ReactiveDag.Job (reactive_dag v0.17.0-rc.25)

Copy Markdown View Source

What every library-provided Oban job needs from its arguments.

A plan is built from resource modules at runtime, so it cannot ride in a job argument — Oban serialises args to JSON, and a %Plan{} is neither JSON nor cheap to rebuild blindly. So a job carries the name of the function that builds one, and this resolves it.

Shared rather than duplicated because two workers now need it (ScanWorker and ReprocessWorker), and a divergence between them would be the kind of bug that only appears on the job that runs least often.

Summary

Functions

drain_opts(args)

The recompute:/key_rule: a job drains with, defaulting to the Node strategies a DSL-authored graph uses.

module(m)

A module name from a job argument, with or without the Elixir. prefix.

plan(arg1, worker)

The plan for a job: its own "plan_mfa" argument, else the configured default.

Functions

drain_opts(args)

@spec drain_opts(map()) :: keyword()

The recompute:/key_rule: a job drains with, defaulting to the Node strategies a DSL-authored graph uses.

module(m)

@spec module(String.t()) :: module()

A module name from a job argument, with or without the Elixir. prefix.

plan(arg1, worker)

@spec plan(map(), module()) :: ReactiveDag.Plan.t()

The plan for a job: its own "plan_mfa" argument, else the configured default.

Raises with the fix rather than a MatchError, because "no plan" is a configuration mistake and the message is the only place it can be explained.