@psyqueue/plugin-workflows

DAG-based workflow orchestration for PsyQueue. Define multi-step workflows with dependencies and conditional branching.

Installation

npm install @psyqueue/plugin-workflows

Usage

import { workflows, workflow } from '@psyqueue/plugin-workflows'

const q = new PsyQueue()
q.use(sqlite({ path: ':memory:' }))

const wf = workflows()
q.use(wf)

const def = workflow('order.process')
  .step('validate', async (ctx) => ({ valid: true }))
  .step('charge', async (ctx) => ({ chargeId: 'ch_123' }), { after: 'validate' })
  .step('ship', async (ctx) => ({ tracking: 'TRK-1' }), { after: 'charge' })
  .build()

wf.engine.registerDefinition(def)

await q.start()
const workflowId = await q.enqueue('order.process', { items: ['widget'] })

Configuration

The workflows() factory takes no options. Configure by registering workflow definitions.

Depends on: backend

Step Options

| Option | Type | Description | |--------|------|-------------| | after | string \| string[] | Step dependencies | | when | (ctx) => boolean | Conditional execution | | compensate | JobHandler | Saga compensation handler |

Exports

• workflows() -- Plugin factory
• workflow(name) -- Workflow builder factory
• WorkflowBuilder -- Builder class
• WorkflowEngine -- Engine class
• WorkflowStore -- State store

Documentation

See Workflow Plugins for detailed usage, parallel execution, and conditional branching.

License

MIT