API

This page contains the main API reference.

Core Primitives

pmrf.Model(*[, name, metadata])	Base class for RF models.
pmrf.Module(*[, name, metadata])	Base class for parameter-aware objects in ParamRF.
pmrf.Frequency([start, stop, npoints, unit])	A frequency axis used to evaluate models over.
pmrf.Param(*[, value, distribution, ...])	The canonical parameter container for ParamRF.
pmrf.Unconstrained(value, *[, fixed, scale, ...])	Create an unconstrained free parameter.
pmrf.Fixed(value, *[, name, scale, metadata])	Create a fixed parameter.
pmrf.Bounded(lower, upper, *[, value, ...])	Create a free parameter constrained within a specific interval.
pmrf.Constrained(constraint, value, *[, ...])	Create a free parameter constrained to a specific domain.
pmrf.Random(distribution, *[, constraint, ...])	Create a free parameter with an associated probability distribution.

Working with Parameters

pmrf.params(tree[, where, free_only])	Returns the parameters of a model, or any collection of models, by name.
pmrf.values(tree[, where, free_only, space])	Returns the values of a model's parameters by name, in one space.
pmrf.log_prior(tree, *[, space])	Returns the log prior density of a model's parameters, in one space.
pmrf.update(tree, selection, node, /, *, ...)	Returns a copy of a model with the parts a selector picks replaced.
pmrf.tie(tree, target, ~typing.Any], source, ...)	Returns a copy of a model in which one part is derived from another.
pmrf.resolve(tree)	Returns a copy of a tree with its ties applied and its parameters left intact.
pmrf.derived(fn)	Turns a function of a base model and new parameters into a derived model constructor.

Main Modules

pmrf.models	RF models, components, surrogates, and adapters, and their solvers.
pmrf.modules	Parameter-aware wrappers for ParamRF modules.
pmrf.materials	Frequency-dependent materials.
pmrf.parameters	Parameter factories, converters, and field specifiers.
pmrf.stats	Public stats modules and their common entry points.
pmrf.objectives	Public objectives modules and their common entry points.
pmrf.optimize	Non-linear optimization of RF models.
pmrf.infer	Bayesian inference of RF models.
pmrf.fitting	Higher-level fitting utilities.

Other Modules

pmrf.math	Core maths functions.
pmrf.rf	Core RF conversions and system representations.
pmrf.serialization	Model loading and saving.
pmrf.viz	Plotting and vizualization.

Utilities

pmrf.NetworkCollection([networks, name, params])	A container for a collection of scikit-rf Networks.
pmrf.Partial(func, /, *args, **kwargs)	(experimental) Like functools.partial, but JAX-compatible.
pmrf.Attrgetter(*attrs)	(experimental) Like operator.attrgetter, but registered as a PyTree.
pmrf.Pathgetter(*paths)	(experimental) Retrieves values from a PyTree given JAX tree paths.
pmrf.InitVar(type)
pmrf.is_param(x)	Returns if x is an instance of pmrf.Param.
pmrf.as_param([value, constraint, scale, ...])	Coerces a value into a parameter.
pmrf.param(*, default, as_free, as_fixed, ...)	A field specifier for registering parameters within a model.
pmrf.partition(pytree, filter_spec[, is_leaf])	Splits a PyTree into two halves based on a filtering rule.
pmrf.combine(*pytrees[, is_leaf])	Recombines previously partitioned PyTrees into a single, functional PyTree.
pmrf.batch_axes(batched, template, *[, is_leaf])	Generates an in_axes PyTree by comparing a batched model to a template.
pmrf.batch_mask(batched, template, *[, is_leaf])	Generates a boolean PyTree mask identifying which leaves have a batch dimension.
pmrf.field(*[, converter, static])	Equinox supports extra functionality on top of the default dataclasses.
pmrf.load(source)	(experimental) Load a ParamRF Model (or any Parax PyTree) from a file.
pmrf.save(target, tree)	(experimental) Save a ParamRF Model (or any Parax PyTree) to a file.
pmrf.derivative(eval_fn, *args[, space])	Computes the exact derivative of a function.
pmrf.sweep(eval_fn, *args[, grid, template])	Vectorizes a function over the given input arguments.
pmrf.freeze(value)	Freezes/fixes a parameter or entire model and returns the frozen model.
pmrf.unfreeze(value)	Unfreezes/unfixes a potentially frozen parameter or model and returns the unfrozen model.
pmrf.replace(obj, /, **changes)	Returns a copy of a dataclass, such as a model or parameter, with fields replaced.
pmrf.unwrap(tree[, only_if, cascade])	Recursively resolves AbstractUnwrappable nodes within a PyTree.
pmrf.unwrap_self(method)	Decorator for Equinox module methods.