Validate hypertoroidal Fourier coefficient counts

[FlorianPfaff]

Bug

HypertoroidalFourierFilter.__init__(...) normalized n_coefficients with an isinstance(..., int) special case followed by tuple(int(n) for n in n_coefficients).

That produced several inconsistent or silent reinterpretations:

• a tuple such as (11.5,) was silently truncated to (11,);
• True was accepted as a one-coefficient filter because bool is an int subclass;
• the string "11" was interpreted as a two-dimensional (1, 1) coefficient shape;
• an empty tuple reached scalar allocation and could be reinterpreted downstream as a one-dimensional coefficient array;
• a valid NumPy integer scalar such as np.int64(11) failed because it was neither a Python int nor iterable.

These behaviors violate the documented integer coefficient-count contract and can construct a filter with a different dimensionality or resolution than the caller requested.

Fix

Reuse the established _normalize_coefficient_shape(...) contract from HypertoroidalFourierDistribution with require_odd=True before allocating the initial coefficient tensor.

The filter now:

• accepts Python and NumPy integer scalars;
• requires a nonempty sequence of positive integers;
• rejects Boolean, textual, fractional, zero, negative, and even coefficient counts;
• reports errors against the n_coefficients parameter instead of relying on backend allocation or later distribution validation.

Regression coverage

Add focused constructor tests covering:

• acceptance of np.int64(11);
• rejection of values previously reinterpreted silently;
• rejection of nonpositive and even coefficient counts in scalar and multidimensional forms.

Validation

• reconstructed the pre-patch source and verified its Git blob SHA exactly matches a198df740e7f05765f29e9753db2bf0fa61bb21a;
• syntax-compiled the modified implementation and regression module;
• ran an isolated contract harness reproducing the old reinterpretations and verifying the shared validator's corrected behavior;
• branch comparison: 2 commits ahead and 0 behind main;
• final diff is limited to 6 additions / 3 deletions in the constructor and one 39-line regression module.

GitHub Actions is authoritative for the full backend, lint, packaging, documentation, and integration matrix.

[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	0.97s
✅ JSON	v8r	7		0	0	4.62s
✅ MARKDOWN	markdownlint	68	0	0	0	1.82s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	1.14s
✅ PYTHON	black	1866	87	0	0	93.46s
✅ PYTHON	isort	1866	140	0	0	2.36s
✅ REPOSITORY	betterleaks	yes		no	no	2.79s
✅ REPOSITORY	checkov	yes		no	no	53.72s
✅ REPOSITORY	gitleaks	yes		no	no	17.38s
✅ REPOSITORY	git_diff	yes		no	no	0.19s
✅ REPOSITORY	secretlint	yes		no	no	84.06s
✅ REPOSITORY	syft	yes		no	no	5.99s
✅ REPOSITORY	trivy-sbom	yes		no	no	6.11s
✅ REPOSITORY	trufflehog	yes		no	no	32.21s
✅ YAML	prettier	11	0	0	0	0.68s
✅ YAML	v8r	11		0	0	11.54s
✅ YAML	yamllint	11		0	0	0.54s

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