如何使用Spacy Matcher松散地匹配中间的单词

Question

from spacy.matcher import Matcher
matcher = Matcher(nlp.vocab, validate=True)
pattern = [{'LOWER': 'play'},
           {'OP': '*'}, {'OP': '!', 'LOWER': 'store'},
           {'LOWER': {'IN': ["game", "pacman"]}}
           ]
matcher.add('HUNTING', None, pattern)

def extract_patterns(nlp_doc, matcher):
    result_spans = []
    matches = matcher(nlp_doc)
    print("matches:", len(matches))
    for match_id, start, end in matches:
        span = nlp_doc[start:end]
        result_spans.append(span)
    return result_spans

text = ('play store game. \n play with pacman') 
doc = nlp(text)
extract_patterns(doc, matcher=matcher)

以上代码的返回结果如下。[与吃豆人一起玩，玩商店游戏。和吃豆子一起玩]

但是预期的结果是[和pacman一起玩]

是否可以使用Spacy Matcher？

Answer 1

您的样式应该看起来像

pattern = [
   {'LOWER': 'play'},
   {'OP': '!', 'LOWER': 'store'},
   {'LOWER': {'IN': ["game", "pacman"]}}
]

将导致单个匹配，['play with pacman']，因为它不敏感地匹配play令牌大小写，然后不敏感地匹配store以外的令牌，然后不敏感地匹配game或pacman大小写。 >

完整测试代码段：

import spacy
from spacy.matcher import Matcher

nlp = spacy.load("en_core_web_sm")
matcher = Matcher(nlp.vocab, validate=True)
pattern = [
   {'LOWER': 'play'},
   {'OP': '!', 'LOWER': 'store'},
   {'LOWER': {'IN': ["game", "pacman"]}}
]
matcher.add('HUNTING', None, pattern)

def extract_patterns(nlp_doc, matcher):
    result_spans = []
    matches = matcher(nlp_doc)
    print("matches:", len(matches))
    for match_id, start, end in matches:
        span = nlp_doc[start:end]
        result_spans.append(span.text)
    return list(set(result_spans))

doc = nlp("play store game. \n play with pacman")
extract_patterns(doc, matcher=matcher)