pymopac 1.1

a python wrapper to interact with the semiempirical quantum chemistry program MOPAC

pyMOPAC

initial test build with basic functionality

wraps MOPAC to conveniently interact with the program from within python scripts

Dependencies

• MOPAC in $PATH (some search is done for slightly different names, but older versions might not parse)

recommended install via conda install mopac

Optional (skip via pip --no-dependencies, supports Input from pure XYZ data and runs without plotting)

• rdkit

• numpy

• matplotlib

Installation

• via pip

pip install pymopac

• directly from github (nightly)

git clone Acetylsalicylsaeure/pymopac
cd pymopac
pip install .

Usage

At the core, this module implements two classes, MopacInput and MopacOutput. These represent the MOPAC .mop input and .out output files, respectively. MopacInput takes a molecular geometry in various supported formats like SMILES or rdkit Mol. Furthermore, various keywords are accepted that define the MOPAC Header. The input file can be sent to MOPAC via the .run() method, which returns a MopacOutput object.

This Object Class parses the .out file and aims to dynamically extract calculation results.

Minimal working example:

import pymopac


infile = pymopac.MopacInput("CC")
outfile = infile.run()
print(outfile.result)

Getting calculated properties

import pymopac


outfile = pymopac.MopacInput("c1ccccc1").run()
print(outfile.keys())
print(outfile["IONIZATION_POTENTIAL"])

working with Mol objects

The module internally represents the molecule via rdkit Mol objects. They can serve both as inputs and can be accessed as outputs after having their geometry optimized with MOPAC

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


mmff_mol = Chem.MolFromSmiles("c1ccc1")
mmff_mol = AllChem.AddHs(mmff_mol)
AllChem.EmbedMolecule(mmff_mol)
AllChem.MMFFOptimizeMolecule(mmff_mol)

outfile = pymopac.MopacInput(mmff_mol).run()
mopac_mol = outfile.toMol()

print(AllChem.GetBestRMS(mmff_mol, mopac_mol))

Run feedback

3 different keywords are implemented, which offer feedback to a MOPAC run

• verbose=True prints internal messages from the python module to stdout

• stream=True streams the MOPAC .out file to stdout

• plot=True uses matplotlib to plot the progress in gradient and heat of formation

Parsers

Custom parsers can be written, inherinting from input.BaseParser. Every Output class has a list at self.parser which, custom parsers can simply be added. If a parser has been added after Output initialization, parsing can be redone using Output.parseAll()

By convention, custom parsers are able to set attributes on the Ouput class by having the Output passed as an argument at parse time.

AUX Parser

A parser working on the MOPAC AUX file is called when a .aux file is passed, as is standard when creating the output via the MopacInput.run() method. Using this, all properties can be parsed in an unsupervised manner. Results of this can be found under Output.auxDict.

API

Starting with MOPAC v23, there exist API headers which are now also wrapped in pyMOPAC. Significant speedups can be observed for small molecules, as the computation takes place in memory and writes to disk can be avoided.

Tests

done for Ubuntu 24 LTS and Fedora 40,

Python 3.9-12

Roadmap

• add parsing for further keywords