Block Utilities

Low-level utilities for managing tensor block structure.

Description

The blocks module provides internal utilities for reasoning about tensor block dictionaries. Most users won't need these directly.

Key Concepts

BlockKey

A BlockKey is a tuple of charges, one per tensor index:

from nicole.blocks import BlockKey

# For a 2-index U(1) tensor
key1 = (0, 0)      # Both indices have charge 0
key2 = (1, 1)      # Both indices have charge 1

BlockSchema

Static utility class for block operations:

• iter_admissible_keys(): Generate all possible charge combinations
• shape_for_key(): Get block shape from key
• validate_blocks(): Validate block data
• charges_conserved(): Check conservation for a key
• block_add(): Merging or concatenating weights from a pair of blocks
• block_compress(): Remove linearly dependent components from a block

Usage

from nicole.blocks import BlockSchema

# Check if block conserves charge
conserved = BlockSchema.charges_conserved(indices, block_key)

See Also

• Tensor: Main tensor class
• Index: Index structure

Notes

These are internal utilities. Most functionality is accessed through the Tensor class interface.