How 26pb27-final-year-project/t1-fx-risk works
T+1 is a local research engine and React dashboard for simulating settlement failures, training risk and FX-receipt timing models, and comparing cash-funding policies. The user workflow runs from experiment setup through dataset generation or selection, model preparation, budget comparison and saved results, then review and export. The graph includes the documented Trades, Funding, Scenarios and Model health views. Several runtime modules are unsampled; their relationships are included only where the README or SOURCE INDEX supports them.
Main entry points: workflow.py Workflow coordinator · pipeline.py Funding comparison · api.py Local API · training.py Risk training
Components 20
Dashboard
- Dashboard shell
frontend/src/main.tsx - Experiment controls
frontend/src/Experiments.tsx - Evaluation calendar
frontend/src/Evaluation.tsx
Experiment workflow
- Local API
src/t1fx/api.py - Durable jobs
src/t1fx/jobs.py - Workflow coordinator
src/t1fx/workflow.py
Data and models
- Settlement calendars
src/t1fx/calendars.py - Synthetic generation
src/t1fx/generation.py - Dataset validation
src/t1fx/datasets.py - Risk features
src/t1fx/features.py - Risk training
src/t1fx/training.py - FX timing forecasts
src/t1fx/timing.py - Model artifacts
src/t1fx/models.py
Simulation and results
- Funding comparison
src/t1fx/pipeline.py - Funding optimizer
src/t1fx/funding.py - Settlement replay
src/t1fx/simulator.py - Decision snapshots
src/t1fx/snapshots.py - Run artifacts
src/t1fx/storage.py
Other components
- Researcher
- Frankfurter rates
Connections 25
- Researcher → Experiment controls: configures experiment
- Dashboard shell → Experiment controls: renders
frontend/src/main.tsx - Dashboard shell → Evaluation calendar: renders
frontend/src/main.tsx - Experiment controls → Local API: submits requests
- Evaluation calendar → Local API: requests online time
- Local API → Workflow coordinator: invokes
src/t1fx/api.py - Local API → Durable jobs: uses job state
src/t1fx/api.py - Durable jobs → Workflow coordinator: dispatches work
src/t1fx/jobs.py - Workflow coordinator → Synthetic generation: generates data
src/t1fx/workflow.py - Workflow coordinator → Dataset validation: validates datasets
src/t1fx/workflow.py - Workflow coordinator → Risk training: trains models
src/t1fx/workflow.py - Workflow coordinator → FX timing forecasts: prepares timing model
src/t1fx/workflow.py - Workflow coordinator → Funding comparison: compares policies
src/t1fx/workflow.py - Workflow coordinator → Settlement calendars: checks dates
src/t1fx/workflow.py - Synthetic generation → Settlement calendars: uses calendars
src/t1fx/generation.py - Synthetic generation → Frankfurter rates: fetches reference rates
src/t1fx/generation.py - Risk training → Dataset validation: builds training tables
src/t1fx/training.py - Risk training → Risk features: fits features
src/t1fx/training.py - Risk training → Model artifacts: publishes models
src/t1fx/training.py - Funding comparison → Settlement replay: replays scenario
src/t1fx/pipeline.py - Funding comparison → Decision snapshots: builds snapshots
src/t1fx/pipeline.py - Funding comparison → Funding optimizer: optimizes reserves
src/t1fx/pipeline.py - Funding comparison → Run artifacts: publishes runs
src/t1fx/pipeline.py - Settlement replay → Run artifacts: stores replay outputs
- Local API → Run artifacts: reads results
src/t1fx/api.py