pydistsim.metrics.MetricCollector

class MetricCollector

Bases: NodeNetworkObserver, AlgorithmObserver, SimulationObserver

MetricCollector is an observer that collects metrics from the simulation.

### Custom metrics Extend this class and implement the desired event methods to collect custom metrics. For registering events, call the _add_metric method. Add an instance of your custom collector to the simulation observers for it to work.

Even so, you can use the events attribute to register the events you want to listen to. This includes new custom events that you would trigger from an algorithm.

Example of implementing a custom metric collector:

class ExampleCustomMetricCollector(MetricCollector):
    events = ["example_custom_event"]

    class CustomMetricEventType(StrEnum):
        "Definition if this enum is optional. It helps to avoid typos in the event names."
        EXAMPLE_CUSTOM_EVENT_ZERO = "EXAMPLE_CUSTOM_EVENT_ZERO"
        ...

    def on_example_custom_event(self, a, b, c):
        self._add_metric(
            self.CustomMetricEventType.EXAMPLE_CUSTOM_EVENT_ZERO,
            {"a": a, "b": b, "c": c}
        )

Methods

__init__
create_report	Returns a dictionary with the collected metrics.
notify
on_added	Called when the observer is added to an observable object.
on_algorithm_finished
on_algorithm_started
on_message_delivered
on_message_sent
on_network_changed
on_node_status_changed
on_sim_state_changed
on_state_changed
on_step_done

Attributes

events

create_report()

Returns a dictionary with the collected metrics.

Currently, it returns the number of messages sent, messages delivered, and quantity of status changes. Future versions may include more metrics, with better reporting capabilities.