Mpboundary tracking

[subravvr]

This PR focuses on leveraging geometric information from the Lamé curve describing a melt pool boundary. Previously, a simple inside/outside check was performed to assign a voxel a melting state of 1 (inside) or 0 (outside).

However, with the Lamé curve for a boundary, we can compute the distance from the center of the melt pool reference frame to the boundary using Newton's method. Comparing this distance to the distance of the voxel from the center of the melt pool reference frame allows us not only to check if a point is inside or outside, but also on the boundary.

By tracking the melt pool boundaries, we can generate computational representations of repeated melting, akin to etched and polished surfaces. Additionally, we can track the intersections between boundaries to understand connectedness of melt pools. Currently, we are using this approach to try and calibrate melt pool dimensions in an inverse problem.

Main changes

• Newton's method for distance to boundary calculation
• Voxel state transition logic (melted -> boundary, unmelted-> boundary, boundary -> melted).
• New output data structure for porosity is a numpy int8 grid of 0 (unmelted), 1 (melted, interior), 2 (melted, boundary), and 3 (melted, intersection region).

Future work

• Need to rewrite some tests in order to account for the new output datatype.

[colemanjs]

@subravvr, this looks great! A few notes:

1. pre-commit run --all-files
2. Need to update readme regarding the new field output. Also, we may want to update the figure. This gif slicing through an RVE would be very informative for the user.

[colemanjs]

@subravvr, the visualization function needs to be updated. This is the current view:

[colemanjs]

@subravvr, Can you make the GIF in the README larger and a higher dpi to make the phases readable.

[colemanjs]

@subravvr, the api example works, but the synthetic_api example gives "no phases present in this dataset" upon visualization. We may want to default this behavior. This should be returning a fully melted mesh, right?