如何避免在Linux上杀死我的代码？

Question

我正在尝试运行应该计算分子描述符的代码，但是即使（我认为？）我有足够的内存来处理它，但它被设置的错误杀死。如何编辑代码以避免被OOM杀死？文件“ data.csv”包含〜62K行，其中包含字符串数据。

import deepchem
import os
import numpy as np
import pandas as pd

import tempfile

from rdkit import Chem
from rdkit.Chem import AllChem
import deepchem as dc

from deepchem.utils import download_url, load_from_disk
from simtk.openmm.app import PDBFile
from pdbfixer import PDBFixer

from deepchem.utils.vina_utils import prepare_inputs
import pandas as pd
import os
data = pd.read_csv('data.csv')

pdbid = data['pdb_id'].iloc[1]
ligand = data['smiles'].iloc[1]

fixer = PDBFixer(pdbid=pdbid)
PDBFile.writeFile(fixer.topology, fixer.positions, open('%s.pdb' % (pdbid), 'w'))

p, m = None, None
# fix protein, optimize ligand geometry, and sanitize molecules
try:
    p, m = prepare_inputs('%s.pdb' % (pdbid), ligand)
except:
    print('%s failed PDB fixing' % (pdbid)) 

if p and m:  # protein and molecule are readable by RDKit
    print(pdbid, p.GetNumAtoms())
    Chem.rdmolfiles.MolToPDBFile(p, '%s.pdb' % (pdbid))
    Chem.rdmolfiles.MolToPDBFile(m, 'ligand_%s.pdb' % (pdbid))

pdbids = data['pdb_id'].values
ligand_smiles = data['smiles'].values


l = []
for (pdbid, ligand) in zip(pdbids, ligand_smiles):
  fixer = PDBFixer(url='https://files.rcsb.org/download/%s.pdb' % (pdbid))
  PDBFile.writeFile(fixer.topology, fixer.positions, open('%s.pdb' % (pdbid), 'w'))
  
  p, m = None, None
  # skip pdb fixing for speed
  try:
    p, m = prepare_inputs('%s.pdb' % (pdbid), ligand, replace_nonstandard_residues=False,
                          remove_heterogens=False, remove_water=False,
                          add_hydrogens=False)
  except:
    print('%s failed sanitization' % (pdbid)) 

  if p and m:  # protein and molecule are readable by RDKit
    one = Chem.rdmolfiles.MolToPDBFile(p, '%s.pdb' % (pdbid))
    two = Chem.rdmolfiles.MolToPDBFile(m, 'ligand_%s.pdb' % (pdbid))
    l.append(p)
    l.append(m)

proteins = [f for f in os.listdir('.') if len(f) == 8 and f.endswith('.pdb')]
ligands = [f for f in os.listdir('.') if f.startswith('ligand') and f.endswith('.pdb')]

# Handle failed sanitizations
failures = set([f[:-4] for f in proteins]) - set([f[7:-4] for f in ligands])
for pdbid in failures:
  proteins.remove(pdbid + '.pdb')

pdbids = [f[:-4] for f in proteins]
small_dataset = data[data['pdb_id'].isin(pdbids)]
labels = small_dataset.labels
fp_featurizer = dc.feat.CircularFingerprint(size=2048)
features = fp_featurizer.featurize([Chem.MolFromPDBFile(l) for l in ligands])
dataset = dc.data.NumpyDataset(X=features, y=labels, ids=pdbids)
a = dataset.to_dataframe()
a.to_csv('descr.csv')

OOM错误和内存状态的屏幕截图