Struct Subscripts 

pub struct Subscripts {
    pub inputs: Vec<Vec<u32>>,
    pub output: Vec<u32>,
}

Einsum subscripts using integer labels (omeinsum-rs compatible).

Each dimension is represented by a u32 label. Labels shared across multiple input tensors are contracted (summed over). Labels present only in the output are free indices.

Examples

use tenferro_einsum::Subscripts;

// Matrix multiplication: C_{ik} = Σ_j A_{ij} * B_{jk}
let subs = Subscripts::new(&[&[0, 1], &[1, 2]], &[0, 2]);
assert_eq!(subs.inputs.len(), 2);
assert_eq!(subs.output, vec![0, 2]);

use tenferro_einsum::Subscripts;

// Parse from string notation
let subs = Subscripts::parse("ij,jk->ik").unwrap();
assert_eq!(subs.inputs.len(), 2);

Fields

inputs: Vec<Vec<u32>>

Index labels for each input tensor.

output: Vec<u32>

Index labels for the output tensor.

Implementations

impl Subscripts

pub fn new(inputs: &[&[u32]], output: &[u32]) -> Self

Create subscripts from integer label arrays.

Arguments

• inputs — Index labels for each input tensor
• output — Index labels for the output tensor

pub fn parse(notation: &str) -> Result<Self>

Parse subscripts from NumPy/PyTorch-style string notation.

Each character (a–z, A–Z) represents a dimension label. Input tensors are separated by commas, and -> separates inputs from the output.

Parentheses can be used to specify contraction order explicitly. Grouped operands are contracted first, enabling manual control over the pairwise contraction sequence without using ContractionTree::from_pairs.

Examples

• "ij,jk->ik" — matrix multiplication
• "ii->i" — diagonal extraction
• "ijk->" — full contraction (scalar result)
• "ij,(jk,kl)->il" — contract B and C first, then with A

Errors

Returns an error if the notation is malformed.

Trait Implementations

impl Clone for Subscripts

fn clone(&self) -> Subscripts

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Subscripts

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

Formats the value using the given formatter.