The 2.2 release brings support for defining the times/phases at which a dataset should be computed during run_compute separately from the times at which the dataset itself is defined. This is particularly useful for computing residuals between a dataset with a large number of timepoints and a model computed evenly in phase space, for example.
Besides the introduction of this new ability, the only significant change between 2.1 and 2.2 is that the mesh dataset and orb dataset no longer have times parameters. The times parameter has been replaced by the compute_times parameter (as neither the mesh or orbits have the ability to attach observable data to the dataset). Passing times to add_dataset (for example: b.add_dataset('mesh', times=[...])) will raise a warning in the logger and apply that value to the compute_times parameter. However, any scripts that attempts to find the times parameter will need to be updated to use the compute_times parameter instead.
For a more thorough explanation of the new behavior, see the new Compute Times & Phases tutorial.
In [ ]: