在Spacy中创建和加载自定义管道

pip install spacy

from spacy.tokens import Doc

class UppercaseComponent:
    def __init__(self):
        pass

    def __call__(self, doc: Doc) -> Doc:
        doc.text = doc.text.upper()
        return doc

import spacy
from .uppercase_component import UppercaseComponent

nlp = spacy.blank("en")
uppercase_component = UppercaseComponent()
nlp.add_pipe(uppercase_component, name="uppercase", first=True)

nlp.to_disk("custom_pipeline")

import spacy

nlp = spacy.load("custom_pipeline")

doc = nlp("This is a test.")
print(doc.text)  # 输出：THIS IS A TEST.