fix: guard degenerate ZStackLoop when calculating home index

[tlambert03]

Fixes #305

The bug

_calc_zstack_home_index divides by abs(high_um - low_um) with no guard. NIS-Elements writes a 1-plane ZStackLoop (uiCount=1, dZStep=0, dZLow == dZHigh) for single-plane acquisitions on systems with a piezo Z device configured, so .experiment, .metadata, .sizes and .asarray() all raise ZeroDivisionError on otherwise perfectly valid 2D files.

Note that _parse_z_stack_loop only recomputes a nonzero step when count > 1, so within nd2 the step_um <= 0 branch is essentially only reachable in the degenerate configurations (count <= 1, dZLow == dZHigh, or a negative dZStep) — it is unexercised by all 102 files in tests/data, which is why this went unnoticed.

Changes

1. count <= 1 returns 0 up front. Also stops the fallback branch returning -1 for count == 0.
2. Zero-width range in the step_um <= 0 branch returns 0 — every plane sits at the same z, so no index is more "home" than another.
3. Separate latent bug, same three lines: hrange = (inverted and home_range_i) or home_range_f doesn't do what it looks like. When inverted is True but home_range_i is exactly 0.0, the and yields falsy 0.0 and it falls through to the non-inverted range. Replaced with real ternaries (which is what limnd2 uses).

Tests

6 new tests in tests/test_parse.py: the issue's exact uLoopPars through _parse_z_stack_loop, the count > 1 zero-range route into the divide, and the inverted-with-zero-range case for both type pairs (2/3 and 6/7 — old code returned 4, correct is 0).

All 6 fail on main and pass here. Full suite green, and no golden homeIndex in tests/samples_metadata.json shifted, so the ternary change is a no-op on every real file in the corpus.

Caveat for review

The step_um <= 0 branch still has zero coverage from real files, so the values it returns for a non-degenerate loop (negative dZStep) remain unvalidated against the SDK. This PR makes the degenerate cases well-defined; it doesn't confirm the surviving arithmetic. @lanery's offer of a sample file for the test suite is still worth taking up.

🤖 Generated with Claude Code

[codspeed]

Merging this PR will not alter performance

✅ 13 untouched benchmarks

---

_{Comparing fix-degenerate-zstack-home-index (7b26c04) with main (a5b3caf)}

[codecov]

Codecov Report

❌ Patch coverage is 87.50000% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 93.70%. Comparing base (a5b3caf) to head (7b26c04).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/nd2/_parse/_parse.py	87.50%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #306      +/-   ##
==========================================
+ Coverage   93.65%   93.70%   +0.05%     
==========================================
  Files          22       22              
  Lines        2617     2622       +5     
==========================================
+ Hits         2451     2457       +6     
+ Misses        166      165       -1     

☔ View full report in Codecov by Harness.
📢 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.