GluRPC Client - Reference Implementation

Official reference client implementation for GluRPC, the glucose prediction server. This package demonstrates best practices for integrating with GluRPC and provides both a Python SDK and a production-ready Gradio web interface for uploading CGM data and visualizing glucose predictions.

Features

🎯 Reference Client Capabilities

• 🌐 Full GluRPC API Coverage: Complete Python SDK with sync and async support
• 🖥️ Production-Ready Web Interface: Beautiful Gradio UI with real-time updates
• 📊 Advanced Visualization: Interactive Plotly charts with confidence intervals
• 🔍 Real-time Health Monitoring: Live server metrics with historical graphs
• ⚡ Smart Request Management: Automatic request cancellation and client-side optimization
• 💾 Format Conversion: Upload raw CGM files, get unified format with data quality warnings
• 🎚️ Interactive Exploration: Slider-based navigation through prediction samples
• 🔐 Secure Authentication: API key support for protected servers
• 📥 Export Capabilities: Download unified CSV and interactive HTML plots
• 🧪 Development Tools: Built-in test server for local development
• 📈 Performance Tracking: Client-side request timing and server queue monitoring

Installation

Using uv (Recommended)

uv sync

Using pip

pip install glucosedao-client

Quick Start

Quickest Start: Using uv Scripts

# Run client + server together (recommended for development)
uv run dev

# Run only the client (connect to existing server)
uv run client

# Run only the server
uv run server

Option 1: Connect to Existing Server

If you have a GluRPC server running elsewhere:

# Start the client interface
glucosedao-client launch

# Or with custom host/port
glucosedao-client launch --host 0.0.0.0 --port 7860

Then open your browser and:

1. Enter the server URL (e.g., http://your-server:8000)
2. Enter API key if required
3. Upload your CGM data file
4. View predictions!

Option 2: Launch with Local Test Server

For development or testing, start both client and server together:

# Start both client and server
glucosedao-client launch --with-server

# With custom ports
glucosedao-client launch --with-server --port 7860 --server-port 8001

Option 3: Start Only the Server

If you want to run just the server:

glucosedao-client server --host 0.0.0.0 --port 8000

CLI Commands

launch - Start the Web Interface

glucosedao-client launch [OPTIONS]

Options:

• --share: Create a public Gradio share link
• --host HOST: Hostname to bind to (default: 127.0.0.1)
• --port PORT: Port to bind to (default: 7860)
• --with-server: Start GluRPC server locally for testing
• --server-host HOST: GluRPC server host when using --with-server (default: 127.0.0.1)
• --server-port PORT: GluRPC server port when using --with-server (default: 8000)

Examples:

# Basic launch
glucosedao-client launch

# Launch with local test server
glucosedao-client launch --with-server

# Create public share link
glucosedao-client launch --share

# Custom configuration
glucosedao-client launch --host 0.0.0.0 --port 8080 --with-server --server-port 8001

server - Start Only the Server

glucosedao-client server [OPTIONS]

Options:

• --host HOST: Host to bind to (default: 0.0.0.0)
• --port PORT: Port to bind to (default: 8000)

Example:

glucosedao-client server --host 0.0.0.0 --port 8000

check - Health Check

glucosedao-client check [OPTIONS]

Options:

• --url URL: Server URL to check (default: http://localhost:8000)

Example:

glucosedao-client check --url http://your-server:8000

Python API

You can also use the client programmatically:

from glucosedao_client import GluRPCClient, GluRPCConfig

# Initialize client
config = GluRPCConfig(
    base_url="http://localhost:8000",
    api_key="your-api-key"  # Optional
)
client = GluRPCClient(config)

# Convert raw CGM file to unified format
convert_response = client.convert_to_unified("path/to/dexcom.csv")

if not convert_response.error:
    # Process the data
    process_response = client.process_unified(convert_response.csv_content)
    
    if not process_response.error:
        # Generate plot for a specific sample
        png_bytes = client.draw_plot(
            handle=process_response.handle,
            index=0
        )
        
        # Save plot
        with open("prediction.png", "wb") as f:
            f.write(png_bytes)

# Check server health
health = client.health()
print(f"Server status: {health.status}")
print(f"Models initialized: {health.models_initialized}")
print(f"Total requests: {health.total_requests_processed}")

# Clean up
client.close()

Async API

For async applications:

from glucosedao_client import AsyncGluRPCClient, GluRPCConfig

async def main():
    config = GluRPCConfig(base_url="http://localhost:8000")
    
    async with AsyncGluRPCClient(config) as client:
        # Convert file
        convert_response = await client.convert_to_unified("data.csv")
        
        # Process data
        process_response = await client.process_unified(
            convert_response.csv_content
        )
        
        # Get plot
        png_bytes = await client.draw_plot(
            process_response.handle,
            index=0
        )

Using the Web Interface

1. Configure Server: Enter the GluRPC server URL and API key (if required)
2. Check Health: Click "Check Server Health" to verify connectivity
3. Upload Data: Upload a CGM data file (Dexcom, LibreView, etc.)
4. View Predictions: The app will automatically process and show predictions
5. Explore Samples: Use the slider to view predictions for different time windows

Supported Data Formats

The client supports any CGM data format that GluRPC server supports:

• Dexcom
• LibreView (Freestyle Libre)
• Nightscout exports
• And more...

The server handles all format conversion automatically.

Configuration

Environment Variables

You can set default configuration using environment variables:

export GLURPC_URL="http://localhost:8000"
export GLURPC_API_KEY="your-api-key"

Logging

Logs are saved to logs/app_<timestamp>.log with both console and file output. The client uses Python's standard logging library with DEBUG level for files and INFO level for console output.

Development

Setting up Development Environment

# Clone the repository
git clone https://github.com/glucosedao/glucosedao_client.git
cd glucosedao_client

# Install dependencies with uv
uv sync

# Run in development mode
uv run python -m glucosedao_client.app

Running Tests

# Install test dependencies
uv add --dev pytest pytest-asyncio httpx

# Run tests
uv run pytest

Architecture

This reference implementation demonstrates proper client architecture:

Core Components

• client.py: Complete HTTP client for GluRPC API

  • Sync (GluRPCClient) and async (AsyncGluRPCClient) implementations
  • Type-safe response models with Pydantic
  • Connection pooling and timeout management
  • Error handling and retry logic

• app.py: Production-ready Gradio web interface (1000+ LOC)

  • Real-time server health monitoring with live graphs
  • Automatic request cancellation for smooth slider interaction
  • Data quality warnings from server validation
  • Format conversion and unified CSV export
  • Interactive Plotly visualizations
  • Client-side performance tracking

• server.py: Development server utilities

  • Local test server management
  • Process lifecycle handling

• cli.py: Typer-based CLI

  • Multiple launch modes (client-only, with-server, server-only)
  • Health check utilities
  • Flexible configuration options

Client-Server Interaction

The reference implementation demonstrates:

1. Efficient API Usage:

   • File upload → Convert to unified format
   • Process unified data → Get handle + warnings
   • Draw plots on demand with caching

2. Request Management:

   • Client-side request ID tracking
   • Automatic cancellation of stale requests
   • Server-side concurrent request handling

3. Real-time Monitoring:

   • Health polling with 1-second updates
   • Historical metrics visualization
   • Queue size and processing time tracking

4. User Experience:

   • Automatic plot generation on file upload
   • Smooth slider interaction with debouncing
   • Clear status messages and error handling
   • Download options for data and plots

Troubleshooting

Server Not Responding

# Check if server is running
glucosedao-client check --url http://your-server:8000

# Try starting a local test server
glucosedao-client launch --with-server

API Key Errors

Make sure you're providing the correct API key in the web interface or:

config = GluRPCConfig(
    base_url="http://localhost:8000",
    api_key="your-api-key"
)

Port Already in Use

If the default port is in use:

# Use a different port
glucosedao-client launch --port 8080

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

Apache 2.0 - See LICENSE file for details.

Links

• GluRPC Server
• Documentation
• Issue Tracker

Why This is a Reference Implementation

This client serves as the official example of how to properly integrate with GluRPC:

✅ Complete API Coverage: Shows how to use all GluRPC endpoints
✅ Best Practices: Demonstrates proper error handling, caching, and request management
✅ Type Safety: Uses Pydantic models for all API interactions
✅ Production Ready: Includes monitoring, logging, and performance tracking
✅ Async Support: Shows both sync and async usage patterns
✅ User Experience: Implements smart features like request cancellation and live updates

Use this client as a template when building your own GluRPC integrations!

Technology Stack

• Gradio 4.x - Interactive web interface with real-time updates
• httpx - Modern async HTTP client
• Typer - CLI framework with type hints
• Plotly - Interactive scientific visualization
• Pydantic 2 - Data validation and settings management
• cgm-format - CGM data format handling