RuntimeError when using calc.pandas

[alsalehf]

description

I get stuck in a loop when using pandas.clac and results in runtime error
RuntimeError:
An attempt has been made to start a new process before the
current process has finished its bootstrapping phase

Below is the code im using for testing:
minimal reproduction code
from rdkit import Chem
from mordred import Calculator, descriptors
#import pandas as pd
import unicodedata

components = ["CCO"]

def s2d(smiles_list):
final_list = [unicodedata.normalize("NFKD", ls) for ls in smiles_list]

mols = [Chem.MolFromSmiles(smi) for smi in final_list]
calc = Calculator(descriptors, ignore_3D=True)
df3 = calc.pandas(mols)
return df3

s2d(components)

df = s2d(components)
print(df)

Please fill me if possible.

environment

I'm running the code in a windows 10 machine with a venv environment.

Please fill me.

conda or pip

pip.

python version

Python 3.10.4

library version

Please execute the command and paste result.

• pip

Package Version

---

mordred 1.2.0
networkx 2.8.8
numpy 1.24.2
pandas 1.5.3
Pillow 9.4.0
pip 22.0.4
python-dateutil 2.8.2
pytz 2022.7.1
rdkit 2022.9.5
setuptools 58.1.0
six 1.16.0

pip show rdkit
Name: rdkit
Version: 2022.9.5
Summary: A collection of chemoinformatics and machine-learning software written in C++ and Python
Home-page: https://github.com/kuelumbus/rdkit-pypi
Author: Christopher Kuenneth
Author-email: [email protected]
License: BSD-3-Clause
Location: c:\users\admin\chemslenv\lib\site-packages
Requires: numpy, Pillow
Required-by:

[ismorphism]

@alsalehf Hello! That's very important question, but somehow the developers do not have a time to answer it. The solution is to drop all "bad" descriptors - they do not work for you due to problems with your molecules' stereochemistry or smth else. That's my idea, which it's based on my experience:

from mordred import Calculator, PBF, MomentOfInertia, TopologicalCharge, MolecularDistanceEdge, MoRSE, GravitationalIndex, GeometricalIndex, EState, DistanceMatrix, DetourMatrix, CPSA, BaryszMatrix, Autocorrelation, AdjacencyMatrix, descriptors, get_descriptors_from_module

descs = get_descriptors_from_module(descriptors, submodule=True)

# exclude some from descs
descs = filter(lambda d: ((d.__module__ != AdjacencyMatrix.__name__) and 
                          (d.__module__ != Autocorrelation.__name__) and
                          (d.__module__ != DetourMatrix.__name__) and 
                          (d.__module__ != BaryszMatrix.__name__) and 
                          (d.__module__ != CPSA.__name__) and 
                          (d.__module__ != DistanceMatrix.__name__) and 
                          (d.__module__ != EState.__name__) and 
                          (d.__module__ != GeometricalIndex.__name__) and 
                          (d.__module__ != GravitationalIndex.__name__) and 
                          (d.__module__ != MoRSE.__name__) and 
                          (d.__module__ != MolecularDistanceEdge.__name__) and 
                          (d.__module__ != MomentOfInertia.__name__) and 
                          (d.__module__ != PBF.__name__) and 
                          (d.__module__ != TopologicalCharge.__name__)), descs)

calc = Calculator(descs)
calc.pandas(mols)

[CHAOHSUTW]

@alsalehf Hi! I also faced with the same error while using calc.pandas(mols). I just tried to run the calculator for each mol, gather it as a dict, then generate the dataframe. It worked eventually. See if this helps.

calc = Calculator(descriptors, ignore_3D=True)
mols = [Chem.MolFromSmiles(smi) for smi in smiles]
results = []
for mol in mols:
    desc = calc(mol)  # Calculate descriptors
    results.append(dict(desc))  # Convert result to a dictionary

df = pd.DataFrame(results)

[JacksonBurns]

This issue is resolved in our community maintained fork mordred-community