ENH: Add option to select stats to display under coefs in summary2.summary_col

[EBoisseauSierra]

• Linked to Extend summary2.summary_col features #8996
• tests added
• tests passed.
• code/documentation is well formatted.
• properly formatted commit message. See
  NumPy's guide.

Notes:

• It is essential that you add a test when making code changes. Tests are not
  needed for doc changes.
• When adding a new function, test values should usually be verified in another package (e.g., R/SAS/Stata).
• When fixing a bug, you must add a test that would produce the bug in main and
  then show that it is fixed with the new code.
• New code additions must be well formatted. Changes should pass flake8. If on Linux or OSX, you can
  verify you changes are well formatted by running

  git diff upstream/main -u -- "*.py" | flake8 --diff --isolated
  

  assuming flake8 is installed. This command is also available on Windows
  using the Windows System for Linux once flake8 is installed in the
  local Linux environment. While passing this test is not required, it is good practice and it help
  improve code quality in statsmodels.
• Docstring additions must render correctly, including escapes and LaTeX.

This PR introduces a new optional argument to statsmodels.iolib.summary2.summary_col, which allows choosing which stat to display under the coefficients:

  def summary_col(
      results,
      float_format: str ='%.4f',
      model_names=(),
      stars=False,
      info_dict=None,
      regressor_order=(),
      drop_omitted=False,
      include_r2=True,
+     coef_stats: str="std_error",
  ):

Example:

• std_error (default, current behaviour):

=======================
                  y    
-----------------------
const          0.5000  
               (0.6708)
x1             0.6000  
               (0.2449)
R-squared      0.3000  
R-squared Adj. 0.2500  
=======================
Standard errors in
parentheses.

• t_stat:

=======================
                  y    
-----------------------
const          0.5000  
               (0.7454)
x1             0.6000  
               (2.4495)
R-squared      0.3000  
R-squared Adj. 0.2500  
=======================
t-statistics in
parentheses.

• z_stat:

=======================
                  y    
-----------------------
const          0.5000  
               (0.7454)
x1             0.6000  
               (2.4495)
R-squared      0.3000  
R-squared Adj. 0.2500  
=======================
z-statistics in
parentheses.

• "" or None:

=====================
                 y   
---------------------
const          0.5000
x1             0.6000
R-squared      0.3000
R-squared Adj. 0.2500
=====================

• "foobar":

Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/Users/ebosi/perso/statsmodels/statsmodels/iolib/summary2.py", line 543, in summary_col
    raise ValueError("coef_stats must be one of 'std_error', 't_stat', "
ValueError: coef_stats must be one of 'std_error', 't_stat', 'z_stat', or an empty string. Got 'foobar' instead.

[josef-pkt]

is this still WIP?

I'm now working on merging the other two summary_col PRs
This and the other 2 PRs have merge conflicts, that I want to work through in sequence. I would like to add this PR last if it is finished.

[josef-pkt]

typing makes signatures really difficult to read

[EBoisseauSierra]

is this still WIP?

I was waiting for both other PRs to be merged to rebase this branch and ensure it's green before requesting a formal review because I'm aware there are conflicts.

typing makes signatures really difficult to read

YMMV. I felt better trying to increase type hints "coverage" within the codebase — which I find very helpful when contributing. But I don't mind dropping them; let me know what you want.

[josef-pkt]

I don't like type hints in code that I regularly maintain and enhance. Type hints are much too distracting, annoying to read and slows down skimming code.
I don't mind type hints in code that I have to look at only every few years, summary is one of those.

Are you using type hints as part of the code documentation or for type checking in editors?