guardrails-genie / guardrails_genie /guardrails /entity_recognition /regex_entity_recognition_guardrail.py
ash0ts's picture
Make PII evals work
3caf047
raw
history blame
6.87 kB
from typing import Dict, Optional, ClassVar, List
import weave
from pydantic import BaseModel
from ...regex_model import RegexModel
from ..base import Guardrail
import re
class RegexEntityRecognitionResponse(BaseModel):
contains_entities: bool
detected_entities: Dict[str, list[str]]
explanation: str
anonymized_text: Optional[str] = None
@property
def safe(self) -> bool:
return not self.contains_entities
class RegexEntityRecognitionSimpleResponse(BaseModel):
contains_entities: bool
explanation: str
anonymized_text: Optional[str] = None
@property
def safe(self) -> bool:
return not self.contains_entities
class RegexEntityRecognitionGuardrail(Guardrail):
regex_model: RegexModel
patterns: Dict[str, str] = {}
should_anonymize: bool = False
DEFAULT_PATTERNS: ClassVar[Dict[str, str]] = {
"EMAIL": r'\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Z|a-z]{2,}\b',
"TELEPHONENUM": r'\b(\+\d{1,3}[-.]?)?\(?\d{3}\)?[-.]?\d{3}[-.]?\d{4}\b',
"SOCIALNUM": r'\b\d{3}[-]?\d{2}[-]?\d{4}\b',
"CREDITCARDNUMBER": r'\b\d{4}[-\s]?\d{4}[-\s]?\d{4}[-\s]?\d{4}\b',
"DATEOFBIRTH": r'\b(0[1-9]|1[0-2])[-/](0[1-9]|[12]\d|3[01])[-/](19|20)\d{2}\b',
"DRIVERLICENSENUM": r'[A-Z]\d{7}', # Example pattern, adjust for your needs
"ACCOUNTNUM": r'\b\d{10,12}\b', # Example pattern for bank accounts
"ZIPCODE": r'\b\d{5}(?:-\d{4})?\b',
"GIVENNAME": r'\b[A-Z][a-z]+\b', # Basic pattern for first names
"SURNAME": r'\b[A-Z][a-z]+\b', # Basic pattern for last names
"CITY": r'\b[A-Z][a-z]+(?:[\s-][A-Z][a-z]+)*\b',
"STREET": r'\b\d+\s+[A-Z][a-z]+\s+(?:Street|St|Avenue|Ave|Road|Rd|Boulevard|Blvd|Lane|Ln|Drive|Dr)\b',
"IDCARDNUM": r'[A-Z]\d{7,8}', # Generic pattern for ID cards
"USERNAME": r'@[A-Za-z]\w{3,}', # Basic username pattern
"PASSWORD": r'[A-Za-z0-9@#$%^&+=]{8,}', # Basic password pattern
"TAXNUM": r'\b\d{2}[-]\d{7}\b', # Example tax number pattern
"BUILDINGNUM": r'\b\d+[A-Za-z]?\b' # Basic building number pattern
}
def __init__(self, use_defaults: bool = True, should_anonymize: bool = False, show_available_entities: bool = False, **kwargs):
patterns = {}
if use_defaults:
patterns = self.DEFAULT_PATTERNS.copy()
if kwargs.get("patterns"):
patterns.update(kwargs["patterns"])
if show_available_entities:
self._print_available_entities(patterns.keys())
# Create the RegexModel instance
regex_model = RegexModel(patterns=patterns)
# Initialize the base class with both the regex_model and patterns
super().__init__(
regex_model=regex_model,
patterns=patterns,
should_anonymize=should_anonymize
)
def text_to_pattern(self, text: str) -> str:
"""
Convert input text into a regex pattern that matches the exact text.
"""
# Escape special regex characters in the text
escaped_text = re.escape(text)
# Create a pattern that matches the exact text, case-insensitive
return rf"\b{escaped_text}\b"
def _print_available_entities(self, entities: List[str]):
"""Print available entities"""
print("\nAvailable entity types:")
print("=" * 25)
for entity in entities:
print(f"- {entity}")
print("=" * 25 + "\n")
@weave.op()
def guard(self, prompt: str, custom_terms: Optional[list[str]] = None, return_detected_types: bool = True, aggregate_redaction: bool = True, **kwargs) -> RegexEntityRecognitionResponse | RegexEntityRecognitionSimpleResponse:
"""
Check if the input prompt contains any entities based on the regex patterns.
Args:
prompt: Input text to check for entities
custom_terms: List of custom terms to be converted into regex patterns. If provided,
only these terms will be checked, ignoring default patterns.
return_detected_types: If True, returns detailed entity type information
Returns:
RegexEntityRecognitionResponse or RegexEntityRecognitionSimpleResponse containing detection results
"""
if custom_terms:
# Create a temporary RegexModel with only the custom patterns
temp_patterns = {term: self.text_to_pattern(term) for term in custom_terms}
temp_model = RegexModel(patterns=temp_patterns)
result = temp_model.check(prompt)
else:
# Use the original regex_model if no custom terms provided
result = self.regex_model.check(prompt)
# Create detailed explanation
explanation_parts = []
if result.matched_patterns:
explanation_parts.append("Found the following entities in the text:")
for entity_type, matches in result.matched_patterns.items():
explanation_parts.append(f"- {entity_type}: {len(matches)} instance(s)")
else:
explanation_parts.append("No entities detected in the text.")
if result.failed_patterns:
explanation_parts.append("\nChecked but did not find these entity types:")
for pattern in result.failed_patterns:
explanation_parts.append(f"- {pattern}")
# Updated anonymization logic
anonymized_text = None
if getattr(self, 'should_anonymize', False) and result.matched_patterns:
anonymized_text = prompt
for entity_type, matches in result.matched_patterns.items():
for match in matches:
replacement = "[redacted]" if aggregate_redaction else f"[{entity_type.upper()}]"
anonymized_text = anonymized_text.replace(match, replacement)
if return_detected_types:
return RegexEntityRecognitionResponse(
contains_entities=not result.passed,
detected_entities=result.matched_patterns,
explanation="\n".join(explanation_parts),
anonymized_text=anonymized_text
)
else:
return RegexEntityRecognitionSimpleResponse(
contains_entities=not result.passed,
explanation="\n".join(explanation_parts),
anonymized_text=anonymized_text
)
@weave.op()
def predict(self, prompt: str, return_detected_types: bool = True, aggregate_redaction: bool = True, **kwargs) -> RegexEntityRecognitionResponse | RegexEntityRecognitionSimpleResponse:
return self.guard(prompt, return_detected_types=return_detected_types, aggregate_redaction=aggregate_redaction, **kwargs)