WellMate Health Assistant System

Overview

WellMate is a comprehensive health assistant system that integrates physical health monitoring and mental wellness support. The system provides intelligent health consultation services through AI-powered agents, supporting both mental health counseling and physical health guidance.

Note: This repository was originally the backend repository. The frontend application was migrated from https://github.com/Mint-green/WellMate-App.git and the agent services were migrated from https://github.com/mooooohan/Health-Agent.git.

System Architecture & Technology Stack

WellMate adopts a microservices architecture with the following core components and technology stack:

Architecture Diagram

flowchart TD
    %% First Layer: Client App at the top
    A[📱 <b style="font-size:1.2em">WellMate Client App</b><br/>User Authentication, Health Chat Interface, Data Visualization & Management]
    
    %% Second Layer: Backend and MySQL on the same level
    B[🚀 <b style="font-size:1.2em">Python Backend Service</b><br/>API Gateway, Authentication, Session & AI Integration]
    D[💾 <b style="font-size:1.2em">MySQL Database</b><br/>User Data & Chat History]
    
    %% Third Layer: Language-Bind, Physical Agent and Mental Agent on the same level
    C[🔤 <b style="font-size:1.2em">Language-Bind Service</b><br/>Mood Analysis & Data Tagging]
    E1[💪 <b style="font-size:1.2em">Physical Health Agent</b><br/>Physical Health Advice<br/>Exercise & Diet Recommendations]
    E2[🧠 <b style="font-size:1.2em">Mental Health Agent</b><br/>Emotional Support<br/>Psychological Counseling]
    
    %% Style definitions
    classDef client fill:#e3f2fd,stroke:#1976d2
    classDef backend fill:#e8f5e8,stroke:#388e3c
    classDef database fill:#fff3e0,stroke:#f57c00
    classDef external fill:#fce4ec,stroke:#c2185b
    
    class A client
    class B,C backend
    class D database
    class E1,E2 external
    
    %% Data flow arrows
    A -->|API Request| B
    B <-->|Data Operations| D
    B <-->|Data Pre-Processing| C
    B -->|Health Consultation| E1
    B -->|Mental Support| E2

Loading

Data Flow Sequence:

1. Client Request: WellMate App sends REST API requests to Python backend
2. Authentication: Backend validates JWT tokens and user permissions
3. Database Operations: Backend performs CRUD operations on MySQL database
4. Data Pre-processing: Backend sends data to Language-Bind service for multimodal data recognition and tagging (optional)
5. AI Service Integration: Backend calls appropriate Coze agents based on health domain, using pre-processed data if available
6. Response: Processed results returned to client through backend

Language-Bind Service Functionality:

• Multimodal Data Recognition: Text, audio, and image data processing
• Emotion Analysis: Mood detection and emotional state classification
• Data Tagging: Automatic labeling and categorization of health data
• Pre-processing Pipeline: Data cleaning, normalization, and feature extraction

Frontend Application

• Technology Stack: uni-app x framework + TypeScript (UTS)
• Platform: Cross-platform mobile application (iOS, Android, Web)
• Core Features:
  • User authentication and profile management
  • Health conversation interface with AI assistants
  • Health data visualization and management
  • Session history management
  • Multi-language support (English/Chinese)

Backend Service

• Technology Stack: Flask + Gunicorn + Python 3.11+
• Deployment: Docker containerized deployment
• Core Features:
  • RESTful API gateway and JWT authentication
  • Session and conversation management
  • AI service integration and routing
  • Database operations and data persistence

AI Service Agents

Mental Health Agent

• Service: Professional mental health and emotional support
• Technology Stack: FastAPI + Coze platform AI agent
• Features: Emotional counseling, psychological support, stress management, emotional analysis

Physical Health Agent

• Service: Physical health consultation and advice
• Technology Stack: Coze platform AI agent
• Features: Exercise recommendations, dietary guidance, body metrics analysis

Language-Bind Service

• Service: Advanced text processing and emotional analysis
• Technology Stack: Custom multimodal AI processing
• Features: Emotion recognition, data tagging, classification

Database & Infrastructure

• Database: MySQL 8.0 (Docker container)
• Container Runtime: Docker
• Orchestration: Docker container management scripts (individual service management)
• Monitoring: Application logging
• Deployment: CI/CD pipeline

Core Features

🤖 Intelligent Health Consultation

• AI-driven health conversations with professional agents
• Support for both physical and mental health topics
• Context-aware conversation management
• Real-time streaming responses

💬 Session Management

• Automatic session creation and management
• Complete conversation history storage
• Session recovery and continuity support
• Multi-device session synchronization

📊 Health Data Management

• Comprehensive health metrics tracking
• Data visualization and analysis
• Privacy-focused data storage
• Cross-platform data synchronization

🔐 Security & Authentication

• JWT-based token authentication
• Secure password hashing
• Role-based access control
• Encrypted data transmission

🌐 Multi-language Support

• Support for English and Chinese
• Automatic language detection
• Localized user interface
• Cultural adaptation

Project Structure

WellMate/
├── backend/                 # Python Flask backend service
├── frontend-app/            # uni-app x frontend application
├── mental_agent/            # Mental health AI service agent
├── physical_agent/          # Physical health AI service agent
├── LanguageBind+Audio_to_text/  # Training code and models
├── README.md                # Project documentation (English)
└── README_zh.md             # Project documentation (Chinese)

Prerequisites

• Python 3.11+
• MySQL 8.0+
• Docker 20.10+ (optional)
• Node.js (for frontend development)