Feature/point cloud utils

[DamienGilliard]

This PR concerns the addition of point cloud utilities to diffCheck: cropping, substraction (one PC - other PC by comparing the distances of their points using KDSearch)

Cropping:

This component crops a point cloud based on a bbox or boundary values (set a minimum value for z coordinate, for example

Using bbox:

Using boundary to coordinates:

Subtracting

This component removes points from a point cloud when they are too close to another point cloud

In the image hereunder, the blue point cloud is the one we subtract from, in red the one we subtract with, and in green the resulting point cloud:

[DamienGilliard]

• Use 3D domain instead of individual boundaries

[DamienGilliard]

• Use 3D domain instead of individual boundaries

I did not find 3D domain, but I think that by providing the possibility to give individual boundary coordinates, I am solving a problem that does not exist and no one asked for it, so I have remove it !

[DamienGilliard]

• For point clouds, get all 3 boolean operations: intersect, union, difference. But no volume, so how do you define the boolean velues, only by distance ? TBD
• for cropping, have one case if a box is given and use o3d, or if any brep given, other case with .IsPointInside method of Rhino.Geometry
• for Cropping, output the points IN and Out separately

[DamienGilliard]

Cropping

Update: as discussed, the cropping now keeps the "in" and "out" (point 3 of comment above), and breps are supported (point 2 of comment above):

Point cloud booleans:

We now have point cloud booleans (point 1 of previouis comment): intersection, union, exclusion:

[DamienGilliard]

@eleniv3d I Believe this PR is ready for review ;)
I'll make another PR for the pose comparison component

[eleniv3d]

• Use 3D domain instead of individual boundaries

I did not find 3D domain, but I think that by providing the possibility to give individual boundary coordinates, I am solving a problem that does not exist and no one asked for it, so I have remove it !

Yeah, i guess one would use the domain box for this

[eleniv3d]

Left some first comments. Let me know what you think :)

[eleniv3d]

I believe the metadata doesn't match the current implementation.. inputs/outputs are different :)

I would also suggest a slight refactoring.. Since we are now outputting both the point clouds inside the box/brep, i would consider renaming the component to sth like DF Cloud Split and then I think inputs could be:

1. Cloud “The point cloud to split.”
2. Boundary “Closed Box or Brep that defines the inside region.”

with outputs:

1. Cloud Inside
2. Cloud Outside

That would also mean that we check internally if its is a box or brep. it just feels a bit weird to me to have inputs that are not used between box and brep

[DamienGilliard]

Yes, and actually I noticed that if you feed in a non-axis-aligned bounbding box, it computes the axis aligned bb of that box, so there is some improvement there too...

[DamienGilliard]

I improved it like this, let me know what you think !

[DamienGilliard]

@eleniv3d , I believed I answered to your comments. Sorry the metadata was quite sloppy.
Asking for a second review following the changes ;)
Thanks a lot !