perf: Don't rechunk when converting DataFrame to numpy/ndarray

[ritchie46]

closes #16267

[codecov]

Codecov Report

Attention: Patch coverage is 95.45455% with 1 lines in your changes are missing coverage. Please review.

Project coverage is 80.79%. Comparing base (bbe73bd) to head (f520057).

Files	Patch %	Lines
crates/polars-core/src/chunked_array/ndarray.rs	95.45%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16288      +/-   ##
==========================================
- Coverage   80.81%   80.79%   -0.02%     
==========================================
  Files        1393     1393              
  Lines      179337   179329       -8     
  Branches     2921     2921              
==========================================
- Hits       144923   144897      -26     
- Misses      33911    33929      +18     
  Partials      503      503              

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

[abstractqqq]

the polars_ensure! call happens now before the none_to_nan, which means that a float64 column with nulls can cause error. The docs says the contrary, and nulls should be filled by NaN automatically. Previously, this was avoided because the columns were pre-processed.