Color the 3D structure by a property #303
Conversation
|
I should note that I've implemented a color_function that includes a check for undefined values (such as when taking the square root of a negative value of the selected property). In such cases, the atom is assigned a gray color. |
|
I think you should start by adding a minimal example to the |
|
Another observation is that IMO we REALLY need to provide a rudimentary color scale, otherwise it's very hard to read the figures. I read a bit and it seems the best way to do this would be to add an HTML entity outside the widget, a bit like the floating buttons in the top-right cornier. |
The idea was to us a custom label, for which we can set the background to an arbitrary HTML canvas, for which doing a gradient would be fairly easy. But let's make sure the core of the functionality works first, merge it and then work on the color bar in a separate PR to reduce the size of the individual PRs. |
Co-Authored-By: Angelo Rossi <[email protected]>
MaximeScope
force-pushed
the
3dPropColor
branch
from
803f1ad
to
b711fc9
Compare
The coloring is now dependent of the selected structure. Added a support of missing properties. Atoms with missing property value are colored in dark gray when the property is selected.
I have addressed the issue, and the code should now work correctly. Initially, the code was selecting the first atoms independently from the structure. With the help of the environnements, the visualizers should now accurately assign colors based on the correct atoms property value. |
'RWB' should really be bwr
|
OK I have added an example that generates a simple color-coded figure. I have also renamed the RWB color scheme because it's blue first. I think one important thing is we can do to make it possible to have a reference for the colors is to make sure we have color schemes that match the plotly one. BWR is very similar to seismic (and this I have used in the example) but it'd be nice to have exactly the same schemes - at least viridis and magma for instance. |
|
Just commenting that this Fixes #118 for good |
This kind of comment needs to be in the first message of the PR for github to automatically close the issue |
python/examples/colors.py
Outdated
| for a in frames: | ||
| a.arrays["polarizability"] = np.array( | ||
| [ | ||
| (axx + ayy + azz) / 3 | ||
| for (axx, ayy, azz) in zip( | ||
| a.arrays["axx"], a.arrays["ayy"], a.arrays["azz"] | ||
| ) | ||
| ] | ||
| ) | ||
|
|
||
| # one possible measure of anisotropy... | ||
| a.arrays["alpha_eigenvalues"] = np.array( | ||
| [ | ||
| np.linalg.eigvalsh([[axx, axy, axz], [axy, ayy, ayz], [axz, ayz, azz]]) | ||
| for (axx, ayy, azz, axy, axz, ayz) in zip( | ||
| a.arrays["axx"], | ||
| a.arrays["ayy"], | ||
| a.arrays["azz"], | ||
| a.arrays["axy"], | ||
| a.arrays["axz"], | ||
| a.arrays["ayz"], | ||
| ) | ||
| ] | ||
| ) | ||
|
|
||
| a.arrays["anisotropy"] = ( | ||
| a.arrays["alpha_eigenvalues"][:, 2] - a.arrays["alpha_eigenvalues"][:, 0] | ||
| ) |
There was a problem hiding this comment.
Do we really need this in the "atomic coloring" example?
There was a problem hiding this comment.
I don't know what you mean by that.
There was a problem hiding this comment.
well, I thought it was better to avoid adding new data files, but we need scalar properties to display.
I'll add a comment so it's clear it's not like one needs to do this in order to use the coloring.
|
merged! 🎉 Thanks @MaximeScope and @angerossi for your work implementing this! |
It should finally be possible to merge the coloring of the 3d structure with master.
There is an additional option to color the background atoms by property
=> (Environments -> more options -> color: prop, style: ball and stick)
EDIT:
Fixes #118