Module sparse 

Sparse tensor support for TensorLogic.

This module provides comprehensive sparse tensor representations and operations:

• CSR (Compressed Sparse Row) format for efficient row operations
• CSC (Compressed Sparse Column) format for efficient column operations
• COO (Coordinate) format for flexible construction
• Sparse-dense hybrid operations
• Automatic sparsity detection and conversion
• Sparse matrix multiplication and linear algebra

Example

use tensorlogic_infer::{SparseFormat, SparseTensor, SparseMatrix};

// Create a sparse matrix in COO format
let mut builder = SparseTensor::builder(vec![100, 100], SparseFormat::COO);
builder.add_entry(vec![5, 10], 3.14);
builder.add_entry(vec![20, 30], 2.71);
let sparse = builder.build()?;

// Convert to CSR for efficient operations
let csr = sparse.to_csr()?;

// Sparse-dense multiplication
let dense = vec![1.0; 100];
let result = csr.multiply_dense(&dense)?;

// Detect sparsity
let sparsity = sparse.sparsity_ratio();
println!("Sparsity: {:.2}%", sparsity * 100.0);

Structs

SparseCOO

Sparse matrix in COO (Coordinate) format.

SparseCSC

Sparse matrix in CSC (Compressed Sparse Column) format.

SparseCSR

Sparse matrix in CSR (Compressed Sparse Row) format.

SparseTensorBuilder

Builder for sparse tensors.

Enums

SparseError

Sparse tensor errors.

SparseFormat

Sparse tensor storage format.

SparseTensor

Sparse tensor representation.

Functions

detect_sparsity

Detect sparsity in a dense tensor.

to_sparse_if_beneficial

Convert dense tensor to sparse if beneficial.