feat(python): Add replace_all expression to complement replace

[stinodego]

Ref #14302

Changes

• Add replace_all, which is exactly the same as replace was, with the exception that defaults to filling unmatched values with null.
• Deprecate the default and return_dtype parameters for replace.

This should be followed up with a breaking change that makes sure replace always keeps the original data type, instead of casting to the supertype of the original and the new values.

[codecov]

Codecov Report

Attention: Patch coverage is 90.47619% with 2 lines in your changes are missing coverage. Please review.

Project coverage is 81.48%. Comparing base (cb40bbd) to head (c6382fd).
Report is 11 commits behind head on main.

Files	Patch %	Lines
py-polars/polars/expr/expr.py	89.47%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16557      +/-   ##
==========================================
+ Coverage   81.47%   81.48%   +0.01%     
==========================================
  Files        1412     1414       +2     
  Lines      185474   185579     +105     
  Branches     2996     3002       +6     
==========================================
+ Hits       151119   151225     +106     
+ Misses      33839    33837       -2     
- Partials      516      517       +1     

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

[ritchie46]

This should be added on the rust side as well.

[stinodego]

This should be added on the rust side as well.

I'll add that when I'm making the breaking change to casting behavior for 1.0 - it'll be easier then!

EDIT: I think I can do some of the work already, working on it...

EDIT2: Nope, it's a hassle. Let's merge like this and I can do this without jumping through hoops to avoid breaking something.