AbstractFastHartleyTransforms.jl

A common Julia interface for real Hartley transforms. This package defines transform functions, plan types, inverse normalization, FFT-backed wrappers, and optional EnzymeCore rules. Concrete backends supply planning methods and execution.

For everyday transforms, install FastHartleyTransform.jl, which implements the interface using FFTW and optional CUDA/Reactant paths. Loading the abstract package alone does not provide a general CPU planner.

Read next

• Interface and conventions: mathematical definition, axes, normalization, and mutation.
• Implementing a backend: a complete small reference backend and the extension protocol.
• Differentiation contract: optional CPU Enzyme rules and their limitations.
• API reference: shared transform and planning functions.

Julia 1.10 or later is required. The Enzyme behavior described here requires AbstractFastHartleyTransforms 0.1.3 or later.