Chapel Support for MCP

A Model-Context-Protocol (MCP) server for the Chapel programming language, providing tools for working with Chapel code, accessing primers and examples, and integrating Chapel functionality with AI assistants and other tools.

What is Chapel?

Chapel is an open-source parallel programming language designed for productive parallel computing at scale. It aims to improve the programmability of parallel computers while matching or beating the performance and portability of current programming models like MPI, OpenMP, and CUDA.

Features

This MCP server provides the following Chapel support functionality:

• Chapel Primer Access: Browse and access Chapel's educational primer examples
• Code Compilation: Compile Chapel code directly through the API
• Linting: Check Chapel code for style and best practices using chplcheck and apply automatic fixes
• Smart CHPL_HOME Detection: Automatically locate Chapel's installation directory

Prerequisites

• Python 3.13 or higher
• Chapel programming language installed (see Chapel installation guide)
• (Optional) chplcheck for linting functionality

Installation

1. Clone this repository:

   git clone <repository-url>
   cd chapel-support
   

2. Create and activate a virtual environment with UV:

   uv venv
   source .venv/bin/activate  # On Windows: .venv\Scripts\activate
   

3. Synchronize the environment with project dependencies:

   uv sync
   

Configuration

The MCP server needs to know the location of your Chapel installation (CHPL_HOME). It will try to find it in this order:

1. From the CHPL_HOME environment variable
2. From a .env file in the project root
3. By running chpl --print-chpl-home if the Chapel compiler is in your PATH

To use a .env file, create one in the project root with:

CHPL_HOME=/path/to/your/chapel/installation

See .env.example for a template.

Usage

Running the MCP Server

uv run chapel-support.py

This will start the MCP server in stdio transport mode using your virtual environment.

Integrating with AI Assistants or Tools

To use this MCP server with AI assistants or other tools, configure them to connect to this server. For example, in a client configuration file:

{
  "context_servers": {
    "chapel-support": {
      "command": {
        "path": "uv",
        "args": [
          "run",
          "--directory",
          "/path/to/chapel-support",
          "chapel-support.py"
        ],
        "env": {}
      },
      "settings": {}
    }
  }
}

Note: Adjust the directory path to the location of your chapel-support installation.

Available Tools

list_primers()

Gets the list of available Chapel primers.

Returns: A list of paths to primer files relative to CHPL_HOME.

get_primer(path: str)

Retrieves the content of a specific Chapel primer.

Parameters:

• path: The path to the primer, as returned by list_primers()

Returns: The content of the primer as a string.

compile_program(program_text: str, program_name: str = "program.chpl")

Compiles a Chapel program.

Parameters:

• program_text: The Chapel code to compile
• program_name: Optional name for the program file (default: "program.chpl")

Returns: A tuple containing:

• Success status (boolean)
• Compiler output/errors (string)

list_chapel_lint_rules()

Lists all available Chapel linting rules from chplcheck.

Returns: A list of dictionaries with rule information:

• name: Rule name
• description: Rule description
• is_default: Whether the rule is enabled by default

lint_chapel_code(program_text: str, program_name: str = "program.chpl", fix: bool = False, custom_rules: Optional[List[str]] = None)

Lints Chapel code and optionally applies fixes.

Parameters:

• program_text: The Chapel code to lint
• program_name: Optional name for the program file (default: "program.chpl")
• fix: Whether to apply automatic fixes (default: False)
• custom_rules: List of specific rules to enable (default: None, uses default rules)

Returns: A dictionary containing:

• warnings: String containing linting warnings
• fixed_code: The fixed code if fix=True
• error: Error message if something went wrong
• stats: Statistics about the linting process

Contributing

Contributions are welcome! Please feel free to submit pull requests or open issues.