NLP4LRL Annotation Platform

A collaborative data annotation platform for NLP research in low-resource language settings. Built on top of doccano and customized for the NLP4LRL project.

Live platform: annotate.nlp4lrl.com

Features

• Collaborative annotation with role-based access (admin / annotator)
• Multi-language support — including low-resource and morphologically rich languages
• Annotation types: NER / sequence labeling, text classification, sequence-to-sequence
• Mobile support
• Dark theme
• RESTful API

Running Locally (from source)

Requirements: Python 3.10+, Poetry, Node 18+, Yarn

git clone https://github.com/NLP4LRL/doccano.git
cd doccano

Backend:

cd backend
poetry install
poetry run python manage.py migrate
poetry run python manage.py create_roles
poetry run python manage.py create_admin --noinput \
  --username admin --email admin@example.com --password yourpassword
poetry run python manage.py runserver

In a second terminal, start the task worker:

cd backend
poetry run celery --app=config worker --loglevel=INFO --concurrency=1

Frontend:

cd frontend
yarn install
yarn dev   # http://localhost:3000

Production Deployment (Docker Compose)

See Doccano_Deployment_Notes_NLP4LRL.md for the full deployment guide including HTTPS setup, nginx configuration, and the Docker image build workflow for the VPS.

cp docker/.env.example docker/.env
# Edit docker/.env with your credentials
docker compose -f docker/docker-compose.prod.yml up -d

The production stack includes: Django + Gunicorn, PostgreSQL, RabbitMQ, Celery, Flower, and Nginx.

Deploying Frontend Changes

The VPS (2 GB RAM) cannot build Docker images. Build locally and push to Docker Hub:

docker buildx build \
  --platform linux/amd64 \
  -f docker/Dockerfile.nginx \
  -t billofosuhene/doccano-frontend:nlp4lrl \
  --push \
  .

Then on the VPS:

docker compose -f docker/docker-compose.prod.yml pull nginx
docker compose -f docker/docker-compose.prod.yml up -d nginx

Active Branch

UI customizations live on the frontend/nlp4lrl-ui-rebrand branch. See Doccano_Deployment_Notes_NLP4LRL.md for a full list of changes made.

Attribution

This platform is a customized fork of doccano by Hiroki Nakayama et al. If you use doccano in your research, please cite:

@misc{doccano,
  title={{doccano}: Text Annotation Tool for Human},
  url={https://github.com/doccano/doccano},
  note={Software available from https://github.com/doccano/doccano},
  author={
    Hiroki Nakayama and
    Takahiro Kubo and
    Junya Kamura and
    Yasufumi Taniguchi and
    Xu Liang},
  year={2018},
}

Contact

For questions about the NLP4LRL annotation platform, visit nlp4lrl.com.