ARCHITECTURE.md

suRxit Architecture# System Architecture

suRxit is built as a set of loosely coupled microservices, each responsible for a distinct stage in the risk intelligence pipeline. This modularity enables scalability, maintainability, and rapid iteration.MedKG-Rx is built as a set of loosely coupled microservices, each responsible for a distinct stage in the clinical document-to-risk pipeline. The architecture ensures scalability, modularity, and compliance with healthcare data standards.

Microservices Flow## Microservice Flow

1. OCR: Converts scanned documents to machine-readable text.
2. NER: Extracts medical entities from text.
3. Standardizer: Maps entities to standard vocabularies (e.g., SNOMED, RxNorm).
4. Knowledge Graph: Links entities and relationships.
5. FeatureGen: Generates features for downstream models.
6. GNN: Predicts risk using graph neural networks.
7. Recommender: Suggests interventions or next steps.
8. MedLM: Provides LLM-based explanations.
9. Risk: Aggregates and explains risk scores.
10. API-Gateway: Orchestrates service calls and enforces security.
11. Frontend: Presents results and explanations to clinicians.

Diagram## Architecture Diagram

mermaidmermaid

flowchart LRflowchart LR

A[OCR] --> B[NER]    A[OCR] --> B[NER]

B --> C[Standardizer]    B --> C[Standardizer]

C --> D[Knowledge Graph]    C --> D[KG]

D --> E[FeatureGen]    D --> E[FeatureGen]

E --> F[GNN]    E --> F[GNN]

F --> G[Recommender]    F --> G[Recommender]

G --> H[MedLM]    G --> H[MedLM]

H --> I[Risk]    H --> I[Risk]

I --> J[API-Gateway]    I --> J[API-Gateway]

J --> K[Frontend]    J --> K[Frontend]