test(confidence-engine): extract focused target input for deterministic verification
This commit is contained in:
@@ -828,3 +828,7 @@ Returns JSON-serializable record with: five deconstruction arrays (`observations
|
||||
**What remains uncertain:** Whether explicit dependency information in other scenario shapes (non-recruitment) similarly improves cross-area relationship representation. Single-run limitation applies.
|
||||
|
||||
**Evidence file:** `.evidence-temp/headless-semantic/HS-03C-focused-E9.json`
|
||||
|
||||
### Iteration 3 — Extract Focused Target Input (apparatus recovery)
|
||||
|
||||
The focused-investigation target-node extraction was purified into a standalone pure function `extractFocusedTargetInput(situationGraph, targetNodeId)` in the canonical helper. Deterministic test verified: known-node exact-match (1/1), unknown-node empty-string fallback (2/2), graph-immutable-side-effect (1/1), source-pattern integration verification (1/1) — total 5/5 PASS. Zero live model calls. Extraction was tested deterministically; the original Experiment 9 runtime graph and root cause remain unestablished.
|
||||
|
||||
Reference in New Issue
Block a user