Coba

This release contains:

* Added more detail to the `where_best` docstring
* Fixed a bug that could prevent plots of unsorted values
* Fixed some docstring typos

This release contains:

* Fixed maxchunksperchild bug when setting in Experiment.run
* Fixed Result.where_best bug for partial experiments
* Improved error messages for params properties
* Corrected some type hints

This release contains:

* Fixed grammatical errors in the documentation

This release contains:

* Updated project links for pypi

This release contains:

* Improved performance for LinUCBLearner
* Improved contrast plot labels
* Fixed result bug with unsortable mixed types
* Fixed result bug with padded Learners
* Fixed inconsistent RandomLearner params

This release contains:

* Improved the performance of all synthetic datasets
* Improved Environments string representation
* Improved post-hoc hyperparameter optimization in Result
* Added Missing type for tables with missing values
* Added Environments create from experiment results
* Added support for rewards list in SimulatedInteractions
* Added warning for IPS evaluation when the evaluation policy is too correlated with logging policy
* Added cloudpickle recommendation message when multiprocessing in Jupyter Notebook
* Added from_feurer to the Environments API
* Added seamless reward serialization when logging rewards during evaluation
* Added support for pytorch serialization when multiprocessing
* Moved all interfaces to the coba.primitives module to reduce dependencies
* Updated all learners to return actions-probs instead of PMFs
* Fixed a bug when plotting x-axis values that were stored in the interactions table
* Fixed many bugs