Add protocol for estimator builder

[abelaba]

This pull request builds upon #1557 and addresses remaining comments by updating DensityEstimatorBuilder Protocol to be more flexible. This change allows the protocol to work with different types of conditional density estimators.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
⚠️ Please upload report for BASE (main@67d7e1c). Learn more about missing BASE report.
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1633   +/-   ##
=======================================
  Coverage        ?   82.56%           
=======================================
  Files           ?      135           
  Lines           ?    10957           
  Branches        ?        0           
=======================================
  Hits            ?     9047           
  Misses          ?     1910           
  Partials        ?        0           

Flag	Coverage Δ
unittests	82.56% <100.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
sbi/inference/trainers/base.py	93.54% <100.00%> (ø)
sbi/inference/trainers/nle/mnle.py	95.65% <100.00%> (ø)
sbi/inference/trainers/nle/nle_a.py	100.00% <100.00%> (ø)
sbi/inference/trainers/nle/nle_base.py	97.82% <100.00%> (ø)
sbi/inference/trainers/npe/mnpe.py	95.65% <100.00%> (ø)
sbi/inference/trainers/npe/npe_a.py	63.91% <100.00%> (ø)
sbi/inference/trainers/npe/npe_b.py	100.00% <100.00%> (ø)
sbi/inference/trainers/npe/npe_base.py	94.77% <100.00%> (ø)
sbi/inference/trainers/npe/npe_c.py	73.33% <100.00%> (ø)
sbi/neural_nets/estimators/base.py	66.10% <100.00%> (ø)

[janfb]

Looks good, thanks! A couple of comments and questions remain.

[janfb]

some last comments, but almost done! 👍

[janfb]

Almost done, please just revert the changes done in sbi_types.py so that it in sync with main, i.e., does not show up in this PR anymore. thanks! 🙏

[janfb]

All good now, well done @abelaba 👏

and thanks again to @ARna06 for starting this!