Implement Modularized MVC structure

[rahibAmin84]

Separate concerns and decouple and modularize where necessary.

[rahibAmin84]

DDD-Inspired Architecture with Real Contexts

1. Bounded Contexts
   1. ROI Calculation Context (Python ML)
      • Handles Bayesian regression, training, inference, and model artifacts.
      • Reads/writes datasets from cloud storage.
      • Provides a CLI or REST API for TypeScript to call.
   2. User Interaction Context (TypeScript MVC)
      • Handles UI, input validation, and displaying predictions.
      • Calls Python ML scripts for inference.
      • Coordinates workflows and orchestrates business logic.
   3. Data Management Context
      • Manages cloud storage for raw datasets and model outputs.
      • Provides abstraction for reading/writing files.
      • Ensures reproducibility and collaboration across environments.

⸻

2. Folder Structure

/ml/ # ROI Calculation Context (Python)
├── train.py # training pipeline
├── inference.py # inference pipeline
├── model_utils.py # Bayesian regression helpers
├── storage.py # cloud storage helpers

/ts_app/ # User Interaction Context (TypeScript)
/investments/
PlatformInvestment.ts # entity representing a platform investment
InvestmentFeatures.ts # input value object for ML
ROIAggregate.ts # predicted ROI + uncertainty
ROIPredictionService.ts # service calling Python and handling aggregates
ROIAggregateFactory.ts # factory to create ROIAggregates
/controllers/
ROIController.ts # handles requests and responses
/views/
roiView.ts # renders predictions and uncertainty
/infrastructure/
pythonService.ts # handles Python CLI or API calls
cloudStorageService.ts # cloud storage abstraction
app.ts # entry point

/data_management/ # Data Management Context
cloud_storage_config.ts # config for S3/GCP/Azure
storage_helpers.ts # utility functions for dataset operations

⸻

3. Example Flow (End-to-End)
   1. User submits input via the UI (TypeScript view).
   2. Controller validates input and calls ROIPredictionService.
   3. Service builds InvestmentFeatures, then calls Python ML via pythonService.ts.
   4. Python ML reads raw data from cloud, runs inference, returns posterior ROI.
   5. Service creates a ROIAggregate and stores it in cloud via cloudStorageService.ts.
   6. Controller passes the aggregate to the view for display (mean + uncertainty).

⸻

4. Communication Between Contexts
   • TypeScript → Python ML: subprocess calls or REST API for inference/training.
   • Python ML → Cloud: reads/writes raw datasets and predictions.
   • TypeScript → Cloud: retrieves output summaries or metadata if needed.

⸻

5. Advantages
   • Clear separation of responsibilities by bounded context.
   • Aligns Python ML with ROI calculation without mixing UI logic.
   • TypeScript UI is modular and scalable.
   • Cloud storage abstracts data management for reproducibility.
   • Supports tactical DDD practices without overloading folder names with concepts.

[rahibAmin84]

Explore React ts front end and hapi for node api server packages

[rahibAmin84]

View itself gets decoupled into a react app modularized and containing its own client services to call APIs to render data in the view