random
================

.. py:module:: jaxns.internals.random

.. rubric:: :code:`jaxns.internals.random`


.. rubric:: Module Contents

.. py:function:: random_ortho_matrix(key, n, special_orthogonal = False)

   Samples a random orthonormal n by n matrix from Stiefels manifold.
   From https://stackoverflow.com/a/38430739

   :param key: PRNG seed
   :param n: Size of matrix, draws from O(num_options) group.

   Returns: random [num_options,num_options] matrix with determinant = +-1


.. py:function:: resample_indicies(key, log_weights = None, S = None, replace = True, num_total = None)

   Get resample indicies according to a given weighting, with or without replacement.

   :param key: PRNGKey
   :param log_weights: Optional log weights
   :param S: Optional number of samples. Computes effective sample size from log weights if not given.
   :param replace: whether to use replacement or not.
   :param num_total: Optional total sample size to use, must be given if `replace=False` and `log_weights=None`

   :returns: index array given the take indicies to resample at.