IR/gen_meta_file.py

import os
import sys
import json
import pandas as pd
import numpy as np
from tqdm import tqdm

import rdkit
from rdkit import Chem
from rdkit.Chem import AllChem

sys.path.append("../../")
from utils.compound_tools import mol_to_geognn_graph_data_MMFF3d


def GetIRMetaFile():
    # get dataset_all, smiles_all, index_all from SMILES in IR dataset

    raw_smiles_all, raw_index_all = [], []
    max_len = -1
    ir_filelist = os.listdir("./qm9_ir_spec/")
    for filename in tqdm(ir_filelist[:max_len]):
        mol_info = json.load(open(os.path.join("./qm9_ir_spec/", filename), "r"))
        raw_smiles_all.append(mol_info['smiles'])
        raw_index_all.append(filename.split('.')[0])

    dataset_all, smiles_all, index_all = [], [], []
    
    for i in tqdm(range(len(raw_smiles_all))): 
        mol = AllChem.MolFromSmiles(raw_smiles_all[i])
        mol = Chem.AddHs(mol)
        AllChem.EmbedMolecule(mol)
        try: 
            data = mol_to_geognn_graph_data_MMFF3d(mol)
            dataset_all.append(data); smiles_all.append(raw_smiles_all[i])
            index_all.append(raw_index_all[i])
        except ValueError:
            print("error in {}".format(i))
    
    result_dict = dict(
        smiles_all=smiles_all, index_all=index_all,
        dataset_all=dataset_all, 
    )
    np.save(f"ir_column_charity_all.npy", result_dict)

if __name__ == "__main__":
    GetIRMetaFile()