v1.4.0

What's Changed

• Add template for DoD on issues by @nabriis in #591
• Update printing for geometry by @jeverink in #628
• Enable callback in Gibbs and update callback interface to pass more information to the callback method by @amal-ghamdi in #623
• Fix geometry printing to account for case where geometry fun_shape is None by @amal-ghamdi in #632
• Fix NumPy 2.0.0 support by @jeverink in #634
• Drop support of Python 3.8 by @chaozg in #636
• Regularized Gaussian - additional constraints by @jeverink in #618
• Add UnboundedUniform distribution by @chaozg in #639
• Update scipy version requirement and related issues by @chaozg in #641
• Add option to use ScipyMinimizer in RegularizedLinearRTO and fix bug by @chaozg in #644
• Multiple Inputs: update the Model module by @amal-ghamdi in #610
• Robust automatic sampler selection/suggestion by @jeverink in #630
• Fix hybrid gibbs ordering by @jeverink in #660
• Multiple inputs support for pde gradient by @amal-ghamdi in #656
• Add example for using recommender class by @amal-ghamdi in #668
• Print for HybridGibbs by @jeverink in #662
• Add header for "task" issue template by @amal-ghamdi in #682
• Multiple input bayesian modeling by @amal-ghamdi in #664
• Allow other options for initial guess in (Regularilzed)LinearRTO steps by @chaozg in #679
• Add TweediePrior and ScorePrior as alias of MoreauYoshidaPrior by @chaozg in #681
• Ensure NUTS statefulness within Gibbs by @amal-ghamdi in #683
• add online thinning option to samplers by @chaozg in #690

Full Changelog: v1.3.0...v1.4.0