jesterTOV.inference.base.prior.Fixed

class Fixed(value, parameter_names)

Bases: Prior

A parameter fixed to a constant value, excluded from the sampling space.

This is not a proper prior distribution — it has no log_prob or sample implementation. Use it in .prior files to pin a parameter to a specific value while keeping the specification co-located with the sampled priors:

lambda_BL = Fixed(0.0, parameter_names=["lambda_BL"])

The parser will extract Fixed entries into a separate fixed_params dict and will not add them to the CombinePrior that defines the sampling space.

Parameters:

• value (float) – The fixed value for the parameter.

• parameter_names (list[str]) – Must contain exactly one parameter name.

__init__(value, parameter_names)

Methods

__init__(value, parameter_names)
add_name(x)	Turn an array into a dictionary.
log_prob(z)
sample(rng_key, n_samples)

Attributes

n_dim
value
parameter_names
composite

log_prob(z)

Return type:

Float

sample(rng_key, n_samples)

Return type:

dict[str, Float[Array, 'n_samples']]

value: float