Rust and Python APIs

Rust package

The idenso crate exposes representation types and syntax macros together with several rewrite families:

• IndexTooling covers canonicalization, conjugation, index wrapping, and dangling-index inspection for Symbolica atoms;
• Cookable, CookSettings, and the cook filters control reversible or flattening encodings;
• SelectiveExpand expands terms by recognized representation families;
• dirac, color, epsilon, and shorthand modules implement algebra-specific rewrites;
• representations::initialize installs the standard representation and tensor symbols.

Representation helper macros such as bis!, cof!, and coad! construct the symbolic forms expected by Spenso and Idenso. The Rust orientation leads to the revision-specific Rustdoc for their accepted forms, return types, feature gates, and source locations. APIs behind bincode, reference-cases, python, and python_stubgen are available only when the matching Cargo feature is enabled.

Python community module

Install the published assembly with python -m pip install --upgrade symbolica, then verify it with python -c "import symbolica.community.idenso". There is no pip install idenso fallback. For a source build, add this crate to the external symbolica-community assembly with the python feature and call IdensoModule’s SymbolicaCommunityModule registration when that extension is assembled. Building the Rust crate alone does not add the community module to an already installed Symbolica package.

The generated Python API records exact signatures and defaults. Its operations group naturally into four phases:

• setup: importing the community module registers its symbols;
• expansion: expand_bis, expand_mink, expand_mink_bis, expand_metrics, and expand_color;
• index preparation: wrap_indices, wrap_dummies, list_dangling, cook_indices, and cook_function;
• algebra: dirac_adjoint, simplify_gamma, to_dots, simplify_metrics, and simplify_color.

from symbolica.community.idenso import list_dangling, simplify_metrics
from symbolica.community.spenso import Representation, TensorName

minkowski = Representation.mink(4)
mu = minkowski("mu")
nu = minkowski("nu")
metric = TensorName.g()
momentum = TensorName("p")
expression = metric(mu, nu) * momentum(mu)

external_indices = list_dangling(expression)
reduced = simplify_metrics(expression)
assert len(external_indices) == 1
assert len(list_dangling(reduced)) == 1

Idenso does not define a second parser syntax: the example constructs a Spenso-compatible Symbolica expression and then applies one Idenso transformation. Keep transformations separate while developing a pipeline so expression growth and convention changes remain observable.

For implementation details, start with the Rust API and the Python binding.