Reset CLEAR continuity across empty frames

[FlorianPfaff]

Bug

evaluate_clear(...) retained previous_timestep_id when a frame had no ground-truth detections or no tracker detections.

The next non-empty frame therefore received the 1000-point continuity bonus from the last non-empty frame rather than from the immediately preceding timestep. After a detection gap, this could force a lower-similarity assignment, suppress legitimate ID switches, and reduce the accumulated MOTP similarity.

Fix

Clear the immediate-timestep association state before continuing from either empty-frame branch. The separate previous_id state is intentionally preserved, so ID switches are still scored against the last historical match after the object reappears.

Regression coverage

The new focused regression covers both gap forms:

• no ground-truth detections with a tracker false positive;
• ground-truth detections with no tracker detections.

In both cases, the post-gap frame must optimize by similarity without stale continuity, while still recording the resulting historical ID switches.

Validation

• isolated reproduction: current behavior chooses stale same-ID assignments with 0 switches and MOTP sum 3.2;
• patched behavior chooses the higher-similarity crossed assignments with 2 switches and MOTP sum 3.8;
• branch comparison: 2 commits ahead, 0 behind main;
• final diff: 2 production additions and one focused regression module.

GitHub Actions is the authoritative full repository validation.

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	24.98s
✅ JSON	prettier	7	0	0	0	0.91s
✅ JSON	v8r	7		0	0	4.17s
✅ MARKDOWN	markdownlint	68	0	0	0	1.67s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	0.68s
✅ PYTHON	black	1866	87	0	0	91.22s
✅ PYTHON	isort	1866	139	0	0	2.35s
✅ REPOSITORY	betterleaks	yes		no	no	2.77s
✅ REPOSITORY	checkov	yes		no	no	50.1s
✅ REPOSITORY	gitleaks	yes		no	no	17.17s
✅ REPOSITORY	git_diff	yes		no	no	0.26s
✅ REPOSITORY	secretlint	yes		no	no	81.94s
✅ REPOSITORY	syft	yes		no	no	6.18s
✅ REPOSITORY	trivy-sbom	yes		no	no	5.87s
✅ REPOSITORY	trufflehog	yes		no	no	31.71s
✅ YAML	prettier	11	0	0	0	0.62s
✅ YAML	v8r	11		0	0	11.24s
✅ YAML	yamllint	11		0	0	0.48s

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

[FlorianPfaff]

Closing as superseded by #5237. Both PRs apply the exact same production change to reset previous_timestep_id across empty frames; #5237 is the newer replacement with the regression integrated into the existing tracking-metrics test module.