History

bmds-ui

Version 25.1

Released on 2025-04-25.

  • Add Nested Dichotomous NCTR Model

  • Add Rao-Scott Transformation for summary Nested Dichotomous data

  • Add cloning analysis action

  • Only show dataset type selector for continuous data

  • Enable schema migration for future backwards compatibility

  • Improve database stability for BMDS Desktop

  • Package versions, security updates, etc.

Version 24.1

Released on 2024-11-25.

  • Added BMDS Desktop mode for execution

  • Added support for Multitumor modeling

  • Added support for Nested Dichotomous modeling

  • Added poly-k adjustment for dichotomous datasets

  • Minimum pybmds version increased to version 24.1

  • Initial release of bmds-ui; project was originally forked from shapiromatron/bmds-server

pybmds

Version 25.2

Released on 2025-10-xx.

  • Added Cochran Armitage trend test for dichotomous data

  • Added Jonckheere-Terpstra trend test for continuous data

  • Added additional plotting functionality for nested dichotomous data

  • Changed restriction for rho parameter in non-constant variance model to allow negative values

  • Added ability to count all parameters in a model automatically for the purpose of calculating AIC and p-values.

Version 25.1

Released on 2025-04-25.

  • Add Nested Dichotomous NCTR Model

  • Add Rao-Scott Transformation for summary Nested Dichotomous data

  • Add CDF to Word Report for dichotomous bayesian model average

  • Add warnings for invalid parameter settings (min, max, initial)

  • Improve plotting ranges for BMDs extrapolated beyond the dose range

  • Fix bug in Quantal Linear plotting

  • Fix bug in Scaled Residual calculations for continuous lognormal distributions

  • Remove reporting of burn-in and # of samples for bayesian modeling since pybmds uses the Laplace Approximation

Version 24.1

Released on 2024-11-25.

  • Added Multitumor and Multistage Cancer model

  • Add Nested Dichotomous models

  • Updated bmdscore library; integration now uses pybind11 for direct integration instead of using a shared library. This should improve performance and stability and make it easier to develop moving forward.

  • Build distributions for windows, mac, and linux using cibuildwheel

  • Initial release of pybmds; the python components of the package were originally taken from shapiromatron/bmds