Amazon Polly Select Voice and Start Synthesis Task

arazzo: 1.0.1
info:
  title: Amazon Polly Select Voice and Start Synthesis Task
  summary: Discover a voice for a language, then start an async synthesis task with it.
  description: >-
    Combines voice discovery with long-form asynchronous synthesis. The workflow
    calls DescribeVoices to find an available voice for the requested engine and
    language, captures the first voice id, and then starts a SpeechSynthesisTask
    that writes its output to an S3 bucket using that voice. Each step spells out
    its request inline, including the AWS Signature Version 4 signing headers, so
    the flow can be read and executed without opening the underlying OpenAPI
    description.
  version: 1.0.0
sourceDescriptions:
- name: pollyApi
  url: ../openapi/amazon-polly-openapi-original.yaml
  type: openapi
workflows:
- workflowId: list-voices-start-synthesis-task
  summary: Pick an available voice and start an asynchronous synthesis task using it.
  description: >-
    Lists voices for the requested engine and language, selects the first match,
    and starts an asynchronous synthesis task to an S3 bucket using that voice.
  inputs:
    type: object
    required:
    - amzDate
    - authorization
    - text
    - outputFormat
    - outputS3BucketName
    properties:
      amzDate:
        type: string
        description: The X-Amz-Date timestamp used to sign the requests.
      authorization:
        type: string
        description: The full SigV4 Authorization header value for the request.
      contentSha256:
        type: string
        description: The X-Amz-Content-Sha256 hex digest of the request payload.
      securityToken:
        type: string
        description: Optional X-Amz-Security-Token for temporary credentials.
      engine:
        type: string
        description: Engine to filter voices and to use for synthesis (standard or neural).
      languageCode:
        type: string
        description: ISO language code to filter voices by (e.g. en-US).
      text:
        type: string
        description: The input text (plain text or SSML) to synthesize.
      outputFormat:
        type: string
        description: The audio output format (mp3, ogg_vorbis, pcm, or json).
      outputS3BucketName:
        type: string
        description: The S3 bucket name to which the output file will be saved.
  steps:
  - stepId: describeVoices
    description: List the voices available for the requested engine and language.
    operationId: DescribeVoices
    parameters:
    - name: Engine
      in: query
      value: $inputs.engine
    - name: LanguageCode
      in: query
      value: $inputs.languageCode
    - name: X-Amz-Date
      in: header
      value: $inputs.amzDate
    - name: Authorization
      in: header
      value: $inputs.authorization
    - name: X-Amz-Content-Sha256
      in: header
      value: $inputs.contentSha256
    - name: X-Amz-Security-Token
      in: header
      value: $inputs.securityToken
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      selectedVoiceId: $response.body#/Voices/0/Id
      voices: $response.body#/Voices
  - stepId: startTask
    description: >-
      Start an asynchronous synthesis task to the supplied S3 bucket using the
      voice selected from the DescribeVoices response.
    operationId: StartSpeechSynthesisTask
    parameters:
    - name: X-Amz-Date
      in: header
      value: $inputs.amzDate
    - name: Authorization
      in: header
      value: $inputs.authorization
    - name: X-Amz-Content-Sha256
      in: header
      value: $inputs.contentSha256
    - name: X-Amz-Security-Token
      in: header
      value: $inputs.securityToken
    requestBody:
      contentType: application/json
      payload:
        Engine: $inputs.engine
        LanguageCode: $inputs.languageCode
        OutputFormat: $inputs.outputFormat
        OutputS3BucketName: $inputs.outputS3BucketName
        Text: $inputs.text
        VoiceId: $steps.describeVoices.outputs.selectedVoiceId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      taskId: $response.body#/SynthesisTask/TaskId
      taskStatus: $response.body#/SynthesisTask/TaskStatus
      outputUri: $response.body#/SynthesisTask/OutputUri
  outputs:
    selectedVoiceId: $steps.describeVoices.outputs.selectedVoiceId
    taskId: $steps.startTask.outputs.taskId
    taskStatus: $steps.startTask.outputs.taskStatus