Python Endpoints Demo

This example demonstrates how to create and use Python-based endpoints in MXCP.

Features Demonstrated

1. Basic Python Functions

analyze_numbers - Statistical analysis with various operations
create_sample_data - Database operations from Python

2. Async Functions

process_time_series - Demonstrates async Python endpoint

3. Database Access

Using mxcp.runtime.db to execute SQL queries
Parameter binding for safe SQL execution

Running the Examples

In a terminal, test the endpoints:

# Create sample data
mxcp run tool create_sample_data --param table_name=test_data --param row_count=100

# Analyze numbers
mxcp run tool analyze_numbers --param numbers="[1, 2, 3, 4, 5]" --param operation=mean

# Process time series (async function)
mxcp run tool process_time_series --param table_name=test_data --param window_days=7

Or, if you prefer, you can also start the MXCP server and use any MCP client to call the tools:

mxcp serve

Project Structure

python-demo/
├── mxcp-site.yml         # Project configuration
├── python/               # Python modules
│   └── data_analysis.py  # Python endpoint implementations
├── tools/                # Tool definitions
│   ├── analyze_numbers.yml
│   ├── create_sample_data.yml
│   └── process_time_series.yml
└── README.md

Key Concepts

Language Declaration: Set language: python in the tool definition
Function Names: The function name must match the tool name
Return Types: Functions must return data matching the declared return type
Database Access: Use db.execute() for SQL queries
Async Support: Both sync and async functions are supported

1.7 KiB Raw Permalink Blame History