Update volume.py

[AstyLavrinenko]

New function to plot the density of diffusion atoms using plotly

[AstyLavrinenko]

I have added the function to plot the density of diffusion atoms using plotly and save as .html. However, I didn't integrate it with the class. Please guys check it

[stefsmeets]

Thanks! No worries that it is not integrated, this is something we will take care of. I'll get back to this.

[stefsmeets]

I have moved the code to the plots submodule. For the isosurface plot I want to use the Volume we have as input directly. I think a nice way to go about this is to do the marching cubes on the existing volume (which is in fractional coordinate system). And then convert the vertices to cartesian coordinates. This avoids double work, and I think this will be a bit more robust than the current implementation.

Usage

from gemdat import Trajectory
from gemdat.volume import trajectory_to_volume

VASP_XML = '/home/stef/md-analysis-matlab-example-short/vasprun.xml'

equilibration_steps = 1250
diffusing_element = 'Li'

trajectory = Trajectory.from_vasprun(VASP_XML)
trajectory = trajectory[equilibration_steps:]

diff_trajectory = trajectory.filter(diffusing_element)

density_resolution = 0.3

vol = trajectory_to_volume(
    trajectory=diff_trajectory,
    resolution=density_resolution,
)

structure = vol.to_structure()

from gemdat import plots
    
plots.density(vol, structure)

Todo

• Add docstrings
• Add isosurface plot
• Check that isosurface aligns well with points
• Apply gaussian filter (should this be an option?)
• See if some code can be moved to Volume

Follow-up

• Add density plot to dashboard
• Prepare notebook demo-ing plot

[stefsmeets]

🚀

[AstyLavrinenko]

Nice! Thank you!

[AstyLavrinenko]

If I understand correctly the points are coordinates of Li during simulation now? Because initially the points were coordinates of sites. I probably can load the structure created from .cif to plot points as sites, but then lattice and cart_coord will not match the simulation cell

[stefsmeets]

It can be whatever you want 😅

site_stucture = load_known_material('argyrodite', supercell=(2,1,1))
plots.density(vol, site_structure)

Yeah, that the lattices don't match is an issue I can't do much about. You could create a new structure with the simulation lattice and the coordinates from the cif.