Reject undefined axial Kalman mean updates

[FlorianPfaff]

Bug

AxialKalmanFilter.update_identity(...) always normalized the Euclidean posterior mean after the Kalman update.

For valid positive-definite state and measurement-noise covariances, the prior and innovation contributions can cancel exactly. The resulting axial mean is mathematically undefined, but floating-point arithmetic leaves a tiny residual. Normalizing that residual amplifies roundoff into an arbitrary unit direction.

A concrete two-dimensional case produces the raw mean

[-1.11022302e-15, -2.47677333e-16]

which current main normalizes to approximately

[-0.9760, -0.2177]

although no meaningful direction is defined.

Fix

• compute the posterior mean norm once;
• reject non-finite norms explicitly;
• reject zero or numerically negligible posterior means before normalization;
• leave the filter state unchanged when the update is undefined.

Regression coverage

Add a focused two-dimensional update with valid positive-definite covariance matrices that cancels the posterior mean. The regression verifies that the update raises a clear ValueError and preserves both the prior mean and covariance.

Validation

• reproduced the arbitrary pre-fix normalized direction in an isolated NumPy calculation;
• executed the exact patched module in a minimal backend-compatible harness;
• verified the undefined update is rejected and state mutation is atomic;
• syntax-compiled both changed Python files;
• passed a focused 88-column line-length check;
• branch is 3 commits ahead and 0 behind main.

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

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	15.59s
✅ JSON	prettier	7	0	0	0	0.59s
✅ JSON	v8r	7		0	0	3.03s
✅ MARKDOWN	markdownlint	68	0	0	0	1.14s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	0.39s
✅ PYTHON	black	1866	89	0	0	56.45s
✅ PYTHON	isort	1866	141	0	0	2.04s
✅ REPOSITORY	betterleaks	yes		no	no	1.53s
✅ REPOSITORY	checkov	yes		no	no	44.8s
✅ REPOSITORY	gitleaks	yes		no	no	10.98s
✅ REPOSITORY	git_diff	yes		no	no	0.16s
✅ REPOSITORY	secretlint	yes		no	no	45.76s
✅ REPOSITORY	syft	yes		no	no	3.67s
✅ REPOSITORY	trivy-sbom	yes		no	no	3.47s
✅ REPOSITORY	trufflehog	yes		no	no	16.76s
✅ YAML	prettier	11	0	0	0	0.5s
✅ YAML	v8r	11		0	0	7.58s
✅ YAML	yamllint	11		0	0	0.38s

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