William Jones
William Jones
Source code inspection for the conda-forge feedstock found that pytables is not referenced anywhere and could be removed as a dependency. Is this true, and if so should it be...
Adding support for periodic boundary conditions in `merge_split_MEST` would be a nice goal during the v1.5.x release cycle. This could be done using the scikit-learn `BallTree` distance search approach used...
Following on from #384 , I've carried out a more substantial refactoring of the PBC distance calculation functions. This includes improved validation of PBC keyword parameters, and optimisation of the...
This turned into a general refactor of `merge_split_MEST`, with the following changes and additions: 1. Adds PBC capabilities, using the same `BallTree` distance search approach as used in feature detection....
In many of our tests, we create test features using the existing routines and then test them against hardcoded values. It would be good to improve this mocking so that...
Cartopy is currently used in the `plotting` module, as well as being a dependency of `iris`. Installing `cartopy` with `pip` is difficult due to its dependency on the `GEOS` library....
It would be nice to add some animations of tracking performed using _tobac_ to the github readme and readthedocs to help promote _tobac_. Something similar to [PyFLEXTRKR](https://github.com/FlexTRKR/PyFLEXTRKR)
I've recently noticed that bulk statistics can run very slowly when applied to data that is non-contiguous. This can happen when slicing dask arrays or broadcasting along the trailing dimension....
As per title. e.g. `np.max` produces values smaller than the threshold. Works correctly when applied to a length 1 slice of the data. I'll try and dig into what is...
Running segmentation over a single timestep array returns a segment cube with no time dimension. This is not normally an issue, but bulk_statistics needs a time dimension so fails to...