How anish1441/spark-engineering-quality-intelligence works
SPARK is a burn-in screening and reliability decision-support application. Engineers upload or select measurement datasets, assess component evidence through trust, applicability, and safety checks, then receive anomaly and forecast evidence, a unified recommendation, an explanation, and a governed QA record. The README documents the screening stages and broader population-monitoring capabilities; sampled source confirms the Flask/UI boundary, dataset and QA persistence, risk fusion, and connection to the preserved Phase-1 pipeline. Some subsystem wiring is documented but not visible in the partial excerpts.
Main entry points: app.py Flask API · pipeline_adapter.py Pipeline Adapter · app.js SPARK Dashboard · qa_manager.py Traceability Ledger
Components 24
User Interface
- SPARK Dashboard
static/js/app.js - Flask API
app.py
Screening Workflow
- Dataset Manager
core/dataset_manager.py - Dataset Analysis
core/analysis_engine.py - Applicability Gate
core/applicability.py - Data Trust Gate
core/gates.py - Engineering Safety
core/gates.py - Rolling Forecast
core/rolling_forecast.py - Reliability Risk
core/risk_engine.py - Decision Explanation
core/explainability.py
Model Evidence
- Pipeline Adapter
core/pipeline_adapter.py - Feature Engineering
bundled_phase1/src/sih26170/features.py - Module A Anomalies
bundled_phase1/src/sih26170/models/module_a_anomaly.py - Module B Forecast
bundled_phase1/src/sih26170/models/module_b_96h.py - Model Artifacts
bundled_phase1/artifacts/models
Population Intelligence
- Lot Intelligence
core/lot_intelligence.py - Commonality Engine
core/commonality.py - Drift Observatory
core/drift.py - Calibration Monitor
core/calibration.py
QA Governance
- QA Governance
core/governance.py - Traceability Ledger
core/qa_manager.py - Feedback Learning
core/feedback_learning.py - Model Registry
core/model_registry.py
Other components
- Reliability Engineer
Connections 25
- Reliability Engineer → SPARK Dashboard: uses
- SPARK Dashboard → Flask API: calls API
- Flask API → Dataset Manager: loads datasets
- Flask API → Dataset Analysis: analyzes data
- Flask API → Applicability Gate: checks applicability
- Flask API → Data Trust Gate: checks evidence
- Flask API → Engineering Safety: checks limits
- Flask API → Pipeline Adapter: requests predictions
- Pipeline Adapter → Feature Engineering: builds features
- Pipeline Adapter → Module A Anomalies: runs anomaly model
- Pipeline Adapter → Module B Forecast: runs forecast model
- Pipeline Adapter → Model Artifacts: loads artifacts
- Flask API → Rolling Forecast: updates trajectory
- Flask API → Reliability Risk: fuses evidence
- Flask API → Decision Explanation: explains decision
- Flask API → Lot Intelligence: summarizes lots
- Flask API → Commonality Engine: finds commonality
- Flask API → Drift Observatory: checks drift
- Flask API → Calibration Monitor: checks calibration
- Flask API → QA Governance: evaluates review
- Flask API → Traceability Ledger: records decisions
- Flask API → Feedback Learning: summarizes feedback
- Flask API → Model Registry: reports model status
- Traceability Ledger → Feedback Learning: provides QA history
- Flask API → SPARK Dashboard: returns results