Geometric constraints in Julia package

[lxvm]

This is an initial implementation of the length scale constraints in the Julia package. There is a high-level api provided by SSP.constraint_solid and SSP.constraint_void that calculates the constraints, with some limitations, and has ChainRulesCore.jl rrules. I also have a low-level api in SSP.Constrain that provides a LengthConstraintProblem with more flexibility regarding the points at which the constraints are imposed and manual adjoints.

An example of the high-level api is in SSP/examples/julia/ssp_constrained.jl and an example of the low-level api is in SSP/examples/julia/constraint.jl

I have tested these on real problems and they are working well.

I still need to write documentation for the new apis. Update: done

[stevengj]

Could @Riarrieta take a look?

[Riarrieta]

Thanks @lxvm.

I checked the geometric constraint formulas and units—they seem correct.
I added the following:

• added a constraint duality test cheking that 'constraint_solid(x) == constraint_void(1-x)'. (8c49e79)
• the geometric constraint decay rate is no longer a magic number (c = 64), now it can be specified to the GeometricConstraint struct. (cc234d2)
• in the pythonic api, now 'constraint_solid' and 'constraint_void' can receive 'conic_radius' and 'constraint_threshold'. (cc234d2)
• added examples/julia/ssp_constrained_paper.jl, which replicates Fig. 1 of [1]. (45e6413)

[1] R. Arrieta, G. Romano, and S. G. Johnson, “Hyperparameter-free minimum-lengthscale constraints for topology optimization,” arXiv.org e-Print archive, 2507.16108, July 2025.