Streamlit Annotator

Download the package from PyPI:

Try the demo on Streamlit Cloud:

Install with pip

pip install st-annotator

st-annotator is a Streamlit component usefull to annotate text, expecially for NLP and Argument Mining purposes. Based on the original project Streamlit Annotation Tools of rmarquet21.

Features:

• 📍 Smart positioning that stays within screen bounds
• 📊 Shows text content, label category, position, and all custom metadata
• 🎨 Matches your custom color scheme
• ⚡ Instant display on hover, disappears on mouse leave
• 🔧 Supports strings, numbers, booleans, lists, and objects

New Features

• Key parameter to text_annotator function
• A special button to show all the annotations together
• 🆕 Hover Popup with Metadata: Hover over annotations to see detailed information including custom metadata

Metadata Support

You can now add custom metadata to each annotation that will appear in a hover popup:

labels = {
    "Sentiment": [
        {
            "start": 0,
            "end": 20,
            "label": "This is amazing!",
            "metadata": {
                "confidence": 0.95,
                "emotion": "Joy",
                "intensity": "High",
                "source": "Customer feedback"
            }
        }
    ]
}

Quick Start

Run the example.py file:

streamlit run examples/example.py