Python API
The Python package provides a working implementation of PME-toolkit, aligned with the MATLAB reference workflow.
Main entry points
CLI
pme-run <case.json>
pme-back <case_back.json>
Python usage
from pme_toolkit import run_case, run_back
run_case("tests/cases/test_glider.json")
run_back("tests/cases/test_glider_back.json")
Scope
The Python implementation supports:
- PME, PI-PME, PD-PME model construction
- JSON-driven workflows
- dimensionality reduction and variance analysis
- backmapping to original variables
- model persistence (
model.npz)
Status
- fully functional and installable via PyPI
- validated against MATLAB reference through regression tests
- recommended for standard usage
MATLAB remains the reference for numerical validation.