feat: Allow specifying Duration time unit

[Nardi Lam (nardi)]

Motivation

I tried to use the library, but our data model requires Durations with a specific time unit. I saw this was implemented for Datetime but not for Duration, so I implemented it in an analogous way.

Changes

1. Added time_unit attribute to dataframely.columns.datetime.Duration.
2. Use time_unit when converting to Polars and Arrow data types.
3. Use time_unit when generating samples.
4. Add test for conversions to pyarrow schema, analogous to Datetime.

[Copilot]

Pull request overview

This PR extends Dataframely’s Duration column type to support specifying a Polars/Arrow duration time_unit, analogous to the existing Datetime support, so schemas can precisely express and export duration precision.

Changes:

• Add a time_unit attribute to dy.Duration and propagate it into Polars dtype and PyArrow dtype conversions.
• Thread time_unit through duration sample generation so sampled series can be created with the requested duration unit.
• Add a PyArrow schema test to confirm Duration(time_unit=...) exports as duration[<unit>].

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
dataframely/columns/datetime.py	Adds time_unit support to the Duration column and uses it for Polars/Arrow dtype + sampling.
dataframely/random.py	Extends Generator.sample_duration with a time_unit argument and casts sampled values accordingly.
tests/columns/test_pyarrow.py	Adds coverage for Duration(time_unit=...) conversion to PyArrow schema.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (6e06c73) to head (ff3a9e9).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #301   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           56        56           
  Lines         3217      3218    +1     
=========================================
+ Hits          3217      3218    +1     

☔ 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.

[Oliver Borchert (borchero)]

Lgtm, thanks!