Stabilize adaptive process-noise source weighting

[FlorianPfaff]

Summary

• normalize positive source weights after scaling by their largest value
• preserve the intended weighted NIS ratio when direct weighted sums would overflow
• validate source-weight scalars through the existing nonnegative-finite scalar path
• add a focused regression using maximum finite floating-point weights

Root cause

RollingNISProcessNoiseAdapter.ratio() accumulated raw weighted sums:

numerator += weight * ratio
denominator += weight

Individually finite source weights can still overflow both accumulators. For example, weights [max_float, max_float / 2] have a valid 2:1 ratio, but the direct calculation produces inf / inf, so the aggregate becomes NaN and downstream adaptive scaling rejects it.

Fix

Collect positive validated weights, divide them by their largest entry, normalize the scaled weights to unit mass, and compute the weighted ratio from those normalized weights. This preserves relative weighting without creating oversized intermediate sums.

Validation

• reproduced the pre-fix result as inf / inf -> NaN
• verified the patched 2:1 weighted ratio is 7/3
• verified the corresponding adaptive scale remains finite
• PYTHONPATH=. python -m pytest -q test_weight_stability.py (1 passed)
• Python compilation passed for the modified module and regression test
• branch is two commits ahead of main, zero behind; only the adapter and focused regression are changed

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	17.85s
✅ JSON	prettier	7	0	0	0	0.8s
✅ JSON	v8r	7		0	0	3.09s
✅ MARKDOWN	markdownlint	68	0	0	0	1.35s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	0.42s
✅ PYTHON	black	1490	183	0	0	68.89s
✅ PYTHON	isort	1490	328	0	0	2.5s
✅ REPOSITORY	betterleaks	yes		no	no	2.02s
✅ REPOSITORY	checkov	yes		no	no	42.36s
✅ REPOSITORY	gitleaks	yes		no	no	10.61s
✅ REPOSITORY	git_diff	yes		no	no	0.32s
✅ REPOSITORY	secretlint	yes		no	no	46.92s
✅ REPOSITORY	syft	yes		no	no	4.99s
✅ REPOSITORY	trivy-sbom	yes		no	no	4.91s
✅ REPOSITORY	trufflehog	yes		no	no	24.38s
✅ YAML	prettier	11	0	0	0	0.47s
✅ YAML	v8r	11		0	0	8.19s
✅ YAML	yamllint	11		0	0	0.49s

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