tensorial.signals package#

Submodules#

tensorial.signals.bases module#

class tensorial.signals.bases.RadialSphericalBasis(irreps)[source]#

Bases: Attr

A combined basis of a set of radial functions and spherical harmonics

Parameters:

irreps (Union[None, Irrep, MulIrrep, str, Irreps, Sequence[str | Irrep | MulIrrep | tuple[int, Union[None, Irrep, MulIrrep, str, Irreps, Sequence[str | Irrep | MulIrrep | tuple[int, IntoIrreps]]]]]])

create_tensor(value)[source]#

Create the signal that represents the expansion of the signal function in this basis

Parameters:

value (array)

Return type:

array

abstractmethod evaluate(value)[source]#

Evaluate the basis at the passed value

class tensorial.signals.bases.SimpleRadialSphericalBasis(radial, spherical)[source]#

Bases: RadialSphericalBasis

Parameters:
evaluate(value)[source]#

Evaluate the basis functions at the given value

expand(x, coefficients)[source]#
Parameters:

  • x (array)

  • coefficients (array)

class tensorial.signals.bases.SphericalBasis(l_max, p_val=1, p_arg=-1)[source]#

Bases: Attr

A set of spherical harmonics basis functions

Parameters:

l_max (int)

create_tensor(value)[source]#
Return type:

array

evaluate(x)[source]#

Evaluate the spherical harmonics at the passed values.

Warning: It is assumed that the values are located on the unit sphere (i.e. normalised vectors), no check is made to enforce this.

Return type:

IrrepsArray

property l_max: int#
property p_arg: int#
property p_val: int#

tensorial.signals.expansion module#

Module for functions performing expansion of functions with a basis

tensorial.signals.expansion.expand(basis, function)[source]#
tensorial.signals.expansion.expand(basis, function)

Expand a function in the given basis

Parameters:
Return type:

array

tensorial.signals.expansion.expand_(basis, function)[source]#
Parameters:
Return type:

array

tensorial.signals.functions module#

class tensorial.signals.functions.DiracDelta(pos, weight=1.0)[source]#

Bases: Function

A Dirac delta with an optional weight

evaluate(x)[source]#

Evaluate the function at point x

class tensorial.signals.functions.Function[source]#

Bases: object

Base class for functions

abstractmethod evaluate(x)[source]#

Evaluate the function at point x

class tensorial.signals.functions.IsotropicGaussian(pos, sigma, weight=1.0)[source]#

Bases: Function

A 3D Gaussian with an optional weight and scalar sigma

evaluate(x)[source]#

Evaluate the function at point x

class tensorial.signals.functions.Sum(functions)[source]#

Bases: Function

A sum of other functions

Parameters:

functions (tuple)

evaluate(x)[source]#

Evaluate the function at point x

tensorial.signals.radials module#

class tensorial.signals.radials.E3nnPolyEnvelope(basis, smoothing_start, n0, n1)[source]#

Bases: RadialBasis

Polynomial envelope that can be used to make a radial basis smoothly approach zero at the cutoff

Parameters:

  • basis (RadialBasis)

  • smoothing_start (float)

  • n0 (int)

  • n1 (int)

evaluate(radius)[source]#

Evaluate the radial basis at r

class tensorial.signals.radials.E3nnRadial(basis, max_radius, number, *, cutoff=None, min_radius=0.0)[source]#

Bases: RadialBasis

Select a radial function from the one-hot linspace built into e3nn-jax

see: https://e3nn-jax.readthedocs.io/en/latest/api/radial.html

Parameters:

  • basis (str)

  • max_radius (float)

  • number (int)

property basis: str#
property cutoff: bool | None#
evaluate(radius)[source]#

Evaluate the radial basis at r

class tensorial.signals.radials.OrthoBasis(radials, n_samples)[source]#

Bases: RadialBasis

Parameters:
area_samples: Array#
evaluate(radius)[source]#

Evaluate the radial basis at r

f_samples: Array#
inner_product(val_a, val_b)[source]#
norm(val)[source]#
radial_samples: Array#
radial_step: Array#
class tensorial.signals.radials.RadialBasis(number, domain=(0.0, inf))[source]#

Bases: Attr

A set of radial basis functions

Parameters:

number (int)

create_tensor(value)[source]#
Return type:

array

property domain: tuple[float, float]#
abstractmethod evaluate(radius)[source]#

Evaluate the radial basis at r

property number: int#

Get the number of radial functions in the basis

Module contents#