Base Metric Decorator

class ibm_watsonx_gov.metrics.base_metric_decorator.BaseMetricDecorator(api_client: APIClient = None, configuration: AgenticAIConfiguration = None, compute_real_time: bool = True, metric_results: list[AgentMetricResult] = [], execution_counts: dict[str, dict[str, int]] = {}, nodes_being_run: dict[str, Set[str]] = {}, lock: allocate_lock = None)

Bases: object

Base class for all metric decorators

compute_helper(*, func: Callable, args: tuple, kwargs: dict[str, Any], configuration: AgenticAIConfiguration, metrics: list[GenAIMetric], metric_inputs: list[EvaluatorFields], metric_outputs: list[EvaluatorFields], metric_references: list[EvaluatorFields] = [], metric_groups: list[MetricGroup] = []) → dict

Helper method for computing metrics.

Does the following:

  1. Computes node latency metric, and appends the result to the AgenticEvaluation.metric_results attribute.

  2. Calls the original node.

  3. Computes the list of metrics given, and appends the result to the AgenticEvaluation.metric_results attribute.

  4. Returns the result of the original node without any changes.

Parameters:

  • func (Callable) – The node on which the metric is to be computed

  • args (tuple) – The tuple of positional arguments passed to the node

  • kwargs (dict[str, Any]) – The dictionary of keyword arguments passed to the node

  • configuration (AgenticAIConfiguration) – The node specific configuration

  • metrics (list[GenAIMetric]) – The list of metrics to compute.

  • metric_inputs (list[EvaluatorFields]) – The list of inputs for the metric.

  • metric_outputs (list[EvaluatorFields]) – The list of outputs for the metric.

  • metric_references (list[EvaluatorFields], optional) – The optional list of references for the metric. Defaults to [].

Raises:

ValueError – If the record id field is missing from the node inputs.

Returns:

The result of the wrapped node.

Return type:

dict

validate(*, func: Callable, metrics: list[GenAIMetric], valid_metric_types: tuple[Any])