Google Forms Harvest Form Responses

arazzo: 1.0.1
info:
  title: Google Forms Harvest Form Responses
  summary: Read a form's question structure, page through its responses since a timestamp, and pull one response in full.
  description: >-
    Form responses key their answers by question id, not by question title, so
    any pipeline that moves responses into a warehouse or a CRM has to read the
    form structure first to build the question id to title map. This workflow
    resolves that structure, lists the responses submitted since a supplied
    timestamp, and then retrieves a single response in full for the detail the
    list view does not carry. Every step spells out its request inline so the
    flow can be read and executed without opening the underlying OpenAPI
    description.
  version: 1.0.0
sourceDescriptions:
- name: googleFormsApi
  url: ../openapi/google-forms-api.yaml
  type: openapi
workflows:
- workflowId: harvest-form-responses
  summary: Pull new form responses with the question structure needed to interpret them.
  description: >-
    Reads the form items to map question ids to titles, lists responses filtered
    by submission timestamp, and fetches the first returned response by id.
  inputs:
    type: object
    required:
    - formId
    - submittedAfter
    properties:
      formId:
        type: string
        description: The id of the form to harvest (e.g. 1BxiMVs0XRA5nFMdLXDSR2r9NphkU4V9qeeW0nQmGR8).
      submittedAfter:
        type: string
        description: RFC 3339 timestamp; only responses submitted after this are returned (e.g. 2026-04-15T00:00:00Z).
      pageSize:
        type: integer
        description: Maximum number of responses to return per page. Defaults to 50 in this flow.
        default: 50
  steps:
  - stepId: readFormStructure
    description: >-
      Read the form to capture its items. Each answer in a response is keyed by
      question id, so this structure is what makes the harvested answers
      interpretable downstream.
    operationId: getForm
    parameters:
    - name: formId
      in: path
      value: $inputs.formId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      items: $response.body#/items
      formTitle: $response.body#/info/title
      linkedSheetId: $response.body#/linkedSheetId
  - stepId: listResponses
    description: >-
      List the responses submitted after the supplied timestamp. The next page
      token is carried out so a caller can resume paging on the next run.
    operationId: listFormResponses
    parameters:
    - name: formId
      in: path
      value: $inputs.formId
    - name: filter
      in: query
      value: "timestamp > $inputs.submittedAfter"
    - name: pageSize
      in: query
      value: $inputs.pageSize
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      responses: $response.body#/responses
      firstResponseId: $response.body#/responses/0/responseId
      nextPageToken: $response.body#/nextPageToken
  - stepId: getResponseDetail
    description: >-
      Retrieve the first returned response by id for the full answer payload,
      including any file upload answers and the respondent email when the form
      collects it.
    operationId: getFormResponse
    parameters:
    - name: formId
      in: path
      value: $inputs.formId
    - name: responseId
      in: path
      value: $steps.listResponses.outputs.firstResponseId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      responseId: $response.body#/responseId
      answers: $response.body#/answers
      respondentEmail: $response.body#/respondentEmail
      lastSubmittedTime: $response.body#/lastSubmittedTime
  outputs:
    formTitle: $steps.readFormStructure.outputs.formTitle
    items: $steps.readFormStructure.outputs.items
    responses: $steps.listResponses.outputs.responses
    nextPageToken: $steps.listResponses.outputs.nextPageToken
    answers: $steps.getResponseDetail.outputs.answers