Using Spenso from Python

The shortest path into Spenso is the community module bundled with Symbolica 2.2.0. This example creates a typed two-dimensional tensor and checks its diagonal entries; it needs no GammaLoop checkout.

Install and verify the module

Create an isolated environment, install the version used by this manual, and verify the community import:

python -m venv .venv
. .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install "symbolica==2.2.0"
python -c "import symbolica.community.spenso as spenso; print(spenso.__name__)"

There is no separate spenso Python wheel. The installed symbolica distribution owns the native symbolica.community.spenso module. Symbolica’s installation and license terms apply; its free restricted mode is enough for this single-process example.

Construct a dense tensor

Save this as spenso_quickstart.py:

from symbolica.community.spenso import Representation, Tensor, TensorIndices

rep = Representation.euc(2)
matrix = Tensor.dense(
    TensorIndices(rep("i"), rep("j")),
    [1.0, 0.0, 0.0, 1.0],
)

assert matrix[0, 0] == 1.0
assert matrix[1, 1] == 1.0
matrix.to_sparse()
print(matrix)

Run python spenso_quickstart.py. TensorIndices gives the data a representation-aware structure, while the four row-major values provide its concrete storage. Converting to sparse storage changes only that storage; it does not change the slots or re-index the tensor.

Continue with the Python tensor-workflow guide, or use the Rust guide to perform a checked native contraction.