Overview

Description

Intended Use

This validator ensures that any given text does not contain PII. This validator uses Microsoft's Presidio (https://github.com/microsoft/presidio) to detect PII in the text. If PII is detected, the validator will fail with a programmatic fix that anonymizes the text. Otherwise, the validator will pass.

Requirements

Dependencies:
- guardrails-ai>=0.4.0
- presidio-analyzer
- presidio-anonymizer

Installation

$ guardrails hub install hub://guardrails/detect_pii

Usage Examples

Validating string output via Python

# Import Guard and Validator
from guardrails.hub import DetectPII
from guardrails import Guard


# Setup Guard
guard = Guard().use(
    DetectPII, ["EMAIL_ADDRESS", "PHONE_NUMBER"], "exception"
)

guard.validate("Good morning!")  # Validator passes
try:
    guard.validate(
        "If interested, apply at not_a_real_email@guardrailsai.com"
    )  # Validator fails
except Exception as e:
    print(e)

Output:

Validation failed for field with errors: The following text in your response contains PII:
If interested, apply at not_a_real_email@guardrailsai.com

Validating JSON output via Python

In this example, we apply the validator to a string field of a JSON output generated by an LLM.

# Import Guard and Validator
from pydantic import BaseModel, Field
from guardrails.hub import DetectPII
from guardrails import Guard

# Initialize Validator
val = DetectPII(pii_entities=["EMAIL_ADDRESS", "PHONE_NUMBER"], on_fail="exception")


# Create Pydantic BaseModel
class UserHistory(BaseModel):
    name: str
    last_msg: str = Field(description="Last message sent by user", validators=[val])


# Create a Guard to check for valid Pydantic output
guard = Guard.from_pydantic(output_class=UserHistory)

# Run LLM output generating JSON through guard
try:
    guard.parse(
        """
    {
        "name": "John Smith",
        "last_msg": "My account isn't working. My username is not_a_real_email@guardrailsai.com"
    }
    """
    )
except Exception as e:
    print(e)

Output:

Validation failed for field with errors: The following text in your response contains PII:
My account isn't working. My username is not_a_real_email@guardrailsai.com

API Reference

__init__(self, pii_entities, on_fail="noop")

Parameters

pii_entities (Union[str, List(str)]): The types of PII entities to filter out. For a full list of entities look at https://microsoft.github.io/presidio/
on_fail (str, Callable): The policy to enact when a validator fails. If str, must be one of reask, fix, filter, refrain, noop, exception or fix_reask. Otherwise, must be a function that is called when the validator fails.

validate(self, value, metadata={}) -> ValidationResult

Note:

This method should not be called directly by the user. Instead, invoke guard.parse(...) where this method will be called internally for each associated Validator.
When invoking guard.parse(...), ensure to pass the appropriate metadata dictionary that includes keys and values required by this validator. If guard is associated with multiple validators, combine all necessary metadata into a single dictionary.