// Reproducibility

Data & Code

Reproducibility Package

Everything needed to reproduce this study from scratch.

Source Code

All simulation, analysis, and visualization code is pure Python with NumPy/SciPy/Matplotlib dependencies. No GPU, no ML, no proprietary tools.

File	Description	Lines
`gs_engine.py`	Core Gray-Scott simulation engine + classification metrics	170
`theory.py`	Linear stability analysis, steady states, Turing conditions	200
`sweep.py`	Parallel parameter sweep across (F, k) space	120
`visualize.py`	Publication-quality figure generation	300

Sweep Results

The full parameter sweep data is available as both CSV and NumPy arrays.

CSV Format (`results.csv`)

Column	Description
`F`	Feed rate
`k`	Kill rate
`coverage`	Fraction of grid where v > 0.1
`mean_v`	Mean V concentration
`std_v`	Standard deviation of V field
`entropy`	Shannon entropy (bits) of V histogram
`wavelength`	Dominant spatial wavelength (pixels)
`pattern_type`	Automated classification

2D Grid Arrays

grid_entropy.npy — 48×48 Shannon entropy
grid_coverage.npy — 48×48 V coverage
grid_std.npy — 48×48 standard deviation
grid_wavelength.npy — 48×48 dominant wavelength
f_values.npy, k_values.npy — axis values

Key Numbers

Simulations:      2,304 (48 × 48 grid)
Time steps:       8,000 per simulation
Grid resolution:  128 × 128
Total compute:    101.8 minutes (4 cores)
Sim rate:         0.38 simulations/second
Random seed:      42

Pattern distribution:
  DEATH:      682  (29.6%)
  PATTERN:    508  (22.0%)
  SATURATION: 1114 (48.4%)

Quick Start

# Install dependencies
pip install numpy scipy matplotlib

# Run a single simulation
python3 gs_engine.py

# Run the full sweep (~2 hours on 4 cores)
python3 sweep.py

# Generate all figures
python3 visualize.py --all

Hardware

This study was computed on a single Linux VM with 4 CPU cores and 23 GB RAM. No GPU acceleration was used. The total compute time of ~102 minutes would have cost thousands of dollars on 1990s supercomputer time; it cost zero dollars on a 2026 cloud VM.

More from Herman — see this project on the main site, alongside other work like Pokemon Emerald: Cathode, the Bobiverse novel, and the dashboard platform.