将SpaCy PhraseMatcher保存到磁盘

pip install spacy dill
python -m spacy download en_core_web_sm

import spacy
print(spacy.__version__)  # 检查版本一致性

import spacy
from spacy.matcher import PhraseMatcher
import dill

# 创建PhraseMatcher
nlp = spacy.blank("en")
phrases = ["hello world", "good morning"]
patterns = [nlp(text) for text in phrases]
matcher = PhraseMatcher(nlp.vocab, attr="LOWER")
matcher.add("PHRASES", patterns)

# 保存到磁盘
with open("matcher.dill", "wb") as file:
    dill.dump(matcher, file)

# 从磁盘加载
with open("matcher.dill", "rb") as file:
    loaded_matcher = dill.load(file)

# 使用加载的匹配器
doc = nlp("hello world! good morning everyone.")
matches = loaded_matcher(doc)
for match_id, start, end in matches:
    print(doc[start:end])