aidsorb.visualize#

Helper functions for visualizing input representations.

Tip

To visualize point clouds or voxels using the CLI:

$ aidsorb visualize path/to/pcd_or_voxels.npy

To visualize a structure you can use ase:

from ase.io import read
from ase.visualize import view

atoms = read('path/to/file')
view(atoms)
aidsorb.visualize.draw_from_file(filename, render=True, **kwargs)[source]#

Visualize point cloud or voxels from a file.

Parameters:

  • filename (str) – Absolute or relative path to a .npy.

  • render (bool, default=True) – Whether to render the data with plotly.io.renderers.default or return the figure object.

  • **kwargs – Valid keyword arguments for draw_pcd() or draw_voxels().

Return type:

plotly.graph_objects.Figure or None

aidsorb.visualize.draw_pcd(pcd, molecular=True, scheme='cpk', size=2.0, feature_to_color=None, colorscale=None)[source]#

Visualize point cloud with Plotly.

Parameters:

  • pcd (array-like of shape (N, 3+C))

  • molecular (bool, default=True) – If True, assume a molecular point cloud. In this case, each atom is sized as \(r_{\text{vdW}}^4\) and colorized based on scheme. If False, assume a generic point cloud and size each point based on size.

  • scheme ({'jmol', 'cpk'}, default='cpk') – Takes effect only if molecular=True and feature_to_color=None.

  • size (float, default=2.) – Controls the size of points.

  • feature_to_color (tuple, optional) – Tuple of the form (index, label), where index is the index of the feature to be colored and label is the text label for the colorbar.

  • colorscale (str, optional) – No effect if feature_to_color=None. For available options, see colorscale.

Return type:

plotly.graph_objects.Figure

Examples

>>> pcd = np.random.randn(10, 3)
>>> fig = draw_pcd(pcd, molecular=False, feature_to_color=(0, 'x coord'), colorscale='viridis')
aidsorb.visualize.draw_voxels(voxels, isomin=-1000.0, isomax=1000.0, opacity=0.15, surface_count=5, colorscale=None, label='Energy / K')[source]#

Visualize voxels with Plotly.

Todo

Add support for multi-channel voxels.

Parameters:

  • voxels (array-like of shape (D, H, W))

  • isomin (float, default=-1000) – Minimum value of the displayed range.

  • isomax (float, default=1000) – Maximum value of the displayed range.

  • opacity (float, default=0.15) – Opacity of the volume rendering.

  • surface_count (int, default=5) – Number of isosurfaces to approximate the volume.

  • colorscale (str, optional) – For available options, see colorscale.

  • label (str, default='Energy / K') – Text label for the colorbar.

Return type:

plotly.graph_objects.Figure

Notes

Default parameter values are chosen to provide a reasonable visualization for energy voxels.

Examples

>>> voxels = np.random.randn(3, 3, 3)
>>> fig = draw_voxels(voxels, colorscale='viridis')
aidsorb.visualize.get_atom_colors(atomic_numbers, scheme='cpk')[source]#

Convert atomic numbers to colors based on scheme.

Parameters:

  • atomic_numbers (array-like of shape (N,))

  • scheme ({'jmol', 'cpk'}, default='cpk')

Returns:

colors

Return type:

array of shape (N,)

aidsorb.visualize.get_atom_names(atomic_numbers)[source]#

Convert atomic numbers to element names.

Parameters:

atomic_numbers (array-like of shape (N,))

Returns:

elements

Return type:

array of shape (N,)

Examples

>>> atomic_numbers = np.array([1, 2, 7])
>>> get_atom_names(atomic_numbers)
array(['Hydrogen', 'Helium', 'Nitrogen'], dtype=object)