jacscanomaly

jacscanomaly is a Python package for scan-based anomaly detection in time-series light curves, with microlensing anomaly searches as the primary use case.

The package is built for the common situation where a smooth baseline model fits most of a light curve, but localized deviations may contain the signal of interest. In microlensing applications the baseline is usually a single-lens model and the deviations are candidate planetary anomalies.

What it does

jacscanomaly provides:

• single-lens fitting utilities for PSPL, FSPL, and parallax variants,

• a residual scan over anomaly center time and effective duration,

• non-overlapping candidate extraction,

• candidate support diagnostics such as n_eff and peak_frac,

• plotting and summary helpers for scripts, CLIs, and notebooks,

• low-memory C++ backends for large PSPL survey scans.

The high-level entry point is jacscanomaly.Finder. A typical workflow is:

1. fit a single-lens baseline model to a light curve,

2. scan the residuals over local anomaly templates,

3. extract non-overlapping anomaly candidates,

4. inspect candidate quality diagnostics such as score and n_eff.

Start here

If you are using the package for the first time:

• read Installation to install the package and optional dependencies,

• follow Quickstart to run a light curve,

• use Method overview to understand dchi2, score, and n_eff,

• use Configuration when tuning a survey run,

• use Results when building candidate tables or downstream analysis.

User guide

• Installation
  • Basic install
  • Optional dependencies
  • FSPL and microjax
  • C++ backend
• Quickstart
  • Run the anomaly finder
  • Use quality criteria
  • Inspect the result
  • Plot the result
  • Fit only the single-lens model
• Examples
  • Minimal synthetic light curve
  • Running only the baseline fit
  • Using candidate criteria
  • Plotting
  • Example notebooks
• Method overview
  • Pipeline
  • Detection statistic
  • Candidate score
  • Effective number of points
  • Other quality diagnostics
  • Season splitting
  • Backends
• Configuration
  • Single-lens model
  • Automatic single-lens initialization
  • Season splitting
  • Anomaly grid
  • Candidate selection
  • Backend selection
• Results
  • Summary methods
  • Important attributes
  • Best candidate
  • Candidate quality fields
  • Working with all candidates

Reference

• API Reference
  • Main workflow
  • Result containers
  • Plotting
  • Single-lens fitting
  • Module reference
• Testing
  • Current scope
• Development
  • Editable install
  • Unit tests
  • Coverage
  • Documentation
  • Continuous integration

Project indices

• Index

• Module Index

• Search Page