Skip to content

Quickstart

This section provides a minimal example to run PME-toolkit and verify that the installation works correctly.

PME-toolkit can be used in two ways:

  • Option A — try a ready-to-run reference case (recommended)
  • Option B — use PME-toolkit on your own data

Option A — Try PME-toolkit in 2 minutes

Clone the repository to access ready-to-run benchmark cases:

git clone https://github.com/cnr-inm-mao/pme-toolkit.git
cd pme-toolkit
pip install -e .

Run a reference case:

pme-run tests/cases/test_glider.json

Run backmapping:

pme-back tests/cases/test_glider_back.json

Option B — Use PME-toolkit on your own data

Install from PyPI:

pip install pme-toolkit

Run PME with your configuration file:

pme-run your_config.json

Run backmapping:

pme-back your_backmapping_config.json

See the Input Format section for details on how to define datasets and variables.

MATLAB

Add the source folder to the path:

addpath(genpath("matlab/src"));

Run a reference case:

run_pme("tests/cases/test_glider.json")

Run backmapping:

run_back("tests/cases/test_glider_back.json")

What happens

Running the example produces:

  • reduced coordinates
  • variance and mode information
  • reconstruction error metrics
  • output files in the results/ directory

Notes

  • the dataset under tests/data/ is self-contained and requires no external downloads
  • JSON configuration files define the full workflow
  • ready-to-run examples are available in the repository under tests/cases/
  • for full benchmark workflows, see the Benchmarks and Datasets sections