Replace instances of pa.PandasDtype enum with python, numpy, pandas types

[cosmicBboy]

Documentation problem

The current documentation demonstrates pandera usage by using the pa.PandasDtype enum, which can make things look a little unfamiliar to new-comers, especially since it now supports the use of python types and numpy scalar types, for example, see:

• documentation example: https://pandera.readthedocs.io/en/stable/#quick-start
• docstrings example: https://github.com/pandera-dev/pandera/blob/master/pandera/schemas.py#L106-L113

Suggested fix for documentation

We should change the documentation and docstrings to reflect ease-of-use and adoption by replace instances of the pa.PandasDtype enum in this order of priority:

• python types, e.g. int, str, float, etc
• numpy types, e.g. np.int64, etc
• pandas types, e.g. pd.StringDtype()

[benkeesey]

Can I do take this?

[cosmicBboy]

• https://github.com/pandera-dev/pandera/blob/master/pandera/schemas.py#L110-L115

[benkeesey]

Things should all be updated in the PR, I did not edit these pa.DataDtypes as I wasn't sure what the best replacement should be. Please let me know if these should also be changed.

"pa.DateTime" https://github.com/pandera-dev/pandera/blob/master/pandera/schemas.py#L114 & https://github.com/pandera-dev/pandera/blob/master/docs/source/lazy_validation.rst#L64

"pa.Category" https://github.com/pandera-dev/pandera/blob/master/docs/source/dataframe_schemas.rst#L624

[cosmicBboy]

cool, thanks! we can leave pa.DateTime and pa.Category as-is for now.

[nathanjmcdougall]

I think this is closed by #581 (c.f. #496).