perf: Remove prepare_series for gather kernels

[orlp]

For left-hand sides with many chunks (most notably in the new-streaming join) the prepare_series was very expensive and it's not necessary at all (we just have to match on a couple more logical types).

[codecov]

Codecov Report

Attention: Patch coverage is 54.76190% with 38 lines in your changes missing coverage. Please review.

Project coverage is 79.00%. Comparing base (c6769c2) to head (88754ef).
Report is 27 commits behind head on main.

Files with missing lines	Patch %	Lines
...tes/polars-ops/src/chunked_array/gather/chunked.rs	54.76%	38 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #20439      +/-   ##
==========================================
- Coverage   79.01%   79.00%   -0.01%     
==========================================
  Files        1563     1563              
  Lines      220290   220340      +50     
  Branches     2488     2488              
==========================================
+ Hits       174055   174079      +24     
- Misses      45662    45688      +26     
  Partials      573      573              

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