Cadences¶

Cadences in TOM Toolkit are re-observation strategies that determine when to observe a target again based on its observation history.

Configured strategies¶

TOM_CADENCE_STRATEGIES = [
    "tom_observations.cadences.retry_failed_observations.RetryFailedObservations",
    "tom_observations.cadences.resume_cadence_after_failure.ResumeCadenceAfterFailure",
    "iac_tom.cadences.mycadence.MyCadenceStrategy",
]

Strategies included in TOM Toolkit¶

RetryFailedObservations¶

Automatically retries observations that have failed. Useful for recovering from transient technical failures (bad weather, telescope fault).

ResumeCadenceAfterFailure¶

Resumes the cadence pattern after a failure, computing the next observation from the last successful state rather than the last failed attempt.

MyCadenceStrategy (customisable)¶

File: iac_tom/cadences/mycadence.py

Empty placeholder for implementing a custom cadence strategy for the RIA.

from tom_observations.cadences import CadenceStrategy

class MyCadenceStrategy(CadenceStrategy):
    name = "My Cadence Strategy"
    description = "Custom cadence for RIA"

    def run(self):
        # Implement cadence logic
        pass

Example: time-since-last-observation cadence¶

A typical implementation for transient follow-up might be:

def run(self):
    from datetime import timedelta
    from django.utils import timezone

    for observation_record in self.observation_group.observation_records.all():
        last_obs = observation_record.created
        interval = timedelta(hours=self.parameters.get("interval_hours", 24))
        if timezone.now() - last_obs >= interval:
            # Submit new observation
            self.submit_observation(observation_record.target)

Relationship with the Scheduler¶

TOM Toolkit cadences and the Scheduler module are complementary mechanisms:

Cadences manage long-term re-observation (days, weeks).
The Scheduler optimises the agenda for a specific night.

In the RIA TOM workflow, the Scheduler can consider the last_scheduled field of SchedulingProfile to increase the priority of targets that have not been observed for a long time.