ajpotts

If casting to float is just to use the `log` trick, can we just cast back to int at the last step?

Associated tasks/tickets: - [ ] https://github.com/Bears-R-Us/arkouda/issues/2927 - [ ] https://github.com/Bears-R-Us/arkouda/issues/2958

Spreadsheet documenting the discrepancies: https://docs.google.com/spreadsheets/d/1HqPh59h3ZhoqqN4cYOxgglzUBRjLpgw4cBmFTzjqd48/edit?usp=sharing Contact ajpotts if you need access.

- [ ] [arkouda.array_api](https://bears-r-us.github.io/arkouda/autoapi/arkouda/array_api/index.html) - [ ] [arkouda.accessor](https://bears-r-us.github.io/arkouda/autoapi/arkouda/accessor/index.html) - [ ] [arkouda.alignment](https://bears-r-us.github.io/arkouda/autoapi/arkouda/alignment/index.html) - [ ] [arkouda.array_view](https://bears-r-us.github.io/arkouda/autoapi/arkouda/array_view/index.html) - [ ] [arkouda.categorical](https://bears-r-us.github.io/arkouda/autoapi/arkouda/categorical/index.html) - [ ] [arkouda.client](https://bears-r-us.github.io/arkouda/autoapi/arkouda/client/index.html) - [ ] [arkouda.client_dtypes](https://bears-r-us.github.io/arkouda/autoapi/arkouda/client_dtypes/index.html) - [...

Spreadsheet documenting the discrepancies: https://docs.google.com/spreadsheets/d/1akI6JuBu9Gj6QsFt1I5VEJUD_uURiZg-yXV5cJrI8-E/edit?usp=sharing Contact ajpotts if you need access.

Relevant pandas docs; https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.count.html

Just a note, to compile this PR you need to set `export ARKOUDA_PYTHON_SERVER=True`.

I can't recreate this problem so I will assume @drculhane is correct and close this issue.

To do this ticket properly it really requires MultiIndex API to match pandas better https://github.com/Bears-R-Us/arkouda/issues/3224 and also we need the dataframe columns property to handle the MultiIndex case.

I researched this a bit and the problem is that the `power_divergence` function in scipy underwent a non-trivial refactor in scipy `1.14.0`. We would need to refactor that function to...