Stabilize particle-scenario weight normalization

[FlorianPfaff]

Summary

• normalize particle-resampling scenario weights after scaling by their largest finite value
• preserve relative particle probabilities while avoiding overflow in the raw Python sum
• add an end-to-end TOML scenario regression using maximum finite floating-point weights

Bug

_normalized_particle_weights() summed validated finite weights directly:

weight_total = sum(weight_values)

Individually finite nonnegative weights can still have a non-finite sum. For example, [sys.float_info.max, sys.float_info.max / 2] has the well-defined ratio 2:1, but its direct sum is inf. Dividing by that total produces [0.0, 0.0], so the particle-resampling scenario passes an invalid probability vector to the backend sampler.

Fix

Scale all weights by their largest entry before summing. The scaled values lie in [0, 1], retain the original ratios, and normalize to a finite unit-mass vector. Existing validation for length, finiteness, non-negativity, and positive mass is preserved.

Validation

• isolated reproduction confirms the old formula returns [0.0, 0.0] for maximum-finite weights
• the scale-first formula returns [2/3, 1/3] with total mass 1.0
• added a public run_scenario() regression that parses the extreme weights from TOML and checks finite normalized metrics
• branch is two commits ahead of current main and zero behind; the production diff is 7 additions and 3 deletions

The full backend test matrix is delegated to GitHub Actions.

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	24.44s
✅ JSON	prettier	7	0	0	0	1.14s
✅ JSON	v8r	7		0	0	5.01s
✅ MARKDOWN	markdownlint	68	0	0	0	1.85s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	0.54s
✅ PYTHON	black	1490	182	0	0	93.47s
✅ PYTHON	isort	1490	327	0	0	2.61s
✅ REPOSITORY	betterleaks	yes		no	no	2.23s
✅ REPOSITORY	checkov	yes		no	no	53.16s
✅ REPOSITORY	gitleaks	yes		no	no	14.85s
✅ REPOSITORY	git_diff	yes		no	no	0.38s
✅ REPOSITORY	secretlint	yes		no	no	59.67s
✅ REPOSITORY	syft	yes		no	no	5.5s
✅ REPOSITORY	trivy-sbom	yes		no	no	7.48s
✅ REPOSITORY	trufflehog	yes		no	no	32.52s
✅ YAML	prettier	11	0	0	0	0.71s
✅ YAML	v8r	11		0	0	12.37s
✅ YAML	yamllint	11		0	0	0.41s

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