Add defect_root_cause template: multi-reasoner quality-defect RCA#89
Open
cafzal wants to merge 5 commits into
Open
Add defect_root_cause template: multi-reasoner quality-defect RCA#89cafzal wants to merge 5 commits into
cafzal wants to merge 5 commits into
Conversation
Backward RCA chain (Graph reachability -> Rules contrast scoring -> Prescriptive set-cover MILP) over a serialized electronics-assembly genealogy. Diagnoses a final-test defect spike to a minimal set of root causes -- the inverse of bom-reachability. Adds the abduction archetype and the first set-cover / minimal-diagnosis MILP to the portfolio; grows Manufacturing 4->5. Engine-validated end-to-end; runbook blind-paste-tested.
|
The docs preview for this pull request has been deployed to Vercel!
|
Each named cause is now reported with its defect signature (paste->cold solder, reflow->bridging), and a kind-specific tell -- supplier and receipt date for a lot, calibration age for a machine -- read back from ontology. Adds Lot.received_date. Diagnosis unchanged; this makes the script produce the corroboration the runbook already interprets.
…porally coherent
Adds a descriptive 'when did the spike start?' examine step (failure rate by build week) ahead of the genealogy chain, and wires the data so both root causes share a week-2 incident onset: the contaminated paste lot is received then, the oven drifts then, and pre-onset units cannot consume the contaminated boards. The timeline is now a real signal (1.9 -> 6.6 -> 8.3%), and the paste lot's receipt date aligns exactly with the onset. Diagnosis unchanged in shape -- {SP-0423, REF-02} explain 120/142 (85%). Regenerated data and indexes; README and runbook resynced.
…sses PR #80 review feedback) Rewrites the README to the sample-template standard the reviewer is applying to new templates: 'What this template is for' is now a broad-audience problem statement and motivation with a bold reasoning-type sentence; adds assumed knowledge, an Access/Tools Prerequisites split, a Model overview, and Customize subsections; and uses words and sentences rather than symbols and fragments throughout the prose. Also drops trailing commas after the last where(...) condition per the cleanup-template-code convention (a no-op; re-run confirms identical output). Validated: sample output matches the live run line-for-line, all code snippets are verbatim from the script, and the runbook reproduced blind end-to-end.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
A new Manufacturing template implementing quality-defect root-cause analysis as a backward multi-reasoner chain — the inverse of
bom-reachability. Given a final-test defect spike, it traces each unit's genealogy backward through the bill of materials, contrast-scores candidate factors against good units, and solves a minimal set-cover MILP for the smallest, most specific set of root causes.Why it's complementary to the portfolio
bom-reachability(same BOM, run backward from failures).The chain (Graph → Rules → Prescriptive)
Unit.touches_lot), attributing a contaminated lot two tiers down to every finished unit that carries it.Factor.lift,Factor.is_suspect), screening out high-coverage / low-lift trunk factors.Factor.is_root_cause,DiagnosisResult), preferring the single deep root over the proximate sub-assemblies that merely carry it.On the bundled corpus (2,500 units, 6.28% failure rate), the chain diagnoses
{SP-0423 (paste lot), REF-02 (reflow oven)}explaining 90% of failures — correctly resisting the high-volume trunk factors that dominate a naive raw-count scan.Verification
py_compileandruffclean.rai-*skills + data alone, never seeing the solution script.Factor.toucheslot incidence independently checked against a pure-Python genealogy closure.Files
v1/defect_root_cause/— README, runbook,pyproject.toml(relationalai==1.13.0),defect_root_cause.py(3-stage pipeline),generate_data.py, and 10 data CSVs.README.mdandv1/README.md.