chore: add missing keywords in blas/ext/base/ndarray cusum packages

[Planeshifter]

Description

Adds the missing "blas" and "extended" keywords to twelve cumulative-sum packages under @stdlib/blas/ext/base/ndarray/. The keywords are inserted after "math" and before "cumulative" in package.json.keywords, matching the canonical ordering observed in conforming siblings.

Namespace summary

• Target: @stdlib/blas/ext/base/ndarray
• Members analyzed: 87 (all non-autogenerated)
• Features extracted: file tree, package.json shape (top-level keys, scripts, stdlib, keywords), README sections (H2/H3 ordered), manifest.json shape, test/benchmark/example file naming, lib/main.js and lib/index.js JSDoc shape (@param, @returns, @throws, @example, ## Notes blocks), function signatures, and require() dependency sets, plus docs/types/index.d.ts and docs/repl.txt section presence
• Features with clear majority (≥75% conformance): file tree (100% on 10 universal files; the 6-package native-binding cluster is a known intentional difference), package.json top-level keys (100%), scripts keys (100%, all {}), test/benchmark/example file naming (100% on test.js / benchmark.js / index.js), README H2 sections Usage (100%) and Examples (100%), JSDoc @param/@returns/@example/## Notes (100%), function name = directory camelCase (100%), four ndarray base accessor require()s (100%), and package.json keywords stdlib (100%), ndarray (100%), blas (86%), extended (86%)
• Features without clear majority (excluded): references README section (31%), stdmath/mathematics/math keywords (99%, but the lone outlier gjoin-between is a string operation where their absence is intentional and matches its strided sibling), and statistics/stats/sum/total/summation keywords (55%, domain-specific and tracks the operation)

Outlier corrections

@stdlib/blas/ext/base/ndarray/dcusum

Add "blas" and "extended" keywords to package.json (86% conformance: 75/87 blas/ext/base/ndarray siblings carry both). The strided sibling @stdlib/blas/ext/base/dcusum and same-family ndarray wrappers (dsum, scusumkbn) carry these keywords; this package was missing them.

@stdlib/blas/ext/base/ndarray/dcusumkbn

Added "blas" and "extended" keywords to package.json. 75 of 87 blas/ext/base/ndarray siblings (86%) carry both keywords; the intra-family inconsistency is direct: scusumkbn and scusumkbn2 ndarray wrappers and the strided sibling @stdlib/blas/ext/base/dcusumkbn all carry them, while this f64 ndarray counterpart did not.

@stdlib/blas/ext/base/ndarray/dcusumkbn2

Adds "blas" and "extended" keywords to package.json, inserted after "math" and before "cumulative". Keyword conformance rises to 86% (75/87), matching the pattern established by the scusumkbn2 ndarray wrapper and the strided sibling @stdlib/blas/ext/base/dcusumkbn2, both of which carry these keywords.

@stdlib/blas/ext/base/ndarray/dcusumors

Adds "blas" and "extended" keywords to package.json, aligning with the keyword schema present in both the strided sibling @stdlib/blas/ext/base/dcusumors and the parallel non-cumulative @stdlib/blas/ext/base/ndarray/dsumors. This package was among the non-conforming entries driving the keyword conformance rate to 86% (75/87) across the ndarray BLAS extension suite.

@stdlib/blas/ext/base/ndarray/dcusumpw

Adds "blas" and "extended" keywords to package.json, inserting them after "math" and before "cumulative". Keyword conformance rises to 86% (75/87). Precedent: sibling packages @stdlib/blas/ext/base/dcusumpw and @stdlib/blas/ext/base/ndarray/dsumpw already carry both keywords.

@stdlib/blas/ext/base/ndarray/gcusum

Added "blas" and "extended" keywords to package.json, inserted after "math" and before "cumulative". Keyword set now conforms to the established pattern carried by the strided sibling @stdlib/blas/ext/base/gcusum and the parallel ndarray package @stdlib/blas/ext/base/ndarray/gsum. Conformance rises to 86% (75/87).

@stdlib/blas/ext/base/ndarray/gcusumkbn

Adds "blas" and "extended" keywords to package.json after "math" and before "cumulative". The package reaches 86% keyword conformance. Precedent is established by the strided sibling @stdlib/blas/ext/base/gcusumkbn and the parallel ndarray wrapper @stdlib/blas/ext/base/ndarray/gsumkbn, both of which carry these keywords.

@stdlib/blas/ext/base/ndarray/gcusumkbn2

Added "blas" and "extended" keywords to package.json after "math" and before "cumulative". Keyword conformance rises to 86% (75/87). Precedent established by the strided sibling and the parallel ndarray wrapper @stdlib/blas/ext/base/ndarray/gsumkbn2.

@stdlib/blas/ext/base/ndarray/gcusumors

Added "blas" and "extended" keywords to package.json after "math" and before "cumulative". Follows precedent set by the strided sibling @stdlib/blas/ext/base/gcusumors and the parallel ndarray wrapper @stdlib/blas/ext/base/ndarray/gsumors, both of which carry these keywords. Conformance: 86% (75/87).

@stdlib/blas/ext/base/ndarray/gcusumpw

Adds "blas" and "extended" keywords to package.json, inserted after "math" and before "cumulative". Keyword conformance rises to 86% (75/87). Matches the ordering established by the strided sibling @stdlib/blas/ext/base/gcusumpw and the parallel ndarray wrapper @stdlib/blas/ext/base/ndarray/gsumpw.

@stdlib/blas/ext/base/ndarray/scusum

Adds "blas" and "extended" keywords to package.json, inserted after "math" and before "cumulative". Keyword conformance across the ndarray wrapper family stands at 86% (75/87); sibling wrappers @stdlib/blas/ext/base/ndarray/scusumkbn and scusumkbn2 carry these keywords, as does the strided counterpart @stdlib/blas/ext/base/scusum. This change brings the package into alignment with established conventions in the extended BLAS ndarray namespace.

@stdlib/blas/ext/base/ndarray/scusumors

Adds "blas" and "extended" keywords to package.json, inserted after "math" and before "cumulative". Keyword conformance across the ndarray BLAS extended family stands at 86% (75/87); the same-family @stdlib/blas/ext/base/ndarray/scusumkbn and the strided sibling @stdlib/blas/ext/base/scusumors both carry these keywords. This change brings scusumors into alignment with established precedent.

Validation

Three independent validation agents reviewed each outlier before any patch was applied:

1. Semantic-review (Opus): for each outlier, compared lib/main.js and the strided sibling against conforming ndarray wrappers (dsum, dsumkbn, gsumors, scusumkbn, etc.). Result for all 12 outliers: confirmed-drift. Within the cumulative-sum family itself, scusumkbn and scusumkbn2 ndarray wrappers carry the keywords while structurally identical siblings do not — this internal inconsistency, combined with universal keyword presence in the strided counterparts, indicates inadvertent omission via copy-paste from an earlier template rather than any semantic intent.
2. Cross-reference (Opus): verified test/test.js and examples/index.js files in each outlier do not reference package.json contents or keywords; the only files mentioning the keyword field are the package.json files themselves. Result for all 12 outliers: safe-to-fix. No cascade into tests, examples, types, or REPL fixtures.
3. Structural-review (Sonnet): confirmed each outlier is a genuine BLAS extension (cumulative summation routines under @stdlib/blas/ext/base/ndarray/) and that the canonical insertion position (after "math", before "cumulative") is determinable from sibling ordering (scusumkbn, dsumkbn). Result for all 12 outliers: confirmed-drift.

Deliberately excluded

• gjoin README missing ## Notes heading (86/87 = 99% have it; the empty <section class="notes"> block is structurally present but lacks content) — fix would require writing package-specific prose, not a mechanical patch.
• gjoin-between lacking stdmath/mathematics/math keywords (99% have them) — gjoin-between is a string-join operation; its strided sibling @stdlib/blas/ext/base/gjoin-between likewise omits these math keywords. Intentional deviation.
• statistics/stats/sum/total/summation keyword distribution (55%) — below the 75% majority threshold and tracks the operation type.
• The 6-package native-binding cluster (csum, dnansum, dsum, snansum, ssum, zsum) extra files (binding.gyp, manifest.json, lib/native.js, etc.) — intentional, only present where a C implementation exists.
• references H2 section (31%) — present where mathematical references warrant it; below threshold.

Related Issues

None.

Questions

None.

Other

• Random seed: 20260506 (deterministic from today's date for reproducibility); namespace pick was blas/ext/base/ndarray from 122 eligible namespaces with ≥8 non-autogenerated members.
• Branch: philipp/drift-blas-ext-base-ndarray-2026-05-06.
• Per-package commits keep the audit trail readable per-outlier rather than per-feature; 12 commits in total.
• No tests, examples, types, or REPL fixtures were modified — package.json.keywords is the only diff.
• Full local report (member list, feature distributions including the 6-package native-binding cluster, agent verdicts, dropped findings) saved at ~/drift-reports/drift-blas-ext-base-ndarray-2026-05-06.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 running the cross-package API drift detection routine: a structural and semantic feature scan over the 87 blas/ext/base/ndarray packages, three independent validation agents (semantic-review, cross-reference, structural-review) per outlier, and a final per-package PR-body refinement pass. Every patch is a two-line addition to package.json.keywords at a deterministic position; no code, signatures, types, tests, examples, or REPL fixtures were changed. A human will audit before promoting from draft.

---

@stdlib-js/reviewers

---

Generated by Claude Code

[stdlib-bot]

Coverage Report

Package	Statements	Branches	Functions	Lines
blas/ext/base/ndarray/dcusum	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/dcusumkbn	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/dcusumkbn2	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/dcusumors	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/dcusumpw	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/gcusum	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/gcusumkbn	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/gcusumkbn2	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/gcusumors	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/gcusumpw	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/scusum	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$
blas/ext/base/ndarray/scusumors	$\color{green}127/127$ $\color{green}+100.00%$	$\color{green}3/3$ $\color{green}+100.00%$	$\color{green}1/1$ $\color{green}+100.00%$	$\color{green}127/127$ $\color{green}+100.00%$

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