Implement SdCondSdGridDistribution for PyRecEst

[Copilot]

Ports the MATLAB SdCondSdGridDistribution class from libDirectional to Python. Represents a conditional distribution f(a|b) on Sd × Sd as a grid of values, where grid_values[i, j] = f(grid[i] | grid[j]).

Conventions

• grid: (n_points, d) — embedding-dimension coordinates (d=3 for S²)
• grid_values: (n_points, n_points) — conditional pdf values
• dim = 2 * d — Cartesian-product embedding dimension (libDirectional convention)

New class: SdCondSdGridDistribution

• __init__ — validates inputs; warns/raises if column-wise normalization check fails. Fixes a latent bug in the MATLAB original where the "maybe wrong order" branch could never trigger.
• normalize() — no-op returning self (matches MATLAB behavior).
• multiply(other) — element-wise grid multiplication with grid-compatibility guard.
• marginalize_out(1|2) — sums over rows or columns, returns a HypersphericalGridDistribution.
• fix_dim(1|2, point) — conditions on an existing grid point, returns a HypersphericalGridDistribution.
• from_function(fun, n, ...) — static factory supporting both a flat paired-call convention and a caller-side Cartesian-product convention.

from pyrecest.distributions.conditional import SdCondSdGridDistribution
import numpy as np

surface = 4 * np.pi  # S² surface area
def vmf_like(a, b):
    # a, b: (n_pairs, 3) — evaluate f(a|b) for each pair
    return np.exp(2.0 * np.sum(a * b, axis=1)) / surface

sc = SdCondSdGridDistribution.from_function(vmf_like, no_of_grid_points=100, dim=6)
# Fix the second argument at the first grid point
cond = sc.fix_dim(2, sc.grid[0])   # → HypersphericalGridDistribution
marg = sc.marginalize_out(1)        # → HypersphericalGridDistribution

Notes

• The get_grid_hypersphere import is deferred inside from_function to avoid a circular-import path through pyrecest.sampling.
• dim argument to from_function follows the Cartesian-product embedding convention (6 for S²×S²); internally converted to manifold dim via dim // 2 - 1.

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	7.74s
✅ JSON	prettier	2	0	0	0	0.45s
✅ JSON	v8r	2		0	0	2.56s
✅ MARKDOWN	markdownlint	1	0	0	0	0.65s
✅ MARKDOWN	markdown-table-formatter	1	0	0	0	0.27s
✅ PYTHON	bandit	249		0	0	3.16s
✅ PYTHON	black	249	1	0	0	4.39s
✅ PYTHON	flake8	249		0	0	1.71s
✅ PYTHON	isort	249	4	0	0	0.48s
✅ PYTHON	mypy	249		0	0	4.19s
✅ PYTHON	pylint	249		0	0	70.57s
✅ PYTHON	ruff	249	4	0	0	0.05s
✅ REPOSITORY	checkov	yes		no	no	22.48s
✅ REPOSITORY	gitleaks	yes		no	no	4.33s
✅ REPOSITORY	git_diff	yes		no	no	0.04s
✅ REPOSITORY	secretlint	yes		no	no	5.57s
✅ REPOSITORY	syft	yes		no	no	3.56s
✅ REPOSITORY	trivy-sbom	yes		no	no	1.7s
✅ REPOSITORY	trufflehog	yes		no	no	15.94s
✅ YAML	prettier	4	0	0	0	0.46s
✅ YAML	v8r	4		0	0	4.83s
✅ YAML	yamllint	4		0	0	0.48s

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@9.4.0 --custom-flavor-setup --custom-flavor-linters PYTHON_PYLINT,PYTHON_BLACK,PYTHON_FLAKE8,PYTHON_ISORT,PYTHON_BANDIT,PYTHON_MYPY,PYTHON_RUFF,COPYPASTE_JSCPD,JSON_V8R,JSON_PRETTIER,MARKDOWN_MARKDOWNLINT,MARKDOWN_MARKDOWN_TABLE_FORMATTER,REPOSITORY_CHECKOV,REPOSITORY_GIT_DIFF,REPOSITORY_GITLEAKS,REPOSITORY_SECRETLINT,REPOSITORY_SYFT,REPOSITORY_TRIVY_SBOM,REPOSITORY_TRUFFLEHOG,YAML_PRETTIER,YAML_YAMLLINT,YAML_V8R

Show us your support by starring ⭐ the repository

[FlorianPfaff]

@copilot Can you fix the linter errors by megalinter?

[Copilot]

@copilot Can you fix the linter errors by megalinter?

Fixed in 2efa63c. The MegaLinter failures were:

• pylint E1101 (no-member on pyrecest.backend.__backend_name__): added # pylint: disable=no-member to each occurrence, matching the pattern used in all other test files.
• pylint W0404 (reimported): removed the local from pyrecest.backend import ones as be_ones inside three test methods; they now use the top-level ones import.
• pylint W0611 / flake8 F401 (unused import): removed the unused zeros import from the test file.
• flake8/ruff F401 (imported but unused in conditional/__init__.py): added __all__ = ["SdCondSdGridDistribution"] to mark the re-export as intentional.