Struct BroadcastShape 

pub struct BroadcastShape {
    pub shape: Vec<usize>,
}

NumPy-style broadcast compatibility checker and stride generator.

Examples

use ipfrs_tensorlogic::slice_view::BroadcastShape;

let bs = BroadcastShape { shape: vec![1, 3] };
assert!(bs.compatible_with(&[4, 3]));

let strides = bs.broadcast_to(&[4, 3]).expect("example: should succeed in docs");
assert_eq!(strides, vec![0, 1]);

Fields

shape: Vec<usize>

The source shape to be broadcast.

Implementations

impl BroadcastShape

pub fn compatible_with(&self, other: &[usize]) -> bool

Check NumPy-style broadcast compatibility with other.

Dimensions are aligned from the right. Each pair (a, b) must satisfy: a == b, a == 1, or b == 1.

pub fn broadcast_to(&self, target: &[usize]) -> Option<Vec<usize>>

Compute the broadcast strides needed to view self as target.

A source dimension of size 1 gets stride 0 (it is repeated along that axis without touching memory). Equal dimensions get the natural row-major stride computed from the source shape.

Returns None if the shapes are not compatible.

Trait Implementations

impl Clone for BroadcastShape

fn clone(&self) -> BroadcastShape

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for BroadcastShape

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Eq for BroadcastShape

impl PartialEq for BroadcastShape

fn eq(&self, other: &BroadcastShape) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl StructuralPartialEq for BroadcastShape