perf: Remove cast to boolean after comparison in optimizer

[nameexhaustion]

• Fixes Bad query plan produced when all_horizontal is used with join_where #21009
  • The expression gets recognized by collapse_joins without issue after we remove the extra cast
• Fixes Incorrect overflowing cast of literals containing Series #21023
  • Updates some old code to propagate options: CastOptions instead of strict: bool

Drive-by renames fn coerce_types() to fn optimize_exprs() under ConversionOptimizer, as it also does other optimizations aside from type coercion.

[codecov]

Codecov Report

Attention: Patch coverage is 93.47826% with 3 lines in your changes missing coverage. Please review.

Project coverage is 79.14%. Comparing base (ea1ea5a) to head (7305d1a).
Report is 6 commits behind head on main.

Files with missing lines	Patch %	Lines
...ars-plan/src/plans/conversion/type_coercion/mod.rs	93.33%	2 Missing ⚠️
...ates/polars-plan/src/plans/conversion/stack_opt.rs	83.33%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #21022      +/-   ##
==========================================
- Coverage   79.20%   79.14%   -0.07%     
==========================================
  Files        1583     1583              
  Lines      225109   225126      +17     
  Branches     2581     2581              
==========================================
- Hits       178303   178179     -124     
- Misses      46216    46357     +141     
  Partials      590      590              

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

[ritchie46]

Do you know why this not returns None? I understand it returns the same aexpr untouched.

[nameexhaustion]

The aexpr in this function is the input expr to the cast (not the cast itself), you can see how it is called if you expand the diff above -

So by returning it here we replace the cast expr with its input, which removes the cast.

[ritchie46]

Ah, now I see. Thanks!