Presidio Python 教程：敏感信息识别、误报漏报与脱敏结果验证

python3.11 -m venv .venv
source .venv/bin/activate
python -m pip install 'presidio-analyzer==2.2.364' 'presidio-anonymizer==2.2.364'

 presidio_analyzer  Pattern, PatternRecognizer
 presidio_anonymizer  AnonymizerEngine
 presidio_anonymizer.entities  OperatorConfig


RULES = [
    PatternRecognizer(
        supported_entity=,
        patterns=[Pattern(, , )],
    ),
    PatternRecognizer(
        supported_entity=,
        patterns=[Pattern(, , )],
    ),
]
OPERATORS = {
    : OperatorConfig(, {: }),
    : OperatorConfig(, {: }),
}
ENGINE = AnonymizerEngine()


 ():
    hits = [h  rule  RULES  h  rule.analyze(text=text, entities=rule.supported_entities)
             h.score >= ]
     (hits, key= h: (h.start, h.end, h.entity_type))


 ():
    previous_end = 
     hit  (hits, key= h: (h.start, h.end)):
          ( <= hit.start < hit.end <= (text)):
             ValueError()
         hit.start < previous_end:
             ValueError()
         hit.entity_type   OPERATORS:
             ValueError()
        previous_end = hit.end


 ():
    hits = detect(text)
    validate_spans(text, hits)
    result = ENGINE.anonymize(text=text, analyzer_results=hits, operators=OPERATORS)
    
     result.text, {: (hits), : }


 ():
    
     (value  value  output  value  values):
         AssertionError()


 ():
    predicted = {(h.entity_type, h.start, h.end)  h  detect(text)}
    expected = (expected)
     {: (predicted & expected), : (predicted - expected),
            : (expected - predicted)}