A Python library for automatic calculation of quantum chemical descriptors for QSAR analysis using ORCA quantum chemistry software.

pip install orca-descriptors

Note: After installation, the orca_descriptors command-line tool will be available in your PATH. If you installed with pip install --user, you may need to add ~/.local/bin to your PATH:

# For bash/zsh (add to ~/.bashrc or ~/.zshrc)
export PATH="$HOME/.local/bin:$PATH"

# For fish shell (add to ~/.config/fish/config.fish)
set -gx PATH $HOME/.local/bin $PATH

After adding to PATH, restart your terminal or run source ~/.bashrc (or source ~/.zshrc).

poetry install

from orca_descriptors import Orca
from rdkit.Chem import MolFromSmiles, AddHs

# Initialize ORCA calculator
orca = Orca(
    script_path="orca",
    functional="PBE0",
    basis_set="def2-SVP",
    method_type="Opt",
    dispersion_correction="D3BJ",
    solvation_model="COSMO(Water)",
    n_processors=8,
)

# Create molecule from SMILES using RDKit
mol = AddHs(MolFromSmiles("C1=CC=CC=C1"))

# Calculate descriptors
homo = orca.homo_energy(mol)
lumo = orca.lumo_energy(mol)
gap = orca.gap_energy(mol)

# Additional descriptors
homo_minus_1 = orca.mo_energy(mol, index=-2)  # HOMO-1 energy
min_h_charge = orca.get_min_h_charge(mol)  # Minimum H charge
xy_shadow = orca.xy_shadow(mol)  # XY projection area
meric = orca.meric(mol)  # Electrophilicity index
logp = orca.m_log_p(mol)  # Log P coefficient
nrot = orca.num_rotatable_bonds(mol)  # Rotatable bonds
wiener = orca.wiener_index(mol)  # Wiener index
sasa = orca.solvent_accessible_surface_area(mol)  # SASA

After installation, you can use orca_descriptors as a command-line tool:

Calibrate time estimation by running a benchmark calculation. The benchmark uses benzene (C1=CC=CC=C1) as a standard test molecule for machine calibration:

orca_descriptors run_benchmark

Estimate calculation time for a molecule without running the actual calculation:

orca_descriptors approximate_time --molecule C1=CC=CC=C1

Automatic Parameter Scaling: The time estimation automatically scales benchmark data for different parameters (number of processors, functional, basis set). You don't need to re-run the benchmark if you change these parameters - the system will automatically recalculate the estimated time based on the existing benchmark data.

For example:

• If benchmark was run with 1 processor, estimation for 4 processors will automatically account for parallel efficiency
• If benchmark used def2-SVP, estimation for def2-TZVP will scale based on basis set size (O(N^3.5) scaling)
• Different functionals are scaled based on their relative computational costs

All parameters from the Orca class are available as command-line arguments:

• --script_path: Path to ORCA executable (default: 'orca')
• --working_dir: Working directory for calculations (default: current directory)
• --output_dir: Directory for output files (default: current directory)
• --functional: DFT functional (default: PBE0)
• --basis_set: Basis set (default: def2-SVP)
• --method_type: Calculation type: Opt, SP, or Freq (default: Opt)
• --dispersion_correction: Dispersion correction, e.g., D3BJ (default: D3BJ). Use 'None' to disable.
• --solvation_model: Solvation model, e.g., 'COSMO(Water)' (default: None). Use 'None' to disable.
• --n_processors: Number of processors (default: 1)
• --max_scf_cycles: Maximum SCF cycles (default: 100)
• --scf_convergence: SCF convergence threshold (default: 1e-6)
• --charge: Molecular charge (default: 0)
• --multiplicity: Spin multiplicity (default: 1)
• --cache_dir: Directory for caching results (default: output_dir/.orca_cache)
• --log_level: Logging level: DEBUG, INFO, WARNING, ERROR (default: INFO)
• --max_wait: Maximum time to wait for output file creation in seconds (default: 300)

# Run benchmark with custom parameters (uses benzene as standard test molecule)
orca_descriptors run_benchmark \
    --functional PBE0 \
    --basis_set def2-SVP \
    --n_processors 4 \
    --working_dir ./calculations

# Estimate time for optimization calculation
orca_descriptors approximate_time \
    --molecule CCO \
    --method_type Opt \
    --n_opt_steps 20 \
    --functional PBE0 \
    --basis_set def2-TZVP \
    --n_processors 8

• Python >= 3.10
• ORCA 6.0.1 installed and available in PATH
• RDKit >= 2023.0.0

See LICENSE.md