Releases: valboz/RTModel
Releases · valboz/RTModel
v3.3
- Adaptive gradients differentiated for each parameter in Levenberg-Marquardt fitting.
- Anomaly detection code to trigger initial conditions with additional lenses (thanks to Antonio Consiglio).
- Triple-lens fitting (only from anomalies on binary lens models) in static configuration, including parallax and/or orbital motion.
- Reader lists the datasets used distiguishing those taken from ground or from a satellite.
- Efficiency much higher than previous versions.
v3.1
v3.0
In v3.0 a whole restructuring of the code has been realized. New features are:
- Astrophotometric fits: it is now possible to include datasets with the measurement of the position of the centroid at each epoch.
- High-Resolution imaging: it is now possible to include constraints from high-resolution observations detecting lens ans source as separate objects.
- Possibility to turn-off the luminosity of the secondary source or the secondary lens, useful to fit systems with dark companions.
- Possibility to change the mass-luminosity-radius relations in binary source and lens models.
- Full access to all parameters ranges used in fitting.
- Preliminary models are now stored in a limited number of files and deleted at the end of the modeling run, unless requested otherwise.
- New plotting functions for the centroid trajectory ans options for further customization.
- Several bugs corrected.
v2.4
Update to v2.4
- Automatic detection of existing installation of VBMicrolensing at build time;
- Output of failing subprocesses shown for easy error detection;
- Bug correction in automatic cleanup of preliminary models;
- Lower boundary of q decreased to 1.e-7;
- Refinement of offset degeneracy calculation;
- New function in plotmodel to convert from fitting parameters to conventional orbital elements.
v2.3
- Update to VBMicrolensing 4.1.2;
- additional accurate initial conditions for planetary events;
- new strategy to search for offset degeneracy partners in binary-lens fits;
- changed maximal steps in each parameter in LevMar;
- ellipsoid increased by delta chi square and more accurate check of duplicates in model selection;
- models above Wilks' theorem threshold are removed in Finalizer;
- accuracy option in plotmodel;
- source trajectory and caustic in templates.show_template().
- several bug corrections in Reader, bumper, ModelSelector, LevMar;
v2.2
v2.2 with new features and optimizations
- Replaced VBBinaryLensing with VBMicrolensing
- Added the possibility to constrain fits
- possibility to choose model categories to fit
- possibility to make just a quick update to previously found models
- added Keplerian orbital motion for binary lenses as model category "LK"
- partial implementation of static triple-lenses, triple-lenses with parallax
- optimization of ModelSelector excluding higher chi square models from check
- possibility to cleanup event directory from preliminary models, thus saving 99% disk space.
v2.1
RTModel v2.1 has several new features and small optimizations:
- Possibility ro recover options from a previous run
- Fit from user-defined initial conditions
- Time limit enforced by killing stuck processes
- New functions in plotmodel subpackage for plotting light curves and caustics separately and for saving figures to file
- Full customizable template library with visualization of templates and computation of peaks for matching
- Minor optimizations for the template library, single-lens-single-source fitting, initial conditions settings.
RTModel 2.0
This is the first official release of RTModel a package for modeling microlensing events. A private version of this software has been operating since 2013 (see RTModel website).
A popular spin-off of this project was VBBinaryLensing, available as a separate public package since 2016.
Development of RTModel will continue over time with new additions in future versions.