nemoguardrails.library.patronusai.actions

Module Contents

Functions

Name	Description
`check_guardrail_pass`	Check if evaluations in the Patronus API response pass based on the success strategy.
`parse_patronus_lynx_response`	Parses the response from the Patronus Lynx LLM and returns a tuple of:
`patronus_api_check_output`	Check the user message, bot response, and/or provided context
`patronus_api_check_output_mapping`	Mapping for patronus_api_check_output.
`patronus_evaluate_request`	Make a call to the Patronus Evaluate API.
`patronus_lynx_check_output_hallucination`	Check the bot response for hallucinations based on the given chunks
`patronus_lynx_check_output_hallucination_mapping`	Mapping for patronus_lynx_check_output_hallucination.

Data

log

API

nemoguardrails.library.patronusai.actions.check_guardrail_pass(
    response: typing.Optional[dict],
    success_strategy: typing.Literal['all_pass', 'any_pass']
) -> bool

Check if evaluations in the Patronus API response pass based on the success strategy. “all_pass” requires all evaluators to pass for success. “any_pass” requires only one evaluator to pass for success.

nemoguardrails.library.patronusai.actions.parse_patronus_lynx_response(
    response: str
) -> typing.Tuple[bool, typing.Union[typing.List[str], None]]

Parses the response from the Patronus Lynx LLM and returns a tuple of:

Whether the response is hallucinated or not.
A reasoning trace explaining the decision.

nemoguardrails.library.patronusai.actions.patronus_api_check_output(
    llm_task_manager: nemoguardrails.llm.taskmanager.LLMTaskManager,
    context: typing.Optional[dict] = None
) -> dict

async

Check the user message, bot response, and/or provided context for issues based on the Patronus Evaluate API

nemoguardrails.library.patronusai.actions.patronus_api_check_output_mapping(
    result: dict
) -> bool

Mapping for patronus_api_check_output.

Block (return True) if “pass” is False.

nemoguardrails.library.patronusai.actions.patronus_evaluate_request(
    api_params: dict,
    user_input: typing.Optional[str] = None,
    bot_response: typing.Optional[str] = None,
    provided_context: typing.Optional[typing.Union[str, typing.List[str]]] = None
) -> typing.Optional[dict]

async

Make a call to the Patronus Evaluate API.

Returns a dictionary of the API response JSON if successful, or None if a server error occurs.

Server errors will cause the guardrail to block the bot response

Raises a ValueError for client errors (400-499), as these indicate invalid requests.

nemoguardrails.library.patronusai.actions.patronus_lynx_check_output_hallucination(
    llm_task_manager: nemoguardrails.llm.taskmanager.LLMTaskManager,
    context: typing.Optional[dict] = None,
    patronus_lynx_llm: typing.Optional[nemoguardrails.types.LLMModel] = None,
    kwargs = {}
) -> dict

async

Check the bot response for hallucinations based on the given chunks using the configured Patronus Lynx model.

nemoguardrails.library.patronusai.actions.patronus_lynx_check_output_hallucination_mapping(
    result: dict
) -> bool

Mapping for patronus_lynx_check_output_hallucination.

Block (return True) if “hallucination” is True.

nemoguardrails.library.patronusai.actions.log = logging.getLogger(__name__)