Skip to main content

Crate scivex_core

Crate scivex_core 

Source
Expand description

scivex-core — Foundation crate for the Scivex ecosystem.

Provides tensors, numeric type traits, linear algebra, FFT, and math primitives. All other scivex-* crates build on top of this one.

§Design

  • Zero external dependencies for math — everything is from scratch.
  • Generic over numeric types via the Scalar / Float / Real trait hierarchy.
  • unsafe is confined to this crate; all higher-level crates use only safe abstractions.

Re-exports§

pub use complex::Complex;
pub use dtype::Float;
pub use dtype::Integer;
pub use dtype::Real;
pub use dtype::Scalar;
pub use error::CoreError;
pub use error::Result;
pub use promote::CastFrom;
pub use promote::DType;
pub use promote::DTypeOf;
pub use promote::promote;
pub use tensor::Tensor;
pub use tensor::named::NamedTensor;
pub use tensor::sparse::SparseTensor;

Modules§

arena
Arena and slab allocators for temporary tensor buffers. Arena and slab allocators for temporary tensor buffers.
complex
Native complex number type. Native complex number type for generic tensor operations.
dtype
Numeric type traits: Scalar, Float, Real, Integer. Numeric type hierarchy for generic math.
error
Core error types.
fft
Fast Fourier Transform (FFT / IFFT / RFFT). Fast Fourier Transform (FFT) module.
jit
Expression JIT for fusing element-wise tensor operations. Expression JIT for element-wise tensor operations.
linalg
Linear algebra: decompositions, solvers, and matrix operations. Linear algebra operations.
math
Elementary mathematical functions. Element-wise tensor math functions (ufuncs).
prelude
Items intended for glob-import: use scivex_core::prelude::*;
promote
Type promotion rules, casting utilities, and runtime dtype tags. Type promotion rules and casting utilities for tensor element types.
random
Pseudo-random number generation. Pseudo-random number generation and random tensor creation.
spatial
Spatial data structures: KD-tree, ball tree. Spatial data structures: KD-tree, ball tree.
tensor
N-dimensional tensor type and operations. N-dimensional tensor type with dynamic shape and contiguous storage.