[REVIEW] Add no_default and adapt Series.reset_index to differentiate None for name parameter

[galipremsagar]

Description

In pandas-2.0 the behavior for name parameter has changed to actually name a column 0 if no value is passed to name. But if name=None, the column will be named None too:

In [1]: import pandas as pd

In [2]: s = pd.Series([10, 11, 23], index=[2, 3, 5])

In [3]: s
Out[3]: 
2    10
3    11
5    23
dtype: int64

In [4]: s.reset_index()
Out[4]: 
   index   0
0      2  10
1      3  11
2      5  23

In [5]: s.reset_index(name=None)
Out[5]: 
   index  None
0      2    10
1      3    11
2      5    23

To achieve the same behavior in cudf, we had to introduce no_default value(which is same as pandas's no_default value).

This also fixes 18 pytests:

= 965 failed, 86325 passed, 2044 skipped, 992 xfailed, 165 xpassed in 508.32s (0:08:28) =

On pandas_2.0_feature_branch:

= 983 failed, 86277 passed, 2034 skipped, 992 xfailed, 165 xpassed in 541.87s (0:09:01) =

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[bdice]

Let's match the pandas API layout.

[bdice]

Can we add tests for the cases newly supported by this PR (None names)?

[galipremsagar]

Can we add tests for the cases newly supported by this PR (None names)?

We do here: https://github.com/rapidsai/cudf/pull/13152/files#diff-5d087b898461a7b9dbabf8cde1c72a8ded397b75fc9402b7108524f24de67adeR1411

[bdice]

Can we add tests for the cases newly supported by this PR (None names)?

We do here: https://github.com/rapidsai/cudf/pull/13152/files#diff-5d087b898461a7b9dbabf8cde1c72a8ded397b75fc9402b7108524f24de67adeR1411

Sorry, I misread the filename and thought that was an implementation and not a test. My mistake. Approving now!