Function dpca_reconstruct 

pub fn dpca_reconstruct(
    data: &FdMatrix,
    argvals: &[f64],
    dpca: &DpcaResult,
) -> Result<DpcaReconstruction, FdarError>

Reconstruct curves from DPCA dynamic scores via inverse dynamic filtering.

For each retained-component count K = 1..=ncomp, reconstructs the interior series X̂_K[t,j] = Σ_{c<K} Σ_l filters[c][l,j] · scores[t'+l, c] and records the integrated-L2 error over the fully-defined interior. The error is monotone non-increasing in K (optimal DPCA projection). fitted holds the full-ncomp reconstruction over the DPCA interior.

Errors

FdarError::InvalidDimension if argvals/data shapes are inconsistent with dpca (grid mismatch or score-length mismatch).