mirror of
https://github.com/allaunthefox/Research-Stack.git
synced 2026-07-31 03:05:21 +00:00
Ports the decision-critical scoring functional from route_repair_v14a.py
into a provable Lean surface:
rank_patches: S = α·specificity − β·cost + γ·success_prior − δ·residual_risk
ALPHA=0.4, BETA=0.3, GAMMA=0.2, DELTA=0.1 (all as Q16_16.ofRatio)
New module: Semantics.PIST.Repair (232 lines)
§1 PatchScoreInputs, PatchWeights structures
§2 rankScore (linear functional), rankScoreDefault, mkInputs, embedResidualRisk
§3 Patch record + mkPatch constructor
§4 rankPatches / rankPatchesDefault (mergeSort desc, tag tie-break)
§5 5 executable #eval witnesses with -- expect: annotations
§6 8 proved invariants (native_decide):
defaultWeights_sum, defaultWeights_pos, defaultWeights_ordered,
rankScore_zero_inputs_negative, embedResidualRisk_one/zero,
rankScore_monotone_specificity_witness, rankScore_zero_lt_full
Full workspace build: 3569 jobs, 0 errors.
route_repair_v14a.py PARTIAL BOUNDARY comment updated to name this module
as the authoritative source for the scoring surface.
Generated with [Devin](https://cli.devin.ai/docs)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
|
||
|---|---|---|
| .. | ||
| core | ||
| lean | ||
| otom | ||
| receipts | ||
| rust/ftdi_gimbal | ||
| README.md | ||
0-Core-Formalism
Purpose: Formal foundations for the entire Research Stack — Lean modules, bind primitive, Triumvirate consensus, core source.
No external dependencies. All other layers depend on this.
Contents (Target)
| Source | Destination |
|---|---|
0-Core-Formalism/lean/Semantics/ |
0-Core-Formalism/lean/Semantics/ |
core/ |
0-Core-Formalism/core/ |
Concepts
- bind — State → (State → Action) → State
- TriumvirateClock — ternary consensus (ADD/PAUSE/SUBTRACT)
- Builder/Judge/Warden — roles mapped to hardware registers
- OTOM — Ordered Transformation & Orchestration Model
Build
cd "0-Core-Formalism"
lake build