dft.grids

Real- and reciprocal-space grids for toy plane-wave DFT.

import mlx_atomistic.dft.grids

Classes

`RealSpaceGrid`

class RealSpaceGrid
    def __init__(shape: Sequence[int], cell: Cell | Sequence[float] | Sequence[Sequence[float]])

Uniform fractional real-space grid in a periodic cell.

Parameters

Name	Type	Default	Description
`shape`	`Sequence[int]`
`cell`	`Cell \| Sequence[float] \| Sequence[Sequence[float]]`

Properties

dv float — Volume represented by one grid point.
lengths mx.array — Cell lengths in atomic units.
ndim int — Number of grid dimensions.
size int — Total number of grid points.
spacing mx.array — Grid-vector step lengths in atomic units.
volume float — Cell volume in bohr cubed.

Methods

`coordinates`

def coordinates() -> mx.array

Return cell-centered Cartesian grid coordinates with shape (*shape, 3).

Returns

mx.array

`ReciprocalGrid`

class ReciprocalGrid
    def __init__(real_grid: RealSpaceGrid, vectors: mx.array, g2: mx.array, zero_mask: mx.array, integer_g: mx.array | None = None, basis_matrix: mx.array | None = None, fingerprint: str | None = None)

Reciprocal-space vectors matching a RealSpaceGrid FFT layout.

Parameters

Name	Type	Default
`real_grid`	`RealSpaceGrid`
`vectors`	`mx.array`
`g2`	`mx.array`
`zero_mask`	`mx.array`
`integer_g`	`mx.array \| None`	`None`
`basis_matrix`	`mx.array \| None`	`None`
`fingerprint`	`str \| None`	`None`

Methods

`from_real_space`

def from_real_space(grid: RealSpaceGrid) -> ReciprocalGrid

Build reciprocal vectors in exact NumPy FFT frequency order.

Parameters

Name	Type	Default	Description
`grid`	`RealSpaceGrid`		Real-space grid whose FFT ordering and cell define the reciprocal descriptor.

Returns

ReciprocalGrid — Reciprocal metadata with exact integer G coordinates and a
ReciprocalGrid — deterministic grid fingerprint.