Adds continuity priors to the spline to ensure smooth transitions between knots.

Computes the domain-normalized $L_p$ norm of the $q(z)$ curvature density (selected by ctype) over the reconstruction range $[0, z_f]$. The case p = 2 yields the root-mean-square curvature; large p approaches the maximum (see ncm_spline_curvature_lp_norm()).

Gets the mean value of $q(z)$ curvature, i.e. the root-mean-square geometric curvature of $q(z)$ over $[0, z_f]$ (the $p = 2$,

NCM_SPLINE_CURVATURE_GEOMETRIC case of nc_hicosmo_qspline_lp_norm()).

Gets the deceleration-acceleration transition redshift, i.e. the smallest $z \in [0, z_f]$ where the reconstructed $q(z)$ crosses zero. Returns NaN when $q(z)$ has no sign change over the reconstruction range.

Computes the weight-normalized $L_p$ norm of the $q(z)$ curvature density (selected by ctype) over the reconstruction range $[0, z_f]$, see ncm_spline_curvature_weighted_lp_norm(). The weight abscissa is the redshift $z$, matching the internal $q(z)$ spline; a large $W(z)$ penalizes curvature near that $z$, a small one tolerates it.