This project is currently under development.
Stave is a fast, lightweight, extensible web-based text annotation and visualization tool designed to support a wide range of data types and NLP tasks. Stave offers the following features:
- Semantic Annotation: supports both template and custom data types
- Multi-document Annotation: supports cross-document annotations and coreference
- Customizable Interface: supports creation of task-specific interface with independently developed plugins
- Machine-Human Collaboration: keeps human in the loop to verify and correct machine suggestions
- Easy Integration: supports seamless integration with pre-built NLP workflows
- Safe Data Serialization: supports a JSON-serializable format for easy data saving, loading and distribution
Stave was originally developed and is actively contributed by Petuum in collaboration with other institutes. A mirror of this repository is maintained by Petuum Open Source.
Packages coming soon ...
The project is tested on:
Python 3.6+ Django 3.0.4 yarn 1.22.4
cd simple-backend- make sure your
pythoncommand is using python3, or use a virtual env by:python3 -m venv venvcreate virtual env (skip if already created). venv/bin/activateuse virtual env
- make sure django is installded, or install it by:
python -m pip install Django
./start-dev.sh- This script will create an example sqlite3 DB from SQL:
db.sqlite3
- This script will create an example sqlite3 DB from SQL:
After the server starts, then simply
- yarn && yarn start
- go to http://127.0.0.1:8000
The default username/password for the demonstration data is admin/admin


