DRGScripts

Shared utility scripts and notebooks for the Dana Research Group — tooling around RMG-Py, ARC, T3, Arkane, and Cantera for building chemical kinetic models and analyzing them.

Most scripts assume you already have RMG-Py / ARC / T3 installed and on your PYTHONPATH (see bashrc/ for the typical environment aliases).

What's in here

Path	What it does
ARC/	Notebooks that generate ARC input files — from a SMILES list, from a Chemkin model (GAV species), or from an Arkane/Explorer input. Includes ARC/CLAUDE.md, a guide for using Claude Code inside the ARC repo.
Kinetics_lib/	K_extract.py — build a non-pressure-dependent RMG kinetic library from literature rate data in CSV.
Pdep/	K_Pdep_extract.py — build a pressure-dependent (PDepArrhenius) RMG kinetic library from CSV.
Cantera/ROP/	ROP.ipynb — rate-of-production analysis on a Cantera mechanism.
FluxDiagrams/JSR/	Generate annotated reaction-flux diagrams (with molecule images) for a JSR simulation.
kinetics/	Notebooks comparing reactions against RMG-database estimates via ARC's functions.
Servers/	Per-server (Atlas / Azure / Zeus) ARC & T3 settings.py, submit.py, and submit-script templates. Placeholders only — copy into ~/.arc / ~/.t3 and fill in your own values.
bashrc/	Example .bashrc aliases for the RMG / ARC / T3 workflow.
data/	Shared reference data used by some of the notebooks.

Usage notes

• Kinetic library extraction (Kinetics_lib/, Pdep/): drop your smiles*.csv and kinetics*.csv inputs into the script's files/ folder, edit the library metadata at the bottom of the script, then run python K_extract.py / python K_Pdep_extract.py. The output reactions.py is a ready-to-use RMG kinetics library.
• Notebooks: open with Jupyter from the notebook's own directory; paths are relative to it. Edit the clearly-marked input paths/placeholders at the top of each notebook.
• Server configs (Servers/): never edit the repo copies with real values. Copy the relevant settings.py / submit.py into your home ~/.arc (or ~/.t3) and fill in your username, server address, and key path there.

Contributing

• Branch off main and open a pull request; don't push directly to main.
• CI runs on every PR (see badges above):
  • a Python lint: python -m compileall byte-compiles all scripts, then ruff fails only on genuine errors (syntax errors, undefined names) — style is intentionally not enforced;
  • a gitleaks secret scan.
• Never commit real secrets. Server settings here are placeholders; keep real credentials in your local ~/.arc / ~/.t3.
• Clear notebook outputs before committing to avoid leaking local paths/usernames and bloating the repo:

  jupyter nbconvert --clear-output --inplace path/to/notebook.ipynb

License

MIT © DanaResearchGroup