feat(python): Add allow_nan parameter to parametric testing series() and dataframes() functions

[mcrumiller]

The allow_nan parameter wasn't being passed through to the underlying floats function via the series and dataframes parametric functions.

I did not update the docstrings because allow_nan is a dtype-specific parameter and it looks like the general docstrings for series and dataframes only include the parameters applicable to all dtypes.

[codecov]

Codecov Report

Attention: Patch coverage is 88.88889% with 1 line in your changes missing coverage. Please review.

Project coverage is 80.83%. Comparing base (7a1be3b) to head (f15bac6).

Files	Patch %	Lines
...olars/polars/testing/parametric/strategies/core.py	83.33%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #17188      +/-   ##
==========================================
- Coverage   80.84%   80.83%   -0.02%     
==========================================
  Files        1464     1464              
  Lines      192016   192022       +6     
  Branches     2743     2745       +2     
==========================================
- Hits       155243   155212      -31     
- Misses      36262    36298      +36     
- Partials      511      512       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[stinodego]

I really don't want to add this at this point. I have to rethink the design a bit and I don't want to take on more complexity before I do. The allow_infinity that you see is more of a backwards compatibility thing - that's not really the way this should work IMO.

We need to go for more of a composable setup. You can already get allow_nan functionality by using st.floats directly as the strategy of a column.

I'm sorry but I'm closing this one for now.