_base

class mlptrain.descriptor._base.Descriptor(name: str)

Bases: ABC

Abstract base class for molecular feature descriptors.

__init__(name: str)

Initializes the descriptor representation.

Parameters:: name (str) – Name of the descriptor. e.g., “ace_descriptor”,”soap_descriptor”,”mace_descriptor”

abstract compute_representation(configurations: Configuration | ConfigurationSet) → ndarray

Compute descriptor representation for a given molecular configuration.

Parameters:: configuration – A molecular structure (e.g., mlptrain.Configuration).
Returns:: The computed descriptor representation as a vector/matrix.
Return type:: np.ndarray

abstract kernel_vector(configuration, configurations, zeta: int = 4) → ndarray

Calculate the kernel matrix between a set of configurations where the kernel is: .. math:

K(p_a, p_b) = (p_a . p_b / (p_a.p_a x p_b.p.b)^1/2 )^ζ