feat: support weekend argument in business_day_count

[MarcoGorelli]

part II of upstreaming business days to Polars

The "grand plan" is:

• part III: add holidays argument to this function (this one may stir up a little more discussion)
• part IV: add add_business_day, which would also take weekend and holidays arguments
• part V: add is_business_day, which would also take weekend and holidays arguments (this one should be trivial)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.16%. Comparing base (33eedfb) to head (7635379).
Report is 4 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #15544      +/-   ##
==========================================
- Coverage   81.17%   81.16%   -0.02%     
==========================================
  Files        1367     1367              
  Lines      175313   175321       +8     
  Branches     2530     2530              
==========================================
- Hits       142314   142297      -17     
- Misses      32524    32550      +26     
+ Partials      475      474       -1     

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

[ritchie46]

Can we put this behind a cached function? Small things to keep import times down. ^^

[MarcoGorelli]

thanks - I've ended up making it much simpler: users just pass an iterable of booleans, and that gets passed to the Rust side as-is

slightly less ergonomic perhaps, but I think realistically people would just define a constant with their week mask and pass that in, and the code's much simpler like this, sorry for having overcomplicated it

[codspeed-hq]

CodSpeed Performance Report

Merging #15544 will not alter performance

_{Comparing MarcoGorelli:business-week-mask (7635379) with main (b91dedb)}

Summary

✅ 22 untouched benchmarks