strainedSMILES2xyz

strainedSMILES2xyz is a Python tool that uses RDKit and ORCA to generate 3D .xyz geometries from SMILES strings — specifically designed for strained molecules. It provides a simple interface for both scripting and command-line use.

📄 Note: A publication describing the method is in preparation. Link will be added here once available.

📦 Installation

Install via pip (requires Python 3.8+):

pip install strainedsmiles2xyz

⚠️ RDKit must be installed separately via conda or similar:

conda install -c conda-forge rdkit=2025.03.1

You will also need ORCA version 6.0 or higher in your PATH or set as ORCA_PATH to enable ORCA based corrections.

🚀 Usage

🔧 From the command line:

python -m strainedsmiles2xyz "SMILES"

or:

strainedsmiles2xyz "SMILES"

This prints an XYZ-formatted structure to stdout.

🐍 From Python:

from strainedsmiles2xyz import smiles_to_xyz

mol, info = smiles_to_xyz(
    "C1=CCC1",                     # SMILES input
    print_level=1,                # log verbosity
    no_chirality_enforced_allowed=True,  # allow fallback if stereochemistry fails
    use_orca_corrections=False,  #  apply ORCA optimization corrections
    n_cores=1                     # number of threads for ORCA
)

• mol is an RDKit Mol object
• info is a dictionary containing metadata and the "xyz_string" (raw XYZ output)

📁 Included Data

• notebooks/: Notebooks demonstrating the inner workings of the method, including how RDKit, Open Babel, and strainedSMILES2xyz perform in different cases
• benchmark/: Benchmarking results — XYZ files generated by different methods and re-optimized using GFN2-xTB

⚙ Requirements

• RDKit ≥ 2025.03.0 (Must be installed via Conda or other method — not available via PyPI)
• ORCA ≥ 6.0 (strongly recommended; used for correction step)

🧪 Example

strainedsmiles2xyz "C1=CCC1" > cyclobutene.xyz

📝 Citation

Coming soon — publication in preparation.

📨 Contact

Maintained by Dennis Svatunek. Contributions and suggestions are welcome.