Skip to main content

Component Breakdown

Page Outline

Component Breakdown​

The DIA service is organized into a modular structure that separates concerns and enables maintainability:

dia_service/
├─ app/
│  ├─ main.py                 # FastAPI application entry point, middleware setup, route registration
│  ├─ api/
│  │  ├─ endpoints.py         # REST API endpoint handlers for /simulate, /analyze, /overview, /history
│  ├─ core/
│  │  ├─ reasoner.py          # LLM-based explainers & summarizers for generating human-readable insights
│  │  ├─ analyzer.py          # Statistical & causal analysis engine for pattern detection
│  │  ├─ simulator.py         # Scenario engine with pluggable models for what-if analysis
│  │  ├─ history.py           # Decision recorder & replay functionality for audit and learning
│  ├─ infra/
│  │  ├─ kafka_client.py      # Async Kafka consumer/producer for event-driven communication
│  │  ├─ ga_client.py         # Governance Agent client for policy validation and token management
│  │  ├─ db.py                # Postgres connection pooling and S3 client helpers
│  │  ├─ vector_store.py      # Chroma / Pinecone wrapper for embedding storage and similarity search
│  ├─ models/
│  │  ├─ schemas.py           # Pydantic request/response models for API validation
│  └─ settings.py             # Configuration management (environment variables, secrets)
├─ tests/                      # Unit and integration tests
├─ Dockerfile                  # Container image definition
├─ helm-chart/                 # Kubernetes deployment manifests
└─ README.md                   # Service documentation

Architecture Rationale:

  • Separation of Concerns: The core/ directory contains business logic, api/ handles HTTP concerns, and infra/ manages external dependencies. This separation makes the codebase easier to test and maintain.

  • Pluggable Design: The simulator uses a pluggable model architecture, allowing different simulation engines (Monte Carlo, deterministic, ML-based) to be swapped without changing the core logic.

  • Async-First: The entire service is built on async/await patterns to handle high concurrency and non-blocking I/O operations, essential for real-time decision-making.