Showcase/Relation Extraction
advancedtext

Relation Extraction

Identify and classify relationships between entities in text (e.g., works-for, located-in, married-to).

📝

text annotation

Configuration Fileconfig.yaml

# Relation Extraction Configuration
# Identify relationships between entities in text

annotation_task_name: "Relation Extraction"

data_files:
  - "data/sentences.json"

item_properties:
  id_key: "id"
  text_display_key: "display"

user_config:
  allow_all_users: true

annotation_schemes:
  - annotation_type: "radio"
    name: "relation_type"
    description: "What is the relationship between the highlighted entities?"
    labels:
      - name: "works_for"
        tooltip: "Subject works for/is employed by object"
        key_value: "w"
      - name: "located_in"
        tooltip: "Subject is located in object (place)"
        key_value: "l"
      - name: "founded_by"
        tooltip: "Subject was founded by object"
        key_value: "f"
      - name: "ceo_of"
        tooltip: "Subject is CEO/leader of object"
        key_value: "c"
      - name: "born_in"
        tooltip: "Subject was born in object (place)"
        key_value: "b"
      - name: "married_to"
        tooltip: "Subject is married to object"
        key_value: "m"
      - name: "subsidiary_of"
        tooltip: "Subject is a subsidiary of object"
        key_value: "s"
      - name: "part_of"
        tooltip: "Subject is part of object"
        key_value: "p"
      - name: "no_relation"
        tooltip: "No clear relationship between entities"
        key_value: "n"
      - name: "other"
        tooltip: "Relationship exists but not in list"
        key_value: "o"

  - annotation_type: "text"
    name: "other_relation"
    description: "If 'other', describe the relationship"
    show_if:
      field: "relation_type"
      value: "other"

  - annotation_type: "radio"
    name: "direction"
    description: "What is the direction of the relation?"
    labels:
      - name: "Subject → Object"
        tooltip: "Relation goes from Subject to Object"
      - name: "Object → Subject"
        tooltip: "Relation goes from Object to Subject"
      - name: "Bidirectional"
        tooltip: "Relation works both ways"

  - annotation_type: "likert"
    name: "confidence"
    description: "How confident are you in this relation?"
    size: 5
    min_label: "Not confident"
    max_label: "Very confident"

output: "annotation_output/"

Sample Datasample-data.json

[
  {
    "id": "rel_001",
    "text": "Tim Cook has been the CEO of Apple since 2011.",
    "subject": "Tim Cook",
    "object": "Apple",
    "subject_type": "PERSON",
    "object_type": "ORGANIZATION",
    "display": "**Text:** Tim Cook has been the CEO of Apple since 2011.\n\n**Subject:** [Tim Cook] (PERSON)\n**Object:** [Apple] (ORGANIZATION)"
  },
  {
    "id": "rel_002",
    "text": "The Eiffel Tower is located in Paris, France.",
    "subject": "Eiffel Tower",
    "object": "Paris",
    "subject_type": "LOCATION",
    "object_type": "LOCATION",
    "display": "**Text:** The Eiffel Tower is located in Paris, France.\n\n**Subject:** [Eiffel Tower] (LOCATION)\n**Object:** [Paris] (LOCATION)"
  }
]

// ... and 2 more items

Get This Design

View on GitHub

Clone or download from the repository

Quick start:

git clone https://github.com/davidjurgens/potato-showcase.git
cd potato-showcase/relation-extraction
potato start config.yaml

Details

Annotation Types

radiospan

Domain

nlpinformation-extraction

Use Cases

knowledge-graphrelation-classification

Tags

relationentitiesknowledge-graphextractionnlp

Found an issue or want to improve this design?

Open an Issue

Related Designs

Reading Comprehension QA

Evaluate question-answer pairs for reading comprehension by verifying answers and rating quality.

radiospan

Adverse Drug Event Extraction (CADEC)

Named entity recognition for adverse drug events from patient-reported experiences, based on the CADEC corpus (Karimi et al., 2015). Annotates drugs, adverse effects, symptoms, diseases, and findings from colloquial health forum posts with mapping to medical vocabularies (SNOMED-CT, MedDRA).

spanradio

Chemical-Disease Relation Extraction (BC5CDR)

Extract chemical-disease relations from biomedical literature. Based on BioCreative V CDR task. Identify chemical and disease entities, then annotate causal relationships between them (chemical induces disease).

spanradio