fix: random choose sample

[Millione]

#37

[codecov]

Codecov Report

Patch coverage: 90.90% and project coverage change: +0.39 🎉

Comparison is base (d1fc4e8) 83.43% compared to head (c93fe98) 83.83%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #42      +/-   ##
==========================================
+ Coverage   83.43%   83.83%   +0.39%     
==========================================
  Files          17       17              
  Lines        1920     1930      +10     
==========================================
+ Hits         1602     1618      +16     
+ Misses        318      312       -6     

Impacted Files	Coverage Δ
src/policy.rs	94.47% <90.90%> (-0.27%)	⬇️

... and 7 files with indirect coverage changes

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[al8n]

Thanks! But let's wait for @alfa07 and @phantomhker's feedback about if the ringbuffer stuff solves the hit ratio problem mentioned in #37 because using rand will bring extra overhead for performance.