Fix accumulated NumPy and PyTorch CI regressions

[FlorianPfaff]

Summary

This PR repairs the accumulated backend regressions that left the repository's main Test workflow red. It keeps the existing strict NumPy/PyTorch/JAX matrix and fixes the implementation or regression contract behind each failure rather than weakening CI.

Root causes and fixes

NumPy and shared distribution contracts

• Restore signed one-dimensional circular/hypertoroidal integration while retaining increasing-bound validation for dimensions greater than one.
• Accept the generic singleton-vector mode representation in VonMisesDistribution.set_mode().
• Replace an over-tight coordinate assertion on a deliberately coarse hyperhemispherical grid with a direct 30-degree measurement-alignment tolerance.

PyTorch backend contracts

• Match SciPy's fftconvolve semantics for Python boolean axes while continuing to reject NumPy boolean scalar axes.
• Accept NumPy and PyTorch zero-dimensional integer axis values in the patched flip helper.
• Accept backend integer scalar orders in WrappedNormalDistribution.trigonometric_moment() while rejecting boolean and non-integral values.
• Compute the quaternion rotation matrix used by SE(3) plotting directly with backend operations instead of calling the unavailable PyTorch spatial.Rotation adapter.
• Compare wrapped-Cauchy and wrapped-Laplace conjugate moments through backend conj/allclose, avoiding NumPy conversion of PyTorch conjugate views and backend-specific tensor methods.
• Let RTS control-sequence normalization fall back to per-entry handling when PyTorch cannot coerce a mixed [None, tensor] sequence.

Regression coverage

• signed scalar and reversed multidimensional integration boundaries
• circular particle-filter update with singleton-vector von Mises mode
• coarse-grid vMF measurement alignment
• SciPy-compatible Python boolean fftconvolve axes and rejected NumPy boolean axes
• NumPy/PyTorch scalar and vector flip axes
• backend scalar integer wrapped-normal moment orders
• backend-independent SE(3) body-axis plotting
• negative wrapped-Cauchy and wrapped-Laplace moment identities
• per-step optional RTS control inputs

Validation

• The diagnostic PyTorch suite before the final patch reported exactly 8 failures, 5,120 passes, and 706 skips.
• NumPy 3.14, NumPy 3.11-3.13, and JAX 3.11-3.13 passed on the diagnostic branch after the earlier fixes.
• The complete repository workflow matrix is running on this updated branch.

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	17.88s
✅ JSON	prettier	7	0	0	0	1.08s
✅ JSON	v8r	7		0	0	3.46s
✅ MARKDOWN	markdownlint	68	0	0	0	1.45s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	0.61s
✅ PYTHON	black	1489	182	0	0	70.6s
✅ PYTHON	isort	1489	327	0	0	1.94s
✅ REPOSITORY	betterleaks	yes		no	no	2.01s
✅ REPOSITORY	checkov	yes		no	no	48.79s
✅ REPOSITORY	gitleaks	yes		no	no	11.18s
✅ REPOSITORY	git_diff	yes		no	no	0.29s
✅ REPOSITORY	secretlint	yes		no	no	49.31s
✅ REPOSITORY	syft	yes		no	no	4.33s
✅ REPOSITORY	trivy-sbom	yes		no	no	4.04s
✅ REPOSITORY	trufflehog	yes		no	no	25.35s
✅ YAML	prettier	11	0	0	0	0.5s
✅ YAML	v8r	11		0	0	8.7s
✅ YAML	yamllint	11		0	0	0.46s

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