Stabilize SO(3) chordal weight normalization

[FlorianPfaff]

Bug

SO3ChordalMeanSmoother._normalize_weight_vector() summed finite nonnegative weights at their original scale. For weights near the active floating-point maximum, the sum could overflow to infinity. Dividing by that total then collapsed every normalized weight to zero, so the projected chordal mean could become an unrelated rotation.

For example, weights proportional to 3:1 at the float64 limit produced the identity rotation instead of the same weighted mean as ordinary [3.0, 1.0] weights.

The same risk applied when finite sample and kernel weights were multiplied during local smoothing.

Fix

• scale every weight vector by its largest positive entry before summation
• normalize the bounded scaled weights when a probability vector is needed
• retain the bounded relative weights for kernel/sample weighting so their products cannot overflow
• preserve existing validation for length, finiteness, nonnegativity, and positive mass

Validation

• reproduced overflow of the old raw weight sum and the resulting identity rotation
• verified scaled normalization yields [0.75, 0.25] for the extreme finite 3:1 case
• added a backend-aware regression comparing extreme weights with ordinary [3.0, 1.0] weights
• branch is 2 commits ahead of main, 0 behind, and changes only the smoother plus the focused regression

Full repository CI is requested by this pull request.

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	22.5s
✅ JSON	prettier	7	0	0	0	0.72s
✅ JSON	v8r	7		0	0	4.52s
✅ MARKDOWN	markdownlint	68	0	0	0	1.96s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	0.99s
✅ PYTHON	black	1543	198	0	0	90.39s
✅ PYTHON	isort	1543	368	0	0	2.58s
✅ REPOSITORY	betterleaks	yes		no	no	2.94s
✅ REPOSITORY	checkov	yes		no	no	52.51s
✅ REPOSITORY	gitleaks	yes		no	no	14.3s
✅ REPOSITORY	git_diff	yes		no	no	0.6s
✅ REPOSITORY	secretlint	yes		no	no	61.44s
✅ REPOSITORY	syft	yes		no	no	5.95s
✅ REPOSITORY	trivy-sbom	yes		no	no	6.28s
✅ REPOSITORY	trufflehog	yes		no	no	31.25s
✅ YAML	prettier	11	0	0	0	0.78s
✅ YAML	v8r	11		0	0	11.45s
✅ YAML	yamllint	11		0	0	0.77s

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