Pandas MCP Server

🚀 Pandas MCP Server

This repository contains a server implemented using the Model Context Protocol (MCP), capable of handling CSV files, executing Pandas code, and generating interactive charts (bar charts and pie charts).

🚀 Quick Start

Prerequisites

• Python 3.11 or higher
• Install the required packages: pip install -r requirements.txt

Steps

1. Install the required Python packages: pip install -r requirements.txt
2. Start the server: Run the main script (e.g., server.py).
3. Call the functions:
   • Pass the file path to the read_metadata function to get file information.
   • Use run_pandas_code to perform Pandas analysis.
   • Call the chart generation functions (bar_chart_to_html, pie_chart_to_html) to create interactive charts.

✨ Features

read_metadata

• Description: Load a CSV file and return its column structure and sample data.
• Parameters:
  • file_path: The path to the CSV file.
• Return Value:
  • A dictionary containing the columns and sample data of the CSV file.
• Notes:
  • Automatically detect the file encoding and delimiter.
  • Limit the file size to 100MB to prevent excessive memory usage.

run_pandas_code

• Description: Execute Pandas code provided as a string.
• Parameters:
  • code: A string containing the Pandas code to be executed.
• Return Value:
  • A dictionary containing the execution results and variables created during execution.
• Security Notes:
  • Prevent the execution of blacklisted operations, such as os., sys., subprocess., open(, exec(, eval(, import os, import sys.
  • Provide detailed error information and suggestions to help users solve problems.

bar_chart_to_html

• Description: Generate an interactive HTML bar chart using the Chart.js template.
• Parameters:
  • categories: A list of category names for the x-axis.
  • values: A list of numerical values for the y-axis.
  • title: The chart title (default: "Interactive Chart").
• Return Value:
  • A dictionary containing the file path and status information.
• Example:

{
    "categories": ["Electronics", "Clothing", "Home Appliances"],
    "values": [120000, 85000, 95000],
    "title": "Q1 Sales by Product Category"
}

pie_chart_to_html

• Description: Generate an interactive HTML pie chart using the Chart.js template.
• Parameters:
  • labels: A list of label names for each pie chart slice.
  • values: A list of numerical values for each slice.
  • title: The chart title (default: "Interactive Pie Chart").
• Return Value:
  • A dictionary containing the file path and status information.
• Example:

{
    "labels": ["Electronics", "Clothing", "Home Appliances"],
    "values": [120000, 85000, 95000],
    "title": "Q1 Sales Distribution"
}

line_chart_to_html

• Description: Generate an interactive HTML line chart using the Chart.js template.
• Parameters:
  • labels: A list of label names for the x-axis.
  • datasets: A list of datasets containing the following:
    • label: The name of the dataset.
    • data: A list of numerical values.
  • title: The chart title (default: "Interactive Line Chart").
• Return Value:
  • A dictionary containing the file path and status information.
• Example:

{
    "status": "SUCCESS",
    "filepath": "/absolute/path/to/chart_1713443200.html"
}

🔧 Technical Details

Workflow

1. Data Loading: Use read_metadata to load and parse the CSV file.
2. Data Analysis: Perform complex Pandas analysis through run_pandas_code.
3. Chart Generation:
   • Call the corresponding chart function based on the analysis results.
   • Example response:

     {
         "status": "SUCCESS",
         "filepath": "/absolute/path/to/chart_1713443200.html"
     }
     

Through these steps, you can effectively process data, perform analysis, and generate interactive visualization results.