Emotion Detector

Tiny Python utility that detects emotions in English text using a Hugging Face Transformers pipeline.

• Model: j-hartmann/emotion-english-distilroberta-base
• Task: text-classification (returns all emotion scores via top_k=None)

Setup

Create/activate a virtual environment, then install dependencies:

python -m venv venv
source venv/bin/activate

pip install --upgrade pip
pip install transformers torch

Notes:

• The first run downloads the model weights (requires internet).
• If you’re on Apple Silicon, torch wheels are supported; if install issues occur, use the official PyTorch install command for your platform.

Usage

The core function is emotion_detector(text) in emotion_detection.py.

Call from Python

from emotion_detection import emotion_detector

print(emotion_detector("I’m really excited about this!"))

One-liner

python -c "from emotion_detection import emotion_detector; print(emotion_detector('I’m really excited about this!'))"

Output

For non-empty input, the function returns a dictionary mapping emotion labels to scores (rounded to 2 decimals), for example:

{
  "joy": 0.87,
  "anger": 0.02,
  "sadness": 0.03
}

For blank input ("" / None), it returns:

({"error": "Input cannot be blank"}, 400)

Project structure

• emotion_detection.py: Loads the model pipeline and exposes emotion_detector(text).

Model / credits

• Hugging Face model: j-hartmann/emotion-english-distilroberta-base
• Transformers library: transformers