Types#

Type definitions and aliases for RAN algorithms.

Overview#

The Types module provides:

  • Array Types - NumPy and JAX array type aliases

  • Scalar Types - Standard dtype definitions for consistency

API Reference#

Centralized array type aliases for PHY (NumPy) and PHY (JAX).

Public API exports only the canonical choices per backend to enforce a single dtype across the codebase and avoid accidental casting:

  • NumPy: IntNP, FloatNP, ComplexNP (always available)

  • JAX: IntJAX, FloatJAX, ComplexJAX (only when ENABLE_MLIR_TRT=ON)

ran.types.arrays.IntNP#

NumPy integer scalar type (int64).

ran.types.arrays.FloatNP#

NumPy floating-point scalar type (float64).

ran.types.arrays.ComplexNP#

NumPy complex scalar type (complex128).

ran.types.arrays.IntJAX#

JAX integer scalar type (int32). Only available when ENABLE_MLIR_TRT=ON.

ran.types.arrays.FloatJAX#

JAX floating-point scalar type (float64). Only available when ENABLE_MLIR_TRT=ON.

ran.types.arrays.ComplexJAX#

JAX complex scalar type (complex128). Only available when ENABLE_MLIR_TRT=ON.