Fix for rejection sampling

[michaeldeistler]

The transforms to unconstrained space used in MCMC broke the rejection sampler. We did not realize because there were no rejection sampling tests due to a bug. This PR...

• ...fixes the bug
• ...adds rejection sampling API tests

[michaeldeistler]

this change is independent from the rest of the PR. The test still followed the API from sbi<0.14.0

[codecov-commenter]

Codecov Report

Merging #561 (23f5db3) into main (2a10717) will decrease coverage by 0.02%.
The diff coverage is 66.66%.

@@            Coverage Diff             @@
##             main     #561      +/-   ##
==========================================
- Coverage   66.95%   66.92%   -0.03%     
==========================================
  Files          56       56              
  Lines        4185     4191       +6     
==========================================
+ Hits         2802     2805       +3     
- Misses       1383     1386       +3     

Flag	Coverage Δ
unittests	66.92% <66.66%> (-0.03%)	⬇️

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

Impacted Files	Coverage Δ
sbi/inference/posteriors/direct_posterior.py	66.42% <66.66%> (-0.24%)	⬇️
...inference/posteriors/likelihood_based_posterior.py	74.39% <66.66%> (-0.61%)	⬇️
sbi/inference/posteriors/ratio_based_posterior.py	80.00% <66.66%> (-0.65%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2a10717...23f5db3. Read the comment docs.

[janfb]

looks good!