mergen-doe 0.1.0

Space-filling Design of Experiments for Python

Multi-dimensional Experimental Run GENerator: space-filling Design of Experiments for Python.

---

What is Mergen?

Every experimental study works under a fixed budget: only a limited number of runs can be afforded, and each one should be as informative as possible. Mergen selects those runs. Given a parameter space (its bounds, resolution, constraints and critical regions), it computes the coordinates of n points that spread evenly through the space, so that no region is left unexplored and no run is wasted.

The result is not a random sample but a mathematically optimised design, delivered together with a statistical quality report suitable for the methods section of a paper. Mergen handles the spaces real studies actually have: mixed factor types, forbidden regions, runs that already exist, and zones that deserve extra attention.

Features

• Factor types: discrete, continuous (linear or log), integer, nominal and ordinal, freely mixed in one space, with feasibility constraints.
• Design control: prescribed points, focus regions, exclusion zones, named extra sets, and resuming from an existing design.
• Optimisation: seven space-filling criteria (including recent developments such as uMaxPro and the stratified L2-discrepancy) and three optimisers over a discrete-grid Latin-hypercube structure, with a compare() sweep for when you are unsure.
• Evidence and export: a Monte Carlo quality report, eight plot types, and CSV, JSON, Markdown, LaTeX, HTML and Excel export.
• Sequential workflows: extend, subsample, reorder and nest designs as a campaign grows.

Installation

git clone https://github.com/acanbay/mergen.git
cd mergen
pip install .

The distribution name is mergen-doe; the import name is mergen. Requires Python 3.9 or newer; the core dependencies (NumPy, SciPy, pandas, matplotlib) are installed automatically.

Quick start

from mergen import ParameterSpace, Sampler

space = ParameterSpace({
    'temperature': range(100, 500, 10),        # discrete
    'pressure':    ('continuous', 0.5, 5.0),   # continuous grid
    'n_layers':    ('integer', 2, 10),         # integer grid
})

sampler = Sampler(space)
sampler.set_design(n_samples=30)
result = sampler.run()

result.summary()
result.quality_report()
result.plot('pairplot', save=True)
result.to_csv()

Documentation

The full documentation lives at mergen.readthedocs.io. It covers tutorials, task guides, the reasoning behind every choice Mergen asks you to make, an executed gallery of fifteen example studies with their complete output and figures, and the full API reference.

To build the documentation locally (for instance when contributing):

pip install -e ".[docs]"
sphinx-build -b html docs docs/_build/html

then open docs/_build/html/index.html.

Examples

Fifteen complete, runnable studies from different domains live in examples/, from a five-line quickstart to CFD, wet-lab and particle-physics campaigns.

Citing

If you use Mergen in academic work, please cite it using the information in CITATION.cff; GitHub renders it under the "Cite this repository" button.

License

MIT © Ali Can Canbay