Adobe Experience Platform Batch Ingestion

arazzo: 1.0.1
info:
  title: Adobe Experience Platform Batch Ingestion
  summary: Create a dataset, open a batch against it, then poll the batch until it finishes loading.
  description: >-
    Drives the Adobe Experience Platform batch ingestion lifecycle. The workflow
    creates a dataset in the Data Catalog, opens an import batch that targets the
    dataset id, and then polls the batch status until it leaves the loading
    state. Each step inlines the sandbox header, bearer token, and API key so the
    flow can be read and executed without opening the underlying OpenAPI
    description.
  version: 1.0.0
sourceDescriptions:
- name: experiencePlatformApi
  url: ../openapi/adobe-experience-platform-api-openapi.yml
  type: openapi
workflows:
- workflowId: batch-ingestion
  summary: Create a dataset, open an ingestion batch, and poll it to completion.
  description: >-
    Creates a dataset bound to an existing schema, creates a parquet import
    batch for that dataset, and polls the batch status until ingestion finishes.
  inputs:
    type: object
    required:
    - authorization
    - apiKey
    - sandboxName
    - datasetName
    - schemaId
    - datasetId
    properties:
      authorization:
        type: string
        description: Bearer access token.
      apiKey:
        type: string
        description: Adobe API key for the x-api-key header.
      sandboxName:
        type: string
        description: The sandbox to operate in.
      datasetName:
        type: string
        description: Name for the new dataset.
      schemaId:
        type: string
        description: Existing XDM schema id the dataset references.
      datasetId:
        type: string
        description: >-
          The dataset id to ingest the batch into. The createDataset call returns
          only a Location header, so the resolved dataset id is supplied as an
          input for the batch step.
      inputFormat:
        type: string
        description: Batch input file format (parquet, json, or csv).
        default: parquet
  steps:
  - stepId: createDataset
    description: Create the dataset that the batch will ingest data into.
    operationId: createDataset
    parameters:
    - name: Authorization
      in: header
      value: $inputs.authorization
    - name: x-api-key
      in: header
      value: $inputs.apiKey
    - name: x-sandbox-name
      in: header
      value: $inputs.sandboxName
    requestBody:
      contentType: application/json
      payload:
        name: $inputs.datasetName
        schemaRef:
          id: $inputs.schemaId
          contentType: standard
    successCriteria:
    - condition: $statusCode == 201
    outputs:
      datasetLocation: $response.headers.Location
  - stepId: createBatch
    description: Open an import batch targeting the dataset for data upload.
    operationId: createBatch
    parameters:
    - name: Authorization
      in: header
      value: $inputs.authorization
    - name: x-api-key
      in: header
      value: $inputs.apiKey
    - name: x-sandbox-name
      in: header
      value: $inputs.sandboxName
    requestBody:
      contentType: application/json
      payload:
        datasetId: $inputs.datasetId
        inputFormat:
          format: $inputs.inputFormat
    successCriteria:
    - condition: $statusCode == 201
    outputs:
      batchId: $response.body#/id
      batchStatus: $response.body#/status
  - stepId: pollBatch
    description: >-
      Poll the batch status. While the batch is still loading the step retries;
      once it reports a terminal status the workflow ends.
    operationId: getBatch
    parameters:
    - name: batchId
      in: path
      value: $steps.createBatch.outputs.batchId
    - name: Authorization
      in: header
      value: $inputs.authorization
    - name: x-api-key
      in: header
      value: $inputs.apiKey
    - name: x-sandbox-name
      in: header
      value: $inputs.sandboxName
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      finalStatus: $response.body#/status
      completed: $response.body#/completed
    onSuccess:
    - name: stillLoading
      type: retry
      retryAfter: 5
      retryLimit: 10
      criteria:
      - context: $response.body
        condition: $.status == "loading"
        type: jsonpath
    - name: ingestionDone
      type: end
      criteria:
      - context: $response.body
        condition: $.status != "loading"
        type: jsonpath
  outputs:
    batchId: $steps.createBatch.outputs.batchId
    finalStatus: $steps.pollBatch.outputs.finalStatus