Skip to content

sweatybridge/streamlit-text-label

Repository files navigation

streamlit-text-label

Custom streamlit component for labelling text selections, based on Label Studio. Demo

Usage

Prerequisite:

  • python >= 3.7
from streamlit_text_label import label_select

selections = label_select(body="lorem ipsum", labels=["Noun", "Verb"])

More details in example directory.

Develop

docker-compose up --build

Or manually, start the frontend

cd streamlit_text_label/frontend
npm install
npm run start

Followed by the example app

pip install -r example/requirements.txt
RELEASE=DEV streamlit run example/app.py

Publish

cd component/frontend
npm run build
cd ../..
python setup.py sdist bdist_wheel

Known Issues

  • Labelling standalone whitespaces will not work.
  • Trailing and leading whitespace will be included in selection but not rendered.
  • Label colors are generated from the first 6 characters of its own MD5 hash, which may clash and have low contrast (native HSL doesn't work fully with LS).
  • Each selection renders best with 1 label only. If you need multiple labels, create multiple selections.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors