calipy.utils

This module provides support functionality related to logging, data organization, preprocessing and other functions not directly related to calipy core domains.

The classes are

CalipyRegistry: Dictionary type class that is used for tracking identity

and uniqueness of objects created during a run and outputs warnings.

The script is meant solely for educational and illustrative purposes. Written by Dr. Jemil Avers Butt, Atlas optimization GmbH, www.atlasoptimization.com.

Functions

check_schema(calipy_dict_obj[, ...])	Example schema validation function:
context_plate_stack(plate_stack)	Context manager to handle multiple nested pyro.plate contexts.
dim_assignment(dim_names[, dim_sizes, ...])	dim_assignment dynamically assigns dimension objects to names and returns them as a DimTuple.
ensure_tuple(item)	Ensures the input is a tuple.
format_mro(cls)
generate_trivial_dims(ndim)
get_params(name)
multi_unsqueeze(input_tensor, dims)
restricted_exec(exec_string, allowed_locals)
robust_meshgrid(tensors[, indexing])	Ensures that meshgrid also works for empty inputs [] of sizes.
safe_eval(expr[, allowed_globals, ...])

Classes

CalipyDim(name[, size, description])	CalipyDim class contains information useful to manage dimensions and is the prime ingredient to DimTuple class which implements arithmentics on dimensions.
DimTuple(input_tuple)	DimTuple is a custom subclass of Python's tuple designed to manage and manipulate tuples of dimension objects, such as those from CalipyDim.
InputSchema(required_keys[, optional_keys, ...])
TorchdimTuple(input_tuple[, superior_dims])	TorchdimTuple is a subclass of the Tuple class that allows esy handling of tuples build from functorchdim.dim.Dim objects.