Tensor plugin#
Introduction#
The spectrochempy-tensor plugin provides tensor decomposition classes backed
by TensorLy. Keeping these algorithms in a plugin lets the core package remain
independent from tensor-specific dependencies while preserving integrated access
when the plugin is installed.
Installation#
python -m pip install "spectrochempy[tensor]"
or directly:
pip install spectrochempy-tensor
Recommended API#
import spectrochempy as scp
model = scp.tensor.CP(n_components=2)
model.fit(dataset)
factors = model.result.factors
weights = model.result.weights
Compatibility aliases#
The historical scp.CP alias is kept as a deprecated compatibility path, but
new documentation and examples should use scp.tensor.CP.
Direct accessors such as model.A, model.B, model.C, and
model.weights remain available.
API Reference#
The generated public API page for the Tensor plugin is listed in
Plugin public API reference. See spectrochempy.tensor.CP.
Examples#
Use the package-level scp.tensor.CP(...) estimator shown above for
TensorLy-backed CP/PARAFAC workflows on existing datasets.
Limitations#
Tensor decomposition implementations live under
spectrochempy_tensor.decompositions. This page currently focuses on the
main CP/PARAFAC workflow exposed through scp.tensor.CP(...).