rubin-data

The long-term record behind the Rubin tab of developerkaleb.github.io/PortfolioWebPage: outer Solar System objects large enough to have at least an 80% chance of having been shaped by their own gravity, assessed from public Rubin Observatory data.

Everything here is written by a scheduled GitHub Action (.github/workflows/collect.yml) running the collector in the site repo (tools/rubin-collect.mjs). Nothing is edited by hand.

Where the data comes from

• JPL Small-Body Database: the list of trans-Neptunian objects and their orbits.
• Fink, a Rubin alert broker: Rubin’s detections of those objects, from which their brightness is measured. Rubin’s alerts are world-public.

Requests are deliberately few and gentle: one at a time, seconds apart, a hard budget per run, backing off when asked, and identified with a contact address. A quiet week is one request to JPL; Fink is asked once a month.

Layout

Path	What
state.json	Last counts and fetch dates
inputs/YYYY-MM/tnos.json	JPL orbits and catalogue H for every TNO that month
inputs/YYYY-MM/detections.json	The Fink detections used that month
digest/YYYY-MM.json	That month’s results: passes, near misses, flags
digest/latest.json	The newest digest, which the site reads
changes/YYYY-MM.json	What changed since the month before
moons.json	The round moons’ orbits for the site’s map, refitted from JPL Horizons once a year

Inputs are kept every month so any future filter can be re-run against the whole history. Each digest records the site commit, cutoffs and thresholds that produced it. JSON files put one object per line, so the history diffs cleanly.

Reading a digest entry

• chance: the chance the object was shaped by its own gravity (evidence reading); passes when at least 0.8.
• chanceGrundy: the same if mid-sized TNOs never compacted (Grundy et al. 2019).
• chanceIfDark: the same if it were as dark as the darkest TNOs measured.
• flags: disputed, largeIfDark, and orbit flags (extremeOrbit, detached, highlyInclined, retrograde, unbound). watch marks a failing object that is both large-if-dark and on an unusual orbit.
• hSource: whether H came from Rubin’s own detections or the catalogue.

Data: JPL/NASA and the Rubin Observatory via the Fink broker; TNO albedos from Johnston’s compilation (NASA PDS, doi:10.26033/y5sn-4t02).

Maintenance windows (maintenance.json, edited by hand)

The only file here edited by hand. When Rubin announces downtime (its forum, www.rubin.community: “Early Operations Update” and “Summit technical progress” posts), add a window. The page says why Rubin is off-sky and, when an end is given, when new Rubin data will next arrive:

{
  "windows": [
    { "start": "2026-07-15", "reason": "Storm recovery and planned maintenance",
      "link": "https://www.rubin.community/t/summit-technical-progress-week-ending-2026-09-25/12773" }
  ]
}

start and end are dates (YYYY-MM-DD); leave end out until one is announced. A window with no end closes itself once Rubin sends alerts again (the page then says when Rubin returned), so forgetting to add the end does no harm. Without this file the page still notices a pause on its own, from Rubin’s nightly alert counts (no alerts for 4 nights or more), but cannot say why or for how long.