chore: update keywords

[Planeshifter]

Description

What is the purpose of this pull request?

This pull request:

• Normalizes two datapackage.json outliers in @stdlib/datasets so their metadata matches the convention used by the other 55 sibling datasets. Two files change; no code, tests, examples, or benchmarks are touched.

Namespace summary

• Target namespace: @stdlib/datasets
• Member count: 57 (all non-autogenerated)
• Features analyzed: full file tree; package.json top-level / scripts / stdlib key sets; datapackage.json top-level shape, resources entry shape, keywords array, license identifier, sources shape; README.md ## / ### heading list; manifest.json presence; test/, benchmark/, examples/ file naming; bin/cli shebang and structure; etc/cli_opts.json shape; docs/repl.txt placeholder set; docs/usage.txt layout; docs/types/index.d.ts shape; lib/index.js re-export pattern; lib/main.js public signature, return kind, validation prologue, error construction, JSDoc shape, and require dependency set.
• Features with clear majority (≥75% conformance): package.json top-level key set (100%), directories value set (100% modulo the scripts sub-key which tracks whether a builder script exists), os / engines / contributors / author / homepage / bugs / repository / main / types (100%), file-tree entries LICENSE, README.md, benchmark/benchmark.js, bin/cli, datapackage.json, docs/repl.txt, docs/types/index.d.ts, docs/types/test.ts, docs/usage.txt, etc/cli_opts.json, examples/index.js, lib/index.js, lib/main.js, package.json, test/test.cli.js, test/test.js (100%), bin/cli shebang #!/usr/bin/env node (100%), README ## Usage / ## Examples / ## CLI / ## License sections (100%), lib/index.js require('./main.js') + module.exports = main pattern (100%), datapackage.json top-level key set {name, version, title, description, resources, sources, keywords, license} (100%), datapackage.json resource entry shape (100%), 40/41 = 98% of PDDL+CC0-licensed datapackages spell the SPDX identifier as PDDL-1.0 AND CC0-1.0, and 56/57 = 98% of datapackages omit the four package-manager-oriented keywords (stdlib, datasets, dataset, data) that belong in package.json rather than datapackage.json.
• Features without clear majority (excluded from drift detection): README optional sections ## Notes (32% at h2, 44% at h3), ## References (49%), ## See Also (47%) — all content-dependent (whether the dataset has bibliographic references or documented siblings). etc/cli_opts.json string key presence (30%) — depends on whether the CLI accepts a --format or similar option. directories.scripts sub-key (25%) — depends on whether the package carries a builder script under scripts/. Copyright year in lib/main.js header (82% 2018, 16% 2019, 2% 2020) — reflects when the package first landed, not drift. datapackage.json license field overall (four distinct values across the namespace) — depends on the underlying data license.

Per outlier package

@stdlib/datasets/cmudict

cmudict/datapackage.json was the sole outlier carrying the four package-manager-oriented keywords stdlib, datasets, dataset, data at the head of its keywords array — a set already present verbatim in cmudict/package.json and correctly omitted by the other 56 sibling datapackages (98% conformance). The remaining nine content-descriptive keywords (words, speech, language, nlp, spelling, pronounciation, dictionary, english, en) are preserved, and the LICENSE-derived "license": "BSD-2-Clause" identifier is untouched.

@stdlib/datasets/stopwords-en

stopwords-en/datapackage.json was the sole PDDL + CC0-licensed sibling using the non-canonical identifier CC0 AND PDDL-1.0 — 40 of the 41 packages sharing this license combination (98% conformance among that subset) use the SPDX-canonical PDDL-1.0 AND CC0-1.0. The prior value dropped the -1.0 version qualifier from CC0 and reversed term order. The LICENSE file explicitly names "Open Data Commons Public Domain Dedication & License 1.0 (PDDL 1.0)" for the database and "Creative Commons Zero v1.0 Universal (CC0 1.0)" for its contents, so the replacement is the SPDX-canonical form of the license the file already documents; AND is commutative in SPDX, so the change is semantically equivalent.

Validation

Checked:

• Structural feature extraction across all 57 members (file tree, package.json shape, datapackage.json top-level shape and resource entries, README section list, test/benchmark/example filenames, bin/cli shebang, etc/cli_opts.json shape, docs/repl.txt placeholder set, docs/types/index.d.ts shape).
• Semantic feature extraction across all 57 lib/main.js files (public signature, return kind, validation prologue, error construction, JSDoc shape, require dependencies). 54/57 datasets are no-argument functions that synchronously read data from disk and rethrow read errors; the 3 options-taking outliers (cmudict, minard-napoleons-march, sotu) have lib/validate.js, use format for error construction, and were ruled intentional (they filter multi-record datasets by user-supplied criteria).
• Adversarial verification of both proposed corrections by independent structural-review and cross-reference agents: confirmed both datapackage.json deviations are drift (not content-legitimate); confirmed no test, example, benchmark, sibling package, or repository-wide tool reads or asserts on datapackage.json keyword or license values; confirmed the pre-commit hook's package.json validator explicitly excludes datapackage.json; confirmed the only cross-references to datapackage.json are per-package scripts/*.js builders that populate their own resources field.
• Cross-run dedup: no open or recently-merged PR targets @stdlib/datasets with the same fixes (searched open PRs and PRs merged in the last 14 days).

Deliberately excluded:

• sotu missing lib/browser.js, benchmark/benchmark.browser.js, test/test.browser.js, test/test.main.js — legitimate architectural difference (multi-file speech corpus; browser support is instead provided via file-specific browser field mapping in its package.json).
• cmudict license: "Apache-2.0 and BSD" in package.json — content-legitimate (data is BSD-2-Clause per the LICENSE file; code is Apache-2.0).
• anscombes-quartet and herndon-venus-semidiameters carrying "string": [] in etc/cli_opts.json while all 40 sibling CLIs without string options omit the key — below the 75% majority threshold when framed at the whole-field granularity; not batched here because it's a cosmetic no-op rather than a metadata correctness fix.
• cmudict keyword typo pronounciation (correct spelling pronunciation) — a spelling fix, not majority-vote drift, and out of scope for a drift-detection run.
• Copyright-year variation in lib/main.js headers — reflects package creation date, not drift.
• Any change to public signatures, tests, benchmarks, examples, or observable behavior.

Related Issues

Does this pull request have any related issues?

No.

Questions

Any questions for reviewers of this pull request?

No.

Other

Any other information relevant to this pull request? This may include screenshots, references, and/or implementation notes.

Random namespace pick, seed 2026-07-11-drift-run. Eligibility criteria: ≥8 non-autogenerated direct child packages under lib/node_modules/@stdlib/ and its sub-namespaces (49 candidate namespaces qualified). The chosen namespace had 57 members, none autogenerated. Full per-feature drift report saved locally alongside the run.

Checklist

Please ensure the following tasks are completed before submitting this pull request.

• Read, understood, and followed the contributing guidelines.

AI Assistance

When authoring the changes proposed in this PR, did you use any kind of 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 running a cross-package drift-detection routine: the namespace (@stdlib/datasets) was picked at random (seed 2026-07-11-drift-run), structural and semantic features were extracted for every member, majority patterns were computed at a 75% conformance threshold, and independent structural-review and cross-reference verification agents confirmed each correction before it was applied. Only two datapackage.json files were changed — four keyword removals in cmudict and a one-line SPDX identifier normalization in stopwords-en.

---

@stdlib-js/reviewers

---

Generated by Claude Code

[stdlib-bot]

Coverage Report

Package	Statements	Branches	Functions	Lines
datasets/cmudict	$\\color{green}333/333$ $\\color{green}+100.00\\%$	$\\color{green}34/34$ $\\color{green}+100.00\\%$	$\\color{green}3/3$ $\\color{green}+100.00\\%$	$\\color{green}333/333$ $\\color{green}+100.00\\%$
datasets/stopwords-en	$\\color{green}145/145$ $\\color{green}+100.00\\%$	$\\color{green}7/7$ $\\color{green}+100.00\\%$	$\\color{green}2/2$ $\\color{green}+100.00\\%$	$\\color{green}145/145$ $\\color{green}+100.00\\%$

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