samples.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from rdkit.Chem import Draw\n",
    "import matplotlib\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import r2_score\n",
    "import pickle\n",
    "import gentrl\n",
    "import numpy as np\n",
    "\n",
    "import susi\n",
    "from susi.SOMPlots import plot_estimation_map, plot_som_histogram, plot_umatrix\n",
    "import gentrl\n",
    "import torch\n",
    "#torch.cuda.set_device(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\mored\\Miniconda3\\envs\\leishnn\\lib\\site-packages\\torch\\nn\\modules\\container.py:434: UserWarning: Setting attributes on ParameterList is not supported.\n",
      "  warnings.warn(\"Setting attributes on ParameterList is not supported.\")\n"
     ]
    }
   ],
   "source": [
    "enc = gentrl.RNNEncoder(latent_size=50)\n",
    "dec = gentrl.DilConvDecoder(latent_input_size=50)\n",
    "model = gentrl.GENTRL(enc, dec, 50 * [('c', 20)], [('c', 20)], beta=0.001)\n",
    "#model.cuda();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\mored\\Miniconda3\\envs\\leishnn\\lib\\site-packages\\torch\\nn\\modules\\container.py:434: UserWarning: Setting attributes on ParameterList is not supported.\n",
      "  warnings.warn(\"Setting attributes on ParameterList is not supported.\")\n"
     ]
    }
   ],
   "source": [
    "# Load from past\n",
    "model.load('modela/4.968215223985663==0.995')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "som_general = pickle.load(open(\"general_leish_som_500K.p\", \"rb\"))\n",
    "som_specific = pickle.load(open(\"specific_leish_som_800K_2.p\", \"rb\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from moses.metrics import mol_passes_filters, SA, logP\n",
    "from moses.metrics.utils import get_n_rings, get_mol\n",
    "\n",
    "from moses.utils import disable_rdkit_log\n",
    "disable_rdkit_log()\n",
    "\n",
    "def get_num_rings_6(mol):\n",
    "    r = mol.GetRingInfo()\n",
    "    return len([x for x in r.AtomRings() if len(x) > 6])\n",
    "\n",
    "\n",
    "def penalized_logP(mol_or_smiles, masked=False, default=-5):\n",
    "    mol = get_mol(mol_or_smiles)\n",
    "    if mol is None:\n",
    "        return default\n",
    "    smiles_tokenized = gentrl.tokenizer.encode([mol_or_smiles])[0][0].tolist()\n",
    "    general_award = 0.03*som_general.predict([smiles_tokenized])[0]\n",
    "    specific_award = 1.4*som_specific.predict([smiles_tokenized])[0]\n",
    "    reward = logP(mol) - SA(mol) - get_num_rings_6(mol) + specific_award + general_award\n",
    "    if masked and not mol_passes_filters(mol):\n",
    "        return default\n",
    "    return reward"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\mored\\Miniconda3\\envs\\leishnn\\lib\\site-packages\\gentrl-0.1-py3.8.egg\\gentrl\\lp.py:108: UserWarning: indexing with dtype torch.uint8 is now deprecated, please use a dtype torch.bool instead. (Triggered internally at  ..\\aten\\src\\ATen/native/IndexingUtils.h:25.)\n"
     ]
    }
   ],
   "source": [
    "generated = []\n",
    "sampled_valid = []\n",
    "while len(generated) < 1000:\n",
    "    sampled = model.sample(100)\n",
    "    sampled_valid = [s for s in sampled if get_mol(s)]\n",
    "    \n",
    "    generated += sampled_valid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "sampled_valid = list(set(sampled_valid))\n",
    "valid = sorted(sampled_valid, key=penalized_logP, reverse=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10.166744239849129"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "smiles_tokenized = gentrl.tokenizer.encode([valid[0]])[0][0].tolist()\n",
    "som_general.predict([smiles_tokenized])[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7.040340000000006"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logP(get_mol(valid[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Draw.MolsToGridImage([get_mol(s) for s in valid[:3]], \n",
    "                     legends=[str(penalized_logP(s)) for s in valid[:3]])"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}