Overview

Description

Intended Use

This validator enforces that an LLM generated output belongs to a subset of acceptable choices.

Requirements

Dependencies:
- guardrails-ai>=0.4.0

Installation

$ guardrails hub install hub://guardrails/valid_choices

Usage Examples

Validating string output via Python

In this example, we’ll use the validator to check if the output belongs to a set of choices: OpenAI, Anthropic, Cohere.

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

# Use the Guard with the validator
guard = Guard().use(
    ValidChoices, choices=["OpenAI", "Anthropic", "Cohere"], on_fail="exception"
)

# Test passing response
guard.validate("OpenAI")

try:
    # Test failing response
    guard.validate("Google")
except Exception as e:
    print(e)

Output:

Validation failed for field with errors: Value Google is not in choices ['OpenAI', 'Anthropic', 'Cohere'].

Validating JSON output via Python

We can use the same validator to confirm that a field in a JSON output belongs to a set of categories. We will use the validator to check for allowed pet types: cat, dog, bird.

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

val = ValidChoices(choices=["cat", "dog", "bird"], on_fail="exception")


# Create Pydantic BaseModel
class PetInfo(BaseModel):
    pet_name: str
    pet_type: str = Field(description="Type of pet", validators=[val])


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

# Run LLM output generating JSON through guard
guard.parse(
    """
    {
            "pet_name": "Caesar",
            "pet_type": "dog"
    }
    """
)

try:
    # Run LLM output generating JSON through guard
    guard.parse(
        """
        {
                "pet_name": "Caesar",
                "pet_type": "fish"
        }
        """
    )
except Exception as e:
    print(e)

Output:

Validation failed for field with errors: Value fish is not in choices ['cat', 'dog', 'bird'].

API Reference

__init__(self, on_fail="noop")

Parameters

choices (List[str]): The list of acceptable choices.
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.

__call__(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.