feat(cad): Corey shape factor (c/sqrt(ab))

[nochallenge]

What

Adds measure::solid_bounding_box_corey_shape_factor and _tol — the Corey shape factor of a solid:

// CSF = c / sqrt(a*b)   (dimensionless, in (0, 1]),  a >= b >= c sorted AABB extents
pub fn solid_bounding_box_corey_shape_factor_tol(solid: &Solid, tol: f64) -> Result<f64, CadError>
pub fn solid_bounding_box_corey_shape_factor(solid: &Solid) -> Result<f64, CadError>

Why

The AABB shape family has the Zingg pair (elongation b/a #375, flatness c/b #381) but not the Corey shape factor — the standard sedimentology / hydraulics particle-shape descriptor that governs the drag and settling velocity of a non-spherical grain. CSF = 1 for an equant cube or sphere and tends to 0 for a flat plate or thin rod. It compounds the two Zingg axes: CSF = flatness * sqrt(elongation). A degenerate solid (zero intermediate or longest extent) returns CadError::Tessellation.

Test

solid_bounding_box_corey_shape_factor_is_c_over_sqrt_ab:

• (a) worked — box(2,4,6) -> 2/sqrt(24) ~= 0.40825;
• (b) threads solid_bounding_box_extents (feat(cad): add solid_bounding_box_extents (raw AABB dimensions) #363) — s0/sqrt(s1*s2);
• (c) cross-check threading flatness (chore(deps)(deps): bump rfd from 0.14.1 to 0.17.2 #381) + elongation (feat(cad): add solid_bounding_box_elongation (Zingg b/a ratio) #375) (non-tautological) — CSF = flatness*sqrt(elongation);
• (d) range + equant — 0 < CSF <= 1; cube and sphere -> 1;
• (e) _tol wrapper agrees with the default.

Both fns re-exported from lib.rs alphabetically. valenx-cad 66 lib tests (was 65); cargo clippy -p valenx-cad --all-targets -- -D warnings clean. Research-grade particle-shape descriptor.