🎤 Emotion Detection System

Project Overview

This project captures microphone audio recordings, performs emotion classification, stores the results in MongoDB, and displays them on a real-time updating dashboard!

The system runs as 3 coordinated containers controlled by Docker Compose:

┌─────────────────────────────┐      ┌──────────────────────────────┐
│     Flask Web App           │      │      ML Client               │
│  - Audio recording UI       │      │  - Watches for new audio     │
│  - Uploads .webm            │─────▶│  - Extracts features         │
│  - Dashboard of results     │◀─────│  - Predicts emotion          │
└─────────────┬───────────────┘      │  - Writes to MongoDB         │
              │                      └─────────────┬────────────────┘
              │                                    │
              ▼                                    ▼
       ┌─────────────────────────────────────────────────┐
       │                   MongoDB                       │
       │ - Stores filename, mean_f0, mean_rms, emotion   │
       │ - Queried by web app for live dashboard         │
       └─────────────────────────────────────────────────┘

Team Members

Name	GitHub
(kaiyuan wu)	https://github.com/qiexian-mf
(Harrison Coon)	https://github.com/hoc2006-code
(Jaylon McDuffie)	https://github.com/jm9908
(Sam Murshed)	https://github.com/SamMurshed

---

Running the Project

1. Install Dependencies

Make sure you have:

• Docker
• Docker Compose
• Python 3.10+ (for linting/tests only)

---

2. Start All Containers

docker compose up --build

This will launch:

• emotion_web — Flask server
• emotion_ml — Machine Learning container
• emotion_mongodb — MongoDB

When running successfully:

• Web app → http://localhost:5050
• MongoDB exposed on → 27017

---

3. View the Dashboard

Visit: http://localhost:5050

From here, you can:

• Click Start Recording
• Record 3 seconds of audio
• View live emotion classification results
• See the latest emotion + emoji
• Watch the table auto-refresh every 5 seconds

---

Testing

Run all tests locally

export PYTHONPATH=$PWD
export TESTING=1
pytest -v

Run tests inside Docker

A helper script is included:

./run_tests.sh

This script:

• Stops old containers
• Rebuilds everything
• Runs CI tests inside the web-app container
• Prints results

---

Continuous Integration (CI)

Two GitHub Actions workflows run automatically:

ml-client-ci.yml

• Lints ML client code using black & pylint
• Runs pytest for ML unit tests
• Reported with a badge in the README

web-app-ci.yml

• Lints Flask server
• Runs Flask unit tests
• Ensures formatting & PEP8 compliance

CI triggers on pull request merge to main.

---

Database Setup (MongoDB)

MongoDB runs via Docker using:

mongodb:
  image: mongo:6.0
  ports:
    - "27017:27017"
  volumes:
    - mongo_data:/data/db

No extra configuration is needed.

---

Environment Variables

Both services use environment variables from docker-compose:

Web App:

MONGO_URI=mongodb://mongodb:27017/emotiondb
UPLOAD_DIR=/data/uploads

ML Client:

MONGO_URI=mongodb://mongodb:27017/emotiondb
AUDIO_DIR=/data/uploads

A template for secrets should include:

.env.example