Normalize positional PyTorch FFT dimensions

[FlorianPfaff]

Bug

The PyTorch FFT compatibility wrappers normalized NumPy-style dimension values only when dim, axis, or axes was supplied by keyword. When the equivalent argument was passed positionally, the wrapper forwarded NumPy scalar or array values directly to Torch.

This made otherwise valid calls fail, for example:

• rfft(values, None, np.array(0))
• fftshift(values, np.array([0]))
• fftn(values, None, np.array([0]))

Torch rejects those positional NumPy values even though the corresponding NumPy FFT calls accept them. Positional empty dimensions also bypassed the wrapper's existing no-op handling for FFT shifts.

Fix

• record the positional index of the Torch dim argument for each wrapper;
• normalize positional and keyword dimensions through the same scalar/sequence helpers;
• inspect the effective positional-or-keyword dimension when applying empty-axis no-op behavior;
• preserve shape normalization, aliases, Boolean-length validation, and ordinary Python integer behavior.

The fix covers rfft, irfft, fftshift, ifftshift, fftn, and ifftn.

Validation

• reproduced the pre-fix Torch TypeError for positional NumPy scalar and array dimensions;
• isolated patched harness passed for all six wrappers against NumPy reference results;
• covered positional np.array(0), np.int64(0), NumPy axis arrays, sequences containing NumPy scalar arrays, and empty positional axes;
• syntax compilation passed for the modified wrapper module;
• branch is 2 commits ahead and 0 behind main, with exactly two changed files.

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	25.61s
✅ JSON	prettier	7	0	0	0	0.66s
✅ JSON	v8r	7		0	0	6.54s
✅ MARKDOWN	markdownlint	68	0	0	0	1.4s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	0.55s
✅ PYTHON	black	1810	56	0	0	85.68s
✅ PYTHON	isort	1810	88	0	0	2.23s
✅ REPOSITORY	betterleaks	yes		no	no	2.86s
✅ REPOSITORY	checkov	yes		no	no	48.96s
✅ REPOSITORY	gitleaks	yes		no	no	15.33s
✅ REPOSITORY	git_diff	yes		no	no	0.22s
✅ REPOSITORY	secretlint	yes		no	no	71.14s
✅ REPOSITORY	syft	yes		no	no	5.89s
✅ REPOSITORY	trivy-sbom	yes		no	no	5.79s
✅ REPOSITORY	trufflehog	yes		no	no	30.02s
✅ YAML	prettier	11	0	0	0	0.6s
✅ YAML	v8r	11		0	0	11.48s
✅ YAML	yamllint	11		0	0	0.36s

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