How trishul4kumar-ui/nayak-quant-lab works
QUANT LAB is a local-first quantitative research and pre-live controlled-execution laboratory, not a broker-connected chatbot. The principal workflow runs from point-in-time market data and integrity checks through research, alpha/model validation, portfolio and risk authorization, controlled execution, and attribution. The README documents this high-level sequence and many supporting labs; source excerpts confirm selected data, feature, regime, adaptive, ensemble, backtest, and desktop-pipeline behavior. Other subsystem links below are documented capabilities, not claims that unsampled implementations were inspected.
Main entry points: features/engine.py Feature engine · store.py PIT data fabric · backtest/engine.py Backtest engine · service.py Paper OMS
Components 26
Desktop workflow
- Desktop app
src/quantlab/app/bootstrap.py - Research pipeline
src/quantlab/app/research_pipeline.py
Data foundation
- Market and research types
src/quantlab/domain/models.py - PIT data fabric
src/quantlab/data/fabric/store.py - Data ingestion
src/quantlab/data/fabric/ingest.py - Integrity and validation
src/quantlab/research/integrity.py
Research engines
- Feature engine
src/quantlab/features/engine.py - Regime engine
src/quantlab/regimes/engine.py - Alpha and labels
src/quantlab/alpha/definition.py - Statistical learning
src/quantlab/learning/engine.py - Adaptive alpha
src/quantlab/adaptive/engine.py - Ensemble engine
src/quantlab/ensemble/engine.py - Alpha discovery
src/quantlab/discovery/search.py - Research control
src/quantlab/orchestration/orchestrator.py - Backtest engine
src/quantlab/backtest/engine.py
Portfolio and controls
- Portfolio construction
src/quantlab/portfolio/construction.py - Capital allocation
src/quantlab/capital/definitions.py - Certification gates
src/quantlab/certification/service.py - Safety gateway
src/quantlab/execution_authorization/service.py
Execution and oversight
- Paper OMS
src/quantlab/paper_oms/service.py - Read-only broker gateway
src/quantlab/broker_gateway/service.py - Digital twin and shadow
src/quantlab/shadow/service.py - Monitoring and attribution
src/quantlab/monitoring/service.py - Research knowledge graph
src/quantlab/knowledge/ingest.py - Live operations
src/quantlab/live_ops/service.py
Other components
- Researcher
Connections 24
- Researcher → Desktop app: uses
- Desktop app → Research pipeline: presents stages
src/quantlab/app/research_pipeline.py - Data ingestion → PIT data fabric: materializes data
src/quantlab/data/fabric/ingest.py - PIT data fabric → Integrity and validation: supplies PIT data
- Integrity and validation → Feature engine: gates research
- Feature engine → Alpha and labels: builds signals
- Feature engine → Adaptive alpha: computes panels
src/quantlab/adaptive/engine.py - Regime engine → Adaptive alpha: provides regimes
src/quantlab/adaptive/engine.py - Alpha and labels → Ensemble engine: combines signals
- Feature engine → Ensemble engine: provides features
src/quantlab/ensemble/engine.py - Alpha discovery → Research control: records hypotheses
- Research control → Backtest engine: coordinates engines
- Backtest engine → Portfolio construction: constructs targets
src/quantlab/backtest/engine.py - Portfolio construction → Capital allocation: feeds allocation
- Capital allocation → Certification gates: submits for promotion
- Certification gates → Safety gateway: gates live boundary
- Safety gateway → Paper OMS: authorizes controlled execution
- Paper OMS → Monitoring and attribution: feeds attribution
- Read-only broker gateway → Digital twin and shadow: supports reconciliation
- Paper OMS → Digital twin and shadow: simulates shadow cycles
src/quantlab/shadow/service.py - Monitoring and attribution → Research knowledge graph: feeds research memory
- Live operations → Digital twin and shadow: observes live health
- PIT data fabric → Market and research types: uses market types
src/quantlab/data/fabric/ingest.py - Backtest engine → Market and research types: consumes bars
src/quantlab/backtest/engine.py