rocket-fft 0.1.4

Rocket-FFT extends Numba by scipy.fft and numpy.fft

Rocket-FFT

Rocket-FFT makes Numba aware of numpy.fft and scipy.fft. It takes its name from the PocketFFT Fast Fourier Transformation library that powers it, and Numba's goal of making your scientific Python code blazingly fast - like a rocket.

Getting Started

The easiest way to get Rocket-FFT is to:

$ pip install rocket-fft

Alternatively, you can build it from source:

$ git clone https://github.com/styfenschaer/rocket-fft.git
$ cd rocket-fft
$ python setup.py install

The latter requires a C++ compiler compatible with your Python installation.

Once installed successfully, the following will work (no import required):

import numba as nb
import numpy as np

@nb.njit
def jit_fft(x):
    return np.fft.fft(x)

a = np.array([2, 7, 1, 8, 2, 8, 1, 8])
jit_fft(a)

Supported NumPy Functions

The whole numpy.fft module is supported, which contains all the functions listed below:

• numpy.fft.fft
• numpy.fft.ifft
• numpy.fft.fft2*
• numpy.fft.ifft2*
• numpy.fft.fftn*
• numpy.fft.ifftn*
• numpy.fft.rfft
• numpy.fft.irfft
• numpy.fft.rfft2
• numpy.fft.irfft2
• numpy.fft.rfftn
• numpy.fft.irfftn
• numpy.fft.hfft
• numpy.fft.ihfft
• numpy.fft.fftfreq
• numpy.fft.rfftfreq
• numpy.fft.fftshift
• numpy.fft.ifftshift

*Rocket-FFT follows SciPy's approach of not allowing duplicate axes

Supported SciPy Functions

If you have SciPy installed, you will have support for most of the scipy.fft module, including:

• scipy.fft.fft
• scipy.fft.ifft
• scipy.fft.fft2
• scipy.fft.ifft2
• scipy.fft.fftn
• scipy.fft.ifftn
• scipy.fft.rfft
• scipy.fft.irfft
• scipy.fft.rfft2
• scipy.fft.irfft2
• scipy.fft.rfftn
• scipy.fft.irfftn
• scipy.fft.hfft
• scipy.fft.ihfft
• scipy.fft.hfft2
• scipy.fft.ihfft2
• scipy.fft.hfftn
• scipy.fft.ihfftn
• scipy.fft.dct
• scipy.fft.dct2
• scipy.fft.dctn
• scipy.fft.idctn
• scipy.fft.dst
• scipy.fft.idst
• scipy.fft.dstn
• scipy.fft.idstn
• scipy.fft.fht
• scipy.fft.ifht
• scipy.fft.fftshift
• scipy.fft.ifftshift
• scipy.fft.fftfreq
• scipy.fft.ifftfreq
• scipy.fft.fhtoffset
• scipy.fft.next_fast_len

Type Conversion

If SciPy is installed, Rocket-FFT follows SciPy's approach to type conversion, otherwise it follows NumPy's approach. You can change the type conversion rule by calling the scipy_like or numpy_like function from the rocket_fft namespace:

from rocket_fft import numpy_like, scipy_like

numpy_like()

Both functions can be used regardless of whether SciPy is installed. Note that this change is irreversible after the compilation of Rocket-FFT's internal functions.

Performance Tip

Rocket-FFT achieves both, flexibility in function signatures similar to scipy.fft and numpy.fft and low compilation times. Compilation takes a few hundred milliseconds in most cases. Calls with default arguments are treated specially and compile fastest.

Low-Level Interface

Rocket-FFT also provides a low-level interface to the PocketFFT library. Using the low-level interface can significantly reduce compile time, minimize overhead and give more flexibility to the user. It also provides some functions that are not available through the SciPy and NumPy interfaces. You can import the functions of the low-level interface from the rocket_fft namespace:

from rocket_fft import c2c, dct, ...

The low-level interface includes the following functions:

def c2c(ain: NDArray[c8] | NDArray[c16], aout: NDArray[c8] | NDArray[c16], axes: NDArray[i8], forward: b1, fct: f4 | f8, nthreads: i8) -> None: ...
def r2c(ain: NDArray[f4] | NDArray[f8], aout: NDArray[c8] | NDArray[c16], axes: NDArray[i8], forward: b1, fct: f4 | f8, nthreads: i8) -> None: ...
def c2r(ain: NDArray[c8] | NDArray[c16], aout: NDArray[f4] | NDArray[f8], axes: NDArray[i8], forward: b1, fct: f4 | f8, nthreads: i8) -> None: ...
def c2c_sym(ain: NDArray[f4] | NDArray[f8], aout: NDArray[c8] | NDArray[c16], axes: NDArray[i8], forward: b1, fct: f4 | f8, nthreads: i8) -> None: ...
def dst(ain: NDArray[f4] | NDArray[f8], aout: NDArray[f4] | NDArray[f8], axes: NDArray[i8], type: i8, fct: f4 | f8, ortho: b1, nthreads: i8) -> None: ...
def dct(ain: NDArray[f4] | NDArray[f8], aout: NDArray[f4] | NDArray[f8], axes: NDArray[i8], type: i8, fct: f4 | f8, ortho: b1, nthreads: i8) -> None: ...
def r2r_separable_hartley(ain: NDArray[f4] | NDArray[f8], aout: NDArray[f4] | NDArray[f8], axes: NDArray[i8], fct: f4 | f8, nthreads: i8) -> None: ...
def r2r_genuine_hartley(ain: NDArray[f4] | NDArray[f8], aout: NDArray[f4] | NDArray[f8], axes: NDArray[i8], fct: f4 | f8, nthreads: i8) -> None: ...
def r2r_fftpack(ain: NDArray[f4] | NDArray[f8], aout: NDArray[f4] | NDArray[f8], axes: NDArray[i8], real2hermitian: b1, forward: b1, fct: f4 | f8, nthreads: i8) -> None: ...
def good_size(target: i8, real: b1) -> i8: ...

Note that the low-level interface does not provide the same level of safety and convenience as the SciPy and NumPy interfaces. There is no safety net, and it is up to the user to ensure proper usage. You may want to look at the original PocketFFT C++ implementation before using it.