Creating Custom Entity Patterns with spaCy#s EntityRuler

This function uses the EntityRuler component of spaCy to create custom entity recognition patterns and adds these patterns to the spaCy NLP pipeline.

Technology Stack : spaCy, EntityRuler

Code Type : Function

Code Difficulty : Intermediate

                
                    
import spacy
from spacy.tokens import Span
from spacy.pipeline import EntityRuler

def create_entity_patterns(entity_patterns):
    """
    Create an EntityRuler component for spaCy and add custom entity patterns.
    
    :param entity_patterns: List of dictionaries, each containing 'label' and 'pattern' keys.
    :return: None
    """
    nlp = spacy.load("en_core_web_sm")
    ruler = EntityRuler(nlp)
    patterns = [EntityRuler.PATTERN(pattern=p["pattern"], label=p["label"]) for p in entity_patterns]
    ruler.add_patterns(patterns)
    nlp.add_pipe(ruler, before="ner")
    return nlp

entity_patterns = [
    {"label": "ORG", "pattern": "Microsoft"},
    {"label": "GPE", "pattern": "New York"},
    {"label": "PERSON", "pattern": "John Doe"}
]

nlp = create_entity_patterns(entity_patterns)
doc = nlp("Microsoft was founded by Bill Gates in New York. John Doe works there.")