fix: handle negative imaginary parts in number_to_string

[Sanjays2402]

Closes #551

number_to_string joined the rendered real and imaginary parts with a hardcoded +, so a negative imaginary part produced a malformed literal like 1.0+-1.0j, which complex() rejects — number_to_string(1-1j, significant_digits=1) raised ValueError, and so did any diff/hash comparison involving such a number. The separator is now omitted when the rendered imaginary part is already negative.

Four cases added to the existing test_number_to_string_complex_digits parametrize list in tests/test_helper.py; they fail on master and pass with the fix.

This change was prepared with AI assistance; the regression test was run locally and fails without the fix.

[seperman]

Static review: #616 fix: handle negative imaginary parts in number_to_string

Summary: Fixes the "1.0+-1.0j" malformed-literal ValueError from #551. In the complex branch of number_to_string, the + separator is now omitted when the rendered imaginary part already starts with -. The fix is correct, but the same fix is already on dev.

Overlap with other PRs

• #598 (merged into dev, also fixes #551) makes a functionally identical change. It computes real and imag separately and writes imag = imag if imag.startswith("-") else "+" + imag (helper.py:497-513 on dev). This PR's sign = '' if str(imag).startswith('-') else '+' (helper.py:511) gives the same output for every input. I found no other open PR for #551.
• This PR targets master, which doesn't have #598 yet. Against dev it conflicts in deepdiff/helper.py and in CHANGELOG.md, where dev already has its own Unreleased section. tests/test_helper.py merges cleanly.

Security

No security concerns found in static review. The change is a local string-formatting refactor plus test cases and a changelog line. There is no I/O, eval or dependency change, and the diff has no non-ASCII or hidden characters.

Correctness

• deepdiff/helper.py:499-514: the logic is correct. -0.0 imaginary parts are normalised by the recursive call's abs() for zero. nan/-inf imaginary parts also produce parseable strings.
• str(imag) at helper.py:511 does nothing, because imag is already the string returned by number_to_string.

Tests

• The four new cases (tests/test_helper.py:187-190) would fail on master. Checked by hand against dev's implementation, they should pass there already. Two of them cover inputs #598's tests don't: "e" notation (1.0e+0-1.0e+0j) and a negative real with a negative imaginary part (-1.0-1.0j). #598 only added "f"-notation cases with a positive real part.

Suggestions

• Since the code fix is already on dev via #598, the most useful remaining part is the extra test coverage. A small PR against dev that adds only the "e"-notation and negative-real cases to test_number_to_string_complex_digits would keep that coverage without conflicts.
• Please target dev for future PRs, and leave CHANGELOG.md edits to the maintainer at release time.

---

AI-assisted static review; the code was not executed.

[Sanjays2402]

Thanks for the review! You're right, the code fix is redundant with #598. I've reworked this into a test-only PR against dev: dropped the helper.py change and the CHANGELOG edit, and kept just the two cases #598 doesn't cover (the "e"-notation case and the negative-real with negative-imaginary case). Tests pass locally.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.07%. Comparing base (59acdd9) to head (43da3ef).
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@           Coverage Diff           @@
##              dev     #616   +/-   ##
=======================================
  Coverage   95.07%   95.07%           
=======================================
  Files          18       18           
  Lines        4918     4918           
=======================================
  Hits         4676     4676           
  Misses        242      242           

☔ 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.