ENH: GLM models now save the names of input Pandas Series

[jmahlik]

Offset, exposure, freq_weights and var_weights have the name of the series saved on the model object. They can be accessed via the class properties.

Closes #9100

• closes Save the offset name in GLM and results wrapper #9100
• tests added / passed.
• code/documentation is well formatted.
• properly formatted commit message. See
  NumPy's guide.

Getting this as a starting point. This enables the feature for GLM's. Happy to make any recommended changes.

I didn't edit the release notes, wasn't sure if this would be worth mentioning. Kept the smaller notes in the commit message for now.

I opted for only doing one dimensional arrays in the helper because the caller would need to check the shape if it returned more than one name. Seems like a separate helper for multi dimensional might be needed unless there's a different way we were expecting to use it.

[josef-pkt]

Thanks, looks good

However, we need a default name if offset or similar are used, i.e. name should be a string and not None.
The name attribute is None should be for the case when the corresponding variable, offset, ..., was not used.

one possibility
add a extra argument to the helper function with the default name, e.g. "offset" or "exposure".
That way the helper function would still be generic, and we don't need to fix up each returned None name.
something like:
def _as_array_with_name(obj, default_name="")

We want to be able to add a string to the summary table

[jmahlik]

The test failure appears unrelated.

FAILED statsmodels/tsa/statespace/tests/test_exact_diffuse_filtering.py::test_nondiagonal_obs_cov failing for an assert_allclose.