Fix naturaldelta sub second precision

[wolf1986]

Fixes #154

Changes proposed in this pull request:

• Convert to float and not to int
• Explicitly drop support from 'inf' & 'nan' that are supported by float
• In tests, assert both float & timedelta variants are equivalent

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.19%. Comparing base (a28cb80) to head (c94f951).
Report is 9 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #169   +/-   ##
=======================================
  Coverage   99.19%   99.19%           
=======================================
  Files           9        9           
  Lines         746      748    +2     
=======================================
+ Hits          740      742    +2     
  Misses          6        6           

Flag	Coverage Δ
macos-latest	97.59% <100.00%> (+<0.01%)	⬆️
ubuntu-latest	97.59% <100.00%> (+<0.01%)	⬆️
windows-latest	96.12% <100.00%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[hugovk]

Thanks!