fix(rust): Return correct Struct dtype from qcut when "include_breaks=True" on empty series

[kimjune01]

Fixes #27284.

When a Series is empty (length 0), qcut with include_breaks=True hits the
early-return path at null_count() == len() (0 == 0) and returns a Categorical
column instead of the expected Struct with breakpoint + category fields.

This returns a properly-typed empty StructChunked with Float64 breakpoint and
Categorical category fields when include_breaks is set in the early-return path.

Tests cover empty series, empty series in lazy context, and all-null series.

[codecov]

Codecov Report

❌ Patch coverage is 93.33333% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 80.55%. Comparing base (a84168f) to head (61d22ef).
⚠️ Report is 91 commits behind head on main.

Files with missing lines	Patch %	Lines
crates/polars-ops/src/series/ops/cut.rs	93.33%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #27561      +/-   ##
==========================================
- Coverage   81.44%   80.55%   -0.90%     
==========================================
  Files        1837     1841       +4     
  Lines      255199   255962     +763     
  Branches     3179     3180       +1     
==========================================
- Hits       207857   206184    -1673     
- Misses      46518    48950    +2432     
- Partials      824      828       +4     

☔ 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.

[orlp]

Please take a look at https://docs.pola.rs/development/contributing/#first-time-contributions.

[orlp]

Our convention is to name the test test_name_27284.

[mkzung]

Hi @kimjune01, while fixing two other qcut panics on degenerate float input (all-NaN and infinite), I found they sit in the same early-return block your PR touches, and a correct all-NaN fix also needs the struct-aware return you added here. I've opened #28220, which includes your empty-series fix (credited) alongside the two extra cases, since they can't cleanly be separated. Didn't want to step on your work; very happy to defer, combine, or adjust however you and the maintainers prefer.

[kimjune01]

I'm happy to have this PR closed and fold it into yours @mkzung

[orlp]

Superseded by #28220.