Compute search index out of the directory metadata

[jviotti]

Signed-off-by: Juan Cruz Viotti jv@jviotti.com

[github-actions]

Benchmark Index (enterprise)

Details

Benchmark suite	Current: cd99d84	Previous: 6d683a9	Ratio
Add one schema (0 existing)	23 ms	27 ms	0.85
Add one schema (100 existing)	34 ms	36 ms	0.94
Add one schema (1000 existing)	154 ms	162 ms	0.95
Add one schema (10000 existing)	1540 ms	1661 ms	0.93
Update one schema (1 existing)	21 ms	22 ms	0.95
Update one schema (101 existing)	33 ms	35 ms	0.94
Update one schema (1001 existing)	154 ms	160 ms	0.96
Update one schema (10001 existing)	1544 ms	1610 ms	0.96
Cached rebuild (1 existing)	12 ms	12 ms	1
Cached rebuild (101 existing)	13 ms	14 ms	0.93
Cached rebuild (1001 existing)	31 ms	32 ms	0.97
Cached rebuild (10001 existing)	228 ms	237 ms	0.96
Index 100 schemas	142 ms	146 ms	0.97
Index 1000 schemas	1418 ms	1175 ms	1.21
Index 10000 schemas	18227 ms	15663 ms	1.16

This comment was automatically generated by workflow using github-action-benchmark.

[augmentcode]

🤖 Augment PR Summary

Summary: This PR changes how the Explorer search index is produced by deriving it from directory listing metadata rather than per-schema metapack extensions.

Changes:

• Removed the global/aggregate SearchIndex build rule and replaced it with a root-only directory rule (search.metapack).
• Introduced DirectoryDependencyKind::AllDirectoryListings and implemented its dependency expansion in the build delta planner.
• Updated GENERATE_EXPLORER_SEARCH_INDEX to read directory.metapack JSON and collect schema entries’ path/title/description from entries.
• Adjusted CLI golden outputs and unit tests to reflect the new build DAG ordering and dependency edges.

Technical Notes: Dependency tracking for search.metapack now flows through directory listing targets, which centralizes schema metadata consumption in directory.metapack rather than reading schema extensions directly.

_{🤖 Was this summary useful? React with 👍 or 👎}

[augmentcode]

Review completed. 3 suggestions posted.

Comment augment review to trigger a new review at any time.

[cubic-dev-ai]

2 issues found across 18 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="src/build/delta.cc">

<violation number="1" location="src/build/delta.cc:988">
P1: `AllDirectoryListings` iterates over `affected_dirs`, which only contains directories derived from dirty/affected schemas. During an incremental rebuild where only one nested schema changes, unrelated directories won't appear in `affected_dirs`, so their `directory.metapack` files won't be listed as dependencies for the `SearchIndex` action. This can cause schemas from those unaffected directories to be silently dropped from `search.metapack`.</violation>

<violation number="2" location="src/build/delta.cc:993">
P3: `DIRECTORY_RULES[0].filename` relies on array ordering to access the `DirectoryList` rule. If the `DIRECTORY_RULES` array is ever reordered, this will silently reference the wrong target filename. Consider looking up the rule by its action type instead of using a magic index.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}

[github-actions]

Benchmark Index (community)

Details

Benchmark suite	Current: cd99d84	Previous: 6d683a9	Ratio
Add one schema (0 existing)	26 ms	22 ms	1.18
Add one schema (100 existing)	39 ms	35 ms	1.11
Add one schema (1000 existing)	195 ms	164 ms	1.19
Add one schema (10000 existing)	2147 ms	1702 ms	1.26
Update one schema (1 existing)	24 ms	19 ms	1.26
Update one schema (101 existing)	39 ms	34 ms	1.15
Update one schema (1001 existing)	191 ms	160 ms	1.19
Update one schema (10001 existing)	2637 ms	1670 ms	1.58
Cached rebuild (1 existing)	12 ms	11 ms	1.09
Cached rebuild (101 existing)	14 ms	13 ms	1.08
Cached rebuild (1001 existing)	37 ms	31 ms	1.19
Cached rebuild (10001 existing)	509 ms	244 ms	2.09
Index 100 schemas	155 ms	154 ms	1.01
Index 1000 schemas	1523 ms	1235 ms	1.23
Index 10000 schemas	19060 ms	15919 ms	1.20

This comment was automatically generated by workflow using github-action-benchmark.