qmctorch.wavefunction.jastrows.distance package

Submodules

qmctorch.wavefunction.jastrows.distance.electron_electron_distance module

class qmctorch.wavefunction.jastrows.distance.electron_electron_distance.ElectronElectronDistance(*args: Any, **kwargs: Any)[source]

Bases: Module

Computes the electron-electron distances

\[r_{ij} = \sqrt{ (x_i-x_j)^2 + (y_i-y_j)^2 + (z_i-z_j)^2}\]

Parameters:

nelec (int) – number of electrons
ndim (int, optional) – number of spatial dimensions. Defaults to 3.
scale (bool, optional) – return scaled values. Defaults to False.
scale_factor (float, optional) – value of the scale factor. Defaults to 0.6.

Examples::

>>> edist = ElectronDistance(2,3)
>>> pos = torch.tensor(500,6)
>>> r = edist(pos)
>>> dr = edist(pos,derivative=1)

forward(input: torch.Tensor, derivative: int = 0) → torch.Tensor[source]

Compute the pairwise distance between the electrons or its derivative.

When required, the derivative is computed wrt to the first electron i.e.

\[\frac{dr_{ij}}{dx_i}\]

which is different from :

\[\frac{d r_{ij}}{dx_j} = -\frac{dr_{ij}}{dx_i}\]

Parameters:

input (torch.Tensor) – position of the electron size : Nbatch x [Nelec x Ndim]
derivative (int, optional) – degre of the derivative. Defaults to 0.

Returns:

distance (or derivative) matrix: Nbatch x Nelec x Nelec if derivative = 0 Nbatch x Ndim x Nelec x Nelec if derivative = 1,2

Return type:

torch.Tensor

safe_sqrt(dist: torch.Tensor) → torch.Tensor[source]

Compute the square root of the electron electron distance matrix.

Parameters:: dist (torch.tensor) – ee distances squared Nbatch x Nelec x Nelec
Returns:: sqrt of dist Nbatch x Nelec x Nelec
Return type:: torch.tensor

get_der_distance(pos: torch.Tensor, dist: torch.Tensor) → torch.Tensor[source]

Get the derivative of the electron electron distance matrix.

\[\frac{d r_{ij}}{d x_i}\]

Parameters:

pos (torch.tensor) – positions of the electrons Nbatch x Nelec x Ndim
dist (torch.tensor) – distance matrix between the elecs Nbatch x Nelec x Nelec

Returns:

[description]

Return type:

[type]

get_second_der_distance(pos: torch.Tensor, dist: torch.Tensor) → torch.Tensor[source]

Get the second derivative of the electron electron distance matrix.

\[\frac{d^2 r_{ij}}{d x_i^2}\]

Parameters:

pos (torch.tensor) – positions of the electrons Nbatch x Nelec x Ndim
dist (torch.tensor) – distance matrix between the elecs Nbatch x Nelec x Nelec

Returns:

[description]

Return type:

[type]

static get_distance_quadratic(pos: torch.Tensor) → torch.Tensor[source]

Compute the distance following a quadratic expansion

Parameters:: ndim] (pos {torch.tensor} -- electron position [nbatch x nelec x)
Returns:: torch.tensor – distance matrices nbatch x nelec x ndim]

static get_difference(pos: torch.Tensor) → torch.Tensor[source]

Compute the difference ri - rj

Parameters:: ndim] (pos {torch.tensor} -- electron position [nbatch x nelec x)
Returns:: torch.tensor – distance matrices nbatch x nelec x nelec x ndim]

qmctorch.wavefunction.jastrows.distance.electron_nuclei_distance module

class qmctorch.wavefunction.jastrows.distance.electron_nuclei_distance.ElectronNucleiDistance(*args: Any, **kwargs: Any)[source]

Bases: Module

Computes the electron-nuclei distances

\[r_{iA} = \sqrt{ (x_i-x_A)^2 + (y_i-y_A)^2 + (z_i-z_A)^2}\]

Parameters:

nelec (int) – number of electrons
atomic_pos (torch.tensor) – positions of the atoms
ndim (int) – number of spatial dimensions
scale (bool, optional) – return scaled values, Defaults to False
scale_factor (float, optional) – value of the scale factor, Defaults to 0.6

Examples::

>>> endist = ElectronNucleiDistance(2,2)
>>> epos = torch.tensor(500,6)
>>> r = edist(pos)
>>> dr = edist(pos,derivative=1)

forward(input: torch.Tensor, derivative: int = 0) → torch.Tensor | Tuple[torch.Tensor, torch.Tensor][source]

Compute the pairwise distances between electrons and atoms or their derivative.

Parameters:

input (torch.Tensor) –
position of the electron

size : Nbatch x [Nelec x Ndim]
derivative (int, optional) –
degre of the derivative.

Defaults to 0.

Returns:

distance (or derivative) matrix

Nbatch x Nelec x Natom if derivative = 0

Nbatch x Ndim x Nelec x Natom if derivative = 1,2

Return type:

Union[torch.Tensor, Tuple[torch.Tensor, torch.Tensor]]

get_der_distance(pos: torch.Tensor, dist: torch.Tensor) → torch.Tensor[source]

Get the derivative of the electron-nuclei distance matrix

\[\frac{d r_{iA}}{d x_i}\]

Parameters:

pos (torch.tensor) – positions of the electrons Nbatch x Nelec x Ndim
dist (torch.tensor) – distance matrix between the elecs Nbatch x Nelec x Nelec

Returns:

[description]

Return type:

[type]

get_second_der_distance(pos: torch.Tensor, dist: torch.Tensor) → torch.Tensor[source]

Get the derivative of the electron-nuclei distance matrix

\[\frac{d^2 r_{iA}}{d x_i^2}\]

Parameters:

pos (torch.tensor) – positions of the electrons Nbatch x Nelec x Ndim
dist (torch.tensor) – distance matrix between the elecs Nbatch x Nelec x Nelec

Returns:

[description]

Return type:

[type]

qmctorch.wavefunction.jastrows.distance.scaling module

qmctorch.wavefunction.jastrows.distance.scaling.get_scaled_distance(kappa: float, r: torch.Tensor) → torch.Tensor[source]

compute the scaled distance

\[u_{ij} = \frac{1-e^{-\kappa r_{ij}}}{\kappa}\]

Parameters:

kappa (float) – scaling factor
r (torch.tensor) – matrix of the e-e distances Nbatch x Nelec x Nelec

Returns:

values of the scaled distance: Nbatch, Nelec, Nelec

Return type:

torch.tensor

qmctorch.wavefunction.jastrows.distance.scaling.get_der_scaled_distance(kappa: float, r: torch.Tensor, dr: torch.Tensor) → torch.Tensor[source]

Returns the derivative of the scaled distances

\[\frac{d u}{d x_i} = \frac{d r_{ij}}{d x_i} e^{-\kappa r_{ij}}\]

Parameters:

kappa (float) – scaling factor
r (torch.tensor) – matrix of the e-e distances Nbatch x Nelec x Nelec
dr (torch.tensor) – matrix of the derivative of the e-e distances Nbatch x Ndim x Nelec x Nelec

Returns:

deriative of the scaled distance: Nbatch x Ndim x Nelec x Nelec

Return type:

torch.tensor

qmctorch.wavefunction.jastrows.distance.scaling.get_second_der_scaled_distance(kappa: float, r: torch.Tensor, dr: torch.Tensor, d2r: torch.Tensor) → torch.Tensor[source]

computes the second derivative of the scaled distances

\[\frac{d^2u_{ij}}{d x_i^2} = \frac{d^2r_{ij}}{d x_i^2} -\kappa \left( \frac{d r_{ij}}{d x_i} \right)^2 e^{-\kappa r_{ij}}\]

Parameters:

kappa (float) – scaling factor
r (torch.tensor) – unsqueezed matrix of the e-e distances Nbatch x Nelec x Nelec
dr (torch.tensor) – matrix of the derivative of the e-e distances Nbatch x Ndim x Nelec x Nelec
d2r (torch.tensor) – matrix of the 2nd derivative of the e-e distances Nbatch x Ndim x Nelec x Nelec

Returns:

second deriative of the scaled distance: Nbatch x Ndim x Nelec x Nelec

Return type:

torch.tensor

Module contents

class qmctorch.wavefunction.jastrows.distance.ElectronElectronDistance(*args: Any, **kwargs: Any)[source]

Bases: Module

Computes the electron-electron distances

\[r_{ij} = \sqrt{ (x_i-x_j)^2 + (y_i-y_j)^2 + (z_i-z_j)^2}\]

Parameters:

nelec (int) – number of electrons
ndim (int, optional) – number of spatial dimensions. Defaults to 3.
scale (bool, optional) – return scaled values. Defaults to False.
scale_factor (float, optional) – value of the scale factor. Defaults to 0.6.

Examples::

>>> edist = ElectronDistance(2,3)
>>> pos = torch.tensor(500,6)
>>> r = edist(pos)
>>> dr = edist(pos,derivative=1)

forward(input: torch.Tensor, derivative: int = 0) → torch.Tensor[source]

Compute the pairwise distance between the electrons or its derivative.

When required, the derivative is computed wrt to the first electron i.e.

\[\frac{dr_{ij}}{dx_i}\]

which is different from :

\[\frac{d r_{ij}}{dx_j} = -\frac{dr_{ij}}{dx_i}\]

Parameters:

input (torch.Tensor) – position of the electron size : Nbatch x [Nelec x Ndim]
derivative (int, optional) – degre of the derivative. Defaults to 0.

Returns:

distance (or derivative) matrix: Nbatch x Nelec x Nelec if derivative = 0 Nbatch x Ndim x Nelec x Nelec if derivative = 1,2

Return type:

torch.Tensor

safe_sqrt(dist: torch.Tensor) → torch.Tensor[source]

Compute the square root of the electron electron distance matrix.

Parameters:: dist (torch.tensor) – ee distances squared Nbatch x Nelec x Nelec
Returns:: sqrt of dist Nbatch x Nelec x Nelec
Return type:: torch.tensor

get_der_distance(pos: torch.Tensor, dist: torch.Tensor) → torch.Tensor[source]

Get the derivative of the electron electron distance matrix.

\[\frac{d r_{ij}}{d x_i}\]

Parameters:

pos (torch.tensor) – positions of the electrons Nbatch x Nelec x Ndim
dist (torch.tensor) – distance matrix between the elecs Nbatch x Nelec x Nelec

Returns:

[description]

Return type:

[type]

get_second_der_distance(pos: torch.Tensor, dist: torch.Tensor) → torch.Tensor[source]

Get the second derivative of the electron electron distance matrix.

\[\frac{d^2 r_{ij}}{d x_i^2}\]

Parameters:

pos (torch.tensor) – positions of the electrons Nbatch x Nelec x Ndim
dist (torch.tensor) – distance matrix between the elecs Nbatch x Nelec x Nelec

Returns:

[description]

Return type:

[type]

static get_distance_quadratic(pos: torch.Tensor) → torch.Tensor[source]

Compute the distance following a quadratic expansion

Parameters:: ndim] (pos {torch.tensor} -- electron position [nbatch x nelec x)
Returns:: torch.tensor – distance matrices nbatch x nelec x ndim]

static get_difference(pos: torch.Tensor) → torch.Tensor[source]

Compute the difference ri - rj

Parameters:: ndim] (pos {torch.tensor} -- electron position [nbatch x nelec x)
Returns:: torch.tensor – distance matrices nbatch x nelec x nelec x ndim]

class qmctorch.wavefunction.jastrows.distance.ElectronNucleiDistance(*args: Any, **kwargs: Any)[source]

Bases: Module

Computes the electron-nuclei distances

\[r_{iA} = \sqrt{ (x_i-x_A)^2 + (y_i-y_A)^2 + (z_i-z_A)^2}\]

Parameters:

nelec (int) – number of electrons
atomic_pos (torch.tensor) – positions of the atoms
ndim (int) – number of spatial dimensions
scale (bool, optional) – return scaled values, Defaults to False
scale_factor (float, optional) – value of the scale factor, Defaults to 0.6

Examples::

>>> endist = ElectronNucleiDistance(2,2)
>>> epos = torch.tensor(500,6)
>>> r = edist(pos)
>>> dr = edist(pos,derivative=1)

forward(input: torch.Tensor, derivative: int = 0) → torch.Tensor | Tuple[torch.Tensor, torch.Tensor][source]

Compute the pairwise distances between electrons and atoms or their derivative.

Parameters:

input (torch.Tensor) –
position of the electron

size : Nbatch x [Nelec x Ndim]
derivative (int, optional) –
degre of the derivative.

Defaults to 0.

Returns:

distance (or derivative) matrix

Nbatch x Nelec x Natom if derivative = 0

Nbatch x Ndim x Nelec x Natom if derivative = 1,2

Return type:

Union[torch.Tensor, Tuple[torch.Tensor, torch.Tensor]]

get_der_distance(pos: torch.Tensor, dist: torch.Tensor) → torch.Tensor[source]

Get the derivative of the electron-nuclei distance matrix

\[\frac{d r_{iA}}{d x_i}\]

Parameters:

pos (torch.tensor) – positions of the electrons Nbatch x Nelec x Ndim
dist (torch.tensor) – distance matrix between the elecs Nbatch x Nelec x Nelec

Returns:

[description]

Return type:

[type]

get_second_der_distance(pos: torch.Tensor, dist: torch.Tensor) → torch.Tensor[source]

Get the derivative of the electron-nuclei distance matrix

\[\frac{d^2 r_{iA}}{d x_i^2}\]

Parameters:

pos (torch.tensor) – positions of the electrons Nbatch x Nelec x Ndim
dist (torch.tensor) – distance matrix between the elecs Nbatch x Nelec x Nelec

Returns:

[description]

Return type:

[type]