xlsindy.catalog module

This file implement catalog classes, it provides a unified way to treat of catalog category through abstract classes

This file replace the catalog_gen function from older version.

class xlsindy.catalog.CatalogCategory(*args, **kwargs)

Bases: ABC

The class that implement each of the subdictionnary for the general SINDy framework.

Some variable are required :

catalog_length (int) : The total lenght of the output of the expanded catalog (should be infered without hard computation)
num_coordinate (int) : The number of coordinate of the system.

abstractmethod create_solution_vector(*args, **kwargs)

This method output the solution vector from any solution related data.

Returns:: an array of shape (catalog_length,1) containing the coefficient that replicate the solution system.
Return type:: np.ndarray

abstractmethod expand_catalog()

This method expand the catalog in a (catalog_length,num_coordinate) matrix.

Returns:: an array of shape (catalog_length,num_coordinate) containing all the function
Return type:: np.ndarray

abstractmethod label(): Return the label of the expanded catalog.

abstractmethod separate_by_mask(mask: ndarray) → tuple

Separate the catalog by a mask. The mask is a boolean array of shape (catalog_length,).

Parameters:: mask (np.ndarray) – a boolean array of shape (catalog_length,) to separate the catalog.
Returns:: a new CatalogCategory with the masked data. CatalogCategory: a new CatalogCategory with the remaining data.
Return type:: CatalogCategory

class xlsindy.catalog.CatalogRepartition(catalog_repartition: List[CatalogCategory])

Bases: object

A class that manage the repartition of the catalog. It is used to create a global catalog from different part of the catalog.

create_solution_vector(solution_data: List[tuple]) → ndarray

Create an unique solution vector from the catalog and the solution data. :param solution_data: the solution data composed of the different information to build the solution vector, each one dependend of the paradigm used [(Lagrangian,substitution),…,(coeff_matrix, binary_matrix)] :type solution_data: List[tuple]

Returns:: the solution vector
Return type:: np.ndarray

expand_catalog()

create a global catalog for the regression system

Returns:: a global catalog of shape (catalog_length,num_coordinate) containing all the function
Return type:: np.ndarray

label_catalog()

WARNING : NOT IMPLEMENTED YET Convert the catalog into label

Parameters:: catalog_repartition (List[tuple]) – a listing of the different part of the catalog used need to follow the following structure : [(“lagrangian”,lagrangian_catalog),…,(“classical”,classical_catalog,expand_matrix)]
Returns:: List of labels for the catalog.
Return type:: List[str]

reunite_solution_by_type(type_mask: List[str], masked_solution: ndarray, remaining_solution: ndarray)

Reunite the solution vector by type. The type is a list of string that match the label of the catalog. the type is the name of the class.

Parameters:

type_mask (List[str]) – a list of string that match the label of the catalog.
masked_solution (np.ndarray) – the masked solution vector to be reunited.
remaining_solution (np.ndarray) – the remaining solution vector to be reunited.

Returns:

the reunited solution vector.

Return type:

np.ndarray

separate_by_mask(mask: ndarray) → tuple

Separate the catalog by a mask. The mask is a boolean array of shape (catalog_length,).

Parameters:: mask (np.ndarray) – a boolean array of shape (catalog_length,) to separate the catalog.
Returns:: two CatalogRepartition with the masked data and the remaining data.
Return type:: tuple

separate_solution_by_type(solution: ndarray, type_mask: List[str]) → tuple

Separate the the solution vector by type. The type is a list of string that match the label of the catalog. the type is the name of the class.

Parameters:

solution (np.ndarray) – the solution vector to be separated.
type_mask (List[str]) – a list of string that match the label of the catalog.

Returns:

two CatalogRepartition with the masked data and the remaining data.

Return type:

tuple

seperate_by_type(type_mask: List[str]) → tuple

Separate the catalog by type. The type is a list of string that match the label of the catalog. the type is the name of the class.

Parameters:: type_mask (List[str]) – a list of string that match the label of the catalog.
Returns:: two CatalogRepartition with the masked data and the remaining data.
Return type:: tuple