I have over 130 repositories – a mix of old code, new code, stable workhorse code, forks, and odds and ends (and to be fair, I should clean a lot of that up). Here are some highlights, split between the public code and private work, the latter where most of my activity lies.
You can check out my company, Nightlight Research, or my publications. Or the SciCoder workshop I created (hey, you should host one!).
These are mature, maintained, and stable packages that make processing and visualizing astronomical data a lot easier. They are foundational building blocks for much more complex infrastructure I’ve created.
The sphere trilogy:
- qlsc – Quadrilateralized spherical cube sky indexing for Python: Q3C's catalog-indexing scheme without needing a PostgreSQL server. Do you have a sphere? Let’s index it.
- cornish – A Pythonic interface to the Starlink AST library: world coordinate systems and regions on the celestial sphere without the C API. What can I say; I love spheres.
- pgAST – Starlink AST as a PostgreSQL extension: WCS and sky-region operations inside the database, next to the data, instead of on the client. Sphere ❤️ database.
Other astronomy code of note.
-
Trillian MCP – Trillian is a broad platform I designed and implemented for working with large amounts of astronomical data. And then I slapped an MCP in front of it. This is a proof of concept - the potential is far, far greater.
-
libfitsverify – libfitsverify takes NASA/HEASARC's
fitsverify— the standard tool for verifying FITS files since 1994 — and refactors it into a modern library that can be embedded in pipelines, called from Python, and run on data in memory. -
ASCL.net – I wrote the latest version of the venerable ASCL.net web site (rebuild partially funded by NASA).
- Claude Sessions – Want to view your past and present Claude Code sessions in excruciating detail, and tag them, and perform full text searches? I did too! Here is a web dashboard you can run locally. Fully tested on every possible platform that is macOS.
- AI Agent Field Guide – Recipes for making your AI agents better.
Install PostgreSQL my way. The SQLAlchemy/database pattern I use in all my Flask apps. macOS/iOS custom code. And then many, many other odds and ends.
You’ll see a lot of GitHub activity in my account. Most of that work falls here.
- FITS Test Suite – I have developed a comprehensive test suite for the cfitsio C library - it is no exaggeration that virtually all astronomy flows through this library. To date it has found over 60 bugs; in the process of reporting them. Driven by me, implemented by AI, generalizable to other C libraries. Paper in preparation.
- ACAT – Automated Concept Assignment for the UAT. ACAT is an AI-driven framework that assigns Unified Astronomy Thesaurus (UAT) concepts to any given text (e.g. astronomy journal papers). It uses a fine-tuned embedding retriever and can utilize a fine-tuned local model or your LLM of choice to assign the most appropriate (i.e. most specific) concepts. Paper in preparation.
- Trillian – An all-sky, multi-wavelength astronomy distributed analysis platform, data interface and visualization. Mature, amazing, ready to do science, impossible to fund. (Ask me how I know.)
- Nightlight – A macOS desktop application for visualization of astronomical data for professionals. Fully modern, fully native, fully accelerated. Previously released, but now needs my attention to update it (which is hard to justify – see “impossible to fund”.)
- The Big AI Project – This is the one I can’t talk about at all as I’m preparing a patent application. Don’t even ask.
