pie/drugprot

DrugProt数据集是用于生物医学信息提取的任务，包含两个版本：drugprot_source和drugprot_bigbio_kb。drugprot_source版本包括文本、ID、元数据、标题和摘要字段，以及实体和关系注释层。drugprot_bigbio_kb版本包括文本、ID和元数据字段，以及段落、实体和关系注释层。数据集提供了预定义的文档转换器，用于将数据转换为不同的文档类型。

The DrugProt dataset is for biomedical information extraction tasks and consists of two versions: drugprot_source and drugprot_bigbio_kb. The drugprot_source version includes fields for text, ID, metadata, title, and abstract, as well as annotation layers for entities and relations. The drugprot_bigbio_kb version includes fields for text, ID, and metadata, and annotation layers for passages, entities, and relations. The dataset provides predefined document converters to transform the data into different document types.

原始信息汇总

PIE Dataset Card for "DrugProt"

数据集版本

• drugprot_source
• drugprot_bigbio_kb

数据集结构

DrugprotDocument for drugprot_source

• 字段:
  • text (str)
  • id (str, optional)
  • metadata (dictionary, optional)
  • title (str, optional)
  • abstract (str, optional)
• 注释层:
  • entities (注释类型: LabeledSpan, 目标: text)
  • relations (注释类型: BinaryRelation, 目标: entities)

DrugprotBigbioDocument for drugprot_bigbio_kb

• 字段:
  • text (str)
  • id (str, optional)
  • metadata (dictionary, optional)
• 注释层:
  • passages (注释类型: LabeledSpan, 目标: text)
  • entities (注释类型: LabeledSpan, 目标: text)
  • relations (注释类型: BinaryRelation, 目标: entities)

文档转换器

• 针对DrugprotDocument提供转换器: pie_modules.documents.TextDocumentWithLabeledSpansAndBinaryRelations
• 针对DrugprotBigbioDocument提供转换器: pie_modules.documents.TextDocumentWithLabeledSpansBinaryRelationsAndLabeledPartitions