refactor: replace isProbability with inline range check in stats/base/dists/bernoulli/mgf

[Planeshifter]

Description

Normalizes the probability-range validation in @stdlib/stats/base/dists/bernoulli/mgf (both lib/main.js and lib/factory.js) from !isProbability( p ) to the inline form isnan( p ) || p < 0.0 || p > 1.0 used by every other JS member of the stats/base/dists/bernoulli namespace, and removes the now-unused @stdlib/math/base/assert/is-probability require from both files.

Namespace summary

• Target: @stdlib/stats/base/dists/bernoulli
• Members analyzed: 13 (cdf, ctor, entropy, kurtosis, mean, median, mgf, mode, pmf, quantile, skewness, stdev, variance)
• Autogenerated members: 0
• Features analyzed: file tree; package.json top-level / scripts / stdlib key sets; manifest.json key set; README ##/### section sequence; test/ / benchmark/ / examples/ filenames; per-package lib/main.js and lib/factory.js public signature, validation prologue, return kind, error-construction style, JSDoc shape, and require dependency set.
• Features with a clear majority (≥75% conformance): full package.json top-level 18-key set (13/13 = 100%); manifest.json presence + 3-key set (12/13 = 92%); the native-bindings file group binding.gyp / include.gypi / src/{Makefile,addon.c,main.c} / lib/native.js / include/.../<name>.h / examples/c/** / benchmark/c/** / benchmark/benchmark.native.js / test/test.native.js (12/13 = 92%); README 5-section sequence ## Usage → ## Examples → ## C APIs → ### Usage → ### Examples (11/13 = 85%); test/fixtures/julia/REQUIRE + runner.jl (11/13 = 85%); validation prologue isnan( p ) || p < 0.0 || p > 1.0 in JS sources where p is checked (12/13 = 92%); value return kind (13/13 = 100%); JSDoc hasExample (13/13 = 100%).
• Features without a clear majority (excluded from drift detection): public signature shape (single-arg name(p) 8/13 vs 2-arg distribution-function packages 4/13 vs ctor 1/13 — no shape ≥75%); error-construction style (only ctor throws; the remaining 12 return NaN, so this is degenerate).

stats/base/dists/bernoulli/mgf

Replaces !isProbability( p ) with the inline isnan( p ) || p < 0.0 || p > 1.0 check in both lib/main.js and lib/factory.js, and drops the unused @stdlib/math/base/assert/is-probability import from each file. The inline form is what the other 12 bernoulli siblings use (92.3% conformance) and what 20 of the 22 stats/base/dists/*/mgf packages across the wider namespace use (90.9%). isProbability( x ) is defined as x >= 0.0 && x <= 1.0, so the two forms agree on NaN, ±Infinity, and all finite inputs; observable behavior is preserved.

Validation

Checked:

• Structural feature extraction (file tree; package.json shape; manifest.json shape; README ##/### section list; test/ / benchmark/ / examples/ filenames) across all 13 members.
• Semantic feature extraction (public signature; validation prologue; return kind; error-construction style; JSDoc shape; require dependency set) across all 13 members by reading each lib/main.js (and lib/factory.js where present) directly.
• Three-agent drift validation on the single surviving correction (bernoulli/mgf validation prologue):
  • Semantic-review confirmed !isProbability( p ) and isnan( p ) || p < 0.0 || p > 1.0 are mathematically equivalent for double-precision input (NaN, ±Infinity, and out-of-range values all map to the false branch of isProbability, which becomes the true branch of !isProbability, matching the disjunction).
  • Cross-reference confirmed test/test.mgf.js boundary cases (NaN, ±Infinity, 2.0, -0.5 for p) all return NaN under both forms, README.md and docs/repl.txt make no reference to isProbability, and no sibling package imports bernoulli/mgf directly.
  • Structural-review is not applicable (semantic-only feature).
• Open-PR check via the GitHub MCP — no open PR targets stats/base/dists/bernoulli or proposes this fix.
• Manual smoke test from a Node REPL against the same boundary inputs the existing test/test.mgf.js exercises; all return NaN as before.

Deliberately excluded:

• ctor — absence of the native-bindings file group, gypfile, manifest.json, and the canonical ## Usage → ## Examples README structure. Intentional: pure-JS class constructor with no native numerical hot path; the constructor's documentation legitimately requires intermediate property/method subsections.
• mode — extra ## Notes section between ## Usage and ## Examples. Intentional: the Bernoulli mode is undefined at p === 0.5; the note documents stdlib's 0 convention. Stripping the section would lose content.
• stdev — uses test/fixtures/python/ instead of test/fixtures/julia/, and delegates stdev( p ) = sqrt( variance( p ) ) rather than carrying the canonical validation prologue. Cascading / intentional: regenerating fixtures and updating the tape test loader is out of scope for a mechanical drift PR; the delegation idiom was the subject of merged PR refactor: reuse variance in stats/base/dists/bernoulli/stdev #11366 and is correct.
• C-side src/main.c for bernoulli/mgf only checks isnan( t ) || isnan( p ), missing the p < 0.0 || p > 1.0 range that cdf/src/main.c and pmf/src/main.c apply, and lib/native.js JSDoc actually documents the missing-check behaviour as expected. Not fixed here: changing observable native-path output and the documented examples is out of scope; surfacing as a separate behavioural follow-up.
• Public signature drift between single-arg property packages and 2-arg distribution-function packages — no signature shape reaches 75%.

Related Issues

This pull request has no related issues.

Questions

No.

Other

stats/base/dists/geometric/mgf carries the identical isProbability drift (the only other distribution-namespace mgf package using the predicate form) and is a clear follow-up candidate. Internal run metadata (random seed, full per-feature distributions, dropped corrections with reasons) lives in the local drift report at ~/drift-reports/drift-stats-base-dists-bernoulli-2026-05-26.md.

Checklist

• Read, understood, and followed the contributing guidelines.

AI Assistance

• Yes
• No

If you answered "yes" above, how did you use AI assistance?

• Code generation (e.g., when writing an implementation or fixing a bug)
• Test/benchmark generation
• Documentation (including examples)
• Research and understanding

Disclosure

This PR was authored by Claude Code via the cross-package drift-detection routine: structural and semantic features were extracted from every member of stats/base/dists/bernoulli, the majority pattern was computed at a 75% conformance threshold, and three independent validation agents (semantic-review, cross-reference, structural-review) confirmed the single surviving correction was high-signal mechanical drift before any file was edited. The edit replaces a predicate-based probability check with the namespace-standard inline check and removes the now-unused require; no test, fixture, or behavioural change.

---

@stdlib-js/reviewers

---

Generated by Claude Code

[stdlib-bot]

Coverage Report

Package	Statements	Branches	Functions	Lines
stats/base/dists/bernoulli/mgf	$\color{green}269/269$ $\color{green}+100.00%$	$\color{green}19/19$ $\color{green}+100.00%$	$\color{green}4/4$ $\color{green}+100.00%$	$\color{green}269/269$ $\color{green}+100.00%$

The above coverage report was generated for the changes in this PR.