184 lines (183 with data), 4.7 kB

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "\n",
    "import numpy as np\n",
    "import rdkit\n",
    "from rdkit import Chem\n",
    "\n",
    "import h5py, ast, pickle\n",
    "\n",
    "# Occupy a GPU for the model to be loaded \n",
    "%env CUDA_DEVICE_ORDER=PCI_BUS_ID\n",
    "# GPU ID, if occupied change to an available GPU ID listed under !nvidia-smi\n",
    "%env CUDA_VISIBLE_DEVICES=2 \n",
    "\n",
    "from ddc_pub import ddc_v3 as ddc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import existing (trained) model\n",
    "# Ignore any warning(s) about training configuration or non-seriazable keyword arguments\n",
    "model_name = \"models/heteroencoder_model\"\n",
    "model = ddc.DDC(model_name=model_name)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load data from dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_name = \"datasets/CHEMBL25_TEST.h5\"\n",
    "npoints = 1000\n",
    "\n",
    "dataset = h5py.File(dataset_name, \"r\")\n",
    "mols    = dataset[\"mols\"][:]\n",
    "# Select random npoints\n",
    "mols_in = mols[np.random.choice(len(mols), npoints, replace=False)]\n",
    "dataset.close()\n",
    "\n",
    "# Get the SMILES behind the binary mols\n",
    "smiles_in = [Chem.MolToSmiles(Chem.Mol(mol)) for mol in mols_in]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Alternatively, use your own SMILES"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Input SMILES to auto-encode\n",
    "smiles_in = ['Cc1cccn2c(CN(C)C3CCCc4ccccc43)c(C(=O)N3CCOCC3)nc12',\n",
    "             'COC(=O)NN=C(c1ccc(O)cc1)C1C(=O)N(C)C(=O)N(C)C1=O',\n",
    "             'CCc1cc(CC)nc(OCCCn2c3c(c4cc(-c5nc(C)no5)ccc42)CC(F)(F)CC3)n1',\n",
    "             'Cc1ccc2c(C(=O)Nc3ccccc3)c(SSc3c(C(=O)Nc4ccccc4)c4ccc(C)cc4n3C)n(C)c2c1',\n",
    "             'Cc1cccc(-c2ccccc2)c1Oc1nc(O)nc(NCc2ccc3occc3c2)n1',\n",
    "             'Cn1nnnc1SCC(=O)NN=Cc1ccc(Cl)cc1',\n",
    "             'COc1cccc(NS(=O)(=O)c2ccc(OC)c(OC)c2)c1',\n",
    "             'COc1ccc(OC)c(S(=O)(=O)n2nc(C)cc2C)c1',\n",
    "             'NCCCn1cc(C2=C(c3ccncc3)C(=O)NC2=O)c2ccccc21',\n",
    "             'CN(C)C(=O)N1CCN(C(c2ccc(Cl)cc2)c2cccnc2)CC1']\n",
    "\n",
    "# MUST convert SMILES to binary mols for the model to accept them (it re-converts them to SMILES internally)\n",
    "mols_in = [Chem.rdchem.Mol.ToBinary(Chem.MolFromSmiles(smiles)) for smiles in smiles_in]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Encode the binary mols into their latent representations\n",
    "latent = model.transform(model.vectorize(mols_in))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert back to SMILES\n",
    "smiles_out = []\n",
    "for lat in latent:   \n",
    "    smiles, _ = model.predict(lat, temp=0)\n",
    "    smiles_out.append(smiles)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# To compare the results, convert smiles_out to CANONICAL\n",
    "for idx, smiles in enumerate(smiles_out):\n",
    "    mol = Chem.MolFromSmiles(smiles)\n",
    "    if mol:\n",
    "        smiles_out[idx] = Chem.MolToSmiles(mol, canonical=True)\n",
    "    else:\n",
    "        smiles_out[idx] = \"INVALID\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "smiles_in"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "smiles_out"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ddc",
   "language": "python",
   "name": "ddc"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}