Emotion Detection Web Application

Final Project for AI-Based Web Application Development and Deployment (Embeddable Watson AI). This app integrates the Watson NLP EmotionPredict API to analyze text and return five emotions (anger, disgust, fear, joy, sadness) plus the dominant emotion. Built with Flask.

---

✨ Features

• Watson NLP EmotionPredict integration (no local model required)
• Clean API wrapper with error handling (blank input → friendly message)
• Flask web UI (GET endpoint compatible with provided index.html + mywebscript.js)
• Unit tests for core behavior
• Passes Pylint static analysis (10/10 with docstrings)

---

📦 Project Structure

emd_ai/
├─ EmotionDetection/   
│ └─ emotion_detection.py # Core Watson API integration + error handling     
│ └─ __init__.py # Core Watson API integration + error handling    
├─ templates/     
│ └─ index.html    
├─ static/     
│ └─ mywebscript.js  
├─ server.py # Flask app (routes: / and /emotionDetector)    
├─ tests/    
│ └─ test_emotion_detection.py # Unit tests
├─requirements.txt
├─LICENSE
├─.gitignore
└─ README.md

---

Installation

Clone the repository:

git clone https://github.com/wkratos/emd_ai.git
cd emb_ai

Create and activate a virtual environment:

python3 -m venv .venv
source .venv/bin/activate

On Windows PowerShell:

python -m venv .venv
.\.venv\Scripts\Activate.ps1

Install the dependencies:

pip install -r requirements.txt

---

🚀 Run

python3 server.py

App runs on http://127.0.0.1:5000.

UI: open / to load index.html.

Response format (example):

For the given statement, the system response is 'anger': 0.006, 'disgust': 0.003,
'fear': 0.009, 'joy': 0.968 and 'sadness': 0.049. The dominant emotion is joy.

---

🧪 Tests

python3 -m unittest test_emotion_detection.py

or

python3 -m unittest tests/test_emotion_detection.py

or

python3 -m unittest discover -s tests

Expected: all tests pass for dominant emotions:

• glad → joy

• mad → anger

• disgusted → disgust

• sad → sadness

• afraid → fear

---

🛡️ Error Handling

Blank or whitespace-only input → backend returns all None from the detector and the server responds with:

Invalid text! Please try again!

If the Watson endpoint returns 400, detector returns the same all-None structure by design.

---

🧹 Static Code Analysis (Pylint)

pylint server.py

• Add module and function docstrings (already included)

• Use snake_case function names

• Avoid unused imports / long lines

---

🔧 Implementation Notes

Detector endpoint & headers

• URL: https://sn-watson-emotion.labs.skills.network/v1/watson.runtime.nlp.v1/NlpService/EmotionPredict

• Header: {"grpc-metadata-mm-model-id": "emotion_aggregated-workflow_lang_en_stock"}

• Payload: {"raw_document": {"text": ""}}

Server routes

• GET / → renders index.html

• GET /emotionDetector?textToAnalyze= → returns formatted plain text