[ENH] pass targets to sklearn as 1D and not 2D if univariate column vector

[fkiraly]

sklearn estimators wrapped in BootstrapRegressor or ResidualDouble complain if uni-output target variables are passed as 2D column array, not 1D.

In order to still allow multioutput regressors but avoid the warning, this PR introduces a utility and applies it to flatten uni-output 2D arrays (column vectors) to 1D, but leave multivariate 2D arrays untouched.

[codecov-commenter]

Codecov Report

Patch coverage: 100.00% and project coverage change: +0.06% 🎉

Comparison is base (54a73a1) 71.35% compared to head (c6c59ec) 71.41%.
Report is 4 commits behind head on main.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #74      +/-   ##
==========================================
+ Coverage   71.35%   71.41%   +0.06%     
==========================================
  Files          92       93       +1     
  Lines        4940     4951      +11     
  Branches      897      898       +1     
==========================================
+ Hits         3525     3536      +11     
  Misses       1179     1179              
  Partials      236      236              

Files Changed	Coverage Δ
skpro/regression/bootstrap.py	98.18% <100.00%> (+0.06%)	⬆️
skpro/regression/residual.py	97.34% <100.00%> (+0.07%)	⬆️
skpro/utils/numpy.py	100.00% <100.00%> (ø)

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