Reset CLEAR continuity after unmatched frames

[FlorianPfaff]

Bug

evaluate_clear(...) keeps two identity caches with different semantics:

• previous_id tracks the last historical match for identity-switch counting;
• previous_timestep_id should only represent matches from the immediately preceding frame for CLEAR continuity assignment.

The immediate-timestep cache was cleared after ordinary matched frames, but not when a frame had no ground-truth detections or no tracker detections. A match from two or more frames ago could therefore receive the +1000 continuity bonus after a gap and override the current similarity-optimal assignment.

In the regression sequence, two tracks are matched diagonally, missed for one frame, and then have stronger crossed similarities. Current main incorrectly preserves the stale diagonal assignment, reports zero identity switches, and accumulates motp_sum=3.2. The correct post-gap assignment is crossed, with two historical identity switches and motp_sum=3.8.

Fix

Clear previous_timestep_id before returning early from either empty-side frame branch. The historical previous_id cache remains unchanged, so identity switches after a gap are still counted correctly.

Regression coverage

Add test_clear_continuity_does_not_cross_unmatched_frames, which verifies:

• four true positives, zero false positives, and two false negatives;
• two identity switches after the gap;
• similarity accumulation of 3.8, proving the post-gap Hungarian assignment follows current similarities rather than stale continuity.

Validation

• independently reproduced the pre-fix result (tp, fp, fn, switches, motp_sum) = (4, 0, 2, 0, 3.2);
• independently exercised the patched logic and obtained (4, 0, 2, 2, 3.8);
• branch comparison: 2 commits ahead, 0 behind main;
• final diff is limited to 2 production additions and a 20-line focused regression in the existing tracking-metrics test module.

GitHub Actions provides the authoritative full test, lint, packaging, documentation, and integration validation.

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	24.09s
✅ JSON	prettier	7	0	0	0	1.14s
✅ JSON	v8r	7		0	0	4.12s
✅ MARKDOWN	markdownlint	68	0	0	0	1.78s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	0.5s
✅ PYTHON	black	1865	87	0	0	89.45s
✅ PYTHON	isort	1865	139	0	0	2.74s
✅ REPOSITORY	betterleaks	yes		no	no	2.44s
✅ REPOSITORY	checkov	yes		no	no	46.5s
✅ REPOSITORY	gitleaks	yes		no	no	16.85s
✅ REPOSITORY	git_diff	yes		no	no	0.29s
✅ REPOSITORY	secretlint	yes		no	no	79.47s
✅ REPOSITORY	syft	yes		no	no	6.03s
✅ REPOSITORY	trivy-sbom	yes		no	no	4.83s
✅ REPOSITORY	trufflehog	yes		no	no	31.53s
✅ YAML	prettier	11	0	0	0	0.69s
✅ YAML	v8r	11		0	0	10.11s
✅ YAML	yamllint	11		0	0	0.54s

Notices

📣 MegaLinter 9.5.0 is out! Discover the new features and security recommendations in the release announcement. (Skip this info by defining SECURITY_SUGGESTIONS: false)

See detailed reports in MegaLinter artifacts

Your project could benefit from a custom flavor, which would allow you to run only the linters you need, and thus improve runtime performances. (Skip this info by defining FLAVOR_SUGGESTIONS: false)

• Documentation: Custom Flavors
• Command: npx mega-linter-runner@9.6.0 --custom-flavor-setup --custom-flavor-linters PYTHON_BLACK,PYTHON_ISORT,COPYPASTE_JSCPD,JSON_V8R,JSON_PRETTIER,MARKDOWN_MARKDOWNLINT,MARKDOWN_MARKDOWN_TABLE_FORMATTER,REPOSITORY_CHECKOV,REPOSITORY_GIT_DIFF,REPOSITORY_GITLEAKS,REPOSITORY_BETTERLEAKS,REPOSITORY_SECRETLINT,REPOSITORY_SYFT,REPOSITORY_TRIVY_SBOM,REPOSITORY_TRUFFLEHOG,YAML_PRETTIER,YAML_YAMLLINT,YAML_V8R

Show us your support by starring ⭐ the repository