README.md

PIE Dataset Card for "ChemProt"

This is a PyTorch-IE wrapper for the ChemProt Huggingface dataset loading script.

Data Schema

There are three versions of the dataset supported, chemprot_full_source, chemprot_shared_task_eval_source and chemprot_bigbio_kb.

ChemprotDocument for chemprot_source and chemprot_shared_task_eval_source

defines following fields:

• text (str)
• id (str, optional)
• metadata (dictionary, optional)

and the following annotation layers:

• entities (annotation type: LabeledSpan, target: text)
• relations (annotation type: BinaryRelation, target: entities)

ChemprotBigbioDocument for chemprot_bigbio_kb

defines following fields:

• text (str)
• id (str, optional)
• metadata (dictionary, optional)

and the following annotation layers:

• passages (annotation type: LabeledSpan, target: text)
• entities (annotation type: LabeledSpan, target: text)
• relations (annotation type: BinaryRelation, target: entities)

See here for the annotation type definitions.

Document Converters

The dataset provides predefined document converters for the following target document types:

• pie_documents.documents.TextDocumentWithLabeledSpansAndBinaryRelations for ChemprotDocument
• pie_documents.documents.TextDocumentWithLabeledSpansBinaryRelationsAndLabeledPartitions for ChemprotBigbioDocument

See here for the document type definitions.