28-Feb-2026 — page-editor: restore enrichment query for cross-hierarchy content paths #170

build 505 — commit 6001af4

Reverts the single raw-regex query approach which broke image resolution for index pages whose child content paths reference pages outside the parent path hierarchy (e.g. test-gallery/ → gallery/...).

• Restore depth-limited regex for the top-level query (direct children)
• Restore enrichment query in enrichIndexColumnsWithImages using child indexes' own content paths (handles cross-hierarchy references)
• Remove allPrefetchedPages parameter — enrichment fetches its own pages
• Keep findFirstImageInPage fallback for TEXT-row-based image sources

Performance

• Replace O(N) recursive resolveIndexContent calls with single batch DB query using raw regex, with client-side filterByMaxPathSegments filtering
• Remove redundant broad prefetch that fetched all sub-pages unnecessarily
• Eliminate separate enrichment and grandchild queries in enrichIndexColumnsWithImages by passing pre-fetched pages from the top-level query — reduces page-content API calls from 3 to 1

Image Resolution

• Fix broken images on deeply nested index pages (e.g. by-year → year → volume → walk) by using unfiltered pre-fetched pages for image search across all depth levels
• Add findFirstImageInPage fallback in enrichIndexColumnsWithImages for pages that use TEXT rows with column-level imageSource instead of CAROUSEL/ALBUM rows (affects Berkshire Weekend Walkers scrapbook)
• Skip album-index rows in findFirstImageInPage to avoid picking up stale/default images from index row column data

Code Cleanup

• Remove dead code: resolveIndexContent, nextVisitedPaths, broad prefetch, visitedPaths/prefetchedPages/depth parameters from public API
• Simplify albumIndexToPageContent signature to just (pageContentRow, rowIndex)
• Add depthLimitedRegex helper for STARTS_WITH with maxPathSegments
• Add batched walk lookup in extractAlbumColumns for album locations

Testing

• Add 54 unit tests for IndexService helper methods covering depthLimitedRegex, findFirstImageInPage, findFirstTextInPage, optimiseIndexImageSource, summarizeTitles, deduplicateByHref, calculateDataCompletenessScore, sortColumns, filterOutExcludedPaths, filterByMaxPathSegments, pageContentFrom

Location Extraction

• Handle album-index rows in extractTitleAndDescription via indexMarkdown

• Skip album-index rows in findFirstImageInPage

• Use match-type-aware regex in index prefetch query so "Contains" paths no longer force a starts-with pattern

• Update filterByMaxPathSegments to respect the configured match type instead of hardcoding startsWith

• Read carousel image titles and dates from gridViewOptions to match the editor's data model

• Add move-to-exclude and move-to-include buttons on content path match rows