Skip to content

System_Architecture.md

Latest commit

 

History

History
59 lines (49 loc) · 2.83 KB

System_Architecture.md

File metadata and controls

59 lines (49 loc) · 2.83 KB

MedCore AI: System Architecture Component

MedCore AI utilizes a robust, multi-layered architecture designed to bridge the gap between Software Engineering and Clinical Health Informatics. The system is built for scalability, context-awareness, and high-performance data processing.

1. High-Level System Architecture

The following diagram illustrates the interaction between the user interface, the application state, and the generative AI layer.

graph TB
    subgraph "Frontend Layer (Vite + React)"
        UI["UI Components (Glassmorphism Dashboard)"]
        RT["React Tabs (Dashboard, Checker, Portal)"]
        CH["Recharts Analytics Engine"]
    end

    subgraph "Application Logic Layer"
        SM["React State Management (Hooks)"]
        CE["Context Extractor (Vitals, Trends, History)"]
    end

    subgraph "AI & Data Layer"
        SDK["@google/generative-ai SDK"]
        API["Gemini 2.5 Flash API"]
        KB["Session Diagnosis Archive"]
    end

    User((Patient/Doctor)) <--> UI
    UI <--> RT
    RT <--> SM
    SM <--> CH
    SM <--> CE
    CE -- "Context-Informed Prompt" --> SDK
    SDK <--> API
    API -- "Structured JSON Response" --> SM
    SM --> KB
Loading

2. Context-Aware Data Flow

MedCore AI's unique "Holistic Diagnostic" capability is achieved through a precisely orchestrated data flow:

  1. Data Collection: The system continuously monitors patient vitals (Heart Rate, BP, SpO2, Glucose).
  2. State Aggregation: React states capture historical trends (e.g., bpmTrend, bpTrend).
  3. Prompt Engineering: Upon symptom input, the Context Extractor compiles these metrics into a comprehensive medical snapshot.
  4. Generative Analysis: The Gemini 2.5 Flash model performs longitudinal analysis, correlating current symptoms with historical trends.
  5. Persistence: The resulting structured JSON is rendered in real-time and archived in the consultation history.

3. Interdisciplinary Integration Points

Layer Domain Integration Detail
Logic Layer Computer Science State encapsulation and async AI service handling.
Analytics Layer Engineering Data normalization for multi-metric chart rendering.
Intelligence Layer Medicine Generative clinical reasoning using longitudinal patient data.
UI Layer Psychology/Design Glassmorphic interface meant to build user trust and reduce clinical anxiety.

4. Component Definitions

  • Analytics Engine: Uses Recharts to transform raw vital arrays into readable trend-lines.
  • Context Extractor: A logic component that serializes multi-dimensional health data for LLM consumption.
  • Consultation Archive: A session-persistent storage layer for medical traceability.
  • AI Sentinel: The Gemini 2.5 core configured with specific safety constraints for medical emergency detection.