Tom Nicholas

@keewis I think this is basically ready, though I don't know why the doctests CI fails

@keewis doctest is failing, but seems to be due to LaTeX formatting in `dequantify`, not because of this PR: ``` ________ [doctest] pint_xarray.accessors.PintDatasetAccessor.dequantify ________ 1117 array([2, 3, 4]) 1118 Dimensions...

@keewis I think this might be ready to merge now finally??

> you'll probably end up with something like Wow a lot of work has gone into that @dcherian ! > What you could do in xhistogram is apply_ufunc the function...

Hi there @miniufo , glad you are finding this package useful. > I had to use time-varying bins @jbusecke is right that changes of the data with time don't necessarily...

I've realised that this feature is kind of inconvenient to try and implement in xhistogram in its current form. xhistogram's numpy API means that all ND bins have to be...

For future reference [this SO discussion](https://stackoverflow.com/questions/5205345/using-numpy-bincount-with-array-weights) is about whether there are clever vectorized ways to do the bincounting when the weights are >1D.

@aaronspring I'm trying to make sure I understand what you're proposing here - is this an accurate summary of what you're suggesting?: If I have data which includes a time...

I'm not sure how the hypothesis dependency should be handled though - if the strategy is exposed publicly then I guess it should be an optional dependency of `dask[array]`? Not...

You might also be interested in some of the suggestions in https://github.com/HypothesisWorks/hypothesis/issues/3433#issuecomment-1212419256