Grids

class torusgrid.Grid(shape: Tuple[int, ...], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

Bases: ABC, Generic[T]

Base class for grids. The generic type T refers to the dtype of the wrapped numpy array (self.psi).

__init__(shape: Tuple[int, ...], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

abstract property isreal: bool: Return whether the field or grid object holds real data

set_psi(psi1: npt.NDArray[T] | FloatLike) → None: Deprecated: use psi[…] = … instead

initialize_fft(*, threads: int = 1, planning_timelimit: Optional[float] = None, effort: Optional[Literal['FFTW_ESTIMATE', 'FFTW_MEASURE', 'FFTW_PATIENT', 'FFTW_EXHAUSTIVE']] = None, wisdom_only: bool = False, destroy_input: bool = False, unaligned: bool = False) → None

Initialize the FFTW forward and backward plans. By default the fourier transform plans are not initialized as it can take a considerable amount of time.

Parameters:

effort: FFTW planning effort

wisdom_only: Raise an error if no plan for the current transforms is present

destroy_input: Whether to allow input to be destroyed during: transform, note that for certain transforms the input is destroyed anyway

unaligned: Alignment of the data will not be assumed.

fft(): Run forward FFT on psi and store results to psi_k If FFT plans are not initialized, a TypeError will be raised

ifft(): Run backward FFT on psi_k and store results to psi If FFT plans are not initialized, a TypeError will be raised

fft_initialized() → bool: Return whether the FFTW plans are initialized.

property shape: Tuple[int, ...]: Return the shape in real space (x-space / time domain etc)

property shape_k: Tuple[int, ...]: Return the shape in Fourier (k-space / frequency domain etc)

property rank: Return the number of axes. Same as numpy’s ndim.

property numel: int: Return the number of elements. Same as numpy’s size() or torch’s numel()

property fft_axes: Return the axes along which FFT is performed

property last_fft_axis: Return the last FFT axis. Useful when dealing with RFFTs.

property psi: Real space data To assign values, use .psi[…] = …, not .psi = …

property psi_k: k-space data To assign values, use .psi_k[…] = …, not .psi_k = …

property precision: Return the floating point precision

copy() → Self: Build a new object with the same grid data. Must be overriden if __init__ signature is overriden

metadata() → dict: Everything except the data itself

classmethod concat(*metas: dict, axis: int) → dict: Concatenate metadata dicts.

zero_(): Set all elements of psi to zero psi_k will not be affected

class torusgrid.Grid1D(n: int, /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double')

class torusgrid.Grid1D(shape: Tuple[int], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double')

Bases: Grid[T]

__init__(n: int, /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double')
__init__(shape: Tuple[int], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double')

property n: =shape[0], the number of samples

property N: Deprecated; use self.n instead.

copy() → Self: Build a new object with the same grid data. Must be overriden if __init__ signature is overriden

class torusgrid.Grid2D(nx: int, ny: int, /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

class torusgrid.Grid2D(shape: Tuple[int, int], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

Bases: Grid[T]

__init__(nx: int, ny: int, /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)
__init__(shape: Tuple[int, int], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

property nx: = shape[0], the number of samples in the x direction

property ny: = shape[1], the number of samples in the y direction

copy() → Self: Build a new object with the same grid data. Must be overriden if __init__ signature is overriden

class torusgrid.ComplexGrid(shape: Tuple[int, ...], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

Bases: Grid[complexfloating]

A ComplexGridND object is a complex array of shape (d1, d2, .., dN) equipped with fourier transform. No length scales are associated with the grid.

property isreal: Return whether the field or grid object holds real data

class torusgrid.RealGrid(shape: Tuple[int, ...], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

Bases: Grid[floating]

A RealGridND object is a real array of shape (d1, d2, .., dN) equipped with fourier transform. No length scales are associated with the grid.

The generic type variables are the dtypes of self.psi & self.psi_k.

property isreal: Return whether the field or grid object holds real data

class torusgrid.ComplexGrid1D(n: int, /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double')

class torusgrid.ComplexGrid1D(shape: Tuple[int], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double')

Bases: ComplexGrid, Grid1D[complexfloating]

1D complex grid; fft axis is always axis 0.

class torusgrid.ComplexGrid2D(nx: int, ny: int, /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

class torusgrid.ComplexGrid2D(shape: Tuple[int, int], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

Bases: ComplexGrid, Grid2D[complexfloating]

2D complex grid

class torusgrid.RealGrid1D(n: int, /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double')

class torusgrid.RealGrid1D(shape: Tuple[int], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double')

Bases: RealGrid, Grid1D[floating]

1D real grid; fft axis is always axis 0.

class torusgrid.RealGrid2D(nx: int, ny: int, /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

class torusgrid.RealGrid2D(shape: Tuple[int, int], /, *, precision: Union[FloatingPointPrecision, Literal['SINGLE', 'DOUBLE', 'LONGDOUBLE', 'single', 'double', 'longdouble']] = 'double', fft_axes: Optional[Tuple[int, ...]] = None)

Bases: RealGrid, Grid2D[floating]

2D real grid