Transform columnar data for interactive

[hoxbro]

In the next release of Holoviews 1.16, this will raise an error:

import hvplot.pandas
import pandas as pd

pd.DataFrame([1, 2, 3, 4]).interactive()

Fixes a known issue first shown here #918 and are related to #932.

Other places that use hv.Dataset are here. I don't know if they will error:

[maximlt]

Thanks a lot for reminding us of this, that was completely out of my radar!

I refactored a bit the code and added a test.

About the remaining occurrences of hv.Dataset:

• Scatter Matrix: applied the conversion as it's easy to create a scatter matrix from pd.DataFrame(np.random.random(...)), just to try it out
• Parallel coordinates: did not apply the conversion, didn't really see how one could end up with non string columns that can be used in this context
• Andrew curves: same

[hoxbro]

Love the changes.

A small edge case could be a xarray.DataArray (or similar) with a variable named columns.

import hvplot.xarray
import xarray as xr

xr.DataArray([1, 2, 3, 4], dims=["columns"], name="test").interactive()

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[12], line 4
      1 import hvplot.xarray
      2 import xarray as xr
----> 4 xr.DataArray([1, 2, 3, 4], dims=["columns"], name="test").interactive()

File ~/miniconda3/envs/holoviz/lib/python3.10/site-packages/xarray/core/extensions.py:38, in _CachedAccessor.__get__(self, obj, cls)
     35     pass
     37 try:
---> 38     accessor_obj = self._accessor(obj)
     39 except AttributeError:
     40     # __getattr__ on data object will swallow any AttributeErrors
     41     # raised when initializing the accessor, so we need to raise as
     42     # something else (GH933):
     43     raise RuntimeError(f"error initializing {self._name!r} accessor.")

File ~/Repos/holoviz/hvplot/hvplot/interactive.py:278, in Interactive.__init__(self, obj, transform, fn, plot, depth, loc, center, dmap, inherit_kwargs, max_rows, method, _shared_obj, _current, **kwargs)
    276 self._max_rows = max_rows
    277 self._kwargs = kwargs
--> 278 ds = hv.Dataset(_convert_col_names_to_str(self._obj))
    279 if _current is not None:
    280     self._current_ = _current

File ~/Repos/holoviz/hvplot/hvplot/util.py:558, in _convert_col_names_to_str(data)
    556 if not hasattr(data, 'columns'):
    557     return data
--> 558 renamed = {c: str(c) for c in data.columns if not isinstance(c, str)}
    559 if renamed:
    560     data = data.rename(columns=renamed)

File ~/Repos/holoviz/hvplot/hvplot/util.py:558, in <dictcomp>(.0)
    556 if not hasattr(data, 'columns'):
    557     return data
--> 558 renamed = {c: str(c) for c in data.columns if not isinstance(c, str)}
    559 if renamed:
    560     data = data.rename(columns=renamed)

TypeError: unhashable type: 'DataArray'

For this case it seems to be possible to add a hashable check:

from collections.abc import Hashable

{c: str(c) for c in list(data.columns) if not isinstance(c, str) and isinstance(c, Hashable)}

[maximlt]

Last suggestion by @hoxbro added in 4093310.

Unrelated job failure, merging.