abc
=============

.. py:module:: jaxns.framework.abc

.. rubric:: :code:`jaxns.framework.abc`


.. rubric:: Module Contents

.. py:class:: AbstractPrior

   Bases: :py:obj:`abc.ABC`


   Represents a generative prior.


.. py:class:: AbstractModel

   Bases: :py:obj:`abc.ABC`


   Represents a Bayesian model in terms of a generative prior, and likelihood function.


   .. py:method:: __hash__()
      :abstractmethod:


      Hash of the model.



   .. py:method:: sample_U(key)
      :abstractmethod:


      Sample uniformly from the prior in U-space.

      :param key: PRNGKey

      :returns: U-space sample



   .. py:method:: transform(U)
      :abstractmethod:


      Compute the prior sample.

      :param U: U-space sample

      :returns: prior sample



   .. py:method:: transform_parametrised(U)
      :abstractmethod:


      Compute the parametrised prior variables.

      :param U: U-space sample

      :returns: prior sample



   .. py:method:: forward(U, allow_nan = False)
      :abstractmethod:


      Compute the log-likelihood.

      :param U: U-space sample
      :param allow_nan: whether to allow nans in likelihood

      :returns: log likelihood at the sample



   .. py:method:: log_prob_likelihood(U, allow_nan = False)

      Compute the log-likelihood.

      :param U: U-space sample
      :param allow_nan: whether to allow nans in likelihood

      :returns: log likelihood at the sample



   .. py:method:: log_prob_prior(U)
      :abstractmethod:


      Computes the log-probability of the prior.

      :param U: The U-space sample

      :returns: the log probability of prior



   .. py:method:: log_prob_joint(U, allow_nan)

      Computes the log-joint probability of the model.

      :param U: The U-space sample
      :param allow_nan: whether to allow nans in likelihood

      :returns: the log joint probability of the model



   .. py:method:: prepare_input(U)
      :abstractmethod:


      Prepares the input for the likelihood function.

      :param U: The U-space sample

      :returns: the input to the likelihood function



   .. py:method:: sanity_check(key, S)
      :abstractmethod:


      Performs a sanity check on the model.

      :param key: PRNGKey
      :param S: number of samples to check

      :raises AssertionError: if any of the samples are nan.



.. py:class:: AbstractDistribution

   Bases: :py:obj:`abc.ABC`


   Represents adistribution.