Tom Nicholas
Tom Nicholas
Removes dask from the "core dependencies" CI environment file, so that no we have a CI run that tests everything without dask installed. I altered an existing run rather than...
Removing huge amounts of code that the refactor makes redundant! Early stage so far but wanted to push so that it was backed up somewhere other than my laptop. We...
Sketch of how we might refactor the code to pad inside the ufunc, as discussed in #443. The idea is that all the complexity of composing the the padding operation...
@rabernat pointed out that in the grid ufunc refactor, we are currently doing padding with xarray and then applying a numpy operation using `xarray.apply_ufunc` afterwards. This works, but it will...
In theory dask is meant to be optional for xgcm, but in practice we don't actually test that it works without dask installed in the CI. Some of my changes...
Currently we have important pieces of information about the grid just stored as basic python class attributes, e.g. `Grid.metrics`, or `Axis._periodic`. Do we want these to be read-only? i.e. are...
@jbusecke and I think that it might be time to remove the `keep_coords` kwarg, for simplicity's sake. My understanding is that it was originally introduced as a workaround to avoiding...
This correction needs to be applied, but we wanted to be confident that the tests pass with no alteration first. corrected version of `test_grid.py::test_grid_no_coords`: ```python ds, periodic, expected = periodic_1d...
This library is great, you've done a really nice job with it so far @t-makaro . I'm using it pretty regularly, and getting close to having it integrated within xarray...
Attempts to fix #32 by automatically choosing the length of the slider based on the length of the slider label and value text. There's a unit test which interrogates the...