Repository navigation
v3.16.0 — Reconstruction Engine Foundation
Reconstruction Engine Foundation
Closes 4 gaps turning HipCortex from "cognitive state substrate" into "dynamic reconstruction engine".
What's new
| Change | Details |
|---|---|
| Anomaly → axiomatic write-back | `LawExtractor` clusters surprising Intent records by `target_entity`; ≥3 in cluster → writes `MemoryType::Law` (structural equation, no LLM, O(n)); idempotent; calls `route_uncertainty` (T0 self-prompt first) when < 2 causal variables identified |
| Policies as first-class citizens | `PolicyRegistry` stores active `MemoryType::Policy` records by `entity_id` sorted by priority; `DigitalTwin::step_with_policies(action, registry, entity_id)` applies highest-priority matching Policy — `S_{t+1} = f(S_t, Policy)` |
| MDL sparsity pressure on Laws | `CognitiveGC::gc_action_for_law(id, mdl_score)`: score ≥ `MDL_KEEP_THRESHOLD` (0.5) → `Keep`; below → `Archive` if referenced, `Delete` if orphaned |
| REST surfaces | `GET /laws` — all Law records; `GET /policies/:entity_id` — active Policies sorted by priority desc |
Test coverage
379 lib + 546 unit (+4) + 186 integration (+4 SIT) + 59 property + 258 Python = all green, 0 failures
Install
pip install hipcortex==3.16.0VSIX: attach hipcortex-memory-3.16.0.vsix to VS Code manually via Extensions → Install from VSIX.
Full changelog
See README.md