README.md

strided-view

Core dynamic-rank strided tensor view types and metadata-only transformations.

Scope

• StridedView / StridedViewMut
• StridedArray
• Stride/layout helpers (row_major_strides, col_major_strides)
• Zero-copy metadata ops (permute, transpose_2d, broadcast, diagonal_view)

strided-view does not implement heavy compute kernels by itself. Use strided-kernel for map/reduce operations.

Core Types

StridedView<'a, T, Op> / StridedViewMut<'a, T>

Dynamic-rank immutable/mutable views over strided data:

• T: Element type
• Op: Element operation (Identity, Conj, Transpose, Adjoint) -- applied lazily on access

StridedArray<T>

Owned strided multidimensional array with view() and view_mut() methods.

Quick Example

use strided_view::StridedArray;

let a = StridedArray::<f64>::from_fn_row_major(&[2, 3], |idx| (idx[0] * 10 + idx[1]) as f64);
let v = a.view();
let vt = v.permute(&[1, 0]).unwrap();
assert_eq!(vt.dims(), &[3, 2]);
assert_eq!(vt.get(&[2, 1]), 12.0);

View Operations

use strided_view::{StridedArray, StridedView};

let a = StridedArray::<f64>::from_fn_row_major(&[3, 4], |idx| {
    (idx[0] * 10 + idx[1]) as f64
});
let v = a.view();

// Transpose (zero-copy, swaps strides)
let vt = v.transpose_2d().unwrap();
assert_eq!(vt.dims(), &[4, 3]);

// General permutation (zero-copy)
let vp = v.permute(&[1, 0]).unwrap();
assert_eq!(vp.get(&[2, 1]), v.get(&[1, 2]));

// Broadcast (stride-0 for size-1 dims)
let row_data = vec![1.0, 2.0, 3.0];
let row = StridedView::<f64>::new(&row_data, &[1, 3], &[3, 1], 0).unwrap();
let broad = row.broadcast(&[4, 3]).unwrap();