Restructures kNN documentation into a clearer page hierarchy

[kosabogi]

This PR split the kNN search kNN search on Elasticsearch page into a nested structure.

• Landing page (knn.md): intro, prerequisites, and method selection only.
• Moved content (headings preserved):
  • Approximate kNN → foundational mapping/indexing/search + indexing considerations + limitations
  • Optimize → tuning, byte vectors, quantization, BFloat16, oversampling/rescoring
  • Build queries → filtering, hybrid, semantic, multi-field, similarity, nested/chunked
  • Exact kNN → existing exact kNN content

Related issue: #1740
Parent issue: https://github.com/elastic/docs-content-internal/issues/801

[github-actions]

Elastic Docs AI PR menu

Check the box to run an AI review for this pull request.

• Review docs changes (docs-review). Status: queued.

Powered by GitHub Agentic Workflows and docs-actions. For more information, reach out to the docs team.

[github-actions]

Docs review summary

Focus areas

• Style and clarity: One grammar error (missing "you") in optimize-performance-accuracy.md line 209. First-person "we" in two new files (exact-knn.md and build-search-queries.md). Both are introduced by this PR.
• Jargon: No unexplained jargon found in the new pages. HNSW, DiskBBQ, bbq_disk are explained in context.
• Frontmatter and applies_to: High-priority: approximate-knn.md duplicates the mapped_pages already present on knn.md. Duplicate mapped_pages entries for the same legacy URLs across two files will cause a build conflict. See inline comment on approximate-knn.md line 6.
• Content type fit: The four new pages are correctly structured as how-to guides. One structural issue: exact-knn.md has the kNN definition and use cases stranded after the 3-step procedure (lines 83–93), which reads as orphaned content from the original page. See inline comment.
• Parent issue satisfaction: The PR satisfies the restructuring goal described in the PR body. The hierarchy is clear and content is sensibly split. Inline issues are in the details of execution, not the overall approach.

Nits

• exact-knn.md line 92: "Check out our [hands-on tutorial]" uses first-person "our". Consider: "For a hands-on example, refer to Bring your own dense vectors."
• build-search-queries.md line 485: Field names "paragraphs" and "metadata" are wrapped in prose double quotes; prefer backtick monospace: `paragraphs` and `metadata`.

Notes

• The Vale finding at optimize-performance-accuracy.md line 219 ("Use 'and later' instead of 'and higher'") is a false positive — the text refers to compression levels, not version numbers.
• Pre-existing style findings in unchanged files (cross-cluster-search.md, ml-nlp-model-ref.md, etc.) are out of scope for this review.

Generated by Docs review agent for issue #7138 · 78.4 AIC · ⌖ 10.8 AIC · ⊞ 10.7K

[github-actions]

🔍 Preview links for changed docs

• deploy-manage/production-guidance/optimize-performance/approximate-knn-search.md
• explore-analyze/cross-cluster-search.md
• explore-analyze/machine-learning/nlp/ml-nlp-model-ref.md
• explore-analyze/machine-learning/nlp/ml-nlp-text-emb-vector-search-example.md
• solutions/search/semantic-search/semantic-search-inference.md
• solutions/search/vector.md
• solutions/search/vector/dense-versus-sparse-ingest-pipelines.md
• solutions/search/vector/knn.md
• solutions/search/vector/knn/approximate-knn.md
• solutions/search/vector/knn/build-search-queries.md

More links …

• solutions/search/vector/knn/exact-knn.md
• solutions/search/vector/knn/nested-knn-search.md
• solutions/search/vector/knn/optimize-performance-accuracy.md
• solutions/search/vector/vector-search-use-cases.md

[github-actions]

Elastic Docs Style Checker (Vale)

Summary: 3 warnings, 2 suggestions found

⚠️ Warnings (3): Fix when the suggestion improves clarity or correctness.

File	Line	Rule	Message
solutions/search/vector/knn/build-search-queries.md	63	Elastic.QuotesPunctuation	Place punctuation inside closing quotation marks.
solutions/search/vector/knn/optimize-performance-accuracy.md	257	Elastic.Spelling	'rescored' is a possible misspelling.
solutions/search/vector/knn/optimize-performance-accuracy.md	258	Elastic.Spelling	'rescored' is a possible misspelling.

💡 Suggestions (2): Optional style improvements. Apply when helpful.

File	Line	Rule	Message
solutions/search/vector/knn/build-search-queries.md	141	Elastic.Wordiness	Consider using 'all' instead of 'all of '.
solutions/search/vector/knn/optimize-performance-accuracy.md	218	Elastic.Versions	Use 'and later' instead of 'and higher' when referring to versions.

---

The Vale linter checks documentation changes against the Elastic Docs style guide. To use Vale locally or report issues, refer to Elastic style guide for Vale.