Extracting Text Relations with Flair#s Relation Extraction Model

This function uses Flair's relation extraction model to extract relations from a given text. It first loads the pre-trained relation extraction model, then creates a Sentence object from the text, and performs relation extraction on the sentence. Finally, it extracts and prints out the relations.

Technology Stack : Flair, Relation Extraction Model

Code Type : Function

Code Difficulty : Intermediate

                
                    
import random
import flair
import flair.models
import flair.data
import flair.data.relation_extraction

def extract_relations(text):
    """
    Extracts relations from a given text using Flair's relation extraction model.
    """
    # Load the pre-trained relation extraction model
    relation_model = flair.models.relation_extraction.relation_extraction_model('en')
    
    # Create a Sentence object from the text
    sentence = flair.data.Sentence(text)
    
    # Perform relation extraction on the sentence
    relation_model.predict(sentence)
    
    # Extract and print the relations
    relations = [(rel.relation, rel.head, rel.dep, rel.dep_head) for rel in sentence.relations]
    for rel in relations:
        print(f"Relation: {rel[0]}, Head: {rel[1]}, Dep: {rel[2]}, Dep Head: {rel[3]}")
        
    return relations