Dockerise the MCP Run Python server (#1837)

[rohit133]

Summary

This adds Docker support for the MCP Run Python server and sets up automated Docker image publishing via GitHub Actions.

Changes

Dockerfile:

• Adds a Dockerfile to run the MCP Run Python server using the official Deno image.
• Default command runs the server in SSE mode on port 3101, suitable for MCP client integration.

GitHub Actions Workflow:

• Adds docker-publish.yml to build and publish the Docker image to Docker Hub (or another registry) on push to main.
• Requires DOCKERHUB_USERNAME and DOCKERHUB_TOKEN secrets to be set in the repository.

Test Client:

• Adds test_mcp_client.py to verify the server is running and can execute Python code via the MCP protocol.

How to test

1. Build and run the Docker image:

docker build -t mcp-run-python ./mcp-run-python
docker run -d -p 3101:3101 mcp-run-python

2. Run the test client:

python3 mcp-run-python/test_mcp_client.py

You should see:

<status>success</status>
<output>
   hello world
</output>

[Kludex]

I think we can publish the images on github instead of dockerhub.

[rohit133]

I think we can publish the images on github instead of dockerhub.

I have updated it to publish the images on GitHub.

[Copilot]

Pull Request Overview

Adds Docker support and CI workflow for the MCP Run Python server and introduces an integration test client.

• Introduces a Dockerfile to containerize the MCP Run Python server using the official Deno image.
• Adds a GitHub Actions workflow to build and publish the Docker image on push to main.
• Adds an integration test client (test_mcp_client.py) to verify the server functionality.

Reviewed Changes

Copilot reviewed 4 out of 5 changed files in this pull request and generated 2 comments.

File	Description
mcp-run-python/test_mcp_client.py	Adds a simple client script to verify server SSE endpoint behavior
mcp-run-python/deno.json	Updates nodeModulesDir configuration value
mcp-run-python/Dockerfile	New Dockerfile for building and running the server container
.github/workflows/docker-publish.yml	Workflow to build and push the Docker image

Comments suppressed due to low confidence (3)

.github/workflows/docker-publish.yml:19

• The workflow logs in to GitHub Container Registry but tags the image for Docker Hub. Replace this step with a Docker Hub login using docker/login-action and the DOCKERHUB_USERNAME and DOCKERHUB_TOKEN secrets.

      - name: Login to GitHub Container Registry

.github/workflows/docker-publish.yml:30

• The tag yourdockerhubuser/mcp-run-python:latest is a placeholder. Update this to use your actual Docker Hub namespace or use a matrix/secret to avoid manual edits.

          tags: yourdockerhubuser/mcp-run-python:latest

mcp-run-python/test_mcp_client.py:14

• The test client only prints the output instead of asserting the expected values. Consider adding assertions (e.g., using assert content.text == 'hello world') so failures are detected automatically.

                print(content.text)

[Copilot]

The Deno server needs explicit network permissions to serve requests. Add --allow-net=0.0.0.0:3101 (or --allow-net) to the run command so the container can accept incoming connections.

Suggested change

	CMD ["run", "-N", "-R=node_modules", "-W=node_modules", "--node-modules-dir=auto", "jsr:@pydantic/mcp-run-python", "sse", "--port=3101"]
	CMD ["run", "--allow-net=0.0.0.0:3101", "-N", "-R=node_modules", "-W=node_modules", "--node-modules-dir=auto", "jsr:@pydantic/mcp-run-python", "sse", "--port=3101"]

[Copilot]

[nitpick] Copying the entire context can bloat the image. Consider adding a .dockerignore to exclude unnecessary files (e.g., tests, local configs) and reduce build time and image size.