Gaia DR3 Epoch Photometry — Variable Star Detector

Processes Gaia DR3 epoch photometry data to identify astronomical objects with significant flux variability (>100% change) across BP and RP bands.

How It Works

1. Input: 20 gzipped CSV files in data/in/ containing epoch photometry from Gaia DR3. Each row represents one star with time-series flux measurements packed as arrays.

2. Processing: For each star, the script:

   • Parses the bp_flux and rp_flux observation arrays
   • Filters out NaN/invalid values
   • Computes min and max flux for each band
   • Calculates percentage change: ((max - min) / min) × 100
   • Keeps the larger of BP vs RP percentage change

3. Output: data/out/result.csv containing only stars where percentage change exceeds 100%, with columns: source_id, bp_min_flux, bp_max_flux, rp_min_flux, rp_max_flux, percentage_change

The heavy lifting is done in Python (src/process.py) using multiprocessing.Pool to process all 20 files in parallel. The IRIS routine (src/RunScript.mac) serves as the entry point and timer.

Installation

Import the RunScript.mac file
Import the process.py file

Running

docker-compose exec iris iris session iris -U USER

Then in the IRIS terminal:

USER> do ^RunScript

The script will process all input files and write results to data/out/result.csv.

Project Structure

src/
  RunScript.mac   — Entry point (Cache ObjectScript routine)
  process.py      — Data processing (Python, stdlib only)
data/
  in/             — Input CSV files (gzipped)
  out/            — Output results (generated)