irgroup/labelstudio-to-fonduer

This small module connects Label Studio with Fonduer by creating a fonduer labeling function for gold labels from a label studio export. Documentation: https://irgroup.github.io/labelstudio-to-fonduer/

Label Studio to Fonduer

Label Studio to Fonduer is a small library to transfer annotations between Label Studio and Fonduer. By that, documents annotated in Label Studio can be used as gold labels in Fonduer and vice versa annotations made in Fonduer can easily be accessed by humans in Label Studio.

Label Studio and especially Fonduer create their own representation of an HTML document.

Therefore, the documents need to be converted into a structure that does not need to be modified by Label Studio or Fonduer.

Installation:

pip install git+https://github.com/irgroup/labelstudio-to-fonduer.git#egg=labelstudiotofonduer\&subdirectory=src

Documentation:

The documentation and a more extensive user guide with examples can be found here.

Examples:

From Fonduer to Label Studio:

train_cands = candidate_extractor.get_candidates()

converter.create_export(candidates=train_cands, fonduer_export_path="import.json")

From Label Studio to Fonduer:

converter = ToFonduer(label_studio_export=export, fonduer_session=session)

labeler.apply(
    docs=docs,
    lfs=[[converter.is_gold]],
    table=GoldLabel,
)