Add loo_i function for single observation PSIS-LOO-CV computation

[jordandeklerk]

This adds the loo_i() function to compute PSIS-LOO-CV for a single observation.

---

Resolves #181

---

📚 Documentation preview 📚: https://arviz-stats--182.org.readthedocs.build/en/182/

[codecov-commenter]

Codecov Report

❌ Patch coverage is 75.55556% with 11 lines in your changes missing coverage. Please review.
✅ Project coverage is 70.54%. Comparing base (a40917b) to head (5fba784).
⚠️ Report is 115 commits behind head on main.

Files with missing lines	Patch %	Lines
src/arviz_stats/loo/helper_loo.py	66.66%	8 Missing ⚠️
src/arviz_stats/loo/loo.py	85.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #182      +/-   ##
==========================================
+ Coverage   62.64%   70.54%   +7.89%     
==========================================
  Files          14       39      +25     
  Lines        1925     4345    +2420     
==========================================
+ Hits         1206     3065    +1859     
- Misses        719     1280     +561     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[aloctavodia]

You checked that i is integer

[aloctavodia]

Maybe use idx?

[jordandeklerk]

This is a good point. I initially thought that this function would just be for quick testing mainly with custom log-likelihood functions so the actual indexing wouldn't be relevant. But I think maybe we should account for more calculated options here for the user.

The most recent commit allows for several different ways of doing this. Maybe this is overkill though?

[jordandeklerk]

Most recent change simplifies the way you can select an observation so we're not trying to infer many different ways of doing this.