PyOPSIN

This is a Python wrapper for the OPSIN (Open Parser for Systematic IUPAC Nomenclature) package, which allows you to generate SMILES and CML form the standardized IUPAC names for organic molecules. The original OPSIN package was written in Java, but this wrapper allows you to use OPSIN functionality directly from Python.

Installation

PyOPSIN requires Java 8+. The first time you create a PyOpsin() instance, it downloads the latest OPSIN CLI JAR (opsin-cli-*-jar-with-dependencies.jar) from the official OPSIN GitHub releases.

pip install pyopsin

From a source checkout:

pip install -e .

The JAR is cached under ~/.cache/pyopsin/ on Unix and %LOCALAPPDATA%\pyopsin on Windows. Override the cache directory with the PYOPSIN_JAR_DIR environment variable. You can also pass a local JAR with PyOpsin(path="/path/to/opsin-cli.jar").

Usage

Here's an example of how to use the PyOPSIN to generate SMILES string using the IUPAC name for a molecule:

from pyopsin import PyOpsin

# create an PyOpsin object
opsin = PyOpsin()

# generate the SMILES string from an IUPAC name for a molecule
name = "2,4,6-trinitrotoluene"
smiles = opsin.to_smiles(name)

# print the SMILES string
print(smiles)

This should output the following SMILES string:

["[N+](=O)([O-])C1=C(C)C(=CC(=C1)[N+](=O)[O-])[N+](=O)[O-]"]

Here is another example of how to invoke parallel processing to process a list of IUPAC names:

names = ["2,4,6-trinitrotoluene", "cycloheptene"]
smiles = opsin.to_smiles(names, num_workers=2)
print(smiles)

This should output the following SMILES strings:

["[N+](=O)([O-])C1=C(C)C(=CC(=C1)[N+](=O)[O-])[N+](=O)[O-]", "C1=CCCCCC1"]

With the default capture_message=False, to_smiles returns a list of SMILES strings. Pass capture_message=True to include the OPSIN parse status and message for each name. Each result is a dictionary with these keys:

• chemical_name: the input IUPAC name
• smiles: the SMILES string, or None when parsing fails
• message: the OPSIN message
• status: the OPSIN status name, such as SUCCESS, WARNING, or FAILURE

opsin = PyOpsin(capture_message=True)

names = ["dichlorobenze", "dichlorobenzene"]
results = opsin.to_smiles(names)
print(results)

This should output:

[
    {
        "chemical_name": "dichlorobenze",
        "smiles": None,
        "message": "dichlorobenze is unparsable due to the following being uninterpretable: benze The following was not parseable: e",
        "status": "FAILURE",
    },
    {
        "chemical_name": "dichlorobenzene",
        "smiles": "ClC1=C(C=CC=C1)Cl",
        "message": "APPEARS_AMBIGUOUS: Connection of chloro to benzen",
        "status": "WARNING",
    },
]

to_smiles_single returns the same dictionary for one name.

Tests

python -m unittest test.py

TestPyOpsinClean checks SMILES conversion for valid names. TestPyOpsinAmbiguous checks capture_message=True for an unparsable name (FAILURE) and an ambiguous name (WARNING).

Acknowledgments

The OPSIN Python wrapper is built on top of the OPSIN package, which was developed by the Centre for Molecular Informatics at the University of Cambridge. We would like to thank the developers of OPSIN for creating such a powerful and useful tool.