Boltzmann_Reweighted_Compare.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "09bd0329-5311-4a8f-9368-c637d3ea10b7",
   "metadata": {},
   "source": [
    "A subsequent paper suggested some modifications to the selection procedure.\n",
    "\n",
    "Zhao, H., Nittinger, E. & Tyrchan, C. Enhanced Thompson Sampling by Roulette Wheel Selection for Screening Ultra-Large Combinatorial Libraries. bioRxiv 2024.05.16.594622 (2024) doi:10.1101/2024.05.16.594622\n",
    "\n",
    "We implemented one of the ideas, specifically reweighting the samples in a Boltzmann style fashion and selecting from that distribution rather than the max sampled value as a traditional TS does.\n",
    "\n",
    "This notebook explores the effect of that change on the quinazoline similarity search test case."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "692a7187-beb2-4c39-a418-88bc70f5ccae",
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "9236f7dc-bb26-412d-87e9-bba01006326d",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/priley/coding/TS/evaluators.py:15: UserWarning: Openeye packages not available in this environment; do not attempt to use ROCSEvaluator or FredEvaluator\n",
      "  warnings.warn(f\"Openeye packages not available in this environment; do not attempt to use ROCSEvaluator or \"\n"
     ]
    }
   ],
   "source": [
    "import copy\n",
    "import json\n",
    "from ts_main import run_ts, parse_input_dict\n",
    "from baseline import enumerate_library\n",
    "import pandas as pd\n",
    "from rdkit import Chem\n",
    "from tqdm.auto import tqdm\n",
    "import numpy as np\n",
    "from evaluators import FPEvaluator\n",
    "import seaborn as sns\n",
    "from scipy import stats\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "2e376892-92f9-4fec-aa59-20459b16988c",
   "metadata": {},
   "outputs": [],
   "source": [
    "tqdm.pandas()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0e6b3349-56c7-4f63-bd28-0b9fd3fa6514",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Suppress RDKit warnings\n",
    "from rdkit import RDLogger \n",
    "RDLogger.DisableLog('rdApp.*')  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "c2f5f543-08dc-4f66-a6dd-c42d16ce0c40",
   "metadata": {},
   "outputs": [],
   "source": [
    "TS_BASE_DIR = \".\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "e82041b7-0057-40e4-bb33-db0aedd5b29e",
   "metadata": {},
   "outputs": [],
   "source": [
    "quinazoline_json = \"\"\"{\n",
    "\"reagent_file_list\": [\n",
    "        \"TS_BASE_DIR/data/aminobenzoic_100.smi\",\n",
    "        \"TS_BASE_DIR/data/primary_amines_100.smi\",\n",
    "        \"TS_BASE_DIR/data/carboxylic_acids_100.smi\"\n",
    "    ],\n",
    "    \"reaction_smarts\": \"N[c:4][c:3]C(O)=O.[#6:1][NH2].[#6:2]C(=O)[OH]>>[C:2]c1n[c:4][c:3]c(=O)n1[C:1]\",\n",
    "    \"num_warmup_trials\": 10,\n",
    "    \"num_ts_iterations\": 5000,\n",
    "    \"evaluator_class_name\": \"FPEvaluator\",\n",
    "    \"evaluator_arg\": {\"query_smiles\" : \"COc1cccc2c(=O)n([C@@H]3C[C@@H](O)[C@@H](O)C3)c([C@@H](N)CNC(N)=O)nc12\"},\n",
    "    \"ts_mode\": \"maximize\",\n",
    "    \"log_filename\": \"ts_logs.txt\",\n",
    "    \"results_filename\": \"ts_results.csv\"\n",
    "}\"\"\".replace(\"TS_BASE_DIR\", TS_BASE_DIR)\n",
    "quinazoline_dict = json.loads(quinazoline_json)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "34e9d970-51f1-4d84-a294-e35bd1785a28",
   "metadata": {},
   "source": [
    "## Do the full enumeration and evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "1c8d91e8-a874-42b4-ad53-8bce3a2ae260",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"enumerate.json\", \"w\") as f:\n",
    "    json.dump(quinazoline_dict, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "3cbd6f4c-fe10-493f-be21-c0f600ee428d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9.70e+05 products\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4e5b2242be144ff88fde528794acb1eb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/970299 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "enumerate_library(\"enumerate.json\",\"quinazoline_100_enum.csv\",-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "172fc797-e310-4f87-9dca-211da608497d",
   "metadata": {},
   "outputs": [],
   "source": [
    "query_smiles = quinazoline_dict['evaluator_arg']['query_smiles']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "d25325e6-4eb0-4068-9cae-e9bcf2c973b1",
   "metadata": {},
   "outputs": [],
   "source": [
    "enum_df = pd.read_csv(\"quinazoline_100_enum.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "4823e662-351c-4ca4-b4c7-347ae79a0220",
   "metadata": {},
   "outputs": [],
   "source": [
    "fp_eval = FPEvaluator({\"query_smiles\": query_smiles})\n",
    "class SmilesEvaluator:\n",
    "    def __init__(self, eval):\n",
    "        self.eval = eval\n",
    "        \n",
    "    def eval_smiles(self,smiles):\n",
    "        mol = Chem.MolFromSmiles(smiles)\n",
    "        return self.eval.evaluate(mol)\n",
    "smiles_eval = SmilesEvaluator(fp_eval)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "ecdb9802-241d-4791-b790-7fe615d27ae2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "eebb02958c394ad3af61ef7e735a8241",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/970299 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "enum_df['truth'] = enum_df.SMILES.progress_apply(smiles_eval.eval_smiles)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "d1161f51-9816-4a4e-84e3-753fa051a05d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SMILES</th>\n",
       "      <th>Name</th>\n",
       "      <th>truth</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>456890</th>\n",
       "      <td>COCc1ncc2c(=O)n([C@H]3CO[C@H]4[C@@H]3OC[C@@H]4...</td>\n",
       "      <td>95883519_104600679_895342</td>\n",
       "      <td>0.302326</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>554872</th>\n",
       "      <td>O=C(O)c1ccc2c(=O)n([C@H]3CN[C@H](C(=O)O)C3)c(C...</td>\n",
       "      <td>389863_95831405_19735199</td>\n",
       "      <td>0.208791</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>660649</th>\n",
       "      <td>C[C@@H](OC(=O)CO)c1nc2c(cnn2C(C)(C)C)c(=O)n1C[...</td>\n",
       "      <td>137316859_1225225_59569926</td>\n",
       "      <td>0.164948</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13654</th>\n",
       "      <td>Cn1c(=O)c2c(=O)n(C(=O)[C@@H]3CNC(=O)C3)c(CNC(=...</td>\n",
       "      <td>26895819_4301264_895314</td>\n",
       "      <td>0.202247</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51924</th>\n",
       "      <td>NC(=O)N[C@@H](CO)c1nc2nc(N)c(Cl)nc2c(=O)n1C(=O...</td>\n",
       "      <td>21298040_8672865_2626461</td>\n",
       "      <td>0.191489</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   SMILES  \\\n",
       "456890  COCc1ncc2c(=O)n([C@H]3CO[C@H]4[C@@H]3OC[C@@H]4...   \n",
       "554872  O=C(O)c1ccc2c(=O)n([C@H]3CN[C@H](C(=O)O)C3)c(C...   \n",
       "660649  C[C@@H](OC(=O)CO)c1nc2c(cnn2C(C)(C)C)c(=O)n1C[...   \n",
       "13654   Cn1c(=O)c2c(=O)n(C(=O)[C@@H]3CNC(=O)C3)c(CNC(=...   \n",
       "51924   NC(=O)N[C@@H](CO)c1nc2nc(N)c(Cl)nc2c(=O)n1C(=O...   \n",
       "\n",
       "                              Name     truth  \n",
       "456890   95883519_104600679_895342  0.302326  \n",
       "554872    389863_95831405_19735199  0.208791  \n",
       "660649  137316859_1225225_59569926  0.164948  \n",
       "13654      26895819_4301264_895314  0.202247  \n",
       "51924     21298040_8672865_2626461  0.191489  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "enum_df.sample(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fdd7821b-51aa-42af-88fb-01cb7621a4e0",
   "metadata": {},
   "source": [
    "# Do the TS run with the original selection algorithm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "4eba17d0-647b-406e-8267-510292bda527",
   "metadata": {},
   "outputs": [],
   "source": [
    "parse_input_dict(quinazoline_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "cee21c4b-1761-4f2d-8489-99d1571ca53c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "aa3e13b419c34441b4698b691e0952e7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup 1 of 3:   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3fca4a434d2a41ccad1bbf53922ed499",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup 2 of 3:   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a46496f93488481493b54be295a8f8bc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup 3 of 3:   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7abb51e693c24cf08bfdf88b29204508",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Cycle:   0%|          | 0/5000 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         score                                             SMILES  \\\n",
      "0     1.000000  COc1cccc2c(=O)n([C@@H]3C[C@@H](O)[C@@H](O)C3)c...   \n",
      "2     0.847458  COc1cccc2c(=O)n([C@@H]3C[C@@H](O)[C@@H](O)C3)c...   \n",
      "1     0.847458  COc1cccc2c(=O)n([C@@H]3C[C@@H](O)[C@@H](O)C3)c...   \n",
      "17    0.779661  COc1cccc2c(=O)n([C@@H]3C[C@@H](O)[C@@H](O)C3)c...   \n",
      "8     0.779661  COc1cccc2c(=O)n([C@@H]3C[C@@H](O)[C@@H](O)C3)c...   \n",
      "1095  0.777778  COc1cccc2c(=O)n(C3CC(O)(C(=O)O)C3)c([C@@H](N)C...   \n",
      "1593  0.777778  COc1cccc2c(=O)n([C@@H]3CC[C@@H](O)[C@@H]3O)c([...   \n",
      "497   0.765625  COc1cccc2c(=O)n([C@@H]3CN(C)C[C@H]3O)c([C@@H](...   \n",
      "995   0.765625  COc1cccc2c(=O)n([C@H]3COC[C@@H]3O)c([C@@H](N)C...   \n",
      "799   0.765625  COc1cccc2c(=O)n([C@@H]3COC[C@H]3O)c([C@@H](N)C...   \n",
      "\n",
      "                          Name  \n",
      "0     389612_146501451_1531037  \n",
      "2     389612_146501451_1576365  \n",
      "1     389612_146501451_2384694  \n",
      "17    389612_146501451_1532556  \n",
      "8     389612_146501451_1529313  \n",
      "1095  389612_114118812_1531037  \n",
      "1593   389612_34071169_1531037  \n",
      "497    389612_34278490_1531037  \n",
      "995    389612_12957834_1531037  \n",
      "799    389612_12957829_1531037  \n"
     ]
    }
   ],
   "source": [
    "ts_df = run_ts(quinazoline_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "c20c3867-bb4e-4857-9237-6d198e6e7b06",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    5000.000000\n",
       "mean        0.430676\n",
       "std         0.065912\n",
       "min         0.321429\n",
       "25%         0.387500\n",
       "50%         0.417722\n",
       "75%         0.459459\n",
       "max         1.000000\n",
       "Name: score, dtype: float64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ts_df[\"score\"].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "292971b4-4ee8-4d04-bd70-667d7c3e9cfc",
   "metadata": {},
   "source": [
    "# Do the TS run with Boltzmann reweighted selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "f76e9b10-e089-45f9-bd87-91ef08639996",
   "metadata": {},
   "outputs": [],
   "source": [
    "quinazoline_dict_boltzmann = copy.copy(quinazoline_dict)\n",
    "quinazoline_dict_boltzmann[\"ts_mode\"] = \"maximize_boltzmann\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "18b607a6-450e-4e99-b1d4-d27a8019e028",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "146614122c034dfa82ac0d697bbe43f6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup 1 of 3:   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0a0332de10984e249d41b4f0ef86b2bc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup 2 of 3:   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6846fe9b87a54fc1bb0f1f94bc8ed55a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup 3 of 3:   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8af0c7e69e2b43548c491481a9241d11",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Cycle:   0%|          | 0/5000 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         score                                             SMILES  \\\n",
      "4368  0.138298  Nc1nnn(Cc2nc3nc(N)c(Cl)nc3c(=O)n2[C@@H](CCO)C(...   \n",
      "97    0.153846  O=C1CNC[C@@H](c2nc3nsnc3c(=O)n2C[C@@H](CO)C(=O...   \n",
      "4843  0.159091  O=c1c2[nH]ncc2nc([C@H]2NCC[C@@H]2O)n1[C@@H]1CO...   \n",
      "4945  0.160920  Cc1nsc2nc(C[C@@H](O)CN)n([C@@H](C(=O)O)[C@@H](...   \n",
      "4781  0.163043  Nc1ncc2c(=O)n(C(=O)[C@H]3CNC(=O)C3)c([C@H]3CNC...   \n",
      "4869  0.166667  O=C1NCCC[C@@H]1C(=O)n1c([C@H]2CNC[C@H]2O)nc2no...   \n",
      "3747  0.166667  CN1C[C@@H](O)[C@H](n2c([C@@H]3CNC(=O)CN3)nc3no...   \n",
      "3362  0.168421  NC[C@H](CO)c1nc2c(sc3nccnc32)c(=O)n1C(=O)c1cnc...   \n",
      "4189  0.168539              O=c1c2nccnc2nc(Cn2ncnn2)n1C1CN(CCO)C1   \n",
      "2386  0.168539  NCC(CN)c1nc2nc(=O)[nH]cc2c(=O)n1[C@@H](C(=O)O)...   \n",
      "\n",
      "                            Name  \n",
      "4368     21298040_895146_4294607  \n",
      "97    17327931_12502811_38540697  \n",
      "4843   26895189_12957829_4096639  \n",
      "4945      3349006_3875092_901811  \n",
      "4781   31613702_4301265_38540696  \n",
      "4869     967421_5382647_14507981  \n",
      "3747    967421_34278490_76576478  \n",
      "3362   12367501_8672865_12502813  \n",
      "4189    967253_82601602_31319659  \n",
      "2386   1699911_3995585_114812654  \n"
     ]
    }
   ],
   "source": [
    "ts_boltzmann_df = run_ts(quinazoline_dict_boltzmann)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b193de15-74b1-4427-8d5c-ff64909550da",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    5000.000000\n",
       "mean        0.420552\n",
       "std         0.098175\n",
       "min         0.138298\n",
       "25%         0.362500\n",
       "50%         0.421370\n",
       "75%         0.478873\n",
       "max         1.000000\n",
       "Name: score, dtype: float64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ts_boltzmann_df[\"score\"].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fd0c712e-a737-43db-ace7-ade3e4b802c8",
   "metadata": {},
   "source": [
    "# Compare the results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "8f3f6237-3fdc-460e-b324-d0360c370258",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Scores of top 200 compounds')"
      ]
     },
     "execution_count": 56,
     "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": [
    "N = 200\n",
    "ts_compare = pd.concat([ts_df.sort_values(\"score\", ascending=False).head(N).assign(method=\"original\"),\n",
    "                        ts_boltzmann_df.sort_values(\"score\", ascending=False).head(N).assign(method=\"boltzmann\"),\n",
    "                        enum_df.sort_values(\"truth\", ascending=False).head(N).rename(columns={\"truth\": \"score\"}).assign(method=\"truth\"),\n",
    "                       ],\n",
    "                       ignore_index=True)\n",
    "ax = sns.stripplot(x=\"method\", y=\"score\", hue=\"method\", dodge=True, data=ts_compare)\n",
    "ax.set_title(f\"Scores of top {N} compounds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "46bde905-1489-482c-99e7-023a4b693e37",
   "metadata": {},
   "source": [
    "Build a merged dataframe with columns for whether each of the original and reweighted found this top molecule"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "7072e1e8-8ee5-4ce5-b7a8-a09dcd2ebaa9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SMILES</th>\n",
       "      <th>Name</th>\n",
       "      <th>truth</th>\n",
       "      <th>original</th>\n",
       "      <th>boltzmann</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>177</th>\n",
       "      <td>COc1cccc2c(=O)n(Cc3nc(CO)n[nH]3)c([C@@H](N)CNC...</td>\n",
       "      <td>389612_521398993_1531037</td>\n",
       "      <td>0.605634</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>278</th>\n",
       "      <td>COc1cccc2c(=O)n([C@@H]3CN(C)C[C@H]3O)c(CNC(N)=...</td>\n",
       "      <td>389612_34278490_1464311</td>\n",
       "      <td>0.579710</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63</th>\n",
       "      <td>COc1cccc2c(=O)n(C3CN(CCO)C3)c([C@H](N)CNC(=N)N...</td>\n",
       "      <td>389612_82601602_1576365</td>\n",
       "      <td>0.652174</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119</th>\n",
       "      <td>COc1cccc2c(=O)n([C@H]3CO[C@H]4[C@@H]3OC[C@@H]4...</td>\n",
       "      <td>389612_104600679_1576365</td>\n",
       "      <td>0.625000</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>COc1cccc2c(=O)n(C3CN(CCO)C3)c([C@@H](N)CNC(N)=...</td>\n",
       "      <td>389612_82601602_1531037</td>\n",
       "      <td>0.765625</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>325</th>\n",
       "      <td>COc1cccc2c(=O)n([C@@H]3CC[C@@H](O)[C@@H]3O)c([...</td>\n",
       "      <td>389612_34071169_52986906</td>\n",
       "      <td>0.573529</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>258</th>\n",
       "      <td>COc1cccc2c(=O)n(C3CN(CCO)C3)c([C@H](N)CN)nc12</td>\n",
       "      <td>389612_82601602_19363627</td>\n",
       "      <td>0.582090</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>197</th>\n",
       "      <td>COc1cccc2c(=O)n([C@@H]3C[C@@H](O)[C@@H](O)C3)c...</td>\n",
       "      <td>389612_146501451_4181821</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>256</th>\n",
       "      <td>COc1cccc2c(=O)n(C3CN(CCO)C3)c([C@H](N)CCO)nc12</td>\n",
       "      <td>389612_82601602_1742215</td>\n",
       "      <td>0.582090</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>260</th>\n",
       "      <td>Cc1cnc2nc([C@@H](N)CNC(N)=O)n([C@@H]3C[C@@H](O...</td>\n",
       "      <td>26441149_146501451_1531037</td>\n",
       "      <td>0.582090</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                SMILES  \\\n",
       "177  COc1cccc2c(=O)n(Cc3nc(CO)n[nH]3)c([C@@H](N)CNC...   \n",
       "278  COc1cccc2c(=O)n([C@@H]3CN(C)C[C@H]3O)c(CNC(N)=...   \n",
       "63   COc1cccc2c(=O)n(C3CN(CCO)C3)c([C@H](N)CNC(=N)N...   \n",
       "119  COc1cccc2c(=O)n([C@H]3CO[C@H]4[C@@H]3OC[C@@H]4...   \n",
       "11   COc1cccc2c(=O)n(C3CN(CCO)C3)c([C@@H](N)CNC(N)=...   \n",
       "325  COc1cccc2c(=O)n([C@@H]3CC[C@@H](O)[C@@H]3O)c([...   \n",
       "258      COc1cccc2c(=O)n(C3CN(CCO)C3)c([C@H](N)CN)nc12   \n",
       "197  COc1cccc2c(=O)n([C@@H]3C[C@@H](O)[C@@H](O)C3)c...   \n",
       "256     COc1cccc2c(=O)n(C3CN(CCO)C3)c([C@H](N)CCO)nc12   \n",
       "260  Cc1cnc2nc([C@@H](N)CNC(N)=O)n([C@@H]3C[C@@H](O...   \n",
       "\n",
       "                           Name     truth original boltzmann  \n",
       "177    389612_521398993_1531037  0.605634     True      True  \n",
       "278     389612_34278490_1464311  0.579710     True      True  \n",
       "63      389612_82601602_1576365  0.652174     True      True  \n",
       "119    389612_104600679_1576365  0.625000     True      True  \n",
       "11      389612_82601602_1531037  0.765625     True      True  \n",
       "325    389612_34071169_52986906  0.573529     True      True  \n",
       "258    389612_82601602_19363627  0.582090     True      True  \n",
       "197    389612_146501451_4181821  0.600000     True      True  \n",
       "256     389612_82601602_1742215  0.582090     True      True  \n",
       "260  26441149_146501451_1531037  0.582090      NaN      True  "
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "N = 500\n",
    "df_top_enum = (\n",
    "    enum_df\n",
    "    .sort_values(\"truth\", ascending=False)\n",
    "    .head(N)\n",
    ")\n",
    "df_top_orig = (\n",
    "    ts_df\n",
    "    .sort_values(\"score\", ascending=False)\n",
    "    .head(N)\n",
    "    [[\"Name\"]]\n",
    "    .assign(original=True)\n",
    ")\n",
    "df_top_boltzmann = (\n",
    "    ts_boltzmann_df\n",
    "    .sort_values(\"score\", ascending=False)\n",
    "    .head(N)\n",
    "    [[\"Name\"]]\n",
    "    .assign(boltzmann=True)\n",
    ")\n",
    "df_top = pd.merge(pd.merge(df_top_enum, df_top_orig, how=\"left\", on=\"Name\"),\n",
    "                  df_top_boltzmann, how=\"left\", on=\"Name\")\n",
    "df_top.sample(10)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "a38f0f40-8081-498c-87c9-4cb260d369cd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "top_ns = [10, 25, 50, 100, 200, 300, 400, 500]\n",
    "for col in [\"original\", \"boltzmann\"]:\n",
    "    ax.plot(top_ns, [df_top.head(n)[col].sum() / n for n in top_ns], label=col, marker=\"o\")\n",
    "ax.set_xlabel(\"top N\")\n",
    "ax.set_ylabel(\"fraction_found\")\n",
    "ax.axhline(1, color=\"k\", linestyle=\"--\", zorder=0)\n",
    "ax.set_title(\"quinazoline 100: Frac of top N found\")\n",
    "ax.legend()\n",
    "pass\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b49bdc94-3b41-4189-9e2e-358f122f9e61",
   "metadata": {},
   "source": [
    "Excellent. This is consistent with the publication that reweighting finds more of the top performing molecules."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ts_ext",
   "language": "python",
   "name": "ts_ext"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}