Skip to content


A library for computational chemistry in python. Featuring support for molecules, crystals, Hirshfeld & promolecule density isosurfaces, spherical harmonic shape descriptors and much more...


While the library is intended to be flexible and make it easy to build complex pipelines or properties, the following is a brief summary of intended features:

  • Load crystal structures from .cif, .res, POSCAR files, and molecules from .xyz, .sdf files.
  • Evaluate promolecule and procrystal electron densities.
  • Easily generate Hirshfeld or promolecule isosurfaces and associated properties.
  • Easily generate spherical harmonic shape descriptors for atoms, molecules, or molecular fragments.
  • Efficiently calculate crystal slabs, periodic connectivity and more...

It should also serve as a simple, easy to read library for learning how to represent crystal structures, molecules etc. and evaluate scientifically relevant information quickly and efficiently using python.

At a glance

Crystal structures and molecules

Loading a crystal structure from a CIF (.cif) or SHELX (.res) file, or a molecule from an XMOL (.xyz) file is straightforward:

>>> from chmpy import Crystal, Molecule
>>> c = Crystal.load("tests/test_files/acetic_acid.cif")
>>> c
<Crystal C2H4O2 Pna2_1>

# Calculate the unique molecules in this crystal
>>> c.symmetry_unique_molecules()
[<Molecule: C2H4O2(2.12,1.15,0.97)>]

>>> m = Molecule.load("tests/test_files/")
>>> m
<Molecule: H2O(-0.67,-0.00,0.01)>

Hirshfeld and promolecule density isosurfaces

Generation of surfaces with the default settings can be done with minimal hassle, simply by using the corresponding members of the Crystal class:

>>> c = Crystal.load("tests/test_files/acetic_acid.cif")

# This will generate a high resolution surface
# for each symmetry unique molecule in the crystal
>>> surfaces = c.hirshfeld_surfaces()
>>> surfaces
[<trimesh.Trimesh(vertices.shape=(3598, 3), faces.shape=(7192, 3))>]

# We can generate lower resolution surfaces with the separation parameter
>>> surfaces = c.hirshfeld_surfaces(separation=0.5)
>>> surfaces
[<trimesh.Trimesh(vertices.shape=(584, 3), faces.shape=(1164, 3))>]

# Surfaces can be saved via trimesh
>>> surfaces[0].export("acetic_acid_trimesh.ply", "ply")
# or a utility function provided in chmpy
>>> from chmpy.util.mesh import save_mesh
>>> save_mesh(surfaces[0], "acetic_acid.ply")