REPRODUCIBILITY.md

Reproducibility guide

What you need

• Python 3.10+ recommended
• Jupyter Notebook or JupyterLab
• Public magnetogram data (NOAA / NSO-GONG) or your own local frames

Quick start (pip)

1. Create and activate a virtual environment:

• macOS/Linux:
  • python3 -m venv .venv
  • source .venv/bin/activate
• Windows:
  • python -m venv .venv
  • .venv\Scripts\activate

2. Install dependencies:

• pip install -r requirements.txt

3. Launch Jupyter:

• jupyter lab (or jupyter notebook)

4. Open and run:

• DLSFH_Entropy_Diagnostic_NOAA.ipynb (Run All)

Expected outputs (canonical notebook)

When executed successfully, the notebook should generate:

• Node overlay / partition visualization (20-node DLSFH layout)
• Node-wise entropy values and ψ⋆s = exp(-S)
• Fragmentation map (ψ⋆s < ψcrit)
• Entropy ring detection result (adjacency-based)
• Composite risk score Rflare

Notes on data inputs

• If the notebook pulls data from remote sources, ensure you have network access.
• If running offline, place magnetogram frames in the notebook’s expected input path and update the input configuration cell accordingly.

Sensitivity checks

For parameter robustness:

• Run DLSFH_PhysicsEntropy_Enhanced.ipynb
• Use the parameter cells to vary:
  • ψcrit
  • minimum ring size
  • entropy histogram binning

Troubleshooting

• If you see missing-package errors: re-run pip install -r requirements.txt
• If plots are blank: confirm notebook kernel is the same environment where packages were installed
• If files are not found: verify the configured input path and filenames