Improve concordance computation and avoid compiled code

[vandenman]

• concordance no longer computed via C++ and drops Rcpp as a direct dependency
• uses an $\mathcal{O}(N\log N)$ algorithm instead of $\mathcal{O}(N^2)$.

See https://refined-github-html-preview.kidonng.workers.dev/vandenman/jaspRegression/raw/8c7f3fae4744ef4aae7a353b0da9e50f18d6d8ef/benchmarks/concordance.html a benchmark.

There are tests, though the quarto benchmark is definitely more extensive.

Inspired by some chats with llms.

@RensDofferhoff I'm not sure who should review this 🤷‍♂️

[codecov-commenter]

Codecov Report

❌ Patch coverage is 97.93814% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 75.94%. Comparing base (ec8a0a0) to head (a0eae8e).

Files with missing lines	Patch %	Lines
R/concordance.R	97.89%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #426      +/-   ##
==========================================
+ Coverage   75.71%   75.94%   +0.22%     
==========================================
  Files          21       21              
  Lines        8080     8163      +83     
==========================================
+ Hits         6118     6199      +81     
- Misses       1962     1964       +2     

☔ 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.

[vandenman]

See https://refined-github-html-preview.kidonng.workers.dev/vandenman/jaspRegression/raw/ce6b51f10f19768fe919da77b45850454b51636b/benchmarks/concordance.html for the new results. Basically the C++ Fenwick approach is always faster than any alternative once N ~ 20.