Skip to main content

​
Parameters

ParameterTypeDefaultDescription
agent_under_testUnion[Agent, Graph, Team]RequiredThe agent, graph, or team to be performance tested
taskUnion[Task, List[Task]]RequiredThe task or list of tasks to execute during performance testing
num_iterationsint10Number of measurement iterations to run
warmup_runsint2Number of warmup runs before measurements begin

​
Functions

​
__init__

Initialize the PerformanceEvaluator. Parameters:
  • agent_under_test (Union[Agent, Graph, Team]): The agent, graph, or team to be performance tested
  • task (Union[Task, List[Task]]): The task or list of tasks to execute during performance testing
  • num_iterations (int): Number of measurement iterations to run
  • warmup_runs (int): Number of warmup runs before measurements begin
Raises:
  • TypeError: If agent_under_test is not an Agent, Graph, or Team instance
  • TypeError: If task is not a Task or list of Task objects
  • ValueError: If num_iterations is not a positive integer
  • ValueError: If warmup_runs is not a non-negative integer

​
run

Execute the end-to-end performance profiling workflow. This method will:
  1. Perform warmup runs to ensure steady-state measurements
  2. Execute the component for a set number of iterations, capturing high-precision latency and memory metrics for each run
  3. Aggregate the metrics into detailed statistics
  4. Return a final PerformanceEvaluationResult object
Parameters:
  • print_results (bool): If True, prints a formatted summary of the results to the console
Returns:
  • PerformanceEvaluationResult: A PerformanceEvaluationResult object with detailed statistics

​
_execute_component

Internal helper that executes the component to be tested. Parameters:
  • agent (Union[Agent, Graph, Team]): The agent, graph, or team to execute
  • task (Union[Task, List[Task]]): The task or tasks to execute

​
_calculate_stats

Calculate and return a dictionary of statistics for a list of numbers. Parameters:
  • data (List[float]): List of numerical data to calculate statistics for
Returns:
  • Dict[str, float]: Dictionary containing statistical measures (average, median, min, max, std_dev)

​
_aggregate_results

Aggregate raw run data into the final result object. Parameters:
  • run_results (List[PerformanceRunResult]): List of performance run results
Returns:
  • PerformanceEvaluationResult: The aggregated performance evaluation result
Print a rich, formatted table of the performance results. Parameters:
  • result (PerformanceEvaluationResult): The performance evaluation results to print

​
Description

The main user-facing profiler for measuring the latency and memory footprint of Upsonic agents, graphs, or teams.
⌘I