Crate tenferro_einsum 

High-level einsum with N-ary contraction tree optimization.

This crate provides:

• String notation: "ij,jk->ik" (NumPy/PyTorch compatible)
• Parenthesized notation: "ij,(jk,kl)->il" respects user-specified contraction order via NestedEinsum
• Integer label notation: using u32 labels
• N-ary contraction: Automatic or manual optimization of pairwise contraction order via ContractionTree
• v2 builder: build_einsum_fragment lowers einsum into a compute graph fragment using DotGeneral, ReduceSum, Transpose, etc.

Examples

use computegraph::fragment::FragmentBuilder;
use computegraph::types::ValRef;
use tenferro_ops::input_key::TensorInputKey;
use tenferro_ops::std_tensor_op::StdTensorOp;
use tenferro_einsum::{ContractionTree, Subscripts, build_einsum_fragment};

let subs = Subscripts::parse("ij,jk->ik").unwrap();
let tree = ContractionTree::optimize(&subs, &[&[2, 3], &[3, 4]]).unwrap();

let mut builder = FragmentBuilder::<StdTensorOp>::new();
let a = builder.add_input(TensorInputKey::User { id: 0 });
let b = builder.add_input(TensorInputKey::User { id: 1 });

let result = build_einsum_fragment(
    &mut builder,
    &tree,
    &[ValRef::Local(a), ValRef::Local(b)],
    &[vec![2, 3], vec![3, 4]],
);

Re-exports

pub use builder::build_einsum_fragment;

Modules

builder

planning

syntax

Structs

ContractionOptimizerOptions

Public options for automatic contraction-path optimization.

ContractionTree

Contraction tree determining pairwise contraction order for N-ary einsum.

Subscripts

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

Enums

NestedEinsum

Recursive einsum tree that preserves parenthesized grouping.

Functions

build_size_dict

Build a label -> size mapping from subscripts and input shapes.

compute_output_shape

Compute output shape from output subscripts and size dictionary.

eager_einsum

Eager N-ary einsum on concrete Tensor values.

typed_eager_einsum

Execute eager einsum over typed tensors and return a typed result.