Examples and Scripts

DyMAD has two complementary learning surfaces:

  • examples/ is the starting point. The published Examples page is a notebook gallery for learning the main ideas step by step.

  • scripts/ is the broader runnable library. It contains more workflows, variants, utilities, and topic coverage than the notebook gallery.

Use them as a progression rather than expecting a strict mirror. The notebooks are introductory tutorials, while the scripts are the place to look when you want more depth, more model variants, or a runnable reference for a specific feature. Not every script has a matching notebook, and this documentation does not assume a one-to-one mapping.

How to move from notebooks to scripts

Start with the notebook gallery when you want a guided walkthrough of the core DyMAD workflow. Then move into scripts/ when you want to:

  • rerun a workflow from Python or the command line

  • inspect the YAML configuration used for a model or dataset

  • explore model families or analysis workflows that are not covered in the notebooks

  • adapt an existing example to your own system

In many script folders, *_cli.py files are the most direct command-line entry points, nearby .py files show the underlying runnable example, and neighboring .yaml files provide the configs those runs use.

Where to look in scripts/

The current script tree is broad, so it is easiest to navigate by topic:

  • Introductory training workflows similar to the notebook material: scripts/linear_time_invariant/, scripts/linear_graph/, scripts/2d_koopman/

  • Kernel-based examples: scripts/ker_lti/, scripts/ker_s1/, scripts/ker_s1u/, scripts/ker_lco/

  • Discrete-time, delayed, or variant system setups: scripts/lti_dt/, scripts/ltg_dt/, scripts/ltg_dt_tv/, scripts/lti_delay/, scripts/lti_1s/, scripts/lti_vlen/

  • Reduced-order and PIROM workflows: scripts/pirom_dyn/, scripts/pirom_res/, scripts/pirom_res_dt/

  • Spectral-analysis-oriented examples: scripts/sa_lti/, scripts/sa_lco/, scripts/sa_2dk/

  • Data preparation, denoising, and post-processing utilities: scripts/denoise/, scripts/kuramoto/, scripts/vortex/

If you began with a specific notebook topic, these are good next stops:

  • After the linear and Koopman notebooks, compare the broader training and sweep examples under scripts/linear_time_invariant/, scripts/linear_graph/, and scripts/2d_koopman/.

  • After the spectral analysis notebooks, explore scripts/sa_lti/, scripts/sa_lco/, and scripts/sa_2dk/ for additional analysis-oriented runs.

  • After the vortex notebooks, look at scripts/vortex/vor_train_cli.py, scripts/vortex/vor_proc_cli.py, and scripts/vortex/vor_post.py for the surrounding training and processing workflow.

What this guide does not require

You do not need a matching notebook for every runnable script. The intended structure is a curated set of notebook tutorials in examples/ plus a larger reference/demo surface in scripts/.