SCIP precision tier and measured resolution regression baseline #361

Merged
toasterson merged 6 commits from claude/wi-019fa8ea-wi-2-scip-precision-tier-the-resolution into main 2026-10-06 19:26:15 +00:00
Owner

Implements a reproducible compiler comparison for Rust resolution on Akh-Medu. The ingester preserves SCIP and stack-graph provenance together; scoring compares exact reference/definition spans, retains parser gaps as false negatives, and enumerates ambiguous or otherwise unjudgeable sites. The recorded full-corpus baseline is replayed through the current rules and through deliberately broken rules that must fail with source-located disagreements.

Work item: 019fa8ea-daf9-7432-b322-86845cbc445e (WI-263).

Measured all 1,020 documents emitted by rust-analyzer at d4d46b598f: 996 indexed and 24 parser gaps. Resolution precision/recall is 92.954743% / 2.730278%; file-dependency precision/recall is 62.407132% / 6.517691%. Coverage, source/index/rule hashes, resource measurements and complete disagreements are committed under crates/shend/fixtures/akh-medu. The parent plan records the decision to require fresh unambiguous SCIP evidence or explicit degraded behavior; these low recall figures do not authorize downstream stack-graph consumers.

Validation: 12 unit tests, 3 focused parser/import tests, the actual full current/broken-rules replay, and strict all-target Clippy passed. Compiler capture took 293.965s; scoring 349.315s; full replay 525.141s. These are local host/default-feature measurements with explicit compiler limitations, not all-feature coverage or Solstice capacity proof. The workflow replays the recorded corpus on harness changes; expensive compiler regeneration is manual/on-demand. Current-head canonical review and native CI remain required.

Implements a reproducible compiler comparison for Rust resolution on Akh-Medu. The ingester preserves SCIP and stack-graph provenance together; scoring compares exact reference/definition spans, retains parser gaps as false negatives, and enumerates ambiguous or otherwise unjudgeable sites. The recorded full-corpus baseline is replayed through the current rules and through deliberately broken rules that must fail with source-located disagreements. Work item: 019fa8ea-daf9-7432-b322-86845cbc445e (WI-263). Measured all 1,020 documents emitted by rust-analyzer at d4d46b598f8c9b7597b4222f2ed2903a20ccabfa: 996 indexed and 24 parser gaps. Resolution precision/recall is 92.954743% / 2.730278%; file-dependency precision/recall is 62.407132% / 6.517691%. Coverage, source/index/rule hashes, resource measurements and complete disagreements are committed under crates/shend/fixtures/akh-medu. The parent plan records the decision to require fresh unambiguous SCIP evidence or explicit degraded behavior; these low recall figures do not authorize downstream stack-graph consumers. Validation: 12 unit tests, 3 focused parser/import tests, the actual full current/broken-rules replay, and strict all-target Clippy passed. Compiler capture took 293.965s; scoring 349.315s; full replay 525.141s. These are local host/default-feature measurements with explicit compiler limitations, not all-feature coverage or Solstice capacity proof. The workflow replays the recorded corpus on harness changes; expensive compiler regeneration is manual/on-demand. Current-head canonical review and native CI remain required.
WI 2 — SCIP precision tier + the resolution accuracy harness that gates everything downstream
All checks were successful
CI / check-seshat (pull_request) Successful in 8m35s
CI / check-akh-medu (pull_request) Successful in 13m18s
CI / publish-chart (pull_request) Successful in 1m39s
CI / docker-seshd (pull_request) Successful in 9m7s
CI / test-akh-medu (pull_request) Successful in 18m18s
c88dd9b526
Anima implementation session.
feat(shend): add SCIP precision harness tier
All checks were successful
CI / check-akh-medu (pull_request) Successful in 7m18s
CI / test-session-platform (pull_request) Successful in 15m26s
CI / docker-seshd (pull_request) Successful in 8m7s
CI / test-akh-medu (pull_request) Successful in 10m40s
CI / check-seshat (pull_request) Successful in 14m47s
CI / publish-chart (pull_request) Successful in 5m47s
CI / test-sokar (pull_request) Successful in 18m51s
CI / docker-akh-transcripts (pull_request) Successful in 6m33s
CI / docker-sokard (pull_request) Successful in 14m9s
6430b6c944
fix(shend): build lexical scopes once per syntax node
All checks were successful
CI / check-seshat (pull_request) Successful in 6m12s
CI / test-akh-medu (pull_request) Successful in 6m41s
CI / check-akh-medu (pull_request) Successful in 8m11s
CI / docker-akh-transcripts (pull_request) Successful in 6m14s
CI / test-sokar (pull_request) Successful in 21m6s
CI / publish-chart (pull_request) Successful in 9m9s
CI / docker-sokard (pull_request) Successful in 14m18s
CI / docker-seshd (pull_request) Successful in 15m11s
CI / test-session-platform (pull_request) Successful in 27m11s
7ecad5b406
Author
Owner

Full-lane babysit 2026-10-01 — WI-263.

Exact head 7ecad5b406a39f008eeb3c1b5abddeb21aaf7cdf, current main ad441fe9e2359a1bf315eda28a4b072cc56cee2d. Real merge-tree clean; branch is not already on main.

SCIP ingestion/harness implementation exists and hosted CI green with clean merge-tree. Acceptance remains incomplete: committed harness_baseline.json has watermark:null, files_indexed:0, empty predicates and no disagreements. It is an empty fixture, not measured Akh-Medu precision/recall. Produce actual same-watermark repo measurements and reviewed baseline before unlocking consumers; canonical gates also uncleared.

Individual disposition also recorded in the Anima WI thread. No approval or CI gate was overridden.

Full-lane babysit 2026-10-01 — WI-263. Exact head `7ecad5b406a39f008eeb3c1b5abddeb21aaf7cdf`, current main `ad441fe9e2359a1bf315eda28a4b072cc56cee2d`. Real merge-tree clean; branch is not already on main. SCIP ingestion/harness implementation exists and hosted CI green with clean merge-tree. Acceptance remains incomplete: committed harness_baseline.json has watermark:null, files_indexed:0, empty predicates and no disagreements. It is an empty fixture, not measured Akh-Medu precision/recall. Produce actual same-watermark repo measurements and reviewed baseline before unlocking consumers; canonical gates also uncleared. Individual disposition also recorded in the Anima WI thread. No approval or CI gate was overridden.
Measure real Akh-Medu SCIP accuracy and preserve truthful coverage
Some checks failed
SCIP accuracy / produce-index (pull_request) Has been skipped
CI / test-sokar (pull_request) Successful in 11m40s
CI / check-akh-medu (pull_request) Successful in 15m7s
CI / check-seshat (pull_request) Successful in 31m37s
CI / docker-sokard (pull_request) Successful in 45m40s
CI / publish-chart (pull_request) Successful in 45m35s
CI / docker-seshd (pull_request) Successful in 56m24s
CI / docker-akh-transcripts (pull_request) Successful in 58m37s
SCIP accuracy / recorded-accuracy (pull_request) Failing after 56m0s
CI / test-akh-medu (pull_request) Failing after 2h11m15s
CI / test-session-platform (pull_request) Successful in 2h32m55s
CI / platform-images-daemons (pull_request) Has been skipped
CI / platform-images-core (pull_request) Has been skipped
CI / platform-images-home (pull_request) Has been skipped
CI / platform-images-fleet (pull_request) Has been skipped
CI / platform-images-collect (pull_request) Has been skipped
1a1adda337
toasterson changed title from WIP: WI 2 — SCIP precision tier + the resolution accuracy harness that gates everything downstream to SCIP precision tier and measured resolution regression baseline 2026-10-01 21:53:45 +00:00
Sign in to join this conversation.
No reviewers
No labels
follow-up
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
toasterson/akh-medu!361
No description provided.