fix: added path compression to UnionFind

[soondubu137]

Main Changes

The current implementation of the union find data structure claims its amortized time complexity is $O(\alpha(n))$, which is incorrect because its Find method has no optimization. I added path compression as the necessary optimization. Path splitting and path halving are also valid optimizations, but I think path compression is the most common one.

I changed union-by-size to union-by-rank, which complements path compression effectively.

In addition, I changed the receivers of Find and Union to pointers. Along with this, I removed the unnecessary return value in Union.

Other Changes

I fixed a small typo in the algorithm description and used more consistent variable names.

graph/kruskal.go uses UnionFind, so I also modified that file to conform to the updated Union method.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 87.68%. Comparing base (24c7f1f) to head (648e1b4).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #734      +/-   ##
==========================================
+ Coverage   87.59%   87.68%   +0.08%     
==========================================
  Files         208      209       +1     
  Lines        5524     5547      +23     
==========================================
+ Hits         4839     4864      +25     
+ Misses        543      542       -1     
+ Partials      142      141       -1     

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