Jobs¶

A job is a named, schedulable workload: it declares what to materialize and, for a cron job, when. The core carries the declaration and compiles it to the same DAG every other entry point runs. Acting on the schedule is the job of the platform scheduler, outside this site.

Job¶

import interloper as il

job = il.Job(
    targets=[Shop(account_id="act_1"), Finance()],
    destinations=[warehouse],
    resources={"gcp": gcp_connection},
    tags=["daily"],
)
il.DAG(job).materialize(partition)

Field	Meaning
`targets`	Sources and assets to materialize. Their operations are flattened into the DAG.
`destinations`	Defaults for any target that declares none.
`resources`	Fill the empty resource slots of targets and destinations, by name then by type.
`enabled`	A disabled job is kept but not scheduled.
`tags`	Free-form labels.

Cascading works exactly like a source's: a job-level destination reaches every target without one, and a job-level connection reaches every target and destination with a matching empty slot.

JobState (next_run_at, last_run_at) is the job's machine-owned state, written by the scheduler.

CronJob¶

CronJob adds the trigger:

job = il.CronJob(
    cron="0 6 * * *",
    timezone="Europe/Berlin",
    lookback=3,
    offset=1,
    targets=[Shop(account_id="act_1")],
)

Field	Default	Meaning
`cron`	required	When the job runs, on the wall clock of `timezone`.
`timezone`	`"UTC"`	IANA zone the schedule is evaluated in. Unknown names are rejected.
`lookback`	`1`	How many partitions each run covers.
`offset`	`1`	How many partitions back from the current one the window ends.

For daily targets the defaults mean "yesterday, in the job's timezone". offset=3, lookback=3 covers the three days ending three days ago, for a vendor whose numbers settle late. The window itself is computed with TimePartitionWindow.lookback; hourly targets always use UTC windows because hour ids are UTC labels. Whether a job is partitioned is derived from its targets, never stored.

Jobs as specs¶

A job is the natural unit for a declarative run. Its spec lists targets with their configuration and the workload-level defaults:

path: interloper.job.base.Job
init:
  resources:
    gcp:
      key: google_cloud_connection
      init: { service_account_key: ${GCP_KEY} }
  destinations:
    - key: bigquery_destination
  targets:
    - key: facebook_ads
      init: { account_id: act_1, select: [campaigns, ads_stats] }

interloper run -f job.yaml --date 2026-01-15 reconstructs it and runs its DAG. See Specs and serialization and CLI. A scheduled job is the same document with key: cron_job and the trigger fields.

Relations¶

A job's relation vocabulary is target (sources and assets), destination and resource. A target is an orchestration pointer, not an input: deleting a target shrinks the job rather than blocking the deletion. Relation semantics are described in the component model.