "│ Nc2nc(NCCNC( ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ =O)C(=C)…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 439  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 0.000037    │\n",
       "│ Nc2nc(NCC(=O ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ )NCC=C)n…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 196  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 0.000037    │\n",
       "│ Nc2nc(NCC(C) ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ OCC=C)nc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 253  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 0.000037    │\n",
       "│ Nc2nc(NCCCOC ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C=C)nc(N…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 1219 ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 0.000037    │\n",
       "│ Nc2nc(NCCOCC ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ =C)nc(N[…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 604  ┆ 0    ┆ 0   ┆ 0   ┆ 543172       ┆ 842          ┆ 0.001188    │\n",
       "│ Nc2nc(Nc3cnn ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ (C4CCN(C…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 121  ┆ 0    ┆ 0   ┆ 0   ┆ 2571428      ┆ 864          ┆ 0.001157    │\n",
       "│ Nc2nc(Nc3n[n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ H]c4c3CN…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ …            ┆ …    ┆ …    ┆ …    ┆ …    ┆ …   ┆ …   ┆ …            ┆ …            ┆ …           │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 536  ┆ 0    ┆ 0   ┆ 0   ┆ 4068980      ┆ 274          ┆ 0.00365     │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1367 ┆ 0    ┆ 0   ┆ 0   ┆ 2088996      ┆ 20           ┆ 0.05        │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 329  ┆ 0    ┆ 0   ┆ 0   ┆ 3192765      ┆ 260          ┆ 0.003846    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 757  ┆ 0    ┆ 0   ┆ 0   ┆ 4085046      ┆ 20           ┆ 0.05        │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2noc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 335  ┆ 0    ┆ 0   ┆ 0   ┆ 4087757      ┆ 20           ┆ 0.05        │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1792 ┆ 0    ┆ 0   ┆ 0   ┆ 4674025      ┆ 20           ┆ 0.05        │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2noc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1178 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      ┆ 60           ┆ 0.016667    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2nnc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1699 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      ┆ 60           ┆ 0.016667    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2nnc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1308 ┆ 0    ┆ 0   ┆ 0   ┆ 3617024      ┆ 20           ┆ 0.05        │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2nnn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1415 ┆ 0    ┆ 0   ┆ 0   ┆ 268351       ┆ 20           ┆ 0.05        │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2noc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "└──────────────┴──────┴──────┴──────┴──────┴─────┴─────┴──────────────┴──────────────┴─────────────┘"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = train_df.with_columns(\n",
    "    (1.0 / pl.col('mol_per_scaf')).alias('sample_weight')\n",
    ")\n",
    "train_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "f433e7d6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BRD4 0.004247076070714362\n",
      "HSA 0.004290767990588202\n",
      "sEH 0.0058066215393041505\n",
      "Union: 0.013846545726438873\n"
     ]
    }
   ],
   "source": [
    "analyze_sample_weight(train_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "baaa68b9",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df.select('sample_weight').write_csv('/home/dangnh36/datasets/competitions/leash_belka/processed/sample_weights/scaffold_v0.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b69997a4",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b4e897cd",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "006076ef",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "f988c91f",
   "metadata": {},
   "source": [
    "# Sample weights by mol_per_scaffold v2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "575c07f7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (98_415_610, 10)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>molecule</th><th>bb1</th><th>bb2</th><th>bb3</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>scaffold_idx</th><th>mol_per_scaf</th><th>sample_weight</th></tr><tr><td>str</td><td>u16</td><td>u16</td><td>u16</td><td>u8</td><td>u8</td><td>u8</td><td>i64</td><td>u32</td><td>f64</td></tr></thead><tbody><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>765</td><td>0</td><td>0</td><td>0</td><td>4283326</td><td>843</td><td>0.034442</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>205</td><td>0</td><td>0</td><td>0</td><td>4486646</td><td>9415</td><td>0.010306</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1653</td><td>0</td><td>0</td><td>0</td><td>1015728</td><td>9006</td><td>0.010537</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>146</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0.006114</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>439</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0.006114</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>196</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0.006114</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>253</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0.006114</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1219</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0.006114</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>604</td><td>0</td><td>0</td><td>0</td><td>543172</td><td>842</td><td>0.034462</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>121</td><td>0</td><td>0</td><td>0</td><td>2571428</td><td>864</td><td>0.034021</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>536</td><td>0</td><td>0</td><td>0</td><td>4068980</td><td>274</td><td>0.060412</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1367</td><td>0</td><td>0</td><td>0</td><td>2088996</td><td>20</td><td>0.223607</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>329</td><td>0</td><td>0</td><td>0</td><td>3192765</td><td>260</td><td>0.062017</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>757</td><td>0</td><td>0</td><td>0</td><td>4085046</td><td>20</td><td>0.223607</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>335</td><td>0</td><td>0</td><td>0</td><td>4087757</td><td>20</td><td>0.223607</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1792</td><td>0</td><td>0</td><td>0</td><td>4674025</td><td>20</td><td>0.223607</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1178</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td><td>0.129099</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1699</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td><td>0.129099</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1308</td><td>0</td><td>0</td><td>0</td><td>3617024</td><td>20</td><td>0.223607</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1415</td><td>0</td><td>0</td><td>0</td><td>268351</td><td>20</td><td>0.223607</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (98_415_610, 10)\n",
       "┌──────────────┬──────┬──────┬──────┬──────┬─────┬─────┬──────────────┬──────────────┬─────────────┐\n",
       "│ molecule     ┆ bb1  ┆ bb2  ┆ bb3  ┆ BRD4 ┆ HSA ┆ sEH ┆ scaffold_idx ┆ mol_per_scaf ┆ sample_weig │\n",
       "│ ---          ┆ ---  ┆ ---  ┆ ---  ┆ ---  ┆ --- ┆ --- ┆ ---          ┆ ---          ┆ ht          │\n",
       "│ str          ┆ u16  ┆ u16  ┆ u16  ┆ u8   ┆ u8  ┆ u8  ┆ i64          ┆ u32          ┆ ---         │\n",
       "│              ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆ f64         │\n",
       "╞══════════════╪══════╪══════╪══════╪══════╪═════╪═════╪══════════════╪══════════════╪═════════════╡\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 765  ┆ 0    ┆ 0   ┆ 0   ┆ 4283326      ┆ 843          ┆ 0.034442    │\n",
       "│ Nc2nc(NCC3CC ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ CN3c3ccc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 205  ┆ 0    ┆ 0   ┆ 0   ┆ 4486646      ┆ 9415         ┆ 0.010306    │\n",
       "│ Nc2nc(NCc3cc ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ cc(Br)n3…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 1653 ┆ 0    ┆ 0   ┆ 0   ┆ 1015728      ┆ 9006         ┆ 0.010537    │\n",
       "│ Nc2nc(NCc3cc ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(OCC#C)…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 146  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 0.006114    │\n",
       "│ Nc2nc(NCCNC( ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ =O)C(=C)…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 439  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 0.006114    │\n",
       "│ Nc2nc(NCC(=O ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ )NCC=C)n…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 196  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 0.006114    │\n",
       "│ Nc2nc(NCC(C) ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ OCC=C)nc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 253  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 0.006114    │\n",
       "│ Nc2nc(NCCCOC ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C=C)nc(N…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 1219 ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 0.006114    │\n",
       "│ Nc2nc(NCCOCC ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ =C)nc(N[…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 604  ┆ 0    ┆ 0   ┆ 0   ┆ 543172       ┆ 842          ┆ 0.034462    │\n",
       "│ Nc2nc(Nc3cnn ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ (C4CCN(C…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 121  ┆ 0    ┆ 0   ┆ 0   ┆ 2571428      ┆ 864          ┆ 0.034021    │\n",
       "│ Nc2nc(Nc3n[n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ H]c4c3CN…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ …            ┆ …    ┆ …    ┆ …    ┆ …    ┆ …   ┆ …   ┆ …            ┆ …            ┆ …           │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 536  ┆ 0    ┆ 0   ┆ 0   ┆ 4068980      ┆ 274          ┆ 0.060412    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1367 ┆ 0    ┆ 0   ┆ 0   ┆ 2088996      ┆ 20           ┆ 0.223607    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 329  ┆ 0    ┆ 0   ┆ 0   ┆ 3192765      ┆ 260          ┆ 0.062017    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 757  ┆ 0    ┆ 0   ┆ 0   ┆ 4085046      ┆ 20           ┆ 0.223607    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2noc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 335  ┆ 0    ┆ 0   ┆ 0   ┆ 4087757      ┆ 20           ┆ 0.223607    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1792 ┆ 0    ┆ 0   ┆ 0   ┆ 4674025      ┆ 20           ┆ 0.223607    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2noc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1178 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      ┆ 60           ┆ 0.129099    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2nnc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1699 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      ┆ 60           ┆ 0.129099    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2nnc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1308 ┆ 0    ┆ 0   ┆ 0   ┆ 3617024      ┆ 20           ┆ 0.223607    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2nnn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1415 ┆ 0    ┆ 0   ┆ 0   ┆ 268351       ┆ 20           ┆ 0.223607    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2noc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "└──────────────┴──────┴──────┴──────┴──────┴─────┴─────┴──────────────┴──────────────┴─────────────┘"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = train_df.with_columns(\n",
    "    (1.0 / (pl.col('mol_per_scaf') ** 0.5) ).alias('sample_weight')\n",
    ")\n",
    "train_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "15a7f5c8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BRD4 0.004435922243997791\n",
      "HSA 0.004351191914480281\n",
      "sEH 0.007742457807451675\n",
      "Union: 0.015901166740355292\n"
     ]
    }
   ],
   "source": [
    "analyze_sample_weight(train_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "ccfdd131",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df.select('sample_weight').write_csv('/home/dangnh36/datasets/competitions/leash_belka/processed/sample_weights/scaffold_v2.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "17e0ab7c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "42db610a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0b129abe",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d0fea60d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "9d579695",
   "metadata": {},
   "source": [
    "# Sample weights by mol_per_scaffold v3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "06181d9e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (98_415_610, 10)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>molecule</th><th>bb1</th><th>bb2</th><th>bb3</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>scaffold_idx</th><th>mol_per_scaf</th><th>sample_weight</th></tr><tr><td>str</td><td>u16</td><td>u16</td><td>u16</td><td>u8</td><td>u8</td><td>u8</td><td>i64</td><td>u32</td><td>f64</td></tr></thead><tbody><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>765</td><td>0</td><td>0</td><td>0</td><td>4283326</td><td>843</td><td>0.000001</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>205</td><td>0</td><td>0</td><td>0</td><td>4486646</td><td>9415</td><td>1.1281e-8</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1653</td><td>0</td><td>0</td><td>0</td><td>1015728</td><td>9006</td><td>1.2329e-8</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>146</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>1.3974e-9</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>439</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>1.3974e-9</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>196</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>1.3974e-9</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>253</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>1.3974e-9</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1219</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>1.3974e-9</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>604</td><td>0</td><td>0</td><td>0</td><td>543172</td><td>842</td><td>0.000001</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>121</td><td>0</td><td>0</td><td>0</td><td>2571428</td><td>864</td><td>0.000001</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>536</td><td>0</td><td>0</td><td>0</td><td>4068980</td><td>274</td><td>0.000013</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1367</td><td>0</td><td>0</td><td>0</td><td>2088996</td><td>20</td><td>0.0025</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>329</td><td>0</td><td>0</td><td>0</td><td>3192765</td><td>260</td><td>0.000015</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>757</td><td>0</td><td>0</td><td>0</td><td>4085046</td><td>20</td><td>0.0025</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>335</td><td>0</td><td>0</td><td>0</td><td>4087757</td><td>20</td><td>0.0025</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1792</td><td>0</td><td>0</td><td>0</td><td>4674025</td><td>20</td><td>0.0025</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1178</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td><td>0.000278</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1699</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td><td>0.000278</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1308</td><td>0</td><td>0</td><td>0</td><td>3617024</td><td>20</td><td>0.0025</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1415</td><td>0</td><td>0</td><td>0</td><td>268351</td><td>20</td><td>0.0025</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (98_415_610, 10)\n",
       "┌──────────────┬──────┬──────┬──────┬──────┬─────┬─────┬──────────────┬──────────────┬─────────────┐\n",
       "│ molecule     ┆ bb1  ┆ bb2  ┆ bb3  ┆ BRD4 ┆ HSA ┆ sEH ┆ scaffold_idx ┆ mol_per_scaf ┆ sample_weig │\n",
       "│ ---          ┆ ---  ┆ ---  ┆ ---  ┆ ---  ┆ --- ┆ --- ┆ ---          ┆ ---          ┆ ht          │\n",
       "│ str          ┆ u16  ┆ u16  ┆ u16  ┆ u8   ┆ u8  ┆ u8  ┆ i64          ┆ u32          ┆ ---         │\n",
       "│              ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆ f64         │\n",
       "╞══════════════╪══════╪══════╪══════╪══════╪═════╪═════╪══════════════╪══════════════╪═════════════╡\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 765  ┆ 0    ┆ 0   ┆ 0   ┆ 4283326      ┆ 843          ┆ 0.000001    │\n",
       "│ Nc2nc(NCC3CC ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ CN3c3ccc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 205  ┆ 0    ┆ 0   ┆ 0   ┆ 4486646      ┆ 9415         ┆ 1.1281e-8   │\n",
       "│ Nc2nc(NCc3cc ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ cc(Br)n3…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 1653 ┆ 0    ┆ 0   ┆ 0   ┆ 1015728      ┆ 9006         ┆ 1.2329e-8   │\n",
       "│ Nc2nc(NCc3cc ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(OCC#C)…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 146  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 1.3974e-9   │\n",
       "│ Nc2nc(NCCNC( ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ =O)C(=C)…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 439  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 1.3974e-9   │\n",
       "│ Nc2nc(NCC(=O ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ )NCC=C)n…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 196  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 1.3974e-9   │\n",
       "│ Nc2nc(NCC(C) ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ OCC=C)nc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 253  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 1.3974e-9   │\n",
       "│ Nc2nc(NCCCOC ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C=C)nc(N…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 1219 ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        ┆ 1.3974e-9   │\n",
       "│ Nc2nc(NCCOCC ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ =C)nc(N[…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 604  ┆ 0    ┆ 0   ┆ 0   ┆ 543172       ┆ 842          ┆ 0.000001    │\n",
       "│ Nc2nc(Nc3cnn ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ (C4CCN(C…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ C#CCOc1ccc(C ┆ 1640 ┆ 1653 ┆ 121  ┆ 0    ┆ 0   ┆ 0   ┆ 2571428      ┆ 864          ┆ 0.000001    │\n",
       "│ Nc2nc(Nc3n[n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ H]c4c3CN…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ …            ┆ …    ┆ …    ┆ …    ┆ …    ┆ …   ┆ …   ┆ …            ┆ …            ┆ …           │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 536  ┆ 0    ┆ 0   ┆ 0   ┆ 4068980      ┆ 274          ┆ 0.000013    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1367 ┆ 0    ┆ 0   ┆ 0   ┆ 2088996      ┆ 20           ┆ 0.0025      │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 329  ┆ 0    ┆ 0   ┆ 0   ┆ 3192765      ┆ 260          ┆ 0.000015    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 757  ┆ 0    ┆ 0   ┆ 0   ┆ 4085046      ┆ 20           ┆ 0.0025      │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2noc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 335  ┆ 0    ┆ 0   ┆ 0   ┆ 4087757      ┆ 20           ┆ 0.0025      │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2ncn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1792 ┆ 0    ┆ 0   ┆ 0   ┆ 4674025      ┆ 20           ┆ 0.0025      │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2noc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1178 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      ┆ 60           ┆ 0.000278    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2nnc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1699 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      ┆ 60           ┆ 0.000278    │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2nnc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1308 ┆ 0    ┆ 0   ┆ 0   ┆ 3617024      ┆ 20           ┆ 0.0025      │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2nnn…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ [N-]=[N+]=NC ┆ 141  ┆ 1415 ┆ 1415 ┆ 0    ┆ 0   ┆ 0   ┆ 268351       ┆ 20           ┆ 0.0025      │\n",
       "│ CC[C@H](Nc1n ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "│ c(Nc2noc…    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              ┆             │\n",
       "└──────────────┴──────┴──────┴──────┴──────┴─────┴─────┴──────────────┴──────────────┴─────────────┘"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = train_df.with_columns(\n",
    "    (1.0 / (pl.col('mol_per_scaf') ** 2) ).alias('sample_weight')\n",
    ")\n",
    "train_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "43000118",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BRD4 0.004212658128955828\n",
      "HSA 0.004414961031634905\n",
      "sEH 0.0032135961826731266\n",
      "Union: 0.011383674971416971\n"
     ]
    }
   ],
   "source": [
    "analyze_sample_weight(train_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "62708889",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df.select('sample_weight').write_csv('/home/dangnh36/datasets/competitions/leash_belka/processed/sample_weights/scaffold_v3.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "86199429",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "08055266",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "00aecd7f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a2d128fe",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9474d675",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f9146260",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "36729ae3",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "18529610",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9ac82df5",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e7489671",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0e0b8ceb",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "24629f0a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ecdcae6e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "16ae3ad2",
   "metadata": {},
   "source": [
    "# Sample weights by mol_per_scaffold v1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "c27b9982",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (98_415_610, 9)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>molecule</th><th>bb1</th><th>bb2</th><th>bb3</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>scaffold_idx</th><th>mol_per_scaf</th></tr><tr><td>str</td><td>u16</td><td>u16</td><td>u16</td><td>u8</td><td>u8</td><td>u8</td><td>i64</td><td>u32</td></tr></thead><tbody><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>765</td><td>0</td><td>0</td><td>0</td><td>4283326</td><td>843</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>205</td><td>0</td><td>0</td><td>0</td><td>4486646</td><td>9415</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1653</td><td>0</td><td>0</td><td>0</td><td>1015728</td><td>9006</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>146</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>439</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>196</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>253</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1219</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>604</td><td>0</td><td>0</td><td>0</td><td>543172</td><td>842</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>121</td><td>0</td><td>0</td><td>0</td><td>2571428</td><td>864</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>536</td><td>0</td><td>0</td><td>0</td><td>4068980</td><td>274</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1367</td><td>0</td><td>0</td><td>0</td><td>2088996</td><td>20</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>329</td><td>0</td><td>0</td><td>0</td><td>3192765</td><td>260</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>757</td><td>0</td><td>0</td><td>0</td><td>4085046</td><td>20</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>335</td><td>0</td><td>0</td><td>0</td><td>4087757</td><td>20</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1792</td><td>0</td><td>0</td><td>0</td><td>4674025</td><td>20</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1178</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1699</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1308</td><td>0</td><td>0</td><td>0</td><td>3617024</td><td>20</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1415</td><td>0</td><td>0</td><td>0</td><td>268351</td><td>20</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (98_415_610, 9)\n",
       "┌────────────────────────────┬──────┬──────┬──────┬──────┬─────┬─────┬──────────────┬──────────────┐\n",
       "│ molecule                   ┆ bb1  ┆ bb2  ┆ bb3  ┆ BRD4 ┆ HSA ┆ sEH ┆ scaffold_idx ┆ mol_per_scaf │\n",
       "│ ---                        ┆ ---  ┆ ---  ┆ ---  ┆ ---  ┆ --- ┆ --- ┆ ---          ┆ ---          │\n",
       "│ str                        ┆ u16  ┆ u16  ┆ u16  ┆ u8   ┆ u8  ┆ u8  ┆ i64          ┆ u32          │\n",
       "╞════════════════════════════╪══════╪══════╪══════╪══════╪═════╪═════╪══════════════╪══════════════╡\n",
       "│ C#CCOc1ccc(CNc2nc(NCC3CCCN ┆ 1640 ┆ 1653 ┆ 765  ┆ 0    ┆ 0   ┆ 0   ┆ 4283326      ┆ 843          │\n",
       "│ 3c3ccc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCc3cccc ┆ 1640 ┆ 1653 ┆ 205  ┆ 0    ┆ 0   ┆ 0   ┆ 4486646      ┆ 9415         │\n",
       "│ (Br)n3…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCc3ccc( ┆ 1640 ┆ 1653 ┆ 1653 ┆ 0    ┆ 0   ┆ 0   ┆ 1015728      ┆ 9006         │\n",
       "│ OCC#C)…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCCNC(=O ┆ 1640 ┆ 1653 ┆ 146  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        │\n",
       "│ )C(=C)…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCC(=O)N ┆ 1640 ┆ 1653 ┆ 439  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        │\n",
       "│ CC=C)n…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCC(C)OC ┆ 1640 ┆ 1653 ┆ 196  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        │\n",
       "│ C=C)nc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCCCOCC= ┆ 1640 ┆ 1653 ┆ 253  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        │\n",
       "│ C)nc(N…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCCOCC=C ┆ 1640 ┆ 1653 ┆ 1219 ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        │\n",
       "│ )nc(N[…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(Nc3cnn(C ┆ 1640 ┆ 1653 ┆ 604  ┆ 0    ┆ 0   ┆ 0   ┆ 543172       ┆ 842          │\n",
       "│ 4CCN(C…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(Nc3n[nH] ┆ 1640 ┆ 1653 ┆ 121  ┆ 0    ┆ 0   ┆ 0   ┆ 2571428      ┆ 864          │\n",
       "│ c4c3CN…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ …                          ┆ …    ┆ …    ┆ …    ┆ …    ┆ …   ┆ …   ┆ …            ┆ …            │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 536  ┆ 0    ┆ 0   ┆ 0   ┆ 4068980      ┆ 274          │\n",
       "│ Nc2ncc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1367 ┆ 0    ┆ 0   ┆ 0   ┆ 2088996      ┆ 20           │\n",
       "│ Nc2ncn…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 329  ┆ 0    ┆ 0   ┆ 0   ┆ 3192765      ┆ 260          │\n",
       "│ Nc2ncn…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 757  ┆ 0    ┆ 0   ┆ 0   ┆ 4085046      ┆ 20           │\n",
       "│ Nc2noc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 335  ┆ 0    ┆ 0   ┆ 0   ┆ 4087757      ┆ 20           │\n",
       "│ Nc2ncn…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1792 ┆ 0    ┆ 0   ┆ 0   ┆ 4674025      ┆ 20           │\n",
       "│ Nc2noc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1178 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      ┆ 60           │\n",
       "│ Nc2nnc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1699 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      ┆ 60           │\n",
       "│ Nc2nnc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1308 ┆ 0    ┆ 0   ┆ 0   ┆ 3617024      ┆ 20           │\n",
       "│ Nc2nnn…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1415 ┆ 0    ┆ 0   ┆ 0   ┆ 268351       ┆ 20           │\n",
       "│ Nc2noc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "└────────────────────────────┴──────┴──────┴──────┴──────┴─────┴─────┴──────────────┴──────────────┘"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = train_df.with_columns(\n",
    "                pl.col('molecule').count().over('scaffold_idx').alias('mol_per_scaf'))\n",
    "train_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "7ecc1dc8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (1, 1)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>scaffold_idx</th></tr><tr><td>u32</td></tr></thead><tbody><tr><td>5711873</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (1, 1)\n",
       "┌──────────────┐\n",
       "│ scaffold_idx │\n",
       "│ ---          │\n",
       "│ u32          │\n",
       "╞══════════════╡\n",
       "│ 5711873      │\n",
       "└──────────────┘"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.select(pl.n_unique('scaffold_idx'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "20efc4a3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (5_711_873, 6)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>scaffold_idx</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>mol_per_scaf</th><th>count</th></tr><tr><td>i64</td><td>f64</td><td>f64</td><td>f64</td><td>u32</td><td>u32</td></tr></thead><tbody><tr><td>3877740</td><td>0.0</td><td>0.0</td><td>0.0</td><td>3</td><td>3</td></tr><tr><td>440695</td><td>0.0</td><td>0.0</td><td>0.0</td><td>2</td><td>2</td></tr><tr><td>4322493</td><td>0.0</td><td>0.0</td><td>0.0</td><td>6</td><td>6</td></tr><tr><td>4137279</td><td>0.0</td><td>0.0</td><td>0.0</td><td>22</td><td>22</td></tr><tr><td>3226929</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>998698</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>5288161</td><td>0.0</td><td>0.0</td><td>0.0</td><td>3</td><td>3</td></tr><tr><td>1175132</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>4614014</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>3376228</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>3367966</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>2633392</td><td>0.0</td><td>0.0</td><td>0.0</td><td>9</td><td>9</td></tr><tr><td>750166</td><td>0.0</td><td>0.0</td><td>0.0</td><td>4</td><td>4</td></tr><tr><td>5833314</td><td>0.0</td><td>0.0</td><td>0.0</td><td>13</td><td>13</td></tr><tr><td>2172514</td><td>0.0</td><td>0.0</td><td>0.0</td><td>141</td><td>141</td></tr><tr><td>3285123</td><td>0.0</td><td>0.0</td><td>0.0</td><td>6</td><td>6</td></tr><tr><td>5502210</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>2287735</td><td>0.0</td><td>0.0</td><td>0.0</td><td>2</td><td>2</td></tr><tr><td>2502180</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>3535090</td><td>0.0</td><td>0.0</td><td>0.0</td><td>5</td><td>5</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (5_711_873, 6)\n",
       "┌──────────────┬──────┬─────┬─────┬──────────────┬───────┐\n",
       "│ scaffold_idx ┆ BRD4 ┆ HSA ┆ sEH ┆ mol_per_scaf ┆ count │\n",
       "│ ---          ┆ ---  ┆ --- ┆ --- ┆ ---          ┆ ---   │\n",
       "│ i64          ┆ f64  ┆ f64 ┆ f64 ┆ u32          ┆ u32   │\n",
       "╞══════════════╪══════╪═════╪═════╪══════════════╪═══════╡\n",
       "│ 3877740      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 3            ┆ 3     │\n",
       "│ 440695       ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 2            ┆ 2     │\n",
       "│ 4322493      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 6            ┆ 6     │\n",
       "│ 4137279      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 22           ┆ 22    │\n",
       "│ 3226929      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 1            ┆ 1     │\n",
       "│ 998698       ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 1            ┆ 1     │\n",
       "│ 5288161      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 3            ┆ 3     │\n",
       "│ 1175132      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 1            ┆ 1     │\n",
       "│ 4614014      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 1            ┆ 1     │\n",
       "│ 3376228      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 1            ┆ 1     │\n",
       "│ …            ┆ …    ┆ …   ┆ …   ┆ …            ┆ …     │\n",
       "│ 3367966      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 1            ┆ 1     │\n",
       "│ 2633392      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 9            ┆ 9     │\n",
       "│ 750166       ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 4            ┆ 4     │\n",
       "│ 5833314      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 13           ┆ 13    │\n",
       "│ 2172514      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 141          ┆ 141   │\n",
       "│ 3285123      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 6            ┆ 6     │\n",
       "│ 5502210      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 1            ┆ 1     │\n",
       "│ 2287735      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 2            ┆ 2     │\n",
       "│ 2502180      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 1            ┆ 1     │\n",
       "│ 3535090      ┆ 0.0  ┆ 0.0 ┆ 0.0 ┆ 5            ┆ 5     │\n",
       "└──────────────┴──────┴─────┴─────┴──────────────┴───────┘"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp_df = train_df.group_by('scaffold_idx').agg(\n",
    "    pl.col('BRD4').mean(),\n",
    "    pl.col('HSA').mean(),\n",
    "    pl.col('sEH').mean(),\n",
    "    pl.col('mol_per_scaf').first(),\n",
    "    pl.col('molecule').count().alias('count')\n",
    ")\n",
    "tmp_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "fcc1eb77",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (301_048, 6)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>scaffold_idx</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>mol_per_scaf</th><th>count</th></tr><tr><td>i64</td><td>f64</td><td>f64</td><td>f64</td><td>u32</td><td>u32</td></tr></thead><tbody><tr><td>3913524</td><td>0.0</td><td>0.005236</td><td>0.0</td><td>191</td><td>191</td></tr><tr><td>923462</td><td>0.0</td><td>0.05</td><td>0.0</td><td>20</td><td>20</td></tr><tr><td>3414635</td><td>0.027235</td><td>0.005777</td><td>0.004677</td><td>3635</td><td>3635</td></tr><tr><td>2380284</td><td>0.0</td><td>0.0</td><td>0.276596</td><td>47</td><td>47</td></tr><tr><td>3492101</td><td>0.0</td><td>0.0</td><td>0.021277</td><td>47</td><td>47</td></tr><tr><td>1951915</td><td>0.0</td><td>0.111111</td><td>0.0</td><td>18</td><td>18</td></tr><tr><td>3058620</td><td>0.0</td><td>0.0</td><td>0.076923</td><td>13</td><td>13</td></tr><tr><td>2333618</td><td>0.0</td><td>0.5</td><td>0.0</td><td>2</td><td>2</td></tr><tr><td>1778887</td><td>0.005556</td><td>0.011111</td><td>0.0</td><td>180</td><td>180</td></tr><tr><td>4656485</td><td>0.0</td><td>0.142857</td><td>0.0</td><td>7</td><td>7</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>276840</td><td>0.0</td><td>0.0</td><td>0.3</td><td>40</td><td>40</td></tr><tr><td>5501028</td><td>0.0</td><td>0.031915</td><td>0.0</td><td>94</td><td>94</td></tr><tr><td>720149</td><td>0.0</td><td>0.0</td><td>0.4</td><td>15</td><td>15</td></tr><tr><td>4131113</td><td>0.0</td><td>0.0</td><td>0.24</td><td>25</td><td>25</td></tr><tr><td>2462335</td><td>0.0</td><td>0.0</td><td>0.3</td><td>20</td><td>20</td></tr><tr><td>607685</td><td>0.0</td><td>0.5</td><td>0.0</td><td>2</td><td>2</td></tr><tr><td>581208</td><td>0.25</td><td>0.0</td><td>0.0</td><td>4</td><td>4</td></tr><tr><td>726443</td><td>1.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>3406897</td><td>0.45</td><td>0.0</td><td>0.0</td><td>20</td><td>20</td></tr><tr><td>540964</td><td>0.010417</td><td>0.0</td><td>0.0</td><td>96</td><td>96</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (301_048, 6)\n",
       "┌──────────────┬──────────┬──────────┬──────────┬──────────────┬───────┐\n",
       "│ scaffold_idx ┆ BRD4     ┆ HSA      ┆ sEH      ┆ mol_per_scaf ┆ count │\n",
       "│ ---          ┆ ---      ┆ ---      ┆ ---      ┆ ---          ┆ ---   │\n",
       "│ i64          ┆ f64      ┆ f64      ┆ f64      ┆ u32          ┆ u32   │\n",
       "╞══════════════╪══════════╪══════════╪══════════╪══════════════╪═══════╡\n",
       "│ 3913524      ┆ 0.0      ┆ 0.005236 ┆ 0.0      ┆ 191          ┆ 191   │\n",
       "│ 923462       ┆ 0.0      ┆ 0.05     ┆ 0.0      ┆ 20           ┆ 20    │\n",
       "│ 3414635      ┆ 0.027235 ┆ 0.005777 ┆ 0.004677 ┆ 3635         ┆ 3635  │\n",
       "│ 2380284      ┆ 0.0      ┆ 0.0      ┆ 0.276596 ┆ 47           ┆ 47    │\n",
       "│ 3492101      ┆ 0.0      ┆ 0.0      ┆ 0.021277 ┆ 47           ┆ 47    │\n",
       "│ 1951915      ┆ 0.0      ┆ 0.111111 ┆ 0.0      ┆ 18           ┆ 18    │\n",
       "│ 3058620      ┆ 0.0      ┆ 0.0      ┆ 0.076923 ┆ 13           ┆ 13    │\n",
       "│ 2333618      ┆ 0.0      ┆ 0.5      ┆ 0.0      ┆ 2            ┆ 2     │\n",
       "│ 1778887      ┆ 0.005556 ┆ 0.011111 ┆ 0.0      ┆ 180          ┆ 180   │\n",
       "│ 4656485      ┆ 0.0      ┆ 0.142857 ┆ 0.0      ┆ 7            ┆ 7     │\n",
       "│ …            ┆ …        ┆ …        ┆ …        ┆ …            ┆ …     │\n",
       "│ 276840       ┆ 0.0      ┆ 0.0      ┆ 0.3      ┆ 40           ┆ 40    │\n",
       "│ 5501028      ┆ 0.0      ┆ 0.031915 ┆ 0.0      ┆ 94           ┆ 94    │\n",
       "│ 720149       ┆ 0.0      ┆ 0.0      ┆ 0.4      ┆ 15           ┆ 15    │\n",
       "│ 4131113      ┆ 0.0      ┆ 0.0      ┆ 0.24     ┆ 25           ┆ 25    │\n",
       "│ 2462335      ┆ 0.0      ┆ 0.0      ┆ 0.3      ┆ 20           ┆ 20    │\n",
       "│ 607685       ┆ 0.0      ┆ 0.5      ┆ 0.0      ┆ 2            ┆ 2     │\n",
       "│ 581208       ┆ 0.25     ┆ 0.0      ┆ 0.0      ┆ 4            ┆ 4     │\n",
       "│ 726443       ┆ 1.0      ┆ 0.0      ┆ 0.0      ┆ 1            ┆ 1     │\n",
       "│ 3406897      ┆ 0.45     ┆ 0.0      ┆ 0.0      ┆ 20           ┆ 20    │\n",
       "│ 540964       ┆ 0.010417 ┆ 0.0      ┆ 0.0      ┆ 96           ┆ 96    │\n",
       "└──────────────┴──────────┴──────────┴──────────┴──────────────┴───────┘"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp_df = tmp_df.filter((pl.col('BRD4') > 0) | (pl.col('HSA') > 0) | (pl.col('sEH') > 0))\n",
    "tmp_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "da18c44f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (17, 7)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>statistic</th><th>scaffold_idx</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>mol_per_scaf</th><th>count</th></tr><tr><td>str</td><td>f64</td><td>f64</td><td>f64</td><td>f64</td><td>f64</td><td>f64</td></tr></thead><tbody><tr><td>&quot;count&quot;</td><td>301048.0</td><td>301048.0</td><td>301048.0</td><td>301048.0</td><td>301048.0</td><td>301048.0</td></tr><tr><td>&quot;null_count&quot;</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td></tr><tr><td>&quot;mean&quot;</td><td>2.9854e6</td><td>0.080581</td><td>0.08141</td><td>0.110171</td><td>185.088335</td><td>185.088335</td></tr><tr><td>&quot;std&quot;</td><td>1.7256e6</td><td>0.221663</td><td>0.214093</td><td>0.200038</td><td>2232.683301</td><td>2232.683301</td></tr><tr><td>&quot;min&quot;</td><td>10.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1.0</td></tr><tr><td>&quot;1%&quot;</td><td>59729.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1.0</td></tr><tr><td>&quot;2.5%&quot;</td><td>149070.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1.0</td></tr><tr><td>&quot;5%&quot;</td><td>297182.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1.0</td></tr><tr><td>&quot;10%&quot;</td><td>594367.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>2.0</td><td>2.0</td></tr><tr><td>&quot;25%&quot;</td><td>1.490325e6</td><td>0.0</td><td>0.0</td><td>0.0</td><td>5.0</td><td>5.0</td></tr><tr><td>&quot;50%&quot;</td><td>2.981351e6</td><td>0.0</td><td>0.0</td><td>0.0</td><td>13.0</td><td>13.0</td></tr><tr><td>&quot;75%&quot;</td><td>4.481635e6</td><td>0.013699</td><td>0.033333</td><td>0.2</td><td>75.0</td><td>75.0</td></tr><tr><td>&quot;90%&quot;</td><td>5.3747e6</td><td>0.25</td><td>0.25</td><td>0.4</td><td>235.0</td><td>235.0</td></tr><tr><td>&quot;95%&quot;</td><td>5.672929e6</td><td>0.5</td><td>0.5</td><td>0.4</td><td>527.0</td><td>527.0</td></tr><tr><td>&quot;97.5%&quot;</td><td>5.823608e6</td><td>1.0</td><td>1.0</td><td>0.6</td><td>1069.0</td><td>1069.0</td></tr><tr><td>&quot;99%&quot;</td><td>5.913977e6</td><td>1.0</td><td>1.0</td><td>1.0</td><td>3331.0</td><td>3331.0</td></tr><tr><td>&quot;max&quot;</td><td>5.971663e6</td><td>1.0</td><td>1.0</td><td>1.0</td><td>499089.0</td><td>499089.0</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (17, 7)\n",
       "┌────────────┬──────────────┬──────────┬──────────┬──────────┬──────────────┬─────────────┐\n",
       "│ statistic  ┆ scaffold_idx ┆ BRD4     ┆ HSA      ┆ sEH      ┆ mol_per_scaf ┆ count       │\n",
       "│ ---        ┆ ---          ┆ ---      ┆ ---      ┆ ---      ┆ ---          ┆ ---         │\n",
       "│ str        ┆ f64          ┆ f64      ┆ f64      ┆ f64      ┆ f64          ┆ f64         │\n",
       "╞════════════╪══════════════╪══════════╪══════════╪══════════╪══════════════╪═════════════╡\n",
       "│ count      ┆ 301048.0     ┆ 301048.0 ┆ 301048.0 ┆ 301048.0 ┆ 301048.0     ┆ 301048.0    │\n",
       "│ null_count ┆ 0.0          ┆ 0.0      ┆ 0.0      ┆ 0.0      ┆ 0.0          ┆ 0.0         │\n",
       "│ mean       ┆ 2.9854e6     ┆ 0.080581 ┆ 0.08141  ┆ 0.110171 ┆ 185.088335   ┆ 185.088335  │\n",
       "│ std        ┆ 1.7256e6     ┆ 0.221663 ┆ 0.214093 ┆ 0.200038 ┆ 2232.683301  ┆ 2232.683301 │\n",
       "│ min        ┆ 10.0         ┆ 0.0      ┆ 0.0      ┆ 0.0      ┆ 1.0          ┆ 1.0         │\n",
       "│ 1%         ┆ 59729.0      ┆ 0.0      ┆ 0.0      ┆ 0.0      ┆ 1.0          ┆ 1.0         │\n",
       "│ 2.5%       ┆ 149070.0     ┆ 0.0      ┆ 0.0      ┆ 0.0      ┆ 1.0          ┆ 1.0         │\n",
       "│ 5%         ┆ 297182.0     ┆ 0.0      ┆ 0.0      ┆ 0.0      ┆ 1.0          ┆ 1.0         │\n",
       "│ 10%        ┆ 594367.0     ┆ 0.0      ┆ 0.0      ┆ 0.0      ┆ 2.0          ┆ 2.0         │\n",
       "│ 25%        ┆ 1.490325e6   ┆ 0.0      ┆ 0.0      ┆ 0.0      ┆ 5.0          ┆ 5.0         │\n",
       "│ 50%        ┆ 2.981351e6   ┆ 0.0      ┆ 0.0      ┆ 0.0      ┆ 13.0         ┆ 13.0        │\n",
       "│ 75%        ┆ 4.481635e6   ┆ 0.013699 ┆ 0.033333 ┆ 0.2      ┆ 75.0         ┆ 75.0        │\n",
       "│ 90%        ┆ 5.3747e6     ┆ 0.25     ┆ 0.25     ┆ 0.4      ┆ 235.0        ┆ 235.0       │\n",
       "│ 95%        ┆ 5.672929e6   ┆ 0.5      ┆ 0.5      ┆ 0.4      ┆ 527.0        ┆ 527.0       │\n",
       "│ 97.5%      ┆ 5.823608e6   ┆ 1.0      ┆ 1.0      ┆ 0.6      ┆ 1069.0       ┆ 1069.0      │\n",
       "│ 99%        ┆ 5.913977e6   ┆ 1.0      ┆ 1.0      ┆ 1.0      ┆ 3331.0       ┆ 3331.0      │\n",
       "│ max        ┆ 5.971663e6   ┆ 1.0      ┆ 1.0      ┆ 1.0      ┆ 499089.0     ┆ 499089.0    │\n",
       "└────────────┴──────────────┴──────────┴──────────┴──────────┴──────────────┴─────────────┘"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp_df.describe(percentiles=PERCENTILES)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "8cd95600",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (17, 10)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>statistic</th><th>molecule</th><th>bb1</th><th>bb2</th><th>bb3</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>scaffold_idx</th><th>mol_per_scaf</th></tr><tr><td>str</td><td>str</td><td>f64</td><td>f64</td><td>f64</td><td>f64</td><td>f64</td><td>f64</td><td>f64</td><td>f64</td></tr></thead><tbody><tr><td>&quot;count&quot;</td><td>&quot;98415610&quot;</td><td>9.841561e7</td><td>9.841561e7</td><td>9.841561e7</td><td>9.841561e7</td><td>9.841561e7</td><td>9.841561e7</td><td>9.841561e7</td><td>9.841561e7</td></tr><tr><td>&quot;null_count&quot;</td><td>&quot;0&quot;</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td></tr><tr><td>&quot;mean&quot;</td><td>null</td><td>1088.96044</td><td>1076.355559</td><td>1044.963728</td><td>0.004643</td><td>0.00415</td><td>0.007362</td><td>3.0290e6</td><td>15454.627679</td></tr><tr><td>&quot;std&quot;</td><td>null</td><td>590.597838</td><td>611.779516</td><td>611.650372</td><td>0.067983</td><td>0.064286</td><td>0.085485</td><td>1.7252e6</td><td>58830.300697</td></tr><tr><td>&quot;min&quot;</td><td>&quot;C#CCOc1ccc(CNc…</td><td>7.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1.0</td></tr><tr><td>&quot;1%&quot;</td><td>null</td><td>21.0</td><td>25.0</td><td>16.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>65475.0</td><td>1.0</td></tr><tr><td>&quot;2.5%&quot;</td><td>null</td><td>76.0</td><td>46.0</td><td>44.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>156352.0</td><td>2.0</td></tr><tr><td>&quot;5%&quot;</td><td>null</td><td>159.0</td><td>92.0</td><td>101.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>287990.0</td><td>3.0</td></tr><tr><td>&quot;10%&quot;</td><td>null</td><td>294.0</td><td>155.0</td><td>205.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>609157.0</td><td>8.0</td></tr><tr><td>&quot;25%&quot;</td><td>null</td><td>575.0</td><td>570.0</td><td>519.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>1.548124e6</td><td>50.0</td></tr><tr><td>&quot;50%&quot;</td><td>null</td><td>1097.0</td><td>1118.0</td><td>1059.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>3.034755e6</td><td>387.0</td></tr><tr><td>&quot;75%&quot;</td><td>null</td><td>1618.0</td><td>1587.0</td><td>1546.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>4.567028e6</td><td>4539.0</td></tr><tr><td>&quot;90%&quot;</td><td>null</td><td>1912.0</td><td>1909.0</td><td>1890.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>5.391476e6</td><td>21796.0</td></tr><tr><td>&quot;95%&quot;</td><td>null</td><td>2011.0</td><td>2003.0</td><td>2010.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>5.684119e6</td><td>73117.0</td></tr><tr><td>&quot;97.5%&quot;</td><td>null</td><td>2050.0</td><td>2066.0</td><td>2071.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>5.845813e6</td><td>224511.0</td></tr><tr><td>&quot;99%&quot;</td><td>null</td><td>2061.0</td><td>2090.0</td><td>2096.0</td><td>0.0</td><td>0.0</td><td>0.0</td><td>5.929071e6</td><td>334289.0</td></tr><tr><td>&quot;max&quot;</td><td>&quot;[N-]=[N+]=NCCC…</td><td>2084.0</td><td>2109.0</td><td>2109.0</td><td>1.0</td><td>1.0</td><td>1.0</td><td>5.971684e6</td><td>499089.0</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (17, 10)\n",
       "┌─────────┬─────────┬─────────┬─────────┬─────────┬─────────┬─────────┬─────────┬─────────┬────────┐\n",
       "│ statist ┆ molecul ┆ bb1     ┆ bb2     ┆ bb3     ┆ BRD4    ┆ HSA     ┆ sEH     ┆ scaffol ┆ mol_pe │\n",
       "│ ic      ┆ e       ┆ ---     ┆ ---     ┆ ---     ┆ ---     ┆ ---     ┆ ---     ┆ d_idx   ┆ r_scaf │\n",
       "│ ---     ┆ ---     ┆ f64     ┆ f64     ┆ f64     ┆ f64     ┆ f64     ┆ f64     ┆ ---     ┆ ---    │\n",
       "│ str     ┆ str     ┆         ┆         ┆         ┆         ┆         ┆         ┆ f64     ┆ f64    │\n",
       "╞═════════╪═════════╪═════════╪═════════╪═════════╪═════════╪═════════╪═════════╪═════════╪════════╡\n",
       "│ count   ┆ 9841561 ┆ 9.84156 ┆ 9.84156 ┆ 9.84156 ┆ 9.84156 ┆ 9.84156 ┆ 9.84156 ┆ 9.84156 ┆ 9.8415 │\n",
       "│         ┆ 0       ┆ 1e7     ┆ 1e7     ┆ 1e7     ┆ 1e7     ┆ 1e7     ┆ 1e7     ┆ 1e7     ┆ 61e7   │\n",
       "│ null_co ┆ 0       ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 0.0    │\n",
       "│ unt     ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆        │\n",
       "│ mean    ┆ null    ┆ 1088.96 ┆ 1076.35 ┆ 1044.96 ┆ 0.00464 ┆ 0.00415 ┆ 0.00736 ┆ 3.0290e ┆ 15454. │\n",
       "│         ┆         ┆ 044     ┆ 5559    ┆ 3728    ┆ 3       ┆         ┆ 2       ┆ 6       ┆ 627679 │\n",
       "│ std     ┆ null    ┆ 590.597 ┆ 611.779 ┆ 611.650 ┆ 0.06798 ┆ 0.06428 ┆ 0.08548 ┆ 1.7252e ┆ 58830. │\n",
       "│         ┆         ┆ 838     ┆ 516     ┆ 372     ┆ 3       ┆ 6       ┆ 5       ┆ 6       ┆ 300697 │\n",
       "│ min     ┆ C#CCOc1 ┆ 7.0     ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 1.0    │\n",
       "│         ┆ ccc(CNc ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆        │\n",
       "│         ┆ 2nc(NC( ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆        │\n",
       "│         ┆ =N)CCCN ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆        │\n",
       "│         ┆ )nc(…   ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆        │\n",
       "│ 1%      ┆ null    ┆ 21.0    ┆ 25.0    ┆ 16.0    ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 65475.0 ┆ 1.0    │\n",
       "│ 2.5%    ┆ null    ┆ 76.0    ┆ 46.0    ┆ 44.0    ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 156352. ┆ 2.0    │\n",
       "│         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆ 0       ┆        │\n",
       "│ 5%      ┆ null    ┆ 159.0   ┆ 92.0    ┆ 101.0   ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 287990. ┆ 3.0    │\n",
       "│         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆ 0       ┆        │\n",
       "│ 10%     ┆ null    ┆ 294.0   ┆ 155.0   ┆ 205.0   ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 609157. ┆ 8.0    │\n",
       "│         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆ 0       ┆        │\n",
       "│ 25%     ┆ null    ┆ 575.0   ┆ 570.0   ┆ 519.0   ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 1.54812 ┆ 50.0   │\n",
       "│         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆ 4e6     ┆        │\n",
       "│ 50%     ┆ null    ┆ 1097.0  ┆ 1118.0  ┆ 1059.0  ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 3.03475 ┆ 387.0  │\n",
       "│         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆ 5e6     ┆        │\n",
       "│ 75%     ┆ null    ┆ 1618.0  ┆ 1587.0  ┆ 1546.0  ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 4.56702 ┆ 4539.0 │\n",
       "│         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆ 8e6     ┆        │\n",
       "│ 90%     ┆ null    ┆ 1912.0  ┆ 1909.0  ┆ 1890.0  ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 5.39147 ┆ 21796. │\n",
       "│         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆ 6e6     ┆ 0      │\n",
       "│ 95%     ┆ null    ┆ 2011.0  ┆ 2003.0  ┆ 2010.0  ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 5.68411 ┆ 73117. │\n",
       "│         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆ 9e6     ┆ 0      │\n",
       "│ 97.5%   ┆ null    ┆ 2050.0  ┆ 2066.0  ┆ 2071.0  ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 5.84581 ┆ 224511 │\n",
       "│         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆ 3e6     ┆ .0     │\n",
       "│ 99%     ┆ null    ┆ 2061.0  ┆ 2090.0  ┆ 2096.0  ┆ 0.0     ┆ 0.0     ┆ 0.0     ┆ 5.92907 ┆ 334289 │\n",
       "│         ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆ 1e6     ┆ .0     │\n",
       "│ max     ┆ [N-]=[N ┆ 2084.0  ┆ 2109.0  ┆ 2109.0  ┆ 1.0     ┆ 1.0     ┆ 1.0     ┆ 5.97168 ┆ 499089 │\n",
       "│         ┆ +]=NCCC ┆         ┆         ┆         ┆         ┆         ┆         ┆ 4e6     ┆ .0     │\n",
       "│         ┆ [C@H](N ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆        │\n",
       "│         ┆ c1nc(Nc ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆        │\n",
       "│         ┆ 2noc…   ┆         ┆         ┆         ┆         ┆         ┆         ┆         ┆        │\n",
       "└─────────┴─────────┴─────────┴─────────┴─────────┴─────────┴─────────┴─────────┴─────────┴────────┘"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.describe(percentiles=PERCENTILES)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "2de09045",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BRD4\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HSA\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sEH\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkIAAAGdCAYAAAD+JxxnAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8g+/7EAAAACXBIWXMAAA9hAAAPYQGoP6dpAAA0cElEQVR4nO3df1RVdb7/8dcBOwd0PPgrfn0lJfthJuqIRVRa3lieimmGO84df42RkY4OOAGTImX4o+7gxam0yWRVUzjr6mjelUwDDoaYekvSRBl/JFwtzFp60CbhKCU/9/ePWezxiKk4IMF+Ptbaazj78977vPdnyPNa++y9sRmGYQgAAMCCfDq6AQAAgI5CEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJbVraMb+D5ramrS8ePH1bNnT9lsto5uBwAAXAHDMHTmzBmFhobKx+fS53wIQpdw/PhxhYWFdXQbAADgKnzxxRfq37//JWsIQpfQs2dPSf+YSKfT2cHdAACAK+HxeBQWFmZ+jl8KQegSmr8OczqdBCEAADqZK7mshYulAQCAZRGEAACAZRGEAACAZRGEAACAZRGEAACAZRGEAACAZRGEAACAZRGEAACAZRGEAACAZRGEAACAZRGEAACAZRGEAACAZRGEAACAZRGEAACAZXXr6AasbOC8/I5uodWOLont6BYAAGgznBECAACWRRACAACWRRACAACWRRACAACWRRACAACWRRACAACWRRACAACWRRACAACWRRACAACWRRACAACWRRACAACWRRACAACWRRACAACW1aoglJmZqTvuuEM9e/ZUYGCg4uLiVF5e7lVz7tw5JSYmqm/fvvrBD36g8ePHq7Ky0qvm2LFjio2NVffu3RUYGKg5c+aooaHBq2br1q0aOXKkHA6HbrrpJuXk5LToZ8WKFRo4cKD8/PwUFRWlXbt2tboXAABgXa0KQtu2bVNiYqI++ugjFRYWqr6+XuPGjVNNTY1Zk5KSor/85S9av369tm3bpuPHj+unP/2pOd7Y2KjY2FjV1dVpx44dWrVqlXJycpSRkWHWVFRUKDY2VmPHjlVpaamSk5P1xBNPaNOmTWbNunXrlJqaqgULFmjPnj0aPny4XC6XTp48ecW9AAAAa7MZhmFc7canTp1SYGCgtm3bpjFjxqi6ulrXX3+91qxZo5/97GeSpLKyMt12220qLi7WXXfdpb/+9a/60Y9+pOPHjysoKEiSlJ2drbS0NJ06dUp2u11paWnKz8/XgQMHzPeaOHGiqqqqVFBQIEmKiorSHXfcoVdeeUWS1NTUpLCwMM2ePVvz5s27ol4ux+PxKCAgQNXV1XI6nVc7Td9p4Lz8Nt9nezu6JLajWwAA4JJa8/n9L10jVF1dLUnq06ePJKmkpET19fWKiYkxawYPHqwbbrhBxcXFkqTi4mJFRESYIUiSXC6XPB6PDh48aNacv4/mmuZ91NXVqaSkxKvGx8dHMTExZs2V9HKh2tpaeTwerwUAAHRdVx2EmpqalJycrHvuuUdDhw6VJLndbtntdvXq1curNigoSG6326w5PwQ1jzePXarG4/Ho22+/1VdffaXGxsaL1py/j8v1cqHMzEwFBASYS1hY2BXOBgAA6IyuOgglJibqwIEDWrt2bVv206HS09NVXV1tLl988UVHtwQAANpRt6vZKCkpSXl5edq+fbv69+9vrg8ODlZdXZ2qqqq8zsRUVlYqODjYrLnw7q7mO7nOr7nw7q7Kyko5nU75+/vL19dXvr6+F605fx+X6+VCDodDDoejFTMBAAA6s1adETIMQ0lJSdqwYYO2bNmi8PBwr/HIyEhdd911KioqMteVl5fr2LFjio6OliRFR0dr//79Xnd3FRYWyul0asiQIWbN+ftormneh91uV2RkpFdNU1OTioqKzJor6QUAAFhbq84IJSYmas2aNfrzn/+snj17mtfaBAQEyN/fXwEBAUpISFBqaqr69Okjp9Op2bNnKzo62rxLa9y4cRoyZIimTp2qrKwsud1uzZ8/X4mJiebZmJkzZ+qVV17R3Llz9fjjj2vLli16++23lZ//z7usUlNTFR8fr1GjRunOO+/UsmXLVFNTo2nTppk9Xa4XAABgba0KQitXrpQk3X///V7r33rrLT322GOSpJdeekk+Pj4aP368amtr5XK59Oqrr5q1vr6+ysvL06xZsxQdHa0ePXooPj5eixcvNmvCw8OVn5+vlJQULV++XP3799cbb7whl8tl1kyYMEGnTp1SRkaG3G63RowYoYKCAq8LqC/XCwAAsLZ/6TlCXR3PEWqJ5wgBAL7vrtlzhAAAADozghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALAsghAAALCsVgeh7du365FHHlFoaKhsNptyc3O9xm0220WXpUuXmjUDBw5sMb5kyRKv/ezbt0+jR4+Wn5+fwsLClJWV1aKX9evXa/DgwfLz81NERIQ2btzoNW4YhjIyMhQSEiJ/f3/FxMTo8OHDrT1kAADQRbU6CNXU1Gj48OFasWLFRcdPnDjhtbz55puy2WwaP368V93ixYu96mbPnm2OeTwejRs3TgMGDFBJSYmWLl2qhQsX6rXXXjNrduzYoUmTJikhIUF79+5VXFyc4uLidODAAbMmKytLL7/8srKzs7Vz50716NFDLpdL586da+1hAwCALqhbazd46KGH9NBDD33neHBwsNfrP//5zxo7dqxuvPFGr/U9e/ZsUdts9erVqqur05tvvim73a7bb79dpaWlevHFFzVjxgxJ0vLly/Xggw9qzpw5kqTnnntOhYWFeuWVV5SdnS3DMLRs2TLNnz9fP/nJTyRJf/zjHxUUFKTc3FxNnDixtYcOAAC6mHa9RqiyslL5+flKSEhoMbZkyRL17dtXP/zhD7V06VI1NDSYY8XFxRozZozsdru5zuVyqby8XKdPnzZrYmJivPbpcrlUXFwsSaqoqJDb7faqCQgIUFRUlFlzodraWnk8Hq8FAAB0Xa0+I9Qaq1atUs+ePfXTn/7Ua/2vf/1rjRw5Un369NGOHTuUnp6uEydO6MUXX5Qkud1uhYeHe20TFBRkjvXu3Vtut9tcd36N2+02687f7mI1F8rMzNSiRYuu8mgBAEBn065B6M0339SUKVPk5+fntT41NdX8ediwYbLb7frlL3+pzMxMORyO9mzpktLT071683g8CgsL67B+AABA+2q3r8b+93//V+Xl5XriiScuWxsVFaWGhgYdPXpU0j+uM6qsrPSqaX7dfF3Rd9WcP37+dheruZDD4ZDT6fRaAABA19VuQegPf/iDIiMjNXz48MvWlpaWysfHR4GBgZKk6Ohobd++XfX19WZNYWGhbr31VvXu3dusKSoq8tpPYWGhoqOjJUnh4eEKDg72qvF4PNq5c6dZAwAArK3VX42dPXtWR44cMV9XVFSotLRUffr00Q033CDpH4Fj/fr1euGFF1psX1xcrJ07d2rs2LHq2bOniouLlZKSol/84hdmyJk8ebIWLVqkhIQEpaWl6cCBA1q+fLleeuklcz9PPvmk7rvvPr3wwguKjY3V2rVrtXv3bvMWe5vNpuTkZD3//PO6+eabFR4ermeffVahoaGKi4tr7WEDAIAuqNVBaPfu3Ro7dqz5uvmamvj4eOXk5EiS1q5dK8MwNGnSpBbbOxwOrV27VgsXLlRtba3Cw8OVkpLidW1OQECA3nvvPSUmJioyMlL9+vVTRkaGeeu8JN19991as2aN5s+fr6efflo333yzcnNzNXToULNm7ty5qqmp0YwZM1RVVaV7771XBQUFLa5ZAgAA1mQzDMPo6Ca+rzwejwICAlRdXd0u1wsNnJff5vtsb0eXxHZ0CwAAXFJrPr/5W2MAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyCEIAAMCyWh2Etm/frkceeUShoaGy2WzKzc31Gn/sscdks9m8lgcffNCr5uuvv9aUKVPkdDrVq1cvJSQk6OzZs141+/bt0+jRo+Xn56ewsDBlZWW16GX9+vUaPHiw/Pz8FBERoY0bN3qNG4ahjIwMhYSEyN/fXzExMTp8+HBrDxkAAHRRrQ5CNTU1Gj58uFasWPGdNQ8++KBOnDhhLn/605+8xqdMmaKDBw+qsLBQeXl52r59u2bMmGGOezwejRs3TgMGDFBJSYmWLl2qhQsX6rXXXjNrduzYoUmTJikhIUF79+5VXFyc4uLidODAAbMmKytLL7/8srKzs7Vz50716NFDLpdL586da+1hAwCALshmGIZx1RvbbNqwYYPi4uLMdY899piqqqpanClqdujQIQ0ZMkQff/yxRo0aJUkqKCjQww8/rC+//FKhoaFauXKlnnnmGbndbtntdknSvHnzlJubq7KyMknShAkTVFNTo7y8PHPfd911l0aMGKHs7GwZhqHQ0FD95je/0VNPPSVJqq6uVlBQkHJycjRx4sTLHp/H41FAQICqq6vldDqvZoouaeC8/DbfZ3s7uiS2o1sAAOCSWvP53S7XCG3dulWBgYG69dZbNWvWLP397383x4qLi9WrVy8zBElSTEyMfHx8tHPnTrNmzJgxZgiSJJfLpfLycp0+fdqsiYmJ8Xpfl8ul4uJiSVJFRYXcbrdXTUBAgKKiosyaC9XW1srj8XgtAACg62rzIPTggw/qj3/8o4qKivRf//Vf2rZtmx566CE1NjZKktxutwIDA7226datm/r06SO3223WBAUFedU0v75czfnj5293sZoLZWZmKiAgwFzCwsJaffwAAKDz6NbWOzz/K6eIiAgNGzZMgwYN0tatW/XAAw+09du1qfT0dKWmppqvPR4PYQgAgC6s3W+fv/HGG9WvXz8dOXJEkhQcHKyTJ0961TQ0NOjrr79WcHCwWVNZWelV0/z6cjXnj5+/3cVqLuRwOOR0Or0WAADQdbV7EPryyy/197//XSEhIZKk6OhoVVVVqaSkxKzZsmWLmpqaFBUVZdZs375d9fX1Zk1hYaFuvfVW9e7d26wpKiryeq/CwkJFR0dLksLDwxUcHOxV4/F4tHPnTrMGAABYW6uD0NmzZ1VaWqrS0lJJ/7goubS0VMeOHdPZs2c1Z84cffTRRzp69KiKior0k5/8RDfddJNcLpck6bbbbtODDz6o6dOna9euXfrwww+VlJSkiRMnKjQ0VJI0efJk2e12JSQk6ODBg1q3bp2WL1/u9bXVk08+qYKCAr3wwgsqKyvTwoULtXv3biUlJUn6xx1tycnJev755/Xuu+9q//79evTRRxUaGup1lxsAALCuVl8jtHv3bo0dO9Z83RxO4uPjtXLlSu3bt0+rVq1SVVWVQkNDNW7cOD333HNyOBzmNqtXr1ZSUpIeeOAB+fj4aPz48Xr55ZfN8YCAAL333ntKTExUZGSk+vXrp4yMDK9nDd19991as2aN5s+fr6efflo333yzcnNzNXToULNm7ty5qqmp0YwZM1RVVaV7771XBQUF8vPza+1hAwCALuhfeo5QV8dzhFriOUIAgO+7Dn+OEAAAQGdAEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJZFEAIAAJbV6iC0fft2PfLIIwoNDZXNZlNubq45Vl9fr7S0NEVERKhHjx4KDQ3Vo48+quPHj3vtY+DAgbLZbF7LkiVLvGr27dun0aNHy8/PT2FhYcrKymrRy/r16zV48GD5+fkpIiJCGzdu9Bo3DEMZGRkKCQmRv7+/YmJidPjw4dYeMgAA6KJaHYRqamo0fPhwrVixosXYN998oz179ujZZ5/Vnj179M4776i8vFw//vGPW9QuXrxYJ06cMJfZs2ebYx6PR+PGjdOAAQNUUlKipUuXauHChXrttdfMmh07dmjSpElKSEjQ3r17FRcXp7i4OB04cMCsycrK0ssvv6zs7Gzt3LlTPXr0kMvl0rlz51p72AAAoAuyGYZhXPXGNps2bNiguLi476z5+OOPdeedd+rzzz/XDTfcIOkfZ4SSk5OVnJx80W1WrlypZ555Rm63W3a7XZI0b9485ebmqqysTJI0YcIE1dTUKC8vz9zurrvu0ogRI5SdnS3DMBQaGqrf/OY3euqppyRJ1dXVCgoKUk5OjiZOnHjZ4/N4PAoICFB1dbWcTueVTEmrDJyX3+b7bG9Hl8R2dAsAAFxSaz6/2/0aoerqatlsNvXq1ctr/ZIlS9S3b1/98Ic/1NKlS9XQ0GCOFRcXa8yYMWYIkiSXy6Xy8nKdPn3arImJifHap8vlUnFxsSSpoqJCbrfbqyYgIEBRUVFmzYVqa2vl8Xi8FgAA0HV1a8+dnzt3TmlpaZo0aZJXIvv1r3+tkSNHqk+fPtqxY4fS09N14sQJvfjii5Ikt9ut8PBwr30FBQWZY71795bb7TbXnV/jdrvNuvO3u1jNhTIzM7Vo0aJ/4YgBAEBn0m5BqL6+Xj//+c9lGIZWrlzpNZaammr+PGzYMNntdv3yl79UZmamHA5He7V0Wenp6V69eTwehYWFdVg/AACgfbXLV2PNIejzzz9XYWHhZb+fi4qKUkNDg44ePSpJCg4OVmVlpVdN8+vg4OBL1pw/fv52F6u5kMPhkNPp9FoAAEDX1eZBqDkEHT58WJs3b1bfvn0vu01paal8fHwUGBgoSYqOjtb27dtVX19v1hQWFurWW29V7969zZqioiKv/RQWFio6OlqSFB4eruDgYK8aj8ejnTt3mjUAAMDaWv3V2NmzZ3XkyBHzdUVFhUpLS9WnTx+FhIToZz/7mfbs2aO8vDw1Njaa1+P06dNHdrtdxcXF2rlzp8aOHauePXuquLhYKSkp+sUvfmGGnMmTJ2vRokVKSEhQWlqaDhw4oOXLl+ull14y3/fJJ5/UfffdpxdeeEGxsbFau3atdu/ebd5ib7PZlJycrOeff14333yzwsPD9eyzzyo0NPSSd7kBAADraPXt81u3btXYsWNbrI+Pj9fChQtbXOTc7P3339f999+vPXv26Fe/+pXKyspUW1ur8PBwTZ06VampqV7XB+3bt0+JiYn6+OOP1a9fP82ePVtpaWle+1y/fr3mz5+vo0eP6uabb1ZWVpYefvhhc9wwDC1YsECvvfaaqqqqdO+99+rVV1/VLbfcckXHyu3zLXH7PADg+641n9//0nOEujqCUEsEIQDA99336jlCAAAA31cEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFkEIQAAYFmtDkLbt2/XI488otDQUNlsNuXm5nqNG4ahjIwMhYSEyN/fXzExMTp8+LBXzddff60pU6bI6XSqV69eSkhI0NmzZ71q9u3bp9GjR8vPz09hYWHKyspq0cv69es1ePBg+fn5KSIiQhs3bmx1LwAAwLpaHYRqamo0fPhwrVix4qLjWVlZevnll5Wdna2dO3eqR48ecrlcOnfunFkzZcoUHTx4UIWFhcrLy9P27ds1Y8YMc9zj8WjcuHEaMGCASkpKtHTpUi1cuFCvvfaaWbNjxw5NmjRJCQkJ2rt3r+Li4hQXF6cDBw60qhcAAGBdNsMwjKve2GbThg0bFBcXJ+kfZ2BCQ0P1m9/8Rk899ZQkqbq6WkFBQcrJydHEiRN16NAhDRkyRB9//LFGjRolSSooKNDDDz+sL7/8UqGhoVq5cqWeeeYZud1u2e12SdK8efOUm5ursrIySdKECRNUU1OjvLw8s5+77rpLI0aMUHZ29hX1cjkej0cBAQGqrq6W0+m82mn6TgPn5bf5Ptvb0SWxHd0CAACX1JrP7za9RqiiokJut1sxMTHmuoCAAEVFRam4uFiSVFxcrF69epkhSJJiYmLk4+OjnTt3mjVjxowxQ5AkuVwulZeX6/Tp02bN+e/TXNP8PlfSy4Vqa2vl8Xi8FgAA0HW1aRByu92SpKCgIK/1QUFB5pjb7VZgYKDXeLdu3dSnTx+vmovt4/z3+K6a88cv18uFMjMzFRAQYC5hYWFXcNQAAKCz4q6x86Snp6u6utpcvvjii45uCQAAtKM2DULBwcGSpMrKSq/1lZWV5lhwcLBOnjzpNd7Q0KCvv/7aq+Zi+zj/Pb6r5vzxy/VyIYfDIafT6bUAAICuq02DUHh4uIKDg1VUVGSu83g82rlzp6KjoyVJ0dHRqqqqUklJiVmzZcsWNTU1KSoqyqzZvn276uvrzZrCwkLdeuut6t27t1lz/vs01zS/z5X0AgAArK3VQejs2bMqLS1VaWmppH9clFxaWqpjx47JZrMpOTlZzz//vN59913t379fjz76qEJDQ807y2677TY9+OCDmj59unbt2qUPP/xQSUlJmjhxokJDQyVJkydPlt1uV0JCgg4ePKh169Zp+fLlSk1NNft48sknVVBQoBdeeEFlZWVauHChdu/eraSkJEm6ol4AAIC1dWvtBrt379bYsWPN183hJD4+Xjk5OZo7d65qamo0Y8YMVVVV6d5771VBQYH8/PzMbVavXq2kpCQ98MAD8vHx0fjx4/Xyyy+b4wEBAXrvvfeUmJioyMhI9evXTxkZGV7PGrr77ru1Zs0azZ8/X08//bRuvvlm5ebmaujQoWbNlfQCAACs6196jlBXx3OEWuI5QgCA77sOe44QAABAZ0IQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAlkUQAgAAltXmQWjgwIGy2WwtlsTEREnS/fff32Js5syZXvs4duyYYmNj1b17dwUGBmrOnDlqaGjwqtm6datGjhwph8Ohm266STk5OS16WbFihQYOHCg/Pz9FRUVp165dbX24AACgE2vzIPTxxx/rxIkT5lJYWChJ+o//+A+zZvr06V41WVlZ5lhjY6NiY2NVV1enHTt2aNWqVcrJyVFGRoZZU1FRodjYWI0dO1alpaVKTk7WE088oU2bNpk169atU2pqqhYsWKA9e/Zo+PDhcrlcOnnyZFsfMgAA6KRshmEY7fkGycnJysvL0+HDh2Wz2XT//fdrxIgRWrZs2UXr//rXv+pHP/qRjh8/rqCgIElSdna20tLSdOrUKdntdqWlpSk/P18HDhwwt5s4caKqqqpUUFAgSYqKitIdd9yhV155RZLU1NSksLAwzZ49W/Pmzbui3j0ejwICAlRdXS2n0/kvzMLFDZyX3+b7bG9Hl8R2dAsAAFxSaz6/2/Uaobq6Ov33f/+3Hn/8cdlsNnP96tWr1a9fPw0dOlTp6en65ptvzLHi4mJFRESYIUiSXC6XPB6PDh48aNbExMR4vZfL5VJxcbH5viUlJV41Pj4+iomJMWsupra2Vh6Px2sBAABdV7f23Hlubq6qqqr02GOPmesmT56sAQMGKDQ0VPv27VNaWprKy8v1zjvvSJLcbrdXCJJkvna73Zes8Xg8+vbbb3X69Gk1NjZetKasrOw7+83MzNSiRYuu+ngBAEDn0q5B6A9/+IMeeughhYaGmutmzJhh/hwREaGQkBA98MAD+vTTTzVo0KD2bOey0tPTlZqaar72eDwKCwvrwI4AAEB7arcg9Pnnn2vz5s3mmZ7vEhUVJUk6cuSIBg0apODg4BZ3d1VWVkqSgoODzf9tXnd+jdPplL+/v3x9feXr63vRmuZ9XIzD4ZDD4biyAwQAAJ1eu10j9NZbbykwMFCxsZe+uLa0tFSSFBISIkmKjo7W/v37ve7uKiwslNPp1JAhQ8yaoqIir/0UFhYqOjpakmS32xUZGelV09TUpKKiIrMGAACgXYJQU1OT3nrrLcXHx6tbt3+edPr000/13HPPqaSkREePHtW7776rRx99VGPGjNGwYcMkSePGjdOQIUM0depU/e1vf9OmTZs0f/58JSYmmmdrZs6cqc8++0xz585VWVmZXn31Vb399ttKSUkx3ys1NVWvv/66Vq1apUOHDmnWrFmqqanRtGnT2uOQAQBAJ9QuX41t3rxZx44d0+OPP+613m63a/PmzVq2bJlqamoUFham8ePHa/78+WaNr6+v8vLyNGvWLEVHR6tHjx6Kj4/X4sWLzZrw8HDl5+crJSVFy5cvV//+/fXGG2/I5XKZNRMmTNCpU6eUkZEht9utESNGqKCgoMUF1AAAwLra/TlCnRnPEWqJ5wgBAL7vvjfPEQIAAPg+IwgBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLIggBAADLavMgtHDhQtlsNq9l8ODB5vi5c+eUmJiovn376gc/+IHGjx+vyspKr30cO3ZMsbGx6t69uwIDAzVnzhw1NDR41WzdulUjR46Uw+HQTTfdpJycnBa9rFixQgMHDpSfn5+ioqK0a9eutj5cAADQibXLGaHbb79dJ06cMJcPPvjAHEtJSdFf/vIXrV+/Xtu2bdPx48f105/+1BxvbGxUbGys6urqtGPHDq1atUo5OTnKyMgwayoqKhQbG6uxY8eqtLRUycnJeuKJJ7Rp0yazZt26dUpNTdWCBQu0Z88eDR8+XC6XSydPnmyPQwYAAJ2QzTAMoy13uHDhQuXm5qq0tLTFWHV1ta6//nqtWbNGP/vZzyRJZWVluu2221RcXKy77rpLf/3rX/WjH/1Ix48fV1BQkCQpOztbaWlpOnXqlOx2u9LS0pSfn68DBw6Y+544caKqqqpUUFAgSYqKitIdd9yhV155RZLU1NSksLAwzZ49W/PmzbuiY/F4PAoICFB1dbWcTue/Mi0XNXBefpvvs70dXRLb0S0AAHBJrfn8bpczQocPH1ZoaKhuvPFGTZkyRceOHZMklZSUqL6+XjExMWbt4MGDdcMNN6i4uFiSVFxcrIiICDMESZLL5ZLH49HBgwfNmvP30VzTvI+6ujqVlJR41fj4+CgmJsasuZja2lp5PB6vBQAAdF1tHoSioqKUk5OjgoICrVy5UhUVFRo9erTOnDkjt9stu92uXr16eW0TFBQkt9stSXK73V4hqHm8eexSNR6PR99++62++uorNTY2XrSmeR8Xk5mZqYCAAHMJCwu7qjkAAACdQ7e23uFDDz1k/jxs2DBFRUVpwIABevvtt+Xv79/Wb9em0tPTlZqaar72eDyEIQAAurB2v32+V69euuWWW3TkyBEFBwerrq5OVVVVXjWVlZUKDg6WJAUHB7e4i6z59eVqnE6n/P391a9fP/n6+l60pnkfF+NwOOR0Or0WAADQdbV7EDp79qw+/fRThYSEKDIyUtddd52KiorM8fLych07dkzR0dGSpOjoaO3fv9/r7q7CwkI5nU4NGTLErDl/H801zfuw2+2KjIz0qmlqalJRUZFZAwAA0OZB6KmnntK2bdt09OhR7dixQ//+7/8uX19fTZo0SQEBAUpISFBqaqref/99lZSUaNq0aYqOjtZdd90lSRo3bpyGDBmiqVOn6m9/+5s2bdqk+fPnKzExUQ6HQ5I0c+ZMffbZZ5o7d67Kysr06quv6u2331ZKSorZR2pqql5//XWtWrVKhw4d0qxZs1RTU6Np06a19SEDAIBOqs2vEfryyy81adIk/f3vf9f111+ve++9Vx999JGuv/56SdJLL70kHx8fjR8/XrW1tXK5XHr11VfN7X19fZWXl6dZs2YpOjpaPXr0UHx8vBYvXmzWhIeHKz8/XykpKVq+fLn69++vN954Qy6Xy6yZMGGCTp06pYyMDLndbo0YMUIFBQUtLqAGAADW1ebPEepKeI5QSzxHCADwfdfhzxECAADoDAhCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsghCAADAsrp1dAMAuoaB8/I7uoVWO7oktqNbANDBCELo8viABgB8F74aAwAAltXmQSgzM1N33HGHevbsqcDAQMXFxam8vNyr5v7775fNZvNaZs6c6VVz7NgxxcbGqnv37goMDNScOXPU0NDgVbN161aNHDlSDodDN910k3Jyclr0s2LFCg0cOFB+fn6KiorSrl272vqQAQBAJ9XmQWjbtm1KTEzURx99pMLCQtXX12vcuHGqqanxqps+fbpOnDhhLllZWeZYY2OjYmNjVVdXpx07dmjVqlXKyclRRkaGWVNRUaHY2FiNHTtWpaWlSk5O1hNPPKFNmzaZNevWrVNqaqoWLFigPXv2aPjw4XK5XDp58mRbHzYAAOiE2vwaoYKCAq/XOTk5CgwMVElJicaMGWOu7969u4KDgy+6j/fee0+ffPKJNm/erKCgII0YMULPPfec0tLStHDhQtntdmVnZys8PFwvvPCCJOm2227TBx98oJdeekkul0uS9OKLL2r69OmaNm2aJCk7O1v5+fl68803NW/evLY+dAAA0Mm0+zVC1dXVkqQ+ffp4rV+9erX69eunoUOHKj09Xd988405VlxcrIiICAUFBZnrXC6XPB6PDh48aNbExMR47dPlcqm4uFiSVFdXp5KSEq8aHx8fxcTEmDUAAMDa2vWusaamJiUnJ+uee+7R0KFDzfWTJ0/WgAEDFBoaqn379iktLU3l5eV65513JElut9srBEkyX7vd7kvWeDweffvttzp9+rQaGxsvWlNWVnbRfmtra1VbW2u+9ng8V3nkAACgM2jXIJSYmKgDBw7ogw8+8Fo/Y8YM8+eIiAiFhITogQce0KeffqpBgwa1Z0uXlJmZqUWLFnXY+wMAgGur3b4aS0pKUl5ent5//33179//krVRUVGSpCNHjkiSgoODVVlZ6VXT/Lr5uqLvqnE6nfL391e/fv3k6+t70ZrvujYpPT1d1dXV5vLFF19c4dECAIDOqM2DkGEYSkpK0oYNG7RlyxaFh4dfdpvS0lJJUkhIiCQpOjpa+/fv97q7q7CwUE6nU0OGDDFrioqKvPZTWFio6OhoSZLdbldkZKRXTVNTk4qKisyaCzkcDjmdTq8FAAB0XW3+1VhiYqLWrFmjP//5z+rZs6d5TU9AQID8/f316aefas2aNXr44YfVt29f7du3TykpKRozZoyGDRsmSRo3bpyGDBmiqVOnKisrS263W/Pnz1diYqIcDockaebMmXrllVc0d+5cPf7449qyZYvefvtt5ef/8ynCqampio+P16hRo3TnnXdq2bJlqqmpMe8iAwAA1tbmQWjlypWS/vHQxPO99dZbeuyxx2S327V582YzlISFhWn8+PGaP3++Wevr66u8vDzNmjVL0dHR6tGjh+Lj47V48WKzJjw8XPn5+UpJSdHy5cvVv39/vfHGG+at85I0YcIEnTp1ShkZGXK73RoxYoQKCgpaXEANAACsqc2DkGEYlxwPCwvTtm3bLrufAQMGaOPGjZesuf/++7V3795L1iQlJSkpKemy7wcAAKyHvzUGAAAsiyAEAAAsiyAEAAAsiyAEAAAsiyAEAAAsiyAEAAAsiyAEAAAsiyAEAAAsiyAEAAAsiyAEAAAsiyAEAAAsiyAEAAAsiyAEAAAsq83/+jy6toHz8ju6BQAA2gxnhAAAgGURhAAAgGURhAAAgGURhAAAgGURhAAAgGURhAAAgGURhAAAgGURhAAAgGURhAAAgGURhAAAgGXxJzYAAOgiOuOfQTq6JLZD358gBMCy+NAAwFdjAADAsghCAADAsvhqDPge6oxf2QBAZ8QZIQAAYFkEIQAAYFmWCEIrVqzQwIED5efnp6ioKO3ataujWwIAAN8DXT4IrVu3TqmpqVqwYIH27Nmj4cOHy+Vy6eTJkx3dGgAA6GBdPgi9+OKLmj59uqZNm6YhQ4YoOztb3bt315tvvtnRrQEAgA7Wpe8aq6urU0lJidLT0811Pj4+iomJUXFxcYv62tpa1dbWmq+rq6slSR6Pp136a6r9pl32C6Draq9/j9A1dMbPlfb4nW7ep2EYl63t0kHoq6++UmNjo4KCgrzWBwUFqaysrEV9ZmamFi1a1GJ9WFhYu/UIAK0RsKyjOwDaVnv+Tp85c0YBAQGXrOnSQai10tPTlZqaar5uamrS119/rb59+8pms7Xpe3k8HoWFhemLL76Q0+ls033jn5jna4N5vjaY52uHub422mueDcPQmTNnFBoaetnaLh2E+vXrJ19fX1VWVnqtr6ysVHBwcIt6h8Mhh8Phta5Xr17t2aKcTif/kV0DzPO1wTxfG8zztcNcXxvtMc+XOxPUrEtfLG232xUZGamioiJzXVNTk4qKihQdHd2BnQEAgO+DLn1GSJJSU1MVHx+vUaNG6c4779SyZctUU1OjadOmdXRrAACgg3X5IDRhwgSdOnVKGRkZcrvdGjFihAoKClpcQH2tORwOLViwoMVXcWhbzPO1wTxfG8zztcNcXxvfh3m2GVdybxkAAEAX1KWvEQIAALgUghAAALAsghAAALAsghAAALAsglA7WrFihQYOHCg/Pz9FRUVp165dl6xfv369Bg8eLD8/P0VERGjjxo3XqNPOrTXz/Prrr2v06NHq3bu3evfurZiYmMv+/4J/aO3vc7O1a9fKZrMpLi6ufRvsIlo7z1VVVUpMTFRISIgcDoduueUW/u24Aq2d52XLlunWW2+Vv7+/wsLClJKSonPnzl2jbjun7du365FHHlFoaKhsNptyc3Mvu83WrVs1cuRIORwO3XTTTcrJyWn3PmWgXaxdu9aw2+3Gm2++aRw8eNCYPn260atXL6OysvKi9R9++KHh6+trZGVlGZ988okxf/5847rrrjP2799/jTvvXFo7z5MnTzZWrFhh7N271zh06JDx2GOPGQEBAcaXX355jTvvXFo7z80qKiqM//f//p8xevRo4yc/+cm1abYTa+0819bWGqNGjTIefvhh44MPPjAqKiqMrVu3GqWlpde4886ltfO8evVqw+FwGKtXrzYqKiqMTZs2GSEhIUZKSso17rxz2bhxo/HMM88Y77zzjiHJ2LBhwyXrP/vsM6N79+5Gamqq8cknnxi///3vDV9fX6OgoKBd+yQItZM777zTSExMNF83NjYaoaGhRmZm5kXrf/7znxuxsbFe66Kiooxf/vKX7dpnZ9faeb5QQ0OD0bNnT2PVqlXt1WKXcDXz3NDQYNx9993GG2+8YcTHxxOErkBr53nlypXGjTfeaNTV1V2rFruE1s5zYmKi8W//9m9e61JTU4177rmnXfvsSq4kCM2dO9e4/fbbvdZNmDDBcLlc7diZYfDVWDuoq6tTSUmJYmJizHU+Pj6KiYlRcXHxRbcpLi72qpckl8v1nfW4unm+0DfffKP6+nr16dOnvdrs9K52nhcvXqzAwEAlJCRcizY7vauZ53fffVfR0dFKTExUUFCQhg4dqt/+9rdqbGy8Vm13Olczz3fffbdKSkrMr88+++wzbdy4UQ8//PA16dkqOupzsMs/WbojfPXVV2psbGzx9OqgoCCVlZVddBu3233Rerfb3W59dnZXM88XSktLU2hoaIv/+PBPVzPPH3zwgf7whz+otLT0GnTYNVzNPH/22WfasmWLpkyZoo0bN+rIkSP61a9+pfr6ei1YsOBatN3pXM08T548WV999ZXuvfdeGYahhoYGzZw5U08//fS1aNkyvutz0OPx6Ntvv5W/v3+7vC9nhGBZS5Ys0dq1a7Vhwwb5+fl1dDtdxpkzZzR16lS9/vrr6tevX0e306U1NTUpMDBQr732miIjIzVhwgQ988wzys7O7ujWupStW7fqt7/9rV599VXt2bNH77zzjvLz8/Xcc891dGtoA5wRagf9+vWTr6+vKisrvdZXVlYqODj4otsEBwe3qh5XN8/Nfve732nJkiXavHmzhg0b1p5tdnqtnedPP/1UR48e1SOPPGKua2pqkiR169ZN5eXlGjRoUPs23Qldze9zSEiIrrvuOvn6+prrbrvtNrndbtXV1clut7drz53R1czzs88+q6lTp+qJJ56QJEVERKimpkYzZszQM888Ix8fzim0he/6HHQ6ne12NkjijFC7sNvtioyMVFFRkbmuqalJRUVFio6Ovug20dHRXvWSVFhY+J31uLp5lqSsrCw999xzKigo0KhRo65Fq51aa+d58ODB2r9/v0pLS83lxz/+scaOHavS0lKFhYVdy/Y7jav5fb7nnnt05MgRM2hK0v/93/8pJCSEEPQdrmaev/nmmxZhpzl8Gvy5zjbTYZ+D7XoptoWtXbvWcDgcRk5OjvHJJ58YM2bMMHr16mW43W7DMAxj6tSpxrx588z6Dz/80OjWrZvxu9/9zjh06JCxYMECbp+/Aq2d5yVLlhh2u934n//5H+PEiRPmcubMmY46hE6htfN8Ie4auzKtnedjx44ZPXv2NJKSkozy8nIjLy/PCAwMNJ5//vmOOoROobXzvGDBAqNnz57Gn/70J+Ozzz4z3nvvPWPQoEHGz3/+8446hE7hzJkzxt69e429e/cakowXX3zR2Lt3r/H5558bhmEY8+bNM6ZOnWrWN98+P2fOHOPQoUPGihUruH2+s/v9739v3HDDDYbdbjfuvPNO46OPPjLH7rvvPiM+Pt6r/u233zZuueUWw263G7fffruRn59/jTvunFozzwMGDDAktVgWLFhw7RvvZFr7+3w+gtCVa+0879ixw4iKijIcDodx4403Gv/5n/9pNDQ0XOOuO5/WzHN9fb2xcOFCY9CgQYafn58RFhZm/OpXvzJOnz597RvvRN5///2L/nvbPLfx8fHGfffd12KbESNGGHa73bjxxhuNt956q937tBkG5/UAAIA1cY0QAACwLIIQAACwLIIQAACwLIIQAACwLIIQAACwLIIQAACwLIIQAACwLIIQAACwLIIQAACwLIIQAACwLIIQAACwLIIQAACwrP8P3nBKA8MO1MAAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "\n",
    "for protein in ['BRD4', 'HSA', 'sEH']:\n",
    "    print(protein)\n",
    "    plt.hist(tmp_df[protein])\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "bfb5b6cc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "scatter_plot(tmp_df['sEH'], tmp_df['count'].log())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "a042ef6e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BRD4\n",
      "HSA\n",
      "sEH\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fc740aff520>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for protein in ['BRD4', 'HSA', 'sEH']:\n",
    "    print(protein)\n",
    "    tmp = tmp_df.with_columns((pl.col(protein) // 0.05).alias('bin')).group_by('bin').agg(pl.col('count').mean()).sort('bin')\n",
    "    plt.plot(tmp['bin'] * 0.05, tmp['count'])\n",
    "plt.legend(['BRD4', 'HSA', 'sEH'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "0d614a13",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (7_570, 6)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>scaffold_idx</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>mol_per_scaf</th><th>count</th></tr><tr><td>i64</td><td>f64</td><td>f64</td><td>f64</td><td>u32</td><td>u32</td></tr></thead><tbody><tr><td>5445847</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1</td><td>1</td></tr><tr><td>3334936</td><td>0.0</td><td>1.0</td><td>1.0</td><td>1</td><td>1</td></tr><tr><td>3951145</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1</td><td>1</td></tr><tr><td>2549769</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1</td><td>1</td></tr><tr><td>4063609</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1</td><td>1</td></tr><tr><td>4304537</td><td>0.0</td><td>1.0</td><td>1.0</td><td>1</td><td>1</td></tr><tr><td>785301</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1</td><td>1</td></tr><tr><td>41685</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1</td><td>1</td></tr><tr><td>2156922</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1</td><td>1</td></tr><tr><td>171715</td><td>0.0</td><td>0.0</td><td>1.0</td><td>1</td><td>1</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>520791</td><td>0.0</td><td>0.0</td><td>0.621053</td><td>95</td><td>95</td></tr><tr><td>5783650</td><td>0.0</td><td>0.0</td><td>0.835052</td><td>97</td><td>97</td></tr><tr><td>2505029</td><td>0.0</td><td>0.051546</td><td>0.680412</td><td>97</td><td>97</td></tr><tr><td>4623477</td><td>0.0</td><td>0.0</td><td>0.678261</td><td>115</td><td>115</td></tr><tr><td>2496255</td><td>0.0</td><td>0.0</td><td>0.608333</td><td>120</td><td>120</td></tr><tr><td>578715</td><td>0.0</td><td>0.106383</td><td>0.758865</td><td>141</td><td>141</td></tr><tr><td>4969246</td><td>0.0</td><td>0.0</td><td>0.697561</td><td>205</td><td>205</td></tr><tr><td>2145497</td><td>0.0</td><td>0.0</td><td>0.736585</td><td>205</td><td>205</td></tr><tr><td>3501572</td><td>0.0</td><td>0.004184</td><td>0.615063</td><td>239</td><td>239</td></tr><tr><td>4283950</td><td>0.0</td><td>0.0</td><td>0.616314</td><td>331</td><td>331</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (7_570, 6)\n",
       "┌──────────────┬──────┬──────────┬──────────┬──────────────┬───────┐\n",
       "│ scaffold_idx ┆ BRD4 ┆ HSA      ┆ sEH      ┆ mol_per_scaf ┆ count │\n",
       "│ ---          ┆ ---  ┆ ---      ┆ ---      ┆ ---          ┆ ---   │\n",
       "│ i64          ┆ f64  ┆ f64      ┆ f64      ┆ u32          ┆ u32   │\n",
       "╞══════════════╪══════╪══════════╪══════════╪══════════════╪═══════╡\n",
       "│ 5445847      ┆ 0.0  ┆ 0.0      ┆ 1.0      ┆ 1            ┆ 1     │\n",
       "│ 3334936      ┆ 0.0  ┆ 1.0      ┆ 1.0      ┆ 1            ┆ 1     │\n",
       "│ 3951145      ┆ 0.0  ┆ 0.0      ┆ 1.0      ┆ 1            ┆ 1     │\n",
       "│ 2549769      ┆ 0.0  ┆ 0.0      ┆ 1.0      ┆ 1            ┆ 1     │\n",
       "│ 4063609      ┆ 0.0  ┆ 0.0      ┆ 1.0      ┆ 1            ┆ 1     │\n",
       "│ 4304537      ┆ 0.0  ┆ 1.0      ┆ 1.0      ┆ 1            ┆ 1     │\n",
       "│ 785301       ┆ 0.0  ┆ 0.0      ┆ 1.0      ┆ 1            ┆ 1     │\n",
       "│ 41685        ┆ 0.0  ┆ 0.0      ┆ 1.0      ┆ 1            ┆ 1     │\n",
       "│ 2156922      ┆ 0.0  ┆ 0.0      ┆ 1.0      ┆ 1            ┆ 1     │\n",
       "│ 171715       ┆ 0.0  ┆ 0.0      ┆ 1.0      ┆ 1            ┆ 1     │\n",
       "│ …            ┆ …    ┆ …        ┆ …        ┆ …            ┆ …     │\n",
       "│ 520791       ┆ 0.0  ┆ 0.0      ┆ 0.621053 ┆ 95           ┆ 95    │\n",
       "│ 5783650      ┆ 0.0  ┆ 0.0      ┆ 0.835052 ┆ 97           ┆ 97    │\n",
       "│ 2505029      ┆ 0.0  ┆ 0.051546 ┆ 0.680412 ┆ 97           ┆ 97    │\n",
       "│ 4623477      ┆ 0.0  ┆ 0.0      ┆ 0.678261 ┆ 115          ┆ 115   │\n",
       "│ 2496255      ┆ 0.0  ┆ 0.0      ┆ 0.608333 ┆ 120          ┆ 120   │\n",
       "│ 578715       ┆ 0.0  ┆ 0.106383 ┆ 0.758865 ┆ 141          ┆ 141   │\n",
       "│ 4969246      ┆ 0.0  ┆ 0.0      ┆ 0.697561 ┆ 205          ┆ 205   │\n",
       "│ 2145497      ┆ 0.0  ┆ 0.0      ┆ 0.736585 ┆ 205          ┆ 205   │\n",
       "│ 3501572      ┆ 0.0  ┆ 0.004184 ┆ 0.615063 ┆ 239          ┆ 239   │\n",
       "│ 4283950      ┆ 0.0  ┆ 0.0      ┆ 0.616314 ┆ 331          ┆ 331   │\n",
       "└──────────────┴──────┴──────────┴──────────┴──────────────┴───────┘"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp = tmp_df.filter(pl.col('sEH') >= 0.6).sort('count')\n",
    "scatter_plot(tmp['sEH'], tmp['count'].log())\n",
    "tmp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "5e7af810",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (210_702, 6)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>scaffold_idx</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>mol_per_scaf</th><th>count</th></tr><tr><td>i64</td><td>f64</td><td>f64</td><td>f64</td><td>u32</td><td>u32</td></tr></thead><tbody><tr><td>3198219</td><td>1.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>5132175</td><td>0.0</td><td>1.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>4527765</td><td>1.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>5641681</td><td>0.0</td><td>1.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>1003027</td><td>0.0</td><td>1.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>2952316</td><td>1.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>3453301</td><td>0.0</td><td>1.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>1506897</td><td>1.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>4070034</td><td>1.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>3256419</td><td>1.0</td><td>0.0</td><td>0.0</td><td>1</td><td>1</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>5509544</td><td>0.000901</td><td>0.000227</td><td>0.000266</td><td>176474</td><td>176474</td></tr><tr><td>2963783</td><td>0.010951</td><td>0.00022</td><td>0.001118</td><td>195054</td><td>195054</td></tr><tr><td>4583912</td><td>0.005541</td><td>0.000178</td><td>0.007193</td><td>224511</td><td>224511</td></tr><tr><td>4912182</td><td>0.003531</td><td>0.000272</td><td>0.000052</td><td>231387</td><td>231387</td></tr><tr><td>162518</td><td>0.003213</td><td>0.001577</td><td>0.000902</td><td>237127</td><td>237127</td></tr><tr><td>4867705</td><td>0.002347</td><td>0.000216</td><td>0.000041</td><td>245872</td><td>245872</td></tr><tr><td>5929438</td><td>0.000871</td><td>0.005708</td><td>0.005967</td><td>312198</td><td>312198</td></tr><tr><td>4632750</td><td>0.000739</td><td>0.000236</td><td>0.000018</td><td>334289</td><td>334289</td></tr><tr><td>4861294</td><td>0.001876</td><td>0.000516</td><td>0.000036</td><td>412652</td><td>412652</td></tr><tr><td>2217250</td><td>0.00547</td><td>0.000202</td><td>0.00027</td><td>499089</td><td>499089</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (210_702, 6)\n",
       "┌──────────────┬──────────┬──────────┬──────────┬──────────────┬────────┐\n",
       "│ scaffold_idx ┆ BRD4     ┆ HSA      ┆ sEH      ┆ mol_per_scaf ┆ count  │\n",
       "│ ---          ┆ ---      ┆ ---      ┆ ---      ┆ ---          ┆ ---    │\n",
       "│ i64          ┆ f64      ┆ f64      ┆ f64      ┆ u32          ┆ u32    │\n",
       "╞══════════════╪══════════╪══════════╪══════════╪══════════════╪════════╡\n",
       "│ 3198219      ┆ 1.0      ┆ 0.0      ┆ 0.0      ┆ 1            ┆ 1      │\n",
       "│ 5132175      ┆ 0.0      ┆ 1.0      ┆ 0.0      ┆ 1            ┆ 1      │\n",
       "│ 4527765      ┆ 1.0      ┆ 0.0      ┆ 0.0      ┆ 1            ┆ 1      │\n",
       "│ 5641681      ┆ 0.0      ┆ 1.0      ┆ 0.0      ┆ 1            ┆ 1      │\n",
       "│ 1003027      ┆ 0.0      ┆ 1.0      ┆ 0.0      ┆ 1            ┆ 1      │\n",
       "│ 2952316      ┆ 1.0      ┆ 0.0      ┆ 0.0      ┆ 1            ┆ 1      │\n",
       "│ 3453301      ┆ 0.0      ┆ 1.0      ┆ 0.0      ┆ 1            ┆ 1      │\n",
       "│ 1506897      ┆ 1.0      ┆ 0.0      ┆ 0.0      ┆ 1            ┆ 1      │\n",
       "│ 4070034      ┆ 1.0      ┆ 0.0      ┆ 0.0      ┆ 1            ┆ 1      │\n",
       "│ 3256419      ┆ 1.0      ┆ 0.0      ┆ 0.0      ┆ 1            ┆ 1      │\n",
       "│ …            ┆ …        ┆ …        ┆ …        ┆ …            ┆ …      │\n",
       "│ 5509544      ┆ 0.000901 ┆ 0.000227 ┆ 0.000266 ┆ 176474       ┆ 176474 │\n",
       "│ 2963783      ┆ 0.010951 ┆ 0.00022  ┆ 0.001118 ┆ 195054       ┆ 195054 │\n",
       "│ 4583912      ┆ 0.005541 ┆ 0.000178 ┆ 0.007193 ┆ 224511       ┆ 224511 │\n",
       "│ 4912182      ┆ 0.003531 ┆ 0.000272 ┆ 0.000052 ┆ 231387       ┆ 231387 │\n",
       "│ 162518       ┆ 0.003213 ┆ 0.001577 ┆ 0.000902 ┆ 237127       ┆ 237127 │\n",
       "│ 4867705      ┆ 0.002347 ┆ 0.000216 ┆ 0.000041 ┆ 245872       ┆ 245872 │\n",
       "│ 5929438      ┆ 0.000871 ┆ 0.005708 ┆ 0.005967 ┆ 312198       ┆ 312198 │\n",
       "│ 4632750      ┆ 0.000739 ┆ 0.000236 ┆ 0.000018 ┆ 334289       ┆ 334289 │\n",
       "│ 4861294      ┆ 0.001876 ┆ 0.000516 ┆ 0.000036 ┆ 412652       ┆ 412652 │\n",
       "│ 2217250      ┆ 0.00547  ┆ 0.000202 ┆ 0.00027  ┆ 499089       ┆ 499089 │\n",
       "└──────────────┴──────────┴──────────┴──────────┴──────────────┴────────┘"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp = tmp_df.filter(pl.col('sEH') < 0.1).sort('count')\n",
    "scatter_plot(tmp['sEH'], tmp['count'].log())\n",
    "tmp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "96c4af3f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "ade3c9a7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (98_415_610, 9)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>molecule</th><th>bb1</th><th>bb2</th><th>bb3</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>scaffold_idx</th><th>mol_per_scaf</th></tr><tr><td>str</td><td>u16</td><td>u16</td><td>u16</td><td>u8</td><td>u8</td><td>u8</td><td>i64</td><td>u32</td></tr></thead><tbody><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>765</td><td>0</td><td>0</td><td>0</td><td>4283326</td><td>843</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>205</td><td>0</td><td>0</td><td>0</td><td>4486646</td><td>9415</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1653</td><td>0</td><td>0</td><td>0</td><td>1015728</td><td>9006</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>146</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>439</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>196</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>253</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1219</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>604</td><td>0</td><td>0</td><td>0</td><td>543172</td><td>842</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>121</td><td>0</td><td>0</td><td>0</td><td>2571428</td><td>864</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>536</td><td>0</td><td>0</td><td>0</td><td>4068980</td><td>274</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1367</td><td>0</td><td>0</td><td>0</td><td>2088996</td><td>20</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>329</td><td>0</td><td>0</td><td>0</td><td>3192765</td><td>260</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>757</td><td>0</td><td>0</td><td>0</td><td>4085046</td><td>20</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>335</td><td>0</td><td>0</td><td>0</td><td>4087757</td><td>20</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1792</td><td>0</td><td>0</td><td>0</td><td>4674025</td><td>20</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1178</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1699</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1308</td><td>0</td><td>0</td><td>0</td><td>3617024</td><td>20</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1415</td><td>0</td><td>0</td><td>0</td><td>268351</td><td>20</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (98_415_610, 9)\n",
       "┌────────────────────────────┬──────┬──────┬──────┬──────┬─────┬─────┬──────────────┬──────────────┐\n",
       "│ molecule                   ┆ bb1  ┆ bb2  ┆ bb3  ┆ BRD4 ┆ HSA ┆ sEH ┆ scaffold_idx ┆ mol_per_scaf │\n",
       "│ ---                        ┆ ---  ┆ ---  ┆ ---  ┆ ---  ┆ --- ┆ --- ┆ ---          ┆ ---          │\n",
       "│ str                        ┆ u16  ┆ u16  ┆ u16  ┆ u8   ┆ u8  ┆ u8  ┆ i64          ┆ u32          │\n",
       "╞════════════════════════════╪══════╪══════╪══════╪══════╪═════╪═════╪══════════════╪══════════════╡\n",
       "│ C#CCOc1ccc(CNc2nc(NCC3CCCN ┆ 1640 ┆ 1653 ┆ 765  ┆ 0    ┆ 0   ┆ 0   ┆ 4283326      ┆ 843          │\n",
       "│ 3c3ccc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCc3cccc ┆ 1640 ┆ 1653 ┆ 205  ┆ 0    ┆ 0   ┆ 0   ┆ 4486646      ┆ 9415         │\n",
       "│ (Br)n3…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCc3ccc( ┆ 1640 ┆ 1653 ┆ 1653 ┆ 0    ┆ 0   ┆ 0   ┆ 1015728      ┆ 9006         │\n",
       "│ OCC#C)…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCCNC(=O ┆ 1640 ┆ 1653 ┆ 146  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        │\n",
       "│ )C(=C)…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCC(=O)N ┆ 1640 ┆ 1653 ┆ 439  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        │\n",
       "│ CC=C)n…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCC(C)OC ┆ 1640 ┆ 1653 ┆ 196  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        │\n",
       "│ C=C)nc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCCCOCC= ┆ 1640 ┆ 1653 ┆ 253  ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        │\n",
       "│ C)nc(N…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(NCCOCC=C ┆ 1640 ┆ 1653 ┆ 1219 ┆ 0    ┆ 0   ┆ 0   ┆ 5301385      ┆ 26751        │\n",
       "│ )nc(N[…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(Nc3cnn(C ┆ 1640 ┆ 1653 ┆ 604  ┆ 0    ┆ 0   ┆ 0   ┆ 543172       ┆ 842          │\n",
       "│ 4CCN(C…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ C#CCOc1ccc(CNc2nc(Nc3n[nH] ┆ 1640 ┆ 1653 ┆ 121  ┆ 0    ┆ 0   ┆ 0   ┆ 2571428      ┆ 864          │\n",
       "│ c4c3CN…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ …                          ┆ …    ┆ …    ┆ …    ┆ …    ┆ …   ┆ …   ┆ …            ┆ …            │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 536  ┆ 0    ┆ 0   ┆ 0   ┆ 4068980      ┆ 274          │\n",
       "│ Nc2ncc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1367 ┆ 0    ┆ 0   ┆ 0   ┆ 2088996      ┆ 20           │\n",
       "│ Nc2ncn…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 329  ┆ 0    ┆ 0   ┆ 0   ┆ 3192765      ┆ 260          │\n",
       "│ Nc2ncn…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 757  ┆ 0    ┆ 0   ┆ 0   ┆ 4085046      ┆ 20           │\n",
       "│ Nc2noc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 335  ┆ 0    ┆ 0   ┆ 0   ┆ 4087757      ┆ 20           │\n",
       "│ Nc2ncn…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1792 ┆ 0    ┆ 0   ┆ 0   ┆ 4674025      ┆ 20           │\n",
       "│ Nc2noc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1178 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      ┆ 60           │\n",
       "│ Nc2nnc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1699 ┆ 0    ┆ 0   ┆ 0   ┆ 5240137      ┆ 60           │\n",
       "│ Nc2nnc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1308 ┆ 0    ┆ 0   ┆ 0   ┆ 3617024      ┆ 20           │\n",
       "│ Nc2nnn…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "│ [N-]=[N+]=NCCC[C@H](Nc1nc( ┆ 141  ┆ 1415 ┆ 1415 ┆ 0    ┆ 0   ┆ 0   ┆ 268351       ┆ 20           │\n",
       "│ Nc2noc…                    ┆      ┆      ┆      ┆      ┆     ┆     ┆              ┆              │\n",
       "└────────────────────────────┴──────┴──────┴──────┴──────┴─────┴─────┴──────────────┴──────────────┘"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7c3a14a3",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "39e817f1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (98_415_610, 14)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>molecule</th><th>bb1</th><th>bb2</th><th>bb3</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>scaffold_idx</th><th>mol_per_scaf</th><th>label</th><th>pos_BRD4_mol_per_scaf</th><th>pos_HSA_mol_per_scaf</th><th>pos_sEH_mol_per_scaf</th><th>pos_mol_per_scaf</th></tr><tr><td>str</td><td>u16</td><td>u16</td><td>u16</td><td>u8</td><td>u8</td><td>u8</td><td>i64</td><td>u32</td><td>u8</td><td>i64</td><td>i64</td><td>i64</td><td>i64</td></tr></thead><tbody><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>765</td><td>0</td><td>0</td><td>0</td><td>4283326</td><td>843</td><td>0</td><td>41</td><td>3</td><td>1</td><td>45</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>205</td><td>0</td><td>0</td><td>0</td><td>4486646</td><td>9415</td><td>0</td><td>98</td><td>1</td><td>11</td><td>110</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1653</td><td>0</td><td>0</td><td>0</td><td>1015728</td><td>9006</td><td>0</td><td>190</td><td>1</td><td>74</td><td>265</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>146</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0</td><td>635</td><td>4</td><td>146</td><td>780</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>439</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0</td><td>635</td><td>4</td><td>146</td><td>780</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>196</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0</td><td>635</td><td>4</td><td>146</td><td>780</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>253</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0</td><td>635</td><td>4</td><td>146</td><td>780</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1219</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0</td><td>635</td><td>4</td><td>146</td><td>780</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>604</td><td>0</td><td>0</td><td>0</td><td>543172</td><td>842</td><td>0</td><td>25</td><td>2</td><td>0</td><td>27</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>121</td><td>0</td><td>0</td><td>0</td><td>2571428</td><td>864</td><td>0</td><td>9</td><td>0</td><td>0</td><td>9</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>536</td><td>0</td><td>0</td><td>0</td><td>4068980</td><td>274</td><td>0</td><td>0</td><td>1</td><td>0</td><td>1</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1367</td><td>0</td><td>0</td><td>0</td><td>2088996</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>329</td><td>0</td><td>0</td><td>0</td><td>3192765</td><td>260</td><td>0</td><td>0</td><td>1</td><td>0</td><td>1</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>757</td><td>0</td><td>0</td><td>0</td><td>4085046</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>335</td><td>0</td><td>0</td><td>0</td><td>4087757</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1792</td><td>0</td><td>0</td><td>0</td><td>4674025</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1178</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1699</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1308</td><td>0</td><td>0</td><td>0</td><td>3617024</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1415</td><td>0</td><td>0</td><td>0</td><td>268351</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (98_415_610, 14)\n",
       "┌─────┬─────┬─────┬─────┬─────┬─────┬─────┬────────┬───────┬───────┬───────┬───────┬───────┬───────┐\n",
       "│ mol ┆ bb1 ┆ bb2 ┆ bb3 ┆ BRD ┆ HSA ┆ sEH ┆ scaffo ┆ mol_p ┆ label ┆ pos_B ┆ pos_H ┆ pos_s ┆ pos_m │\n",
       "│ ecu ┆ --- ┆ --- ┆ --- ┆ 4   ┆ --- ┆ --- ┆ ld_idx ┆ er_sc ┆ ---   ┆ RD4_m ┆ SA_mo ┆ EH_mo ┆ ol_pe │\n",
       "│ le  ┆ u16 ┆ u16 ┆ u16 ┆ --- ┆ u8  ┆ u8  ┆ ---    ┆ af    ┆ u8    ┆ ol_pe ┆ l_per ┆ l_per ┆ r_sca │\n",
       "│ --- ┆     ┆     ┆     ┆ u8  ┆     ┆     ┆ i64    ┆ ---   ┆       ┆ r_sca ┆ _scaf ┆ _scaf ┆ f     │\n",
       "│ str ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆ u32   ┆       ┆ f     ┆ ---   ┆ ---   ┆ ---   │\n",
       "│     ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆ ---   ┆ i64   ┆ i64   ┆ i64   │\n",
       "│     ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆ i64   ┆       ┆       ┆       │\n",
       "╞═════╪═════╪═════╪═════╪═════╪═════╪═════╪════════╪═══════╪═══════╪═══════╪═══════╪═══════╪═══════╡\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 765 ┆ 0   ┆ 0   ┆ 0   ┆ 428332 ┆ 843   ┆ 0     ┆ 41    ┆ 3     ┆ 1     ┆ 45    │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 6      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 3CC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CN3 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c3c ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ cc… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 205 ┆ 0   ┆ 0   ┆ 0   ┆ 448664 ┆ 9415  ┆ 0     ┆ 98    ┆ 1     ┆ 11    ┆ 110   │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 6      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ NCc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 3cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ cc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Br) ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ n3… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 165 ┆ 0   ┆ 0   ┆ 0   ┆ 101572 ┆ 9006  ┆ 0     ┆ 190   ┆ 1     ┆ 74    ┆ 265   │\n",
       "│ COc ┆ 0   ┆ 3   ┆ 3   ┆     ┆     ┆     ┆ 8      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ NCc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 3cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(O ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC# ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C)… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 146 ┆ 0   ┆ 0   ┆ 0   ┆ 530138 ┆ 26751 ┆ 0     ┆ 635   ┆ 4     ┆ 146   ┆ 780   │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 5      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ NC( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =O) ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C(= ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C)… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 439 ┆ 0   ┆ 0   ┆ 0   ┆ 530138 ┆ 26751 ┆ 0     ┆ 635   ┆ 4     ┆ 146   ┆ 780   │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 5      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ (=O ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ )NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C=C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ )n… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 196 ┆ 0   ┆ 0   ┆ 0   ┆ 530138 ┆ 26751 ┆ 0     ┆ 635   ┆ 4     ┆ 146   ┆ 780   │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 5      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ (C) ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ OCC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =C) ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 253 ┆ 0   ┆ 0   ┆ 0   ┆ 530138 ┆ 26751 ┆ 0     ┆ 635   ┆ 4     ┆ 146   ┆ 780   │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 5      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ COC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C=C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ )nc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ (N… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 121 ┆ 0   ┆ 0   ┆ 0   ┆ 530138 ┆ 26751 ┆ 0     ┆ 635   ┆ 4     ┆ 146   ┆ 780   │\n",
       "│ COc ┆ 0   ┆ 3   ┆ 9   ┆     ┆     ┆     ┆ 5      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ OCC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =C) ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N[… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 604 ┆ 0   ┆ 0   ┆ 0   ┆ 543172 ┆ 842   ┆ 0     ┆ 25    ┆ 2     ┆ 0     ┆ 27    │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc3 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ cnn ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ (C4 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CCN ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ (C… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 121 ┆ 0   ┆ 0   ┆ 0   ┆ 257142 ┆ 864   ┆ 0     ┆ 9     ┆ 0     ┆ 0     ┆ 9     │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 8      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ Nc3 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ n[n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ H]c ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ 4c3 ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CN… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …      ┆ …     ┆ …     ┆ …     ┆ …     ┆ …     ┆ …     │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 536 ┆ 0   ┆ 0   ┆ 0   ┆ 406898 ┆ 274   ┆ 0     ┆ 0     ┆ 1     ┆ 0     ┆ 1     │\n",
       "│ ]=[ ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆ 0      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ cc… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 136 ┆ 0   ┆ 0   ┆ 0   ┆ 208899 ┆ 20    ┆ 0     ┆ 0     ┆ 0     ┆ 0     ┆ 0     │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 7   ┆     ┆     ┆     ┆ 6      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ cn… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 329 ┆ 0   ┆ 0   ┆ 0   ┆ 319276 ┆ 260   ┆ 0     ┆ 0     ┆ 1     ┆ 0     ┆ 1     │\n",
       "│ ]=[ ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆ 5      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ cn… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 757 ┆ 0   ┆ 0   ┆ 0   ┆ 408504 ┆ 20    ┆ 0     ┆ 0     ┆ 0     ┆ 0     ┆ 0     │\n",
       "│ ]=[ ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆ 6      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ oc… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 335 ┆ 0   ┆ 0   ┆ 0   ┆ 408775 ┆ 20    ┆ 0     ┆ 0     ┆ 0     ┆ 0     ┆ 0     │\n",
       "│ ]=[ ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆ 7      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ cn… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 179 ┆ 0   ┆ 0   ┆ 0   ┆ 467402 ┆ 20    ┆ 0     ┆ 0     ┆ 0     ┆ 0     ┆ 0     │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 2   ┆     ┆     ┆     ┆ 5      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ oc… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 117 ┆ 0   ┆ 0   ┆ 0   ┆ 524013 ┆ 60    ┆ 0     ┆ 0     ┆ 0     ┆ 0     ┆ 0     │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 8   ┆     ┆     ┆     ┆ 7      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 169 ┆ 0   ┆ 0   ┆ 0   ┆ 524013 ┆ 60    ┆ 0     ┆ 0     ┆ 0     ┆ 0     ┆ 0     │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 9   ┆     ┆     ┆     ┆ 7      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nc… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 130 ┆ 0   ┆ 0   ┆ 0   ┆ 361702 ┆ 20    ┆ 0     ┆ 0     ┆ 0     ┆ 0     ┆ 0     │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 8   ┆     ┆     ┆     ┆ 4      ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ nn… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 141 ┆ 0   ┆ 0   ┆ 0   ┆ 268351 ┆ 20    ┆ 0     ┆ 0     ┆ 0     ┆ 0     ┆ 0     │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 5   ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "│ oc… ┆     ┆     ┆     ┆     ┆     ┆     ┆        ┆       ┆       ┆       ┆       ┆       ┆       │\n",
       "└─────┴─────┴─────┴─────┴─────┴─────┴─────┴────────┴───────┴───────┴───────┴───────┴───────┴───────┘"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = train_df.with_columns(\n",
    "    pl.any_horizontal('BRD4', 'HSA', 'sEH').cast(pl.UInt8).alias('label')).with_columns(\n",
    "    pl.col('BRD4').sum().over('scaffold_idx').alias('pos_BRD4_mol_per_scaf'),\n",
    "    pl.col('HSA').sum().over('scaffold_idx').alias('pos_HSA_mol_per_scaf'),\n",
    "    pl.col('sEH').sum().over('scaffold_idx').alias('pos_sEH_mol_per_scaf'),\n",
    "    pl.col('label').sum().over('scaffold_idx').alias('pos_mol_per_scaf'),\n",
    ")\n",
    "\n",
    "train_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "8503ee6d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (1, 2)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>molecule</th><th>scaffold_idx</th></tr><tr><td>u32</td><td>u32</td></tr></thead><tbody><tr><td>42695137</td><td>5410825</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (1, 2)\n",
       "┌──────────┬──────────────┐\n",
       "│ molecule ┆ scaffold_idx │\n",
       "│ ---      ┆ ---          │\n",
       "│ u32      ┆ u32          │\n",
       "╞══════════╪══════════════╡\n",
       "│ 42695137 ┆ 5410825      │\n",
       "└──────────┴──────────────┘"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.filter(pl.col('pos_mol_per_scaf') == 0).select(pl.n_unique('molecule', 'scaffold_idx'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "edaff8e5",
   "metadata": {},
   "source": [
    "5410825/5711873 (94.7294%) scaffolds are all negative"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "00ca1ab5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([5.41023e+06, 4.26000e+02, 1.24000e+02, 2.60000e+01, 6.00000e+00,\n",
       "        1.10000e+01, 1.00000e+00, 0.00000e+00, 0.00000e+00, 1.00000e+00]),\n",
       " array([1.00000e+00, 1.47270e+03, 2.94440e+03, 4.41610e+03, 5.88780e+03,\n",
       "        7.35950e+03, 8.83120e+03, 1.03029e+04, 1.17746e+04, 1.32463e+04,\n",
       "        1.47180e+04]),\n",
       " <BarContainer object of 10 artists>)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tmp_df = train_df.filter(pl.col('pos_mol_per_scaf') == 0).group_by('scaffold_idx').agg(pl.col('mol_per_scaf').first())\n",
    "plt.hist(tmp_df['mol_per_scaf'], log = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "34ccac67",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(7.890688942998526, 2.0, 14718)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp_df['mol_per_scaf'].mean(), tmp_df['mol_per_scaf'].median(), tmp_df['mol_per_scaf'].max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "815409b2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "e760b69e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (301_048, 3)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>scaffold_idx</th><th>mol_per_scaf</th><th>pos_mol_per_scaf</th></tr><tr><td>i64</td><td>u32</td><td>i64</td></tr></thead><tbody><tr><td>3242527</td><td>94</td><td>1</td></tr><tr><td>422340</td><td>36</td><td>1</td></tr><tr><td>991752</td><td>11</td><td>1</td></tr><tr><td>4767770</td><td>4</td><td>1</td></tr><tr><td>2009064</td><td>592</td><td>5</td></tr><tr><td>5045402</td><td>140</td><td>4</td></tr><tr><td>3305967</td><td>5</td><td>1</td></tr><tr><td>4224707</td><td>3</td><td>1</td></tr><tr><td>5945787</td><td>4</td><td>1</td></tr><tr><td>1215638</td><td>60</td><td>8</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>1033434</td><td>28</td><td>1</td></tr><tr><td>5550052</td><td>781</td><td>7</td></tr><tr><td>5384238</td><td>5</td><td>2</td></tr><tr><td>2775486</td><td>5</td><td>2</td></tr><tr><td>4915101</td><td>5</td><td>1</td></tr><tr><td>2887295</td><td>87</td><td>3</td></tr><tr><td>3739707</td><td>44</td><td>1</td></tr><tr><td>2767489</td><td>366</td><td>2</td></tr><tr><td>2260341</td><td>2</td><td>1</td></tr><tr><td>1075902</td><td>7</td><td>1</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (301_048, 3)\n",
       "┌──────────────┬──────────────┬──────────────────┐\n",
       "│ scaffold_idx ┆ mol_per_scaf ┆ pos_mol_per_scaf │\n",
       "│ ---          ┆ ---          ┆ ---              │\n",
       "│ i64          ┆ u32          ┆ i64              │\n",
       "╞══════════════╪══════════════╪══════════════════╡\n",
       "│ 3242527      ┆ 94           ┆ 1                │\n",
       "│ 422340       ┆ 36           ┆ 1                │\n",
       "│ 991752       ┆ 11           ┆ 1                │\n",
       "│ 4767770      ┆ 4            ┆ 1                │\n",
       "│ 2009064      ┆ 592          ┆ 5                │\n",
       "│ 5045402      ┆ 140          ┆ 4                │\n",
       "│ 3305967      ┆ 5            ┆ 1                │\n",
       "│ 4224707      ┆ 3            ┆ 1                │\n",
       "│ 5945787      ┆ 4            ┆ 1                │\n",
       "│ 1215638      ┆ 60           ┆ 8                │\n",
       "│ …            ┆ …            ┆ …                │\n",
       "│ 1033434      ┆ 28           ┆ 1                │\n",
       "│ 5550052      ┆ 781          ┆ 7                │\n",
       "│ 5384238      ┆ 5            ┆ 2                │\n",
       "│ 2775486      ┆ 5            ┆ 2                │\n",
       "│ 4915101      ┆ 5            ┆ 1                │\n",
       "│ 2887295      ┆ 87           ┆ 3                │\n",
       "│ 3739707      ┆ 44           ┆ 1                │\n",
       "│ 2767489      ┆ 366          ┆ 2                │\n",
       "│ 2260341      ┆ 2            ┆ 1                │\n",
       "│ 1075902      ┆ 7            ┆ 1                │\n",
       "└──────────────┴──────────────┴──────────────────┘"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp_df = train_df.filter(pl.col('pos_mol_per_scaf') > 0).group_by('scaffold_idx').agg(pl.col('mol_per_scaf').first(),\n",
    "                                                                                     pl.col('pos_mol_per_scaf').first())\n",
    "tmp_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "33c1e87e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([3.00941e+05, 7.30000e+01, 1.80000e+01, 6.00000e+00, 5.00000e+00,\n",
       "        1.00000e+00, 1.00000e+00, 1.00000e+00, 1.00000e+00, 1.00000e+00]),\n",
       " array([1.0000e+00, 6.5230e+02, 1.3036e+03, 1.9549e+03, 2.6062e+03,\n",
       "        3.2575e+03, 3.9088e+03, 4.5601e+03, 5.2114e+03, 5.8627e+03,\n",
       "        6.5140e+03]),\n",
       " <BarContainer object of 10 artists>)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(tmp_df['pos_mol_per_scaf'], log = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "fd024839",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([129212.,  49459.,  22683.,  41576.,  22827.,   1718.,   1936.,\n",
       "          1262.,    420.,  29955.]),\n",
       " array([9.28246542e-05, 1.00083542e-01, 2.00074260e-01, 3.00064977e-01,\n",
       "        4.00055695e-01, 5.00046412e-01, 6.00037130e-01, 7.00027847e-01,\n",
       "        8.00018565e-01, 9.00009282e-01, 1.00000000e+00]),\n",
       " <BarContainer object of 10 artists>)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(tmp_df['pos_mol_per_scaf'] / tmp_df['mol_per_scaf'], log = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "59641e6c",
   "metadata": {},
   "outputs": [],
   "source": [
    "tmp_df = tmp_df.with_columns((pl.col('pos_mol_per_scaf') / pl.col('mol_per_scaf')).alias('pos_ratio_per_scaf'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "3e80ace6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (1, 4)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>scaffold_idx</th><th>mol_per_scaf</th><th>pos_mol_per_scaf</th><th>pos_ratio_per_scaf</th></tr><tr><td>i64</td><td>u32</td><td>i64</td><td>f64</td></tr></thead><tbody><tr><td>5114213</td><td>65818</td><td>6514</td><td>0.09897</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (1, 4)\n",
       "┌──────────────┬──────────────┬──────────────────┬────────────────────┐\n",
       "│ scaffold_idx ┆ mol_per_scaf ┆ pos_mol_per_scaf ┆ pos_ratio_per_scaf │\n",
       "│ ---          ┆ ---          ┆ ---              ┆ ---                │\n",
       "│ i64          ┆ u32          ┆ i64              ┆ f64                │\n",
       "╞══════════════╪══════════════╪══════════════════╪════════════════════╡\n",
       "│ 5114213      ┆ 65818        ┆ 6514             ┆ 0.09897            │\n",
       "└──────────────┴──────────────┴──────────────────┴────────────────────┘"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp_df.filter(pl.col('pos_mol_per_scaf') == 6514)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "385ae939",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(301048, 4)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "06f2cc59",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.015340848875498511"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df['label'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "82314fc4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (13, 4)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>scaffold_idx</th><th>mol_per_scaf</th><th>pos_mol_per_scaf</th><th>pos_ratio_per_scaf</th></tr><tr><td>i64</td><td>u32</td><td>i64</td><td>f64</td></tr></thead><tbody><tr><td>897637</td><td>87</td><td>82</td><td>0.942529</td></tr><tr><td>2671746</td><td>86</td><td>78</td><td>0.906977</td></tr><tr><td>5689881</td><td>87</td><td>81</td><td>0.931034</td></tr><tr><td>3951295</td><td>87</td><td>86</td><td>0.988506</td></tr><tr><td>5051066</td><td>97</td><td>88</td><td>0.907216</td></tr><tr><td>2217833</td><td>50</td><td>45</td><td>0.9</td></tr><tr><td>267005</td><td>97</td><td>88</td><td>0.907216</td></tr><tr><td>4075663</td><td>87</td><td>83</td><td>0.954023</td></tr><tr><td>1524870</td><td>87</td><td>80</td><td>0.91954</td></tr><tr><td>4339037</td><td>87</td><td>87</td><td>1.0</td></tr><tr><td>5664790</td><td>87</td><td>79</td><td>0.908046</td></tr><tr><td>4345605</td><td>87</td><td>80</td><td>0.91954</td></tr><tr><td>669621</td><td>87</td><td>84</td><td>0.965517</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (13, 4)\n",
       "┌──────────────┬──────────────┬──────────────────┬────────────────────┐\n",
       "│ scaffold_idx ┆ mol_per_scaf ┆ pos_mol_per_scaf ┆ pos_ratio_per_scaf │\n",
       "│ ---          ┆ ---          ┆ ---              ┆ ---                │\n",
       "│ i64          ┆ u32          ┆ i64              ┆ f64                │\n",
       "╞══════════════╪══════════════╪══════════════════╪════════════════════╡\n",
       "│ 897637       ┆ 87           ┆ 82               ┆ 0.942529           │\n",
       "│ 2671746      ┆ 86           ┆ 78               ┆ 0.906977           │\n",
       "│ 5689881      ┆ 87           ┆ 81               ┆ 0.931034           │\n",
       "│ 3951295      ┆ 87           ┆ 86               ┆ 0.988506           │\n",
       "│ 5051066      ┆ 97           ┆ 88               ┆ 0.907216           │\n",
       "│ 2217833      ┆ 50           ┆ 45               ┆ 0.9                │\n",
       "│ 267005       ┆ 97           ┆ 88               ┆ 0.907216           │\n",
       "│ 4075663      ┆ 87           ┆ 83               ┆ 0.954023           │\n",
       "│ 1524870      ┆ 87           ┆ 80               ┆ 0.91954            │\n",
       "│ 4339037      ┆ 87           ┆ 87               ┆ 1.0                │\n",
       "│ 5664790      ┆ 87           ┆ 79               ┆ 0.908046           │\n",
       "│ 4345605      ┆ 87           ┆ 80               ┆ 0.91954            │\n",
       "│ 669621       ┆ 87           ┆ 84               ┆ 0.965517           │\n",
       "└──────────────┴──────────────┴──────────────────┴────────────────────┘"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp_df.filter((pl.col('pos_ratio_per_scaf') >= 0.9) & (pl.col('mol_per_scaf') >= 50))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "06c872c8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5.015077329861019,\n",
       " 1.0,\n",
       " 6514,\n",
       " 185.08833475060456,\n",
       " 13.0,\n",
       " 499089,\n",
       " 0.2627146191906659,\n",
       " 0.16666666666666666)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp_df['pos_mol_per_scaf'].mean(), \\\n",
    "tmp_df['pos_mol_per_scaf'].median(), \\\n",
    "tmp_df['pos_mol_per_scaf'].max(), \\\n",
    "tmp_df['mol_per_scaf'].mean(), \\\n",
    "tmp_df['mol_per_scaf'].median(), \\\n",
    "tmp_df['mol_per_scaf'].max(), \\\n",
    "tmp_df['pos_ratio_per_scaf'].mean(), \\\n",
    "tmp_df['pos_ratio_per_scaf'].median()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1dc10234",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f9891f0e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "7203864a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fc6317dece0>]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import math\n",
    "\n",
    "    \n",
    "def make_intra_group_pos_weight(pos_mol_per_scaf, mol_per_scaf, min_pos_ratio = 0.1, max_pos_ratio = 0.9):\n",
    "    pos_ratio = pos_mol_per_scaf / mol_per_scaf\n",
    "    if pos_ratio == 0:\n",
    "        return 0.0\n",
    "    elif pos_ratio == 1:\n",
    "        return 1.0\n",
    "    else:\n",
    "        return min(max(pos_ratio, min_pos_ratio), max_pos_ratio)\n",
    "    \n",
    "    \n",
    "def make_neg_inter_group_weight(mol_per_scaf, base = 7.890688942998526, mode = 'poly', log_base = None, poly_degree = 0.5):\n",
    "    if mol_per_scaf <= base:\n",
    "        return mol_per_scaf\n",
    "    if mode == 'log':\n",
    "        log_base = log_base or math.e\n",
    "        return base * (math.log(mol_per_scaf / base, log_base) + 1)\n",
    "    elif mode == 'poly':\n",
    "        # 1.0 is linear\n",
    "        poly_degree = poly_degree or 1.0\n",
    "        return base * (mol_per_scaf / base) ** poly_degree\n",
    "    else:\n",
    "        raise ValueError\n",
    "        \n",
    "\n",
    "def make_inter_group_weight(pos_mol_per_scaf, mol_per_scaf):\n",
    "    if pos_mol_per_scaf == 0:\n",
    "        return make_neg_inter_group_weight(mol_per_scaf, base = 7.890688942998526, mode = 'poly', poly_degree = 0.5)\n",
    "    else:\n",
    "        # keep all pos\n",
    "        intra_pos_weight = make_intra_group_pos_weight(pos_mol_per_scaf, mol_per_scaf)\n",
    "        return pos_mol_per_scaf / intra_pos_weight\n",
    "\n",
    "x = np.arange(1, 14000)\n",
    "y = [make_neg_inter_group_weight(e, mode = 'poly', poly_degree = 0.5) for e in x]\n",
    "plt.plot(x, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a830df32",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4959fe99",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3299d2f6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "08b3e422",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "6a17602e",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df = train_df.with_columns(pl.struct(['pos_mol_per_scaf', 'mol_per_scaf']).map_elements(\n",
    "    lambda x: make_inter_group_weight(x['pos_mol_per_scaf'], x['mol_per_scaf']), return_dtype = pl.Float32).alias('inter_group_weight'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "23dfd656",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df = train_df.with_columns(pl.struct(['pos_mol_per_scaf', 'mol_per_scaf']).map_elements(\n",
    "    lambda x: make_intra_group_pos_weight(x['pos_mol_per_scaf'], x['mol_per_scaf']), return_dtype = pl.Float32).alias('intra_group_weight'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "55777a8a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (98_415_610, 17)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>molecule</th><th>bb1</th><th>bb2</th><th>bb3</th><th>BRD4</th><th>HSA</th><th>sEH</th><th>scaffold_idx</th><th>mol_per_scaf</th><th>label</th><th>pos_BRD4_mol_per_scaf</th><th>pos_HSA_mol_per_scaf</th><th>pos_sEH_mol_per_scaf</th><th>pos_mol_per_scaf</th><th>inter_group_weight</th><th>intra_group_weight</th><th>sample_weight</th></tr><tr><td>str</td><td>u16</td><td>u16</td><td>u16</td><td>u8</td><td>u8</td><td>u8</td><td>i64</td><td>u32</td><td>u8</td><td>i64</td><td>i64</td><td>i64</td><td>i64</td><td>f32</td><td>f32</td><td>f64</td></tr></thead><tbody><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>765</td><td>0</td><td>0</td><td>0</td><td>4283326</td><td>843</td><td>0</td><td>41</td><td>3</td><td>1</td><td>45</td><td>450.0</td><td>0.1</td><td>0.507519</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>205</td><td>0</td><td>0</td><td>0</td><td>4486646</td><td>9415</td><td>0</td><td>98</td><td>1</td><td>11</td><td>110</td><td>1100.0</td><td>0.1</td><td>0.106394</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1653</td><td>0</td><td>0</td><td>0</td><td>1015728</td><td>9006</td><td>0</td><td>190</td><td>1</td><td>74</td><td>265</td><td>2650.0</td><td>0.1</td><td>0.272852</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>146</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0</td><td>635</td><td>4</td><td>146</td><td>780</td><td>7800.0</td><td>0.1</td><td>0.270301</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>439</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0</td><td>635</td><td>4</td><td>146</td><td>780</td><td>7800.0</td><td>0.1</td><td>0.270301</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>196</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0</td><td>635</td><td>4</td><td>146</td><td>780</td><td>7800.0</td><td>0.1</td><td>0.270301</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>253</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0</td><td>635</td><td>4</td><td>146</td><td>780</td><td>7800.0</td><td>0.1</td><td>0.270301</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>1219</td><td>0</td><td>0</td><td>0</td><td>5301385</td><td>26751</td><td>0</td><td>635</td><td>4</td><td>146</td><td>780</td><td>7800.0</td><td>0.1</td><td>0.270301</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>604</td><td>0</td><td>0</td><td>0</td><td>543172</td><td>842</td><td>0</td><td>25</td><td>2</td><td>0</td><td>27</td><td>270.0</td><td>0.1</td><td>0.29816</td></tr><tr><td>&quot;C#CCOc1ccc(CNc…</td><td>1640</td><td>1653</td><td>121</td><td>0</td><td>0</td><td>0</td><td>2571428</td><td>864</td><td>0</td><td>9</td><td>0</td><td>0</td><td>9</td><td>90.0</td><td>0.1</td><td>0.094737</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>536</td><td>0</td><td>0</td><td>0</td><td>4068980</td><td>274</td><td>0</td><td>0</td><td>1</td><td>0</td><td>1</td><td>10.0</td><td>0.1</td><td>0.032967</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1367</td><td>0</td><td>0</td><td>0</td><td>2088996</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>12.562395</td><td>0.0</td><td>0.62812</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>329</td><td>0</td><td>0</td><td>0</td><td>3192765</td><td>260</td><td>0</td><td>0</td><td>1</td><td>0</td><td>1</td><td>10.0</td><td>0.1</td><td>0.034749</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>757</td><td>0</td><td>0</td><td>0</td><td>4085046</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>12.562395</td><td>0.0</td><td>0.62812</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>335</td><td>0</td><td>0</td><td>0</td><td>4087757</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>12.562395</td><td>0.0</td><td>0.62812</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1792</td><td>0</td><td>0</td><td>0</td><td>4674025</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>12.562395</td><td>0.0</td><td>0.62812</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1178</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>21.758707</td><td>0.0</td><td>0.362645</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1699</td><td>0</td><td>0</td><td>0</td><td>5240137</td><td>60</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>21.758707</td><td>0.0</td><td>0.362645</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1308</td><td>0</td><td>0</td><td>0</td><td>3617024</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>12.562395</td><td>0.0</td><td>0.62812</td></tr><tr><td>&quot;[N-]=[N+]=NCCC…</td><td>141</td><td>1415</td><td>1415</td><td>0</td><td>0</td><td>0</td><td>268351</td><td>20</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>12.562395</td><td>0.0</td><td>0.62812</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (98_415_610, 17)\n",
       "┌─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┬─────┐\n",
       "│ mol ┆ bb1 ┆ bb2 ┆ bb3 ┆ BRD ┆ HSA ┆ sEH ┆ sca ┆ mol ┆ lab ┆ pos ┆ pos ┆ pos ┆ pos ┆ int ┆ int ┆ sam │\n",
       "│ ecu ┆ --- ┆ --- ┆ --- ┆ 4   ┆ --- ┆ --- ┆ ffo ┆ _pe ┆ el  ┆ _BR ┆ _HS ┆ _sE ┆ _mo ┆ er_ ┆ ra_ ┆ ple │\n",
       "│ le  ┆ u16 ┆ u16 ┆ u16 ┆ --- ┆ u8  ┆ u8  ┆ ld_ ┆ r_s ┆ --- ┆ D4_ ┆ A_m ┆ H_m ┆ l_p ┆ gro ┆ gro ┆ _we │\n",
       "│ --- ┆     ┆     ┆     ┆ u8  ┆     ┆     ┆ idx ┆ caf ┆ u8  ┆ mol ┆ ol_ ┆ ol_ ┆ er_ ┆ up_ ┆ up_ ┆ igh │\n",
       "│ str ┆     ┆     ┆     ┆     ┆     ┆     ┆ --- ┆ --- ┆     ┆ _pe ┆ per ┆ per ┆ sca ┆ wei ┆ wei ┆ t   │\n",
       "│     ┆     ┆     ┆     ┆     ┆     ┆     ┆ i64 ┆ u32 ┆     ┆ r_s ┆ _sc ┆ _sc ┆ f   ┆ ght ┆ ght ┆ --- │\n",
       "│     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ caf ┆ af  ┆ af  ┆ --- ┆ --- ┆ --- ┆ f64 │\n",
       "│     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ --- ┆ --- ┆ --- ┆ i64 ┆ f32 ┆ f32 ┆     │\n",
       "│     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ i64 ┆ i64 ┆ i64 ┆     ┆     ┆     ┆     │\n",
       "╞═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╪═════╡\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 765 ┆ 0   ┆ 0   ┆ 0   ┆ 428 ┆ 843 ┆ 0   ┆ 41  ┆ 3   ┆ 1   ┆ 45  ┆ 450 ┆ 0.1 ┆ 0.5 │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 332 ┆     ┆     ┆     ┆     ┆     ┆     ┆ .0  ┆     ┆ 075 │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆ 6   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 19  │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ 3CC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CN3 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c3c ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ cc… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 205 ┆ 0   ┆ 0   ┆ 0   ┆ 448 ┆ 941 ┆ 0   ┆ 98  ┆ 1   ┆ 11  ┆ 110 ┆ 110 ┆ 0.1 ┆ 0.1 │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 664 ┆ 5   ┆     ┆     ┆     ┆     ┆     ┆ 0.0 ┆     ┆ 063 │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆ 6   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 94  │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ NCc ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ 3cc ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ cc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Br) ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ n3… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 165 ┆ 0   ┆ 0   ┆ 0   ┆ 101 ┆ 900 ┆ 0   ┆ 190 ┆ 1   ┆ 74  ┆ 265 ┆ 265 ┆ 0.1 ┆ 0.2 │\n",
       "│ COc ┆ 0   ┆ 3   ┆ 3   ┆     ┆     ┆     ┆ 572 ┆ 6   ┆     ┆     ┆     ┆     ┆     ┆ 0.0 ┆     ┆ 728 │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆ 8   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 52  │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ NCc ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ 3cc ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(O ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC# ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C)… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 146 ┆ 0   ┆ 0   ┆ 0   ┆ 530 ┆ 267 ┆ 0   ┆ 635 ┆ 4   ┆ 146 ┆ 780 ┆ 780 ┆ 0.1 ┆ 0.2 │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 138 ┆ 51  ┆     ┆     ┆     ┆     ┆     ┆ 0.0 ┆     ┆ 703 │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 01  │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ NC( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ =O) ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C(= ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C)… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 439 ┆ 0   ┆ 0   ┆ 0   ┆ 530 ┆ 267 ┆ 0   ┆ 635 ┆ 4   ┆ 146 ┆ 780 ┆ 780 ┆ 0.1 ┆ 0.2 │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 138 ┆ 51  ┆     ┆     ┆     ┆     ┆     ┆ 0.0 ┆     ┆ 703 │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 01  │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ (=O ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ )NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C=C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ )n… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 196 ┆ 0   ┆ 0   ┆ 0   ┆ 530 ┆ 267 ┆ 0   ┆ 635 ┆ 4   ┆ 146 ┆ 780 ┆ 780 ┆ 0.1 ┆ 0.2 │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 138 ┆ 51  ┆     ┆     ┆     ┆     ┆     ┆ 0.0 ┆     ┆ 703 │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 01  │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ (C) ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ OCC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ =C) ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 253 ┆ 0   ┆ 0   ┆ 0   ┆ 530 ┆ 267 ┆ 0   ┆ 635 ┆ 4   ┆ 146 ┆ 780 ┆ 780 ┆ 0.1 ┆ 0.2 │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 138 ┆ 51  ┆     ┆     ┆     ┆     ┆     ┆ 0.0 ┆     ┆ 703 │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 01  │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ COC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C=C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ )nc ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ (N… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 121 ┆ 0   ┆ 0   ┆ 0   ┆ 530 ┆ 267 ┆ 0   ┆ 635 ┆ 4   ┆ 146 ┆ 780 ┆ 780 ┆ 0.1 ┆ 0.2 │\n",
       "│ COc ┆ 0   ┆ 3   ┆ 9   ┆     ┆     ┆     ┆ 138 ┆ 51  ┆     ┆     ┆     ┆     ┆     ┆ 0.0 ┆     ┆ 703 │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 01  │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ NCC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ OCC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ =C) ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ N[… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 604 ┆ 0   ┆ 0   ┆ 0   ┆ 543 ┆ 842 ┆ 0   ┆ 25  ┆ 2   ┆ 0   ┆ 27  ┆ 270 ┆ 0.1 ┆ 0.2 │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 172 ┆     ┆     ┆     ┆     ┆     ┆     ┆ .0  ┆     ┆ 981 │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 6   │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc3 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ cnn ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ (C4 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CCN ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ (C… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C#C ┆ 164 ┆ 165 ┆ 121 ┆ 0   ┆ 0   ┆ 0   ┆ 257 ┆ 864 ┆ 0   ┆ 9   ┆ 0   ┆ 0   ┆ 9   ┆ 90. ┆ 0.1 ┆ 0.0 │\n",
       "│ COc ┆ 0   ┆ 3   ┆     ┆     ┆     ┆     ┆ 142 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 0   ┆     ┆ 947 │\n",
       "│ 1cc ┆     ┆     ┆     ┆     ┆     ┆     ┆ 8   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 37  │\n",
       "│ c(C ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc2 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc( ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ Nc3 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ n[n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ H]c ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ 4c3 ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CN… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   ┆ …   │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 536 ┆ 0   ┆ 0   ┆ 0   ┆ 406 ┆ 274 ┆ 0   ┆ 0   ┆ 1   ┆ 0   ┆ 1   ┆ 10. ┆ 0.1 ┆ 0.0 │\n",
       "│ ]=[ ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆ 898 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 0   ┆     ┆ 329 │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆ 0   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 67  │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ cc… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 136 ┆ 0   ┆ 0   ┆ 0   ┆ 208 ┆ 20  ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 12. ┆ 0.0 ┆ 0.6 │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 7   ┆     ┆     ┆     ┆ 899 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 562 ┆     ┆ 281 │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆ 6   ┆     ┆     ┆     ┆     ┆     ┆     ┆ 395 ┆     ┆ 2   │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ cn… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 329 ┆ 0   ┆ 0   ┆ 0   ┆ 319 ┆ 260 ┆ 0   ┆ 0   ┆ 1   ┆ 0   ┆ 1   ┆ 10. ┆ 0.1 ┆ 0.0 │\n",
       "│ ]=[ ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆ 276 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 0   ┆     ┆ 347 │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 49  │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ cn… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 757 ┆ 0   ┆ 0   ┆ 0   ┆ 408 ┆ 20  ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 12. ┆ 0.0 ┆ 0.6 │\n",
       "│ ]=[ ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆ 504 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 562 ┆     ┆ 281 │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆ 6   ┆     ┆     ┆     ┆     ┆     ┆     ┆ 395 ┆     ┆ 2   │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ oc… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 335 ┆ 0   ┆ 0   ┆ 0   ┆ 408 ┆ 20  ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 12. ┆ 0.0 ┆ 0.6 │\n",
       "│ ]=[ ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆ 775 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 562 ┆     ┆ 281 │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆ 7   ┆     ┆     ┆     ┆     ┆     ┆     ┆ 395 ┆     ┆ 2   │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ cn… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 179 ┆ 0   ┆ 0   ┆ 0   ┆ 467 ┆ 20  ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 12. ┆ 0.0 ┆ 0.6 │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 2   ┆     ┆     ┆     ┆ 402 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 562 ┆     ┆ 281 │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆ 5   ┆     ┆     ┆     ┆     ┆     ┆     ┆ 395 ┆     ┆ 2   │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ oc… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 117 ┆ 0   ┆ 0   ┆ 0   ┆ 524 ┆ 60  ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 21. ┆ 0.0 ┆ 0.3 │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 8   ┆     ┆     ┆     ┆ 013 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 758 ┆     ┆ 626 │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆ 7   ┆     ┆     ┆     ┆     ┆     ┆     ┆ 707 ┆     ┆ 45  │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 169 ┆ 0   ┆ 0   ┆ 0   ┆ 524 ┆ 60  ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 21. ┆ 0.0 ┆ 0.3 │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 9   ┆     ┆     ┆     ┆ 013 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 758 ┆     ┆ 626 │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆ 7   ┆     ┆     ┆     ┆     ┆     ┆     ┆ 707 ┆     ┆ 45  │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nc… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 130 ┆ 0   ┆ 0   ┆ 0   ┆ 361 ┆ 20  ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 12. ┆ 0.0 ┆ 0.6 │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 8   ┆     ┆     ┆     ┆ 702 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 562 ┆     ┆ 281 │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆ 4   ┆     ┆     ┆     ┆     ┆     ┆     ┆ 395 ┆     ┆ 2   │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ nn… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ [N- ┆ 141 ┆ 141 ┆ 141 ┆ 0   ┆ 0   ┆ 0   ┆ 268 ┆ 20  ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 0   ┆ 12. ┆ 0.0 ┆ 0.6 │\n",
       "│ ]=[ ┆     ┆ 5   ┆ 5   ┆     ┆     ┆     ┆ 351 ┆     ┆     ┆     ┆     ┆     ┆     ┆ 562 ┆     ┆ 281 │\n",
       "│ N+] ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆ 395 ┆     ┆ 2   │\n",
       "│ =NC ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ CC[ ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ C@H ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ ](N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c1n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c(N ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ c2n ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "│ oc… ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     ┆     │\n",
       "└─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┴─────┘"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = train_df.with_columns(\n",
    "    pl.when(pl.col('label') == 1).then(\n",
    "        pl.col('inter_group_weight') * pl.col('intra_group_weight') / pl.col('pos_mol_per_scaf')).otherwise(\n",
    "        pl.col('inter_group_weight') * (1.0 - pl.col('intra_group_weight')) / (pl.col('mol_per_scaf') - pl.col('pos_mol_per_scaf'))\n",
    "    ).alias('sample_weight')\n",
    ")\n",
    "train_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "03e0cf25",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([4.2695137e+07, 5.3968569e+07, 9.1266800e+05, 3.3136300e+05,\n",
       "        2.9932300e+05, 9.9974000e+04, 3.7910000e+04, 1.8902000e+04,\n",
       "        1.1310000e+04, 4.0454000e+04]),\n",
       " array([0.        , 0.1       , 0.2       , 0.30000001, 0.40000001,\n",
       "        0.5       , 0.60000002, 0.69999999, 0.80000001, 0.89999998,\n",
       "        1.        ]),\n",
       " <BarContainer object of 10 artists>)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(train_df['intra_group_weight'], log = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "60962032",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([9.4222456e+07, 1.8949360e+06, 5.8428700e+05, 1.9505400e+05,\n",
       "        1.0170030e+06, 3.8229600e+05, 0.0000000e+00, 0.0000000e+00,\n",
       "        5.3760000e+04, 6.5818000e+04]),\n",
       " array([1.00000000e+00, 6.51489990e+03, 1.30287998e+04, 1.95426992e+04,\n",
       "        2.60565996e+04, 3.25705000e+04, 3.90843984e+04, 4.55983008e+04,\n",
       "        5.21121992e+04, 5.86261016e+04, 6.51400000e+04]),\n",
       " <BarContainer object of 10 artists>)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAicAAAGdCAYAAADJ6dNTAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8g+/7EAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAf9ElEQVR4nO3dfXBU1cHH8V9CyCYpJBAjCcGEqCAYhIAkxFTRMqymQLFlOi3j0DamHVrtMurEWqEvpH8gYbRDqbpCXwbpTB1B24pUFEvDS1QiCYEIGItQQ03VJFAkm6QaIDnPHz6srgmUkA17bu73M7Mz7N4z9549s2y+s7t3N8oYYwQAAGCJ6EhPAAAA4LOIEwAAYBXiBAAAWIU4AQAAViFOAACAVYgTAABgFeIEAABYhTgBAABWiYn0BHqrq6tL77//voYOHaqoqKhITwcAAFwAY4xaW1uVnp6u6OjzvzbiuDh5//33lZGREelpAACAi9DQ0KArrrjivGMcFydDhw6V9MmdS0xMjPBsAADAhQgEAsrIyAj+HT8fx8XJ2bdyEhMTiRMAABzmQj6SwQdiAQCAVYgTAABgFeIEAABYhTgBAABWIU4AAIBVIhInv/rVrzRhwgRlZ2frnnvukTEmEtMAAAAWuuRxcuzYMT3++OOqqanRgQMHVFNTo9dff/1STwMAAFgqIt9zcubMGX388ceSpNOnT2vEiBGRmAYAALBQr185qaio0Ny5c5Wenq6oqCht3Lix2xi/36+srCzFxcUpPz9fVVVVwW2XX365fvSjHykzM1Pp6enyer26+uqr+3QnAADAwNHrOGlvb1dOTo78fn+P2zds2KCSkhKVlpZq7969ysnJUWFhoZqbmyVJH374oV544QUdPXpU7733nnbt2qWKioq+3QsAADBg9DpOZs2apWXLlmnevHk9bl+5cqUWLlyo4uJiZWdna82aNUpISNDatWslSX//+981ZswYJScnKz4+XnPmzDnvZ046OjoUCARCLgAAYOAK6wdiT506pZqaGnm93k8PEB0tr9eryspKSVJGRoZ27dqljz/+WJ2dndqxY4fGjRt3zn2WlZUpKSkpeOEXiQEAGNjCGifHjx9XZ2enUlNTQ25PTU1VY2OjJOmGG27Q7NmzNWXKFE2aNElXX321br/99nPuc8mSJWppaQleGhoawjllAABgmYicrfPQQw/poYceuqCxHo9HHo+nn2f0qazFmy/ZscLl6Io5kZ4CAABhE9ZXTlJSUjRo0CA1NTWF3N7U1KS0tLRwHgoAAAxQYY2T2NhYTZ06VeXl5cHburq6VF5eroKCgnAeCgAADFC9flunra1NR44cCV6vr69XbW2tkpOTlZmZqZKSEhUVFSk3N1fTpk3TqlWr1N7eruLi4j5N1O/3y+/3q7Ozs0/7AQAAdosyvfxhmx07dmjGjBndbi8qKtK6deskSY8//rgeeeQRNTY2avLkyXr00UeVn58flgkHAgElJSWppaVFiYmJYdnnZ/GZEwAAwq83f797HSeRRpx0R5wAAGzXm7/fEflVYgAAgHMhTgAAgFUcEyd+v1/Z2dnKy8uL9FQAAEA/ckyc+Hw+1dXVqbq6OtJTAQAA/cgxcQIAANyBOAEAAFYhTgAAgFWIEwAAYBXHxAln6wAA4A6OiRPO1gEAwB0cEycAAMAdiBMAAGAV4gQAAFiFOAEAAFYhTgAAgFUcEyecSgwAgDs4Jk44lRgAAHdwTJwAAAB3IE4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFUcEyd8zwkAAO7gmDjhe04AAHAHx8QJAABwB+IEAABYhTgBAABWIU4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFUcEyd8QywAAO7gmDjhG2IBAHAHx8QJAABwB+IEAABYhTgBAABWIU4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAVnFMnPDDfwAAuINj4oQf/gMAwB0cEycAAMAdiBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVYgTAABgFeIEAABYhTgBAABWIU4AAIBViBMAAGAV4gQAAFiFOAEAAFZxTJz4/X5lZ2crLy8v0lMBAAD9yDFx4vP5VFdXp+rq6khPBQAA9CPHxAkAAHAH4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVYgTAABgFeIEAABYhTgBAABWIU4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVYgTAABgFeIEAABYhTgBAABWIU4AAIBVLnmcHDp0SJMnTw5e4uPjtXHjxks9DQAAYKmYS33AcePGqba2VpLU1tamrKws3XrrrZd6GgAAwFIRfVtn06ZNmjlzpr7whS9EchoAAMAivY6TiooKzZ07V+np6YqKiurxLRm/36+srCzFxcUpPz9fVVVVPe7rmWee0fz583s9aQAAMHD1Ok7a29uVk5Mjv9/f4/YNGzaopKREpaWl2rt3r3JyclRYWKjm5uaQcYFAQLt27dLs2bPPe7yOjg4FAoGQCwAAGLh6HSezZs3SsmXLNG/evB63r1y5UgsXLlRxcbGys7O1Zs0aJSQkaO3atSHjnn/+ed12222Ki4s77/HKysqUlJQUvGRkZPR2ygAAwEHC+pmTU6dOqaamRl6v99MDREfL6/WqsrIyZOyFvqWzZMkStbS0BC8NDQ3hnDIAALBMWM/WOX78uDo7O5Wamhpye2pqqv7xj38Er7e0tKiqqkp//vOf/+c+PR6PPB5POKcJAAAsdslPJZakpKQkNTU1ReLQAADAcmF9WyclJUWDBg3qFh5NTU1KS0sL56EAAMAAFdY4iY2N1dSpU1VeXh68raurS+Xl5SooKOjTvv1+v7Kzs5WXl9fXaQIAAIv1+m2dtrY2HTlyJHi9vr5etbW1Sk5OVmZmpkpKSlRUVKTc3FxNmzZNq1atUnt7u4qLi/s0UZ/PJ5/Pp0AgoKSkpD7tCwAA2KvXcbJnzx7NmDEjeL2kpESSVFRUpHXr1mn+/Pk6duyYli5dqsbGRk2ePFlbtmzp9iFZAACAnkQZY0ykJ9EbZ185aWlpUWJiYtj3n7V4c9j32d+OrpgT6SkAAHBevfn7HdHf1gEAAPg8x8QJH4gFAMAdHBMnPp9PdXV1qq6ujvRUAABAP3JMnAAAAHcgTgAAgFWIEwAAYBXiBAAAWMUxccLZOgAAuINj4oSzdQAAcAfHxAkAAHAH4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXHxAmnEgMA4A6OiRNOJQYAwB0cEycAAMAdiBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWMUxccKXsAEA4A6OiRO+hA0AAHdwTJwAAAB3IE4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWMUxccI3xAIA4A6OiRO+IRYAAHdwTJwAAAB3IE4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVYgTAABgFcfECT/8BwCAOzgmTvjhPwAA3MExcQIAANyBOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVYgTAABgFeIEAABYhTgBAABWIU4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXHxInf71d2drby8vIiPRUAANCPHBMnPp9PdXV1qq6ujvRUAABAP4qJ9ATQd1mLN0d6Cr12dMWcSE8BAGApx7xyAgAA3IE4AQAAViFOAACAVYgTAABgFeIEAABYhTgBAABWIU4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVYgTAABgFeIEAABYhTgBAABWIU4AAIBViBMAAGAV4gQAAFiFOAEAAFaJSJzU19drxowZys7O1sSJE9Xe3h6JaQAAAAvFROKgd955p5YtW6bp06frxIkT8ng8kZgGAACw0CWPkzfffFODBw/W9OnTJUnJycmXegoAAMBivX5bp6KiQnPnzlV6erqioqK0cePGbmP8fr+ysrIUFxen/Px8VVVVBbcdPnxYQ4YM0dy5c3X99ddr+fLlfboDAABgYOl1nLS3tysnJ0d+v7/H7Rs2bFBJSYlKS0u1d+9e5eTkqLCwUM3NzZKkM2fO6JVXXtETTzyhyspKbd26VVu3bu3bvQAAAANGr9/WmTVrlmbNmnXO7StXrtTChQtVXFwsSVqzZo02b96stWvXavHixRo1apRyc3OVkZEhSZo9e7Zqa2t166239ri/jo4OdXR0BK8HAoHeThlwrazFmyM9hV47umJOpKcAIMLCerbOqVOnVFNTI6/X++kBoqPl9XpVWVkpScrLy1Nzc7M+/PBDdXV1qaKiQtdee+0591lWVqakpKTg5WzUAACAgSmscXL8+HF1dnYqNTU15PbU1FQ1NjZKkmJiYrR8+XLdfPPNmjRpksaOHauvfOUr59znkiVL1NLSErw0NDSEc8oAAMAyETmV+H+9NfRZHo+HU40BAHCRsL5ykpKSokGDBqmpqSnk9qamJqWlpYXzUAAAYIAKa5zExsZq6tSpKi8vD97W1dWl8vJyFRQUhPNQAABggOr12zptbW06cuRI8Hp9fb1qa2uVnJyszMxMlZSUqKioSLm5uZo2bZpWrVql9vb24Nk7F8vv98vv96uzs7NP+wEAAHbrdZzs2bNHM2bMCF4vKSmRJBUVFWndunWaP3++jh07pqVLl6qxsVGTJ0/Wli1bun1Itrd8Pp98Pp8CgYCSkpL6tC8AAGCvXsfJl770JRljzjtm0aJFWrRo0UVPCgAAuFdEfpUYAADgXIgTAABgFcfEid/vV3Z2tvLy8iI9FQAA0I8cEyc+n091dXWqrq6O9FQAAEA/ckycAAAAdyBOAACAVYgTAABgFeIEAABYxTFxwtk6AAC4g2PihLN1AABwB8fECQAAcAfiBAAAWIU4AQAAViFOAACAVYgTAABgFcfECacSAwDgDjGRnsCF8vl88vl8CgQCSkpKivR00EdZizdHegq9dnTFnEhPAQBcwTGvnAAAAHcgTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVRwTJ3zPCQAA7uCYOPH5fKqrq1N1dXWkpwIAAPqRY+IEAAC4A3ECAACsQpwAAACrECcAAMAqxAkAALAKcQIAAKxCnAAAAKsQJwAAwCqOiRO+IRYAAHdwTJzwDbEAALhDTKQnAACflbV4c6Sn0GtHV8yJ9BSAAcUxr5wAAAB3IE4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVRwTJ/zwHwAA7uCYOOGH/wAAcAfHxAkAAHAH4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVYgTAABgFeIEAABYhTgBAABWIU4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFUcEyd+v1/Z2dnKy8uL9FQAAEA/ckyc+Hw+1dXVqbq6OtJTAQAA/Sgm0hMAnCJr8eZITwEAXMExr5wAAAB3IE4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVYgTAABgFeIEAABYhTgBAABWIU4AAIBViBMAAGAV4gQAAFiFOAEAAFYhTgAAgFWIEwAAYBXiBAAAWIU4AQAAViFOAACAVYgTAABglZhIHDQrK0uJiYmKjo7W8OHDtX379khMAwAAWCgicSJJu3bt0pAhQyJ1eAAAYCne1gEAAFbpdZxUVFRo7ty5Sk9PV1RUlDZu3NhtjN/vV1ZWluLi4pSfn6+qqqqQ7VFRUbrllluUl5enp5566qInDwAABp5ex0l7e7tycnLk9/t73L5hwwaVlJSotLRUe/fuVU5OjgoLC9Xc3Bwc8+qrr6qmpkabNm3S8uXLtX///ou/BwAAYEDpdZzMmjVLy5Yt07x583rcvnLlSi1cuFDFxcXKzs7WmjVrlJCQoLVr1wbHjBo1SpI0cuRIzZ49W3v37j3n8To6OhQIBEIuAABg4ArrZ05OnTqlmpoaeb3eTw8QHS2v16vKykpJn7zy0traKklqa2vTtm3bNGHChHPus6ysTElJScFLRkZGOKcMAAAsE9Y4OX78uDo7O5Wamhpye2pqqhobGyVJTU1Nuummm5STk6MbbrhB3/nOd5SXl3fOfS5ZskQtLS3BS0NDQzinDAAALHPJTyW+6qqr9MYbb1zweI/HI4/H048zAgAANgnrKycpKSkaNGiQmpqaQm5vampSWlpaOA8FAAAGqLDGSWxsrKZOnary8vLgbV1dXSovL1dBQUGf9u33+5WdnX3et4AAAIDz9fptnba2Nh05ciR4vb6+XrW1tUpOTlZmZqZKSkpUVFSk3NxcTZs2TatWrVJ7e7uKi4v7NFGfzyefz6dAIKCkpKQ+7QsAANir13GyZ88ezZgxI3i9pKREklRUVKR169Zp/vz5OnbsmJYuXarGxkZNnjxZW7Zs6fYhWQAAgJ5EGWNMpCfRG2dfOWlpaVFiYmLY95+1eHPY9wlgYDu6Yk6kpwBYrzd/v/ltHQAAYBXiBAAAWMUxccLZOgAAuINj4sTn86murk7V1dWRngoAAOhHjokTAADgDsQJAACwCnECAACsQpwAAACrOCZOOFsHAAB3cEyccLYOAADu0Ovf1gEAABfGiT+JYsPPMTjmlRMAAOAOxAkAALAKcQIAAKxCnAAAAKs4Jk44lRgAAHdwTJxwKjEAAO7gmDgBAADuQJwAAACr8CVsAABHcOIXmuHi8MoJAACwCnECAACsQpwAAACrOCZO+J4TAADcwTFxwvecAADgDo6JEwAA4A7ECQAAsApxAgAArEKcAAAAqxAnAADAKsQJAACwCnECAACsQpwAAACrOCZO+IZYAADcwTFxwjfEAgDgDjGRnkBvGWMkSYFAoF/239Xx337ZL4CBq7+ejxCK5+dLo78ez2f3e/bv+PlEmQsZZZF///vfysjIiPQ0AADARWhoaNAVV1xx3jGOi5Ouri69//77Gjp0qKKiosK670AgoIyMDDU0NCgxMTGs+3Yq1qQ71qQ71qQ71qRnrEt3blkTY4xaW1uVnp6u6Ojzf6rEcW/rREdH/8/i6qvExMQB/QC5GKxJd6xJd6xJd6xJz1iX7tywJklJSRc0zjEfiAUAAO5AnAAAAKsQJ5/h8XhUWloqj8cT6alYgzXpjjXpjjXpjjXpGevSHWvSneM+EAsAAAY2XjkBAABWIU4AAIBViBMAAGAV4gQAAFiFOPl/fr9fWVlZiouLU35+vqqqqiI9pYtSUVGhuXPnKj09XVFRUdq4cWPIdmOMli5dqpEjRyo+Pl5er1eHDx8OGXPixAktWLBAiYmJGjZsmL73ve+pra0tZMz+/fs1ffp0xcXFKSMjQw8//HC3uTz77LMaP3684uLiNHHiRL344othv78XoqysTHl5eRo6dKhGjBihr33tazp06FDImI8//lg+n0+XXXaZhgwZoq9//etqamoKGfPuu+9qzpw5SkhI0IgRI/TAAw/ozJkzIWN27Nih66+/Xh6PR2PGjNG6deu6zceGx9rq1as1adKk4Jc+FRQU6KWXXgpud9t69GTFihWKiorSfffdF7zNjevyi1/8QlFRUSGX8ePHB7e7cU0k6b333tO3vvUtXXbZZYqPj9fEiRO1Z8+e4HY3PteGlYFZv369iY2NNWvXrjVvvvmmWbhwoRk2bJhpamqK9NR67cUXXzQ//elPzV/+8hcjyTz33HMh21esWGGSkpLMxo0bzRtvvGFuv/12c+WVV5qPPvooOObLX/6yycnJMa+//rp55ZVXzJgxY8wdd9wR3N7S0mJSU1PNggULzMGDB83TTz9t4uPjzW9+85vgmNdee80MGjTIPPzww6aurs787Gc/M4MHDzYHDhzo9zX4vMLCQvPkk0+agwcPmtraWjN79myTmZlp2tragmPuuusuk5GRYcrLy82ePXvMDTfcYL74xS8Gt585c8Zcd911xuv1mn379pkXX3zRpKSkmCVLlgTHvPPOOyYhIcGUlJSYuro689hjj5lBgwaZLVu2BMfY8ljbtGmT2bx5s3n77bfNoUOHzE9+8hMzePBgc/DgQWOM+9bj86qqqkxWVpaZNGmSuffee4O3u3FdSktLzYQJE8wHH3wQvBw7diy43Y1rcuLECTN69Ghz5513mt27d5t33nnHvPzyy+bIkSPBMW58rg0n4sQYM23aNOPz+YLXOzs7TXp6uikrK4vgrPru83HS1dVl0tLSzCOPPBK87eTJk8bj8Zinn37aGGNMXV2dkWSqq6uDY1566SUTFRVl3nvvPWOMMU888YQZPny46ejoCI558MEHzbhx44LXv/nNb5o5c+aEzCc/P9/84Ac/COt9vBjNzc1Gktm5c6cx5pM1GDx4sHn22WeDY9566y0jyVRWVhpjPom+6Oho09jYGByzevVqk5iYGFyHH//4x2bChAkhx5o/f74pLCwMXrf5sTZ8+HDz+9//3vXr0draasaOHWu2bt1qbrnllmCcuHVdSktLTU5OTo/b3LomDz74oLnpppvOuZ3n2r5z/ds6p06dUk1Njbxeb/C26Ohoeb1eVVZWRnBm4VdfX6/GxsaQ+5qUlKT8/Pzgfa2srNSwYcOUm5sbHOP1ehUdHa3du3cHx9x8882KjY0NjiksLNShQ4f04YcfBsd89jhnx9iwpi0tLZKk5ORkSVJNTY1Onz4dMt/x48crMzMzZF0mTpyo1NTU4JjCwkIFAgG9+eabwTHnu8+2PtY6Ozu1fv16tbe3q6CgwPXr4fP5NGfOnG5zd/O6HD58WOnp6brqqqu0YMECvfvuu5LcuyabNm1Sbm6uvvGNb2jEiBGaMmWKfve73wW381zbd66Pk+PHj6uzszPkP44kpaamqrGxMUKz6h9n78/57mtjY6NGjBgRsj0mJkbJyckhY3rax2ePca4xkV7Trq4u3Xfffbrxxht13XXXSfpkrrGxsRo2bFjI2M+vy8Xe50AgoI8++si6x9qBAwc0ZMgQeTwe3XXXXXruueeUnZ3t2vWQpPXr12vv3r0qKyvrts2t65Kfn69169Zpy5YtWr16terr6zV9+nS1tra6dk3eeecdrV69WmPHjtXLL7+su+++W/fcc4/+8Ic/SOK5Nhwc96vEQF/4fD4dPHhQr776aqSnEnHjxo1TbW2tWlpa9Kc//UlFRUXauXNnpKcVMQ0NDbr33nu1detWxcXFRXo61pg1a1bw35MmTVJ+fr5Gjx6tZ555RvHx8RGcWeR0dXUpNzdXy5cvlyRNmTJFBw8e1Jo1a1RUVBTh2Q0Mrn/lJCUlRYMGDer26fKmpialpaVFaFb94+z9Od99TUtLU3Nzc8j2M2fO6MSJEyFjetrHZ49xrjGRXNNFixbphRde0Pbt23XFFVcEb09LS9OpU6d08uTJkPGfX5eLvc+JiYmKj4+37rEWGxurMWPGaOrUqSorK1NOTo5+/etfu3Y9ampq1NzcrOuvv14xMTGKiYnRzp079eijjyomJkapqamuXJfPGzZsmK655hodOXLEtY+VkSNHKjs7O+S2a6+9Nvh2l9ufa8PB9XESGxurqVOnqry8PHhbV1eXysvLVVBQEMGZhd+VV16ptLS0kPsaCAS0e/fu4H0tKCjQyZMnVVNTExyzbds2dXV1KT8/PzimoqJCp0+fDo7ZunWrxo0bp+HDhwfHfPY4Z8dEYk2NMVq0aJGee+45bdu2TVdeeWXI9qlTp2rw4MEh8z106JDefffdkHU5cOBAyJPJ1q1blZiYGHyS+l/32fbHWldXlzo6Oly7HjNnztSBAwdUW1sbvOTm5mrBggXBf7txXT6vra1N//znPzVy5EjXPlZuvPHGbl9H8Pbbb2v06NGS3PtcG1aR/kSuDdavX288Ho9Zt26dqaurM9///vfNsGHDQj5d7hStra1m3759Zt++fUaSWblypdm3b5/517/+ZYz55PS2YcOGmeeff97s37/ffPWrX+3x9LYpU6aY3bt3m1dffdWMHTs25PS2kydPmtTUVPPtb3/bHDx40Kxfv94kJCR0O70tJibG/PKXvzRvvfWWKS0tjdjpbXfffbdJSkoyO3bsCDkd8r///W9wzF133WUyMzPNtm3bzJ49e0xBQYEpKCgIbj97OuRtt91mamtrzZYtW8zll1/e4+mQDzzwgHnrrbeM3+/v8XRIGx5rixcvNjt37jT19fVm//79ZvHixSYqKsr87W9/M8a4bz3O5bNn6xjjznW5//77zY4dO0x9fb157bXXjNfrNSkpKaa5udkY4841qaqqMjExMeahhx4yhw8fNk899ZRJSEgwf/zjH4Nj3PhcG07Eyf977LHHTGZmpomNjTXTpk0zr7/+eqSndFG2b99uJHW7FBUVGWM+OcXt5z//uUlNTTUej8fMnDnTHDp0KGQf//nPf8wdd9xhhgwZYhITE01xcbFpbW0NGfPGG2+Ym266yXg8HjNq1CizYsWKbnN55plnzDXXXGNiY2PNhAkTzObNm/vtfp9PT+shyTz55JPBMR999JH54Q9/aIYPH24SEhLMvHnzzAcffBCyn6NHj5pZs2aZ+Ph4k5KSYu6//35z+vTpkDHbt283kydPNrGxseaqq64KOcZZNjzWvvvd75rRo0eb2NhYc/nll5uZM2cGw8QY963HuXw+Tty4LvPnzzcjR440sbGxZtSoUWb+/Pkh3+fhxjUxxpi//vWv5rrrrjMej8eMHz/e/Pa3vw3Z7sbn2nCKMsaYyLxmAwAA0J3rP3MCAADsQpwAAACrECcAAMAqxAkAALAKcQIAAKxCnAAAAKsQJwAAwCrECQAAsApxAgAArEKcAAAAqxAnAADAKsQJAACwyv8BmkY6t5Y6sTMAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(train_df['inter_group_weight'], log = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "656a900c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([8.4094985e+07, 1.4320595e+07, 1.8000000e+01, 4.0000000e+00,\n",
       "        1.0000000e+00, 6.0000000e+00, 0.0000000e+00, 0.0000000e+00,\n",
       "        0.0000000e+00, 1.0000000e+00]),\n",
       " array([8.35499443e-04, 9.56307770e-01, 1.91178004e+00, 2.86725231e+00,\n",
       "        3.82272458e+00, 4.77819685e+00, 5.73366912e+00, 6.68914139e+00,\n",
       "        7.64461366e+00, 8.60008593e+00, 9.55555820e+00]),\n",
       " <BarContainer object of 10 artists>)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(train_df['sample_weight'], log = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "d1c58262",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(98415610, 17)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "9a9b0ff2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BRD4 0.014448811387579806\n",
      "HSA 0.012913575375656095\n",
      "sEH 0.02290908303554082\n",
      "Union: 0.047737922495750755\n"
     ]
    }
   ],
   "source": [
    "analyze_sample_weight(train_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "77cae613",
   "metadata": {},
   "outputs": [],
   "source": [
    "!mkdir /home/dangnh36/datasets/competitions/leash_belka/processed/sample_weights/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "524405a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df.select('sample_weight').write_csv('/home/dangnh36/datasets/competitions/leash_belka/processed/sample_weights/scaffold_v1.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "03972fdb",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dd6aae7a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "9c4eda19",
   "metadata": {},
   "source": [
    "# Sample weight by Baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "cec4d3ab",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (98_415_610, 6)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>BRD4</th><th>HSA</th><th>sEH</th><th>BRD4_baseline</th><th>HSA_baseline</th><th>sEH_baseline</th></tr><tr><td>u8</td><td>u8</td><td>u8</td><td>f64</td><td>f64</td><td>f64</td></tr></thead><tbody><tr><td>0</td><td>0</td><td>0</td><td>0.000002</td><td>0.000047</td><td>0.000008</td></tr><tr><td>0</td><td>0</td><td>0</td><td>0.000002</td><td>0.000029</td><td>0.000008</td></tr><tr><td>0</td><td>0</td><td>0</td><td>6.5125e-8</td><td>0.000007</td><td>0.000006</td></tr><tr><td>0</td><td>0</td><td>0</td><td>1.4169e-7</td><td>0.000001</td><td>0.000002</td></tr><tr><td>0</td><td>0</td><td>0</td><td>2.8887e-7</td><td>0.000006</td><td>0.000003</td></tr><tr><td>0</td><td>0</td><td>0</td><td>3.1989e-7</td><td>0.000011</td><td>0.000006</td></tr><tr><td>0</td><td>0</td><td>0</td><td>2.3713e-7</td><td>0.000007</td><td>0.000007</td></tr><tr><td>0</td><td>0</td><td>0</td><td>4.0376e-7</td><td>0.00001</td><td>0.000006</td></tr><tr><td>0</td><td>0</td><td>0</td><td>0.000001</td><td>0.000034</td><td>0.000006</td></tr><tr><td>0</td><td>0</td><td>0</td><td>2.3771e-7</td><td>0.000011</td><td>0.000003</td></tr><tr><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td><td>&hellip;</td></tr><tr><td>0</td><td>0</td><td>0</td><td>3.6129e-7</td><td>0.000018</td><td>0.000003</td></tr><tr><td>0</td><td>0</td><td>0</td><td>0.000001</td><td>0.000073</td><td>0.000004</td></tr><tr><td>0</td><td>0</td><td>0</td><td>0.000001</td><td>0.000029</td><td>0.000003</td></tr><tr><td>0</td><td>0</td><td>0</td><td>0.000001</td><td>0.000027</td><td>0.000003</td></tr><tr><td>0</td><td>0</td><td>0</td><td>0.000002</td><td>0.000063</td><td>0.000006</td></tr><tr><td>0</td><td>0</td><td>0</td><td>9.2532e-8</td><td>0.000002</td><td>0.000001</td></tr><tr><td>0</td><td>0</td><td>0</td><td>9.1673e-7</td><td>0.00002</td><td>0.000002</td></tr><tr><td>0</td><td>0</td><td>0</td><td>1.7096e-7</td><td>0.00016</td><td>0.000011</td></tr><tr><td>0</td><td>0</td><td>0</td><td>0.000003</td><td>0.000112</td><td>0.000007</td></tr><tr><td>0</td><td>0</td><td>0</td><td>2.0285e-7</td><td>0.000008</td><td>0.000003</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (98_415_610, 6)\n",
       "┌──────┬─────┬─────┬───────────────┬──────────────┬──────────────┐\n",
       "│ BRD4 ┆ HSA ┆ sEH ┆ BRD4_baseline ┆ HSA_baseline ┆ sEH_baseline │\n",
       "│ ---  ┆ --- ┆ --- ┆ ---           ┆ ---          ┆ ---          │\n",
       "│ u8   ┆ u8  ┆ u8  ┆ f64           ┆ f64          ┆ f64          │\n",
       "╞══════╪═════╪═════╪═══════════════╪══════════════╪══════════════╡\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 0.000002      ┆ 0.000047     ┆ 0.000008     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 0.000002      ┆ 0.000029     ┆ 0.000008     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 6.5125e-8     ┆ 0.000007     ┆ 0.000006     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 1.4169e-7     ┆ 0.000001     ┆ 0.000002     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 2.8887e-7     ┆ 0.000006     ┆ 0.000003     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 3.1989e-7     ┆ 0.000011     ┆ 0.000006     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 2.3713e-7     ┆ 0.000007     ┆ 0.000007     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 4.0376e-7     ┆ 0.00001      ┆ 0.000006     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 0.000001      ┆ 0.000034     ┆ 0.000006     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 2.3771e-7     ┆ 0.000011     ┆ 0.000003     │\n",
       "│ …    ┆ …   ┆ …   ┆ …             ┆ …            ┆ …            │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 3.6129e-7     ┆ 0.000018     ┆ 0.000003     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 0.000001      ┆ 0.000073     ┆ 0.000004     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 0.000001      ┆ 0.000029     ┆ 0.000003     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 0.000001      ┆ 0.000027     ┆ 0.000003     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 0.000002      ┆ 0.000063     ┆ 0.000006     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 9.2532e-8     ┆ 0.000002     ┆ 0.000001     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 9.1673e-7     ┆ 0.00002      ┆ 0.000002     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 1.7096e-7     ┆ 0.00016      ┆ 0.000011     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 0.000003      ┆ 0.000112     ┆ 0.000007     │\n",
       "│ 0    ┆ 0   ┆ 0   ┆ 2.0285e-7     ┆ 0.000008     ┆ 0.000003     │\n",
       "└──────┴─────┴─────┴───────────────┴──────────────┴──────────────┘"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pl.scan_csv('/home/dangnh36/datasets/competitions/leash_belka/processed/train_v3.csv').select(\n",
    "                                     pl.col('BRD4', 'HSA',\n",
    "                                            'sEH').cast(pl.UInt8),\n",
    "                                     pl.col('BRD4_baseline', 'HSA_baseline',\n",
    "                                            'sEH_baseline')).collect()\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "2a2a0c42",
   "metadata": {},
   "outputs": [],
   "source": [
    "PROTEINS = ['BRD4', 'HSA', 'sEH']\n",
    "BASELINE_COLS = [f'{protein}_baseline' for protein in PROTEINS]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "caba6beb",
   "metadata": {},
   "outputs": [],
   "source": [
    "# df[BASELINE_COLS].to_pandas().hist(bins = 50, log = True, figsize = (20, 15))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "6e1cb702",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><style>\n",
       ".dataframe > thead > tr,\n",
       ".dataframe > tbody > tr {\n",
       "  text-align: right;\n",
       "  white-space: pre-wrap;\n",
       "}\n",
       "</style>\n",
       "<small>shape: (17, 4)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>statistic</th><th>BRD4_baseline</th><th>HSA_baseline</th><th>sEH_baseline</th></tr><tr><td>str</td><td>f64</td><td>f64</td><td>f64</td></tr></thead><tbody><tr><td>&quot;count&quot;</td><td>9.841561e7</td><td>9.841561e7</td><td>9.841561e7</td></tr><tr><td>&quot;null_count&quot;</td><td>0.0</td><td>0.0</td><td>0.0</td></tr><tr><td>&quot;mean&quot;</td><td>0.000017</td><td>0.000083</td><td>0.000019</td></tr><tr><td>&quot;std&quot;</td><td>0.000441</td><td>0.000975</td><td>0.000415</td></tr><tr><td>&quot;min&quot;</td><td>0.0</td><td>0.0</td><td>0.0</td></tr><tr><td>&quot;1%&quot;</td><td>4.2266e-9</td><td>1.7573e-7</td><td>6.1547e-9</td></tr><tr><td>&quot;2.5%&quot;</td><td>1.1823e-8</td><td>4.1410e-7</td><td>1.7956e-8</td></tr><tr><td>&quot;5%&quot;</td><td>3.0127e-8</td><td>9.1994e-7</td><td>6.4347e-8</td></tr><tr><td>&quot;10%&quot;</td><td>9.1774e-8</td><td>0.000002</td><td>3.4208e-7</td></tr><tr><td>&quot;25%&quot;</td><td>4.3897e-7</td><td>0.000006</td><td>0.000001</td></tr><tr><td>&quot;50%&quot;</td><td>0.000002</td><td>0.000015</td><td>0.000004</td></tr><tr><td>&quot;75%&quot;</td><td>0.000006</td><td>0.000041</td><td>0.000008</td></tr><tr><td>&quot;90%&quot;</td><td>0.000021</td><td>0.00011</td><td>0.000016</td></tr><tr><td>&quot;95%&quot;</td><td>0.000045</td><td>0.000226</td><td>0.000026</td></tr><tr><td>&quot;97.5%&quot;</td><td>0.000091</td><td>0.000463</td><td>0.000048</td></tr><tr><td>&quot;99%&quot;</td><td>0.000221</td><td>0.001113</td><td>0.000172</td></tr><tr><td>&quot;max&quot;</td><td>1.0</td><td>1.0</td><td>1.0</td></tr></tbody></table></div>"
      ],
      "text/plain": [
       "shape: (17, 4)\n",
       "┌────────────┬───────────────┬──────────────┬──────────────┐\n",
       "│ statistic  ┆ BRD4_baseline ┆ HSA_baseline ┆ sEH_baseline │\n",
       "│ ---        ┆ ---           ┆ ---          ┆ ---          │\n",
       "│ str        ┆ f64           ┆ f64          ┆ f64          │\n",
       "╞════════════╪═══════════════╪══════════════╪══════════════╡\n",
       "│ count      ┆ 9.841561e7    ┆ 9.841561e7   ┆ 9.841561e7   │\n",
       "│ null_count ┆ 0.0           ┆ 0.0          ┆ 0.0          │\n",
       "│ mean       ┆ 0.000017      ┆ 0.000083     ┆ 0.000019     │\n",
       "│ std        ┆ 0.000441      ┆ 0.000975     ┆ 0.000415     │\n",
       "│ min        ┆ 0.0           ┆ 0.0          ┆ 0.0          │\n",
       "│ 1%         ┆ 4.2266e-9     ┆ 1.7573e-7    ┆ 6.1547e-9    │\n",
       "│ 2.5%       ┆ 1.1823e-8     ┆ 4.1410e-7    ┆ 1.7956e-8    │\n",
       "│ 5%         ┆ 3.0127e-8     ┆ 9.1994e-7    ┆ 6.4347e-8    │\n",
       "│ 10%        ┆ 9.1774e-8     ┆ 0.000002     ┆ 3.4208e-7    │\n",
       "│ 25%        ┆ 4.3897e-7     ┆ 0.000006     ┆ 0.000001     │\n",
       "│ 50%        ┆ 0.000002      ┆ 0.000015     ┆ 0.000004     │\n",
       "│ 75%        ┆ 0.000006      ┆ 0.000041     ┆ 0.000008     │\n",
       "│ 90%        ┆ 0.000021      ┆ 0.00011      ┆ 0.000016     │\n",
       "│ 95%        ┆ 0.000045      ┆ 0.000226     ┆ 0.000026     │\n",
       "│ 97.5%      ┆ 0.000091      ┆ 0.000463     ┆ 0.000048     │\n",
       "│ 99%        ┆ 0.000221      ┆ 0.001113     ┆ 0.000172     │\n",
       "│ max        ┆ 1.0           ┆ 1.0          ┆ 1.0          │\n",
       "└────────────┴───────────────┴──────────────┴──────────────┘"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[BASELINE_COLS].describe(percentiles = PERCENTILES)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "15eea09d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "57149.65025596602"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp = df.filter(pl.col('BRD4_baseline') > 0.0)['BRD4_baseline']\n",
    "tmp.max() / tmp.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "28695c98",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "96b09c57",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a4756580",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "df084132",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "865e6e37",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "571aeb2e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fbef52a4",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b3efee58",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "66385f68",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c5e308fe",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c6dfae79",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "76e19b88",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "50786708",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "aaea70f0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5a1f7b81",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6b21766c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5840521a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1af6c8ee",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1d8f845f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "66b1df99",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f6f61d3f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2f84d46e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ce42f992",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}