Sensible Portfolio Upload URL And Poll

arazzo: 1.0.1
info:
  title: Sensible Portfolio Upload URL And Poll
  summary: Generate a Sensible-signed upload URL for a multi-document portfolio, then poll the portfolio extraction to completion.
  description: >-
    The Sensible-hosted upload variant of portfolio extraction. The workflow
    asks Sensible for a presigned upload_url scoped to the document types
    contained in a portfolio, returns that URL and the portfolio extraction id,
    and then polls the Retrieve extraction by ID endpoint until the portfolio
    reports a COMPLETE status. The PUT of the portfolio PDF bytes to the
    returned upload_url happens out of band against Amazon S3 and is not a
    Sensible API operation, so it is documented as an output expectation rather
    than modeled as a step. Every step spells out its request inline, including
    the Bearer authorization.
  version: 1.0.0
sourceDescriptions:
- name: extractionsApi
  url: ../openapi/sensible-extractions-api-openapi.yml
  type: openapi
workflows:
- workflowId: portfolio-upload-url-and-poll
  summary: Generate a portfolio upload URL for a set of document types and poll the resulting portfolio extraction.
  description: >-
    Requests a presigned upload_url for a portfolio of the supplied document
    types, surfaces the upload_url and portfolio id for an out-of-band PUT, then
    polls the portfolio id until Sensible reports COMPLETE.
  inputs:
    type: object
    required:
    - apiKey
    - types
    properties:
      apiKey:
        type: string
        description: Sensible API key used as the Bearer token.
      types:
        type: array
        description: The document types contained in the portfolio.
        items:
          type: string
      segmentDocumentsWith:
        type: string
        description: How to segment the portfolio page ranges.
        enum:
        - llm
        - fingerprints
        default: fingerprints
  steps:
  - stepId: generatePortfolioUploadUrl
    description: >-
      Request a Sensible-signed upload_url for a portfolio of the supplied
      document types, returning the portfolio extraction id.
    operationId: generate-an-upload-url-for-a-pdf-portfolio
    parameters:
    - name: Authorization
      in: header
      value: "Bearer $inputs.apiKey"
    requestBody:
      contentType: application/json
      payload:
        types: $inputs.types
        segment_documents_with: $inputs.segmentDocumentsWith
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      portfolioId: $response.body#/id
      uploadUrl: $response.body#/upload_url
      status: $response.body#/status
  - stepId: pollPortfolio
    description: >-
      After the portfolio is PUT to the upload_url out of band, poll the
      portfolio extraction by id until Sensible reports the COMPLETE status.
    operationId: retrieving-results
    parameters:
    - name: Authorization
      in: header
      value: "Bearer $inputs.apiKey"
    - name: id
      in: path
      value: $steps.generatePortfolioUploadUrl.outputs.portfolioId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      status: $response.body#/status
      documents: $response.body#/documents
    onSuccess:
    - name: portfolioComplete
      type: end
      criteria:
      - context: $response.body
        condition: $.status == "COMPLETE"
        type: jsonpath
    - name: keepPolling
      type: goto
      stepId: pollPortfolio
      criteria:
      - context: $response.body
        condition: $.status == "WAITING" || $.status == "PROCESSING"
        type: jsonpath
  outputs:
    portfolioId: $steps.generatePortfolioUploadUrl.outputs.portfolioId
    uploadUrl: $steps.generatePortfolioUploadUrl.outputs.uploadUrl
    documents: $steps.pollPortfolio.outputs.documents