AnnApi - Latin Text Annotation Tool

A Flask-based web application for annotating Latin text with linguistic features using spaCy's Latin language model.

Features

• Text Tokenization: Automatically segments Latin text into tokens using spaCy
• Linguistic Annotation: Automatically generates annotations including:
  • Part-of-Speech (POS) tags
  • Lemmatization
  • Morphological features (case, gender, number, aspect, tense, mood, person, voice, verbForm)
• Interactive Editing: Modify generated annotations directly in the web interface
• JSON Export: Save annotations as structured JSON files with timestamps

Requirements

• Python 3.7+
• Flask 2.3.3
• spaCy 3.5.0
• Additional dependencies listed in requirements.txt

Installation

1. Clone or download the repository

2. Install required packages:

   pip install -r requirements.txt

3. Download the Latin language model:

   python -m spacy download la_core_web_lg

Project Structure

AnnApi/
├── api/
│   └── main.py              # Flask application and core NLP functions
├── templates/
│   └── annotate.html        # Web interface template
├── requirements.txt         # Python dependencies
└── README.md               # This file

Usage

1. Start the Flask application:

   python api/main.py

2. Open your browser and navigate to http://localhost:5000

3. Enter Latin text in the text area and click "Annotate"

4. Review and edit the generated annotations as needed

5. Click "Save Modifications as JSON" to export the annotated data

How It Works

Tokenization

The application uses spaCy's Latin model to tokenize input text into individual tokens.

Annotation

Each token receives automatic annotations:

• All tokens: POS tag, lemma
• Nouns, adjectives, determiners, pronouns, conjunctions, particles: case, gender, number
• Verbs and auxiliaries: aspect, tense, verbForm, voice, mood, person

Editing & Export

Annotations are displayed in an editable form where users can modify any field before saving. Export saves all modified annotations as JSON with a timestamp-based filename.

Dependencies

• Flask: Web framework for creating the web interface
• spaCy: Natural Language Processing library with pre-trained Latin model
• Materialize CSS: Frontend framework for responsive UI

Output Format

Annotations are saved as JSON with the following structure:

[
  {
    "token": "word",
    "pos": "NOUN",
    "lemma": "lemma",
    "case": "Nom",
    "gender": "Fem",
    "number": "Sing",
    ...
  }
]

License

Please see the LICENSE of the application.