fix: use standard linear interpolation for Percentile

[montanaflynn]

Summary

• Replaces the non-standard percentile algorithm with the NIST Linear Interpolation Between Closest Ranks method
• Now matches Excel PERCENTILE, Google Sheets, NumPy (default), and other standard tools
• Adds test case from the issue to prevent regression

Algorithm

rank = (p / 100) * (n - 1)
result = data[k] + f * (data[k+1] - data[k])

Where k is the integer part and f is the fractional part of rank.

Reference: https://www.itl.nist.gov/div898/handbook/prc/section2/prc262.htm

Test plan

• All existing tests updated and passing
• Added TestPercentile_Issue92 with the exact values from the issue report
• Verified results match Excel, Google Sheets, and NumPy

Fixes #92

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (8743180) to head (5aaeda5).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##            master       #93   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           28        28           
  Lines          908       905    -3     
=========================================
- Hits           908       905    -3     

☔ 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.