BTYD v0.1beta1

[ColtAllen]

0.1beta1

• Bayesian predictions now supported, enabling entire probability distributions as well as point estimates for predictive outputs.
• Streamlined user API to minimize input arguments. All predictive methods are now also called from a single function.
• model._idata attribute now persisted as an arviz.InferenceData object, and can be saved externally in JSON or CSV format. However, only JSONs can be loaded as of this release.
• Documentation updated to latest versions of sphinx and pydata-sphinx-theme.
• Removed High Level Overview from documentation.
• Added deprecation warning for legacy Lifetimes fitters module.
• Removed extraneous lifetimes import causing build issues.
• Removed psutils library dependency.
• Added numpy >=1.20.0 library dependency.
• utils.posterior_predictive_deviation metric removed pending further evaluation.
• CI/CD pre-commit scripts added.

---

This discussion was created from the release BTYD v0.1beta1.