Assignment 1.5: MCP on HiPerGator

CIS 6930 Data Engineering - Spring 2026

Overview

This repository contains starter code for the in-class MCP activity on HiPerGator. You will build an MCP server that processes the dair-ai/emotion dataset from Hugging Face.

Setup on HiPerGator

1. Clone this repository

cd /blue/cis6930/YOUR_GATORLINK
git clone https://github.com/YOUR_USERNAME/cis6930sp26-assignment1.5.git
cd cis6930sp26-assignment1.5

2. Create environment file

cp .env.example .env
# Edit .env with your Hugging Face token

3. Install dependencies

module load mamba
uv sync
source .venv/bin/activate

Running the MCP Inspector

Option 1: HiPerGator Desktop (Recommended)

Using HiPerGator Desktop is much easier and avoids SSH tunneling complexity.

Go to https://ondemand.rc.ufl.edu
Navigate to Interactive Apps > HiPerGator Desktop
Request a session (1 hour, 4 GB memory, 2 CPUs)
Once the desktop launches, open a terminal

Run:

cd /blue/cis6930/YOUR_GATORLINK/cis6930sp26-assignment1.5
module load mamba
source .venv/bin/activate
mcp dev server.py

Open Firefox in the virtual desktop and go to http://localhost:6274

Option 2: SSH Tunneling from Your Laptop

If you prefer to use your local browser, you'll need to set up SSH tunneling.

Step 1: Start the MCP Inspector on a Compute Node

Run this command to start an interactive job with the MCP inspector:

ssh hpg "srun --partition=hpg-turin --account=cis6930 --qos=cis6930 \
    --cpus-per-task=4 --ntasks=1 --mem-per-cpu=4gb --time=1:00:00 \
    bash -c 'cd /blue/cis6930/YOUR_GATORLINK/cis6930sp26-assignment1.5 && \
    module load mamba && source .venv/bin/activate && mcp dev server.py'"

Step 2: Find Your Compute Node Name

In another terminal, find which compute node your job is running on:

squeue -u $USER

Look for the node name in the NODELIST column (e.g., c0702a-s2).

Step 3: Set Up the SSH Tunnel

The MCP Inspector runs on two ports:

6274 - Web interface
6277 - Proxy server

Open a new terminal on your laptop and run:

ssh -L 6274:localhost:6274 -L 6277:localhost:6277 \
    -J YOUR_GATORLINK@hpg.rc.ufl.edu YOUR_GATORLINK@COMPUTE_NODE

Example:

ssh -L 6274:localhost:6274 -L 6277:localhost:6277 \
    -J jsmith@hpg.rc.ufl.edu jsmith@c0702a-s2

The -J flag (ProxyJump) connects through the login node directly to the compute node.

Step 4: Open the Inspector

Open your browser and go to the url that was output by the the dev server. Is should look like http://localhost:6274/?MCP_PROXY_AUTH_TOKEN=e2a71ba1e83a76dd0ea24fed08b1d62413d5837fbea81cbc41a9233ae169f989 : http://localhost:6274?MCP_PROXY_AUTH_TOKEN={CODE}

You should see the MCP Inspector interface:

MCP Inspector Interface

Passwordless SSH Login

To avoid entering your password multiple times, set up SSH keys for HiPerGator: https://docs.rc.ufl.edu/access/ssh_keys/

Troubleshooting

Problem	Solution
"Connection refused"	Ensure you're using `-J` (ProxyJump) and forwarding both ports (6274 and 6277)
"Address already in use"	Kill processes: `lsof -ti:6274
Host key verification failed	Add `-o StrictHostKeyChecking=no` to the SSH command
Page won't load	Verify `mcp dev` is still running on the compute node

Using the MCP Inspector

Click the Tools tab in the left sidebar
Select a tool from the list (e.g., get_sample)
Fill in the parameter values
Click Run Tool
View the JSON response in the output panel

Tools to Test

Tool	Parameters
`get_sample`	`n: 3`
`count_by_emotion`	`emotion: "joy"`
`search_text`	`query: "happy"`, `limit: 5`
`analyze_emotion_distribution`	(no parameters)

Dataset

The dair-ai/emotion dataset contains English Twitter messages labeled with six emotions:

Label	Emotion
0	sadness
1	joy
2	love
3	anger
4	fear
5	surprise

Submission

After completing the activity, submit to Canvas:

Your outputs.txt file with tool outputs
A brief reflection (2-3 sentences)

cis6930sp26-assignment1.5