Add mean rank and geometric mean rank

[cthoyt]

Summary

This PR follows the style of the implementation of the mean reciprocal rank to add an implementation of the rank helper function, the mean rank, and the geometric mean rank (originally described here).

• [x] Code passes all tests
• [x] Unit tests provided for these changes
• [x] Documentation and docstrings added for these changes

Changes

• Implement rank() helper function similar to reciprocal_rank() helper function
• Implement mean_rank() function, corresponding to the commonly used mean rank (MR) metric in KGEM evaluation
• Implement the gmean_rank() function, corresponding to recently introduced geometric mean rank (GMR) metric useful for KGEM evaluation

Questions

• It would be possible to re-implement the reciprocal_rank() function simply with return 1.0 / rank(relevant_item, recommendation). I'm not sure if there was a specific reason why MRR was implemented but not MR, but the logic is heavily related so it might even be more instructive to have one implemented in terms of the other

[codecov-commenter]

Codecov Report

Merging #23 (5a4b281) into main (6d166ff) will increase coverage by 0.00%. The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main      #23   +/-   ##
=======================================
  Coverage   99.86%   99.87%           
=======================================
  Files          15       15           
  Lines         754      787   +33     
=======================================
+ Hits          753      786   +33     
  Misses          1        1           

Impacted Files	Coverage Δ
rexmex/metrics/ranking.py	100.00% <100.00%> (ø)
tests/unit/test_metrics.py	100.00% <100.00%> (ø)

---

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 6d166ff...5a4b281. Read the comment docs.

[benedekrozemberczki]

Looks great @cthoyt! Thank you - we will add you to the bib.