feat(python): Infer time_unit in pl.duration when nanoseconds is specified

[cojmeister]

Closes #14751

• Made time_unit optional
• If nanoseconds is provided and time unit set to None, then time unit will be automatically set to nanoseconds
• If microseconds is provided (and nanoseconds isn't) and time_unit set to None, then time unit will be automatically set to microseconds
• If milliseconds is provided (and the other two aren't) and time_unit set to None, then time unit will be automatically set to milliseconds
• If none of the above are provided, and time_unit is not provided, it will be set to microseconds - 'us'

@stinodego I hope this is what you meant

[stinodego]

Thanks for the PR!

We don't want to infer ms time unit when a millisecond is specified. The logic for inference should be: if no time unit is specified, and if nanosecond is specified, use ns. Otherwise, use us.

Please also add a test that showcases this new functionality.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.05%. Comparing base (e204928) to head (ed60d83).
Report is 1 commits behind head on main.

❗ Current head ed60d83 differs from pull request most recent head d1ad911. Consider uploading reports for the commit d1ad911 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #14987      +/-   ##
==========================================
- Coverage   81.21%   81.05%   -0.17%     
==========================================
  Files        1348     1338      -10     
  Lines      175318   173832    -1486     
  Branches     2506     2458      -48     
==========================================
- Hits       142390   140892    -1498     
- Misses      32448    32473      +25     
+ Partials      480      467      -13     

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

[cojmeister]

Done with the changes

[stinodego]

Thanks!