TensorLibrary

TensorLibrary

TensorLibrary()

A library for registering and managing tensor templates and structures.

The TensorLibrary provides a centralized registry for tensor definitions that can be reused across tensor networks and expressions. It manages tensor structures with their associated names and can resolve symbolic references to registered tensors.

import symbolica
from symbolica.community.spenso import TensorLibrary, LibraryTensor, TensorStructure, Representation

lib = TensorLibrary()
rep = Representation.euc(3)
name = symbolica.S("my_tensor")
structure = TensorStructure(rep, rep, name=name)
tensor = LibraryTensor.dense(structure, [1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0])
lib.register(tensor)
tensor_ref = lib[name]

Methods

Name Description
__getitem__ Retrieve a registered tensor structure by name
__new__ Create a new empty tensor library
construct Create a new empty tensor library (static method)
hep_lib Create a library pre-loaded with High Energy Physics tensor definitions
hep_lib_atom Create a library pre-loaded with High Energy Physics tensor definitions
register Register a tensor in the library

__getitem__

TensorLibrary.__getitem__(key: Expression | int | str | float | builtins.complex | builtins.str) -> TensorStructure

Related types: Expression, TensorStructure

Retrieve a registered tensor structure by name.

Looks up a previously registered tensor by its name and returns a reference structure that can be used to create new tensor instances.

Parameters

  • key (str or Expression) The tensor name - can be a string or symbolic expression

Returns

  • TensorStructure A TensorStructure representing the registered tensor template

Raises

  • RuntimeError: If the tensor name is not found in the library

Examples

structure = lib["T"]

__new__

TensorLibrary.__new__() -> TensorLibrary

Create a new empty tensor library.

Initializes an empty library ready for registering tensor structures. The library automatically manages internal tensor IDs.

Returns

  • TensorLibrary A new empty tensor library

Examples

from symbolica.community.spenso import TensorLibrary
lib = TensorLibrary()

construct

TensorLibrary.construct() -> TensorLibrary

Create a new empty tensor library (static method).

Initializes an empty library ready for registering tensor structures. The library automatically manages internal tensor IDs.

Returns

  • TensorLibrary A new empty tensor library

Examples

from symbolica.community.spenso import TensorLibrary
lib = TensorLibrary.construct()

hep_lib

TensorLibrary.hep_lib() -> TensorLibrary

Create a library pre-loaded with High Energy Physics tensor definitions.

Returns a library containing standard HEP tensors such as gamma matrices, color generators, metric tensors, and other commonly used structures in particle physics calculations. They are floating point tensors with f64 precision.

Returns

  • TensorLibrary A TensorLibrary pre-populated with HEP tensor definitions

Examples

import symbolica
from symbolica.community.spenso import TensorLibrary, TensorName
hep_lib = TensorLibrary.hep_lib()
gamma_structure = hep_lib[symbolica.S("spenso::gamma")]

hep_lib_atom

TensorLibrary.hep_lib_atom() -> TensorLibrary

Create a library pre-loaded with High Energy Physics tensor definitions.

Returns a library containing standard HEP tensors such as gamma matrices, color generators, metric tensors, and other commonly used structures in particle physics calculations. They are tensors with atom numeric entries.

Returns

  • TensorLibrary A TensorLibrary pre-populated with HEP tensor definitions

Examples

import symbolica
from symbolica.community.spenso import TensorLibrary, TensorName
hep_lib = TensorLibrary.hep_lib()
gamma_structure = hep_lib[symbolica.S("spenso::gamma")]

register

TensorLibrary.register(tensor: Tensor | LibraryTensor) -> None

Related types: Tensor, LibraryTensor

Register a tensor in the library.

Adds a tensor template to the library that can be referenced by name in tensor networks and symbolic expressions. The tensor must have a name set in its structure.

Parameters

  • tensor (LibraryTensor or Tensor) The tensor to register - can be a LibraryTensor or regular Tensor

Examples

import symbolica
from symbolica.community.spenso import TensorLibrary, LibraryTensor, TensorStructure, Representation
lib = TensorLibrary()
rep = Representation.euc(3)
name = symbolica.S("my_tensor")
structure = TensorStructure(rep, rep, name=name)
tensor = LibraryTensor.dense(structure, [1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0])
lib.register(tensor)
tensor_ref = lib[name]