Add Rhodo (Rhombic Dodecahedron) Infill

[AndreVallestero]

The honeycomb pattern has the highest specific strength (strength to weight ratio) in 2D planar loading. The rhombic dodecahedron is the honeycomb of 3d polyhedra, so it should provide the highest isometric (omni-directional) specific strength.

I've verified that it builds and works as intended:

Here's a cross section of what the phases should look like:

[AndreVallestero]

Converting back to draft. I just realized that the phases for permutation 1 are wrong by 1 hexagon width

[AndreVallestero]

Fixed in the latest version and validated that the cross section matches the expected cross section at each phase; it should now be ready for review / merge.

Side note, I create the polylines in the most efficient way possible for print speed, but somewhere later in the slicing process, the polylines are no longer in order, and the resultant path is horribly inefficient. Would appreciate any help to make the ordering "sticky" for faster print times.

[RF47]

How did you manage to make the hexagonal pattern with a single line? Does it have bridges?
By the way, it's an incredible piece of work.

[AndreVallestero]

@RF47 The hexagonal pattern is made with 2 lines per row. And there is no bridging, only overhangs (should be small enough to not cause any issues).

I only recently figured out how the polylines are re-arranged by the slicer, so I'll be doing a pass of print time optimizations to hopefully make the time competitive with cubic and gyroid.

[RF47]

Implement your code in Orca Slicer and I see a problem: the hexagonal cells are made with a single line, so it is not possible to make continuous paths.

Perhaps an approach similar to regular honeycomb infill could be used to make continuous sections, which would make the print much faster.

[AndreVallestero]

Interesting, I didn't realize that honeycomb "doubled up" on some walls. Avoiding that was the main reason that I couldn't do it as one continuous line. I'll try to implement that later tonight.

[RF47]

I tested your latest implementation, and the problem I see now is that sometimes the nozzle goes back and forth along the same line. Not only does this waste time, but it can also drag material and cause a lot of retraction.