AGENT SKILL

exoplanet-workflows

End-to-end workflow for detecting a transiting exoplanet in a stellar-variability-dominated light curve, including final period selection rules to avoid half/double-period aliases.

License 许可
Apache-2.0
最近更新
2026-11-14
Stars
★ 0
Forks
1

01 前言 Front Matter

name
exoplanet-workflows
description
End-to-end workflow for detecting a transiting exoplanet in a stellar-variability-dominated light curve, including final period selection rules to avoid half/double-period aliases.

03 SKILL.md 正文

Steps

  1. Load data and confirm column meanings (time system, normalized flux, quality flag convention, flux uncertainty).
  2. Filter on quality flags (typically flag == 0 for TESS-like exports) and remove NaNs/infs.
  3. Remove outliers conservatively (sigma-clip), then detrend stellar variability with a method that preserves short transits (e.g., lightkurve.flatten with window length longer than the transit duration).
  4. Run a transit-focused period search (TLS preferred; BLS acceptable) over a broad period range.
  5. Identify top candidate periods and always check common harmonics (P, 2P, sometimes P/2) using:
    • odd/even transit-depth consistency,
    • folded-transit shape clarity,
    • SNR/SDE (or BLS depth SNR).
  6. Only after alias resolution, do a narrow, dense-grid refinement around the chosen period to reach the precision needed for reporting.
  7. Write the final period to /root/period.txt as a single value in days rounded to 5 decimals.

Expected Result

A reproducible pipeline that detrends stellar activity, detects the transit period, and finalizes the orbital period after harmonic/odd-even validation, producing correctly formatted output.

04 同仓库更多 Skills