pymiediff.special.pena_D1_n

Contents

pymiediff.special.pena_D1_n#

pymiediff.special.pena_D1_n(n: Tensor | int, z: Tensor, n_add: int = 15, eps: float = 1e-30, precision: str = 'double')#

Compute D_n^(1)(z) = psi'_n(z) / psi_n(z) by downward recurrence.

Parameters:

  • n (int or torch.Tensor) – Maximum order.

  • z (torch.Tensor) – Complex argument(s).

  • n_add (int, default=15) – Extra recurrence depth for stable seeding.

  • eps (float, default=1e-30) – Small denominator safeguard.

  • precision ({"single", "double"}, default="double") – Complex dtype selection.

Returns:

Log-derivative values for orders 0..n.

Return type:

torch.Tensor

Contents