timex¶
Transit timing estimator in Jax — a Python package for robust, efficient, and flexible Bayesian analysis of individual exoplanet transit events.
timex is a JAX/numpyro/jaxoplanet port of timer, which uses PyMC/exoplanet.
The primary goal of timex is to precisely measure transit mid-times from one or more light curves of the same event, with full posterior uncertainties via MCMC. It is designed for the common workflow of fitting a single transit epoch covered by one or more simultaneous datasets (e.g., multi-band photometry from MuSCAT), while incorporating prior knowledge of the system parameters.
Because all key transit parameters can be assigned either Gaussian or uniform priors, timex is also useful for broader transit analyses beyond timing — for example, measuring chromatic radius ratios across multiple photometric bands to validate planet candidates or detect atmospheric features (e.g., Na absorption with narrowband spectrophotometry), or characterizing spot-crossing and flare events that overlap with transits.
Features¶
- Simultaneous fitting of multiple light curves of the same transit event
- Chromatic (per-band) radius ratio fitting for planet validation and spectrophotometry
- Stellar flare modeling (Davenport et al. 2014 analytic profile) with support for multiple flares
- Spot-crossing (Gaussian bump) modeling, including chromatic amplitudes
- Flexible detrending: polynomial trends, B-splines, Gaussian processes (Matern-3/2), and arbitrary covariates (e.g., airmass, centroids)
- Theoretical quadratic limb darkening priors via Claret tables
- MCMC posterior sampling via numpyro (NUTS) with automated MAP initialization
- Sigma-clipping of outliers with automatic re-fitting
- Publication-quality outputs: multi-panel light curve fits, corner plots, trace plots, limb darkening diagnostics
Quick start¶
git clone https://github.com/john-livingston/timex.git
cd timex
pip install -e .
pip install git+https://github.com/john-livingston/limbdark
timex examples/hip67522b
The working directory must contain both sys.yaml (system parameters) and fit.yaml (fit configuration) files.