DelayedTensor

DelayedTensor performs sparse and out-of-core tensor arithmetic and decomposition directly on DelayedArray objects.

Key Features:

• Tensor Arithmetic on DelayedArray: Provides generic functions for tensor arithmetic and decomposition dispatched on the DelayedArray class.
• Tensor Contraction: Supports tensor contraction via an einsum function inspired by numpy einsum.

Scientific Applications:

• Large-Scale Tensor Analysis: Enables out-of-core computation for tensor operations that exceed available memory.
• Sparse Tensor Operations: Supports efficient arithmetic and decomposition of sparse tensor data structures.

Methodology:

DelayedTensor dispatches generic tensor arithmetic and decomposition functions on the DelayedArray class and implements tensor contraction using an einsum function inspired by numpy einsum.