The scientific taggers aims at proposing subject classification for scholarly publications. So the input is a list of publications, only some metadata are taken into account (title, journal title, abstract,keywords and MeSH), not the full-text for the time being. Several classifications are implemented:
- Biomedical subject classification based on Fields of Research documented here. The results can be reproduced with this Jupyter notebook
- Pascal and Francis tags
- Sustainable Development Goals (SDG) from UN, based on Dataset of search queries to map scientific publications to the UN sustainable development goals, Bordignon, 2021
git clone git@github.com:dataesr/scientific-tagger.git
cd scientific-tagger
docker compose pull && docker compose down && docker system prune -f && docker compose upIn your browser, you now have :
- scientific-tagger : http://localhost:5004/
It uses semver.
To create a new release, do
make release VERSION=x.x.x