compas_cgal.triangulation.refined_delaunay_mesh incompatible with numpy >=2.4.0

[cookiedan42]

Describe the bug
having an installation of numpy >= 2.4.0 alongside compas_cgal==0.9.1 causes compas_cgal.triangulation.refined_delaunay_mesh function to fail with error message

E       TypeError: refined_delaunay_mesh(): incompatible function arguments. The following argument types are supported:
E           1. refined_delaunay_mesh(boundary_vertices: numpy.ndarray[dtype=float64, shape=(*, *)], internal_vertices: numpy.ndarray[dtype=float64, shape=(*, *)], holes: compas_cgal._types_std.VectorRowMatrixXd, curves: compas_cgal._types_std.VectorRowMatrixXd, min_angle: float = 0.0, max_length: float = 0.0, is_optimized: bool = True) -> tuple[numpy.ndarray[dtype=float64, shape=(*, *), order='C'], numpy.ndarray[dtype=int32, shape=(*, *), order='C']]
E       
E       Invoked with types: ndarray, ndarray, compas_cgal._types_std.VectorRowMatrixXd, compas_cgal._types_std.VectorRowMatrixXd, float, float, bool

.venv/lib/python3.12/site-packages/compas_cgal/triangulation.py:215: TypeError

To Reproduce
Reproduction of working and broken versions: https://github.com/cookiedan42/compas_cgal_poc

Expected behavior
function should not crash

Screenshots
If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS:
  • Ubuntu under Windows WSL
  • MacOs
• Python version:
  • 3.12 and 3.13
• Python package manager
  • uv

Suggested Solution
include numpy <= 2.4.0 as an explicit dependency in compas_cgal