Reject nonsymmetric candidate mixture covariances

[FlorianPfaff]

Summary

• Reject materially nonsymmetric covariance matrices in GaussianMixtureMeasurementFactor before Cholesky factorization.
• Preserve tolerance for floating-point-scale skew (relative asymmetry up to 1e-12) and symmetrize only that residual numerical noise.
• Add regressions for shared covariances, per-component covariances, and tolerated round-off asymmetry.

Bug

GaussianMixtureMeasurementFactor.__post_init__() currently symmetrizes every covariance before checking positive definiteness. An invalid covariance such as

np.array([[2.0, 1.0], [0.0, 2.0]])

is therefore silently replaced by [[2.0, 0.5], [0.5, 2.0]], after which Cholesky succeeds. This changes the caller's measurement-noise model and can alter Mahalanobis distances, component responsibilities, evidence, and moment-matched outputs instead of rejecting invalid input.

Fix

Compare each covariance against its transpose with a scale-aware 1e-12 relative-asymmetry threshold before _symmetrize(). Material asymmetry now raises ValueError("covariances must be symmetric"); only round-off-scale skew is retained for the existing symmetrization step.

Validation

• Added focused regressions for both broadcast/shared and per-component covariance inputs.
• Added a regression confirming round-off-scale asymmetry remains accepted and symmetrized.
• Branch is based on fc15bed910b324cd0619a2f2b71abc887bc09747 and is 2 commits ahead / 0 behind that base.
• Diff is limited to 11 production additions and one 48-line regression file.
• Full repository test/backend validation is delegated to GitHub Actions because the execution environment cannot clone the repository over the network and does not have PyRecEst installed.

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Max errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no		no	5.91s
✅ JSON	prettier	7	0	0		0	0.49s
✅ JSON	v8r	7		0		0	3.57s
✅ MARKDOWN	markdownlint	68	0	0		0	1.82s
✅ MARKDOWN	markdown-table-formatter	68	0	0		0	0.27s
✅ PYTHON	black	1984	10	0		0	46.34s
✅ PYTHON	isort	1984	20	0		0	2.15s
✅ REPOSITORY	betterleaks	yes		no		no	0.95s
✅ REPOSITORY	checkov	yes		no		no	31.28s
✅ REPOSITORY	git_diff	yes		no		no	0.09s
✅ REPOSITORY	secretlint	yes		no		no	77.14s
✅ REPOSITORY	syft	yes		no		no	2.21s
✅ REPOSITORY	trivy-sbom	yes		no		no	1.7s
✅ YAML	prettier	11	0	0		0	0.42s
✅ YAML	v8r	11		0		0	8.51s
✅ YAML	yamllint	11		0		0	0.71s

Notices

⚠️ Your configuration references items that have been removed from MegaLinter and are ignored: MAKEFILE_CHECKMAKE. See Removed linters to find their replacements.

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@10.0.0 --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_BETTERLEAKS,REPOSITORY_SECRETLINT,REPOSITORY_SYFT,REPOSITORY_TRIVY_SBOM,YAML_PRETTIER,YAML_YAMLLINT,YAML_V8R

Show us your support by starring ⭐ the repository