[codex] Document memory-efficient semantic dedup fitting

[lbliii]

Addresses the semantic-deduplication portion of #2144. The bundled 26.06 release-note disposition is tracked in existing PR #2143.

Summary

• documents KMeansStage.fit_data_fraction, SemanticDeduplicationWorkflow.fit_data_fraction, and TextSemanticDeduplicationWorkflow.kmeans_fit_data_fraction
• explains per-actor file sampling, validation, the two-pass execution path, peak-memory reduction, and added I/O
• documents direct KMeansStage.cache_path centroid persistence and its distinction from workflow cache paths
• corrects the output artifact tree: workflow wrappers do not save kmeans_centroids.npy

Validation

• fern check: 0 errors, 103 pre-existing warnings
• all 11 Python examples parse with ast
• source contract audit confirms all documented parameter names exist on current main
• link check baseline remains the 22 pre-existing errors outside this page

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[greptile-apps]

Greptile Summary

This PR adds documentation for the memory-efficient K-means fitting path in semantic deduplication, covering the new kmeans_fit_data_fraction / fit_data_fraction parameters across all three API layers (TextSemanticDeduplicationWorkflow, SemanticDeduplicationWorkflow, and KMeansStage).

• Adds a "Reduce K-means fitting memory" section with accurate two-pass mechanics, validated against the source implementation in kmeans.py.
• Corrects the output artifact tree: removes kmeans_centroids.npy from the workflow cache path and documents that only a direct KMeansStage(cache_path=…) call persists centroids.

Confidence Score: 5/5

Pure documentation change; no executable code paths are modified, and all documented behaviour was verified against the current source.

Every parameter name (kmeans_fit_data_fraction, fit_data_fraction), default value (None), sampling formula (round(fraction × N_files), minimum 1), memory characterisation, and centroid-persistence caveat (KMeansStage.cache_path vs workflow cache_path=None) was cross-checked against kmeans.py, workflow.py, and semantic.py and found to be correct. The directory tree correction (removing kmeans_centroids.npy from the workflow path) also matches the source. No logic regressions are possible from a docs-only diff.

No files require special attention.

Important Files Changed

Filename	Overview
fern/versions/main/pages/curate-text/process-data/deduplication/semdedup.mdx	Documentation-only change adding accurate descriptions of fit_data_fraction/kmeans_fit_data_fraction; all parameter names, sampling logic, memory semantics, and centroid-persistence behaviour verified against source.

_{Reviews (1): Last reviewed commit: "docs: explain memory-efficient semantic ..." | Re-trigger Greptile}