Amazon Entity Resolution Run Matching Pipeline

arazzo: 1.0.1
info:
  title: Amazon Entity Resolution Run Matching Pipeline
  summary: Stand up a schema mapping and matching workflow, start a job, and poll it to completion.
  description: >-
    The end-to-end record matching pipeline for AWS Entity Resolution. The
    workflow first registers a schema mapping that describes the input customer
    records table, then creates a matching workflow that wires that schema to an
    input source, an output destination, and a resolution technique. It starts a
    matching job for the new workflow and polls the job status until it reaches a
    terminal state, branching on SUCCEEDED versus FAILED. 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: entityResolutionApi
  url: ../openapi/amazon-entity-resolution-openapi.yml
  type: openapi
workflows:
- workflowId: run-matching-pipeline
  summary: Create a schema mapping and matching workflow, run a job, and wait for it to finish.
  description: >-
    Registers a schema mapping, creates a matching workflow referencing it,
    starts a matching job, and polls GetMatchingJob until the job status is
    SUCCEEDED, FAILED, or QUEUED no longer applies.
  inputs:
    type: object
    required:
    - schemaName
    - workflowName
    - inputSourceARN
    - outputS3Path
    - roleArn
    properties:
      schemaName:
        type: string
        description: The name of the schema mapping to create (pattern ^[a-zA-Z_0-9-]*$).
      workflowName:
        type: string
        description: The name of the matching workflow to create (pattern ^[a-zA-Z_0-9-]*$).
      description:
        type: string
        description: Optional description applied to the schema mapping and workflow.
      mappedInputFields:
        type: array
        description: List of SchemaInputAttribute objects (fieldName, type, optional matchKey/groupName), minimum 2.
        items:
          type: object
      inputSourceARN:
        type: string
        description: The Glue table ARN for the input source table.
      outputS3Path:
        type: string
        description: The S3 path to which Entity Resolution writes the output table.
      output:
        type: array
        description: List of OutputAttribute objects selecting columns for the output table.
        items:
          type: object
      attributeMatchingModel:
        type: string
        description: Either ONE_TO_ONE or MANY_TO_MANY for rule-based matching.
      rules:
        type: array
        description: List of Rule objects (ruleName, matchingKeys) for RULE_MATCHING.
        items:
          type: object
      roleArn:
        type: string
        description: The ARN of the IAM role Entity Resolution assumes to run the workflow.
  steps:
  - stepId: createSchemaMapping
    description: >-
      Register the schema mapping that defines the columns of the input customer
      records table and which columns to match on.
    operationId: CreateSchemaMapping
    requestBody:
      contentType: application/json
      payload:
        schemaName: $inputs.schemaName
        description: $inputs.description
        mappedInputFields: $inputs.mappedInputFields
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      schemaArn: $response.body#/schemaArn
      schemaName: $response.body#/schemaName
  - stepId: createMatchingWorkflow
    description: >-
      Create the matching workflow that binds the new schema mapping to its
      input source, output destination, and rule-based resolution technique.
    operationId: CreateMatchingWorkflow
    requestBody:
      contentType: application/json
      payload:
        workflowName: $inputs.workflowName
        description: $inputs.description
        inputSourceConfig:
        - inputSourceARN: $inputs.inputSourceARN
          schemaName: $steps.createSchemaMapping.outputs.schemaName
        outputSourceConfig:
        - outputS3Path: $inputs.outputS3Path
          output: $inputs.output
        resolutionTechniques:
          resolutionType: RULE_MATCHING
          ruleBasedProperties:
            attributeMatchingModel: $inputs.attributeMatchingModel
            rules: $inputs.rules
        roleArn: $inputs.roleArn
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      workflowName: $response.body#/workflowName
      workflowArn: $response.body#/workflowArn
  - stepId: startMatchingJob
    description: >-
      Start a matching job for the newly created workflow and capture the
      returned job id for polling.
    operationId: StartMatchingJob
    parameters:
    - name: workflowName
      in: path
      value: $steps.createMatchingWorkflow.outputs.workflowName
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      jobId: $response.body#/jobId
  - stepId: pollJob
    description: >-
      Poll the job status. While the job is RUNNING or QUEUED, loop back and
      check again; on SUCCEEDED finish, on FAILED route to the failure step.
    operationId: GetMatchingJob
    parameters:
    - name: workflowName
      in: path
      value: $steps.createMatchingWorkflow.outputs.workflowName
    - name: jobId
      in: path
      value: $steps.startMatchingJob.outputs.jobId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      status: $response.body#/status
      metrics: $response.body#/metrics
      errorDetails: $response.body#/errorDetails
    onSuccess:
    - name: jobSucceeded
      type: end
      criteria:
      - context: $response.body
        condition: $.status == "SUCCEEDED"
        type: jsonpath
    - name: jobFailed
      type: goto
      stepId: reportFailure
      criteria:
      - context: $response.body
        condition: $.status == "FAILED"
        type: jsonpath
    - name: stillRunning
      type: goto
      stepId: pollJob
      criteria:
      - context: $response.body
        condition: $.status == "RUNNING" || $.status == "QUEUED"
        type: jsonpath
  - stepId: reportFailure
    description: >-
      Re-read the failed job so its error details are captured in the workflow
      outputs for diagnosis.
    operationId: GetMatchingJob
    parameters:
    - name: workflowName
      in: path
      value: $steps.createMatchingWorkflow.outputs.workflowName
    - name: jobId
      in: path
      value: $steps.startMatchingJob.outputs.jobId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      status: $response.body#/status
      errorDetails: $response.body#/errorDetails
  outputs:
    schemaArn: $steps.createSchemaMapping.outputs.schemaArn
    workflowArn: $steps.createMatchingWorkflow.outputs.workflowArn
    jobId: $steps.startMatchingJob.outputs.jobId
    finalStatus: $steps.pollJob.outputs.status
    metrics: $steps.pollJob.outputs.metrics