feat(stepfunctions-tasks): support databrew startJobRun task (#12532)

This adds support for AWS Glue DataBrew StartJobRun API as a task. Task documentation: https://docs.aws.amazon.com/step-functions/latest/dg/connect-databrew.html ---- *By submitting this pull request, I confirm that my contribution is made under the terms of the Apache-2.0 license*
aws · Jan 25, 2021 · eacd2f7 · eacd2f7
1 parent c2f0e15
commit eacd2f7
Show file tree

Hide file tree

Showing 7 changed files with 589 additions and 0 deletions.
diff --git a/packages/@aws-cdk/aws-stepfunctions-tasks/README.md b/packages/@aws-cdk/aws-stepfunctions-tasks/README.md
@@ -55,6 +55,7 @@ This module is part of the [AWS Cloud Development Kit](https://github.com/aws/aw
   - [Modify Instance Fleet](#modify-instance-fleet)
   - [Modify Instance Group](#modify-instance-group)
 - [Glue](#glue)
+- [Glue DataBrew](#glue-databrew)
 - [Lambda](#lambda)
 - [SageMaker](#sagemaker)
   - [Create Training Job](#create-training-job)
@@ -680,6 +681,18 @@ new GlueStartJobRun(stack, 'Task', {
 });
 ```
 
+## Glue DataBrew
+
+Step Functions supports [AWS Glue DataBrew](https://docs.aws.amazon.com/step-functions/latest/dg/connect-databrew.html) through the service integration pattern.
+
+You can call the [`StartJobRun`](https://docs.aws.amazon.com/databrew/latest/dg/API_StartJobRun.html) API from a `Task` state.
+
+```ts
+new GlueDataBrewStartJobRun(stack, 'Task', {
+  Name: 'databrew-job',
+});
+```
+
 ## Lambda
 
 [Invoke](https://docs.aws.amazon.com/lambda/latest/dg/API_Invoke.html) a Lambda function.

diff --git a/packages/@aws-cdk/aws-stepfunctions-tasks/lib/databrew/start-job-run.ts b/packages/@aws-cdk/aws-stepfunctions-tasks/lib/databrew/start-job-run.ts
@@ -0,0 +1,80 @@
+import * as iam from '@aws-cdk/aws-iam';
+import * as sfn from '@aws-cdk/aws-stepfunctions';
+import * as cdk from '@aws-cdk/core';
+import { Construct } from 'constructs';
+import { integrationResourceArn, validatePatternSupported } from '../private/task-utils';
+
+/**
+ * Properties for starting a job run with StartJobRun
+ * @experimental
+ */
+export interface GlueDataBrewStartJobRunProps extends sfn.TaskStateBaseProps {
+
+  /**
+   * Glue DataBrew Job to run
+   */
+  readonly name: string;
+}
+
+/**
+ * Start a Job run as a Task
+ *
+ * @see https://docs.aws.amazon.com/step-functions/latest/dg/connect-databrew.html
+ * @experimental
+ */
+export class GlueDataBrewStartJobRun extends sfn.TaskStateBase {
+
+  private static readonly SUPPORTED_INTEGRATION_PATTERNS: sfn.IntegrationPattern[] = [
+    sfn.IntegrationPattern.REQUEST_RESPONSE,
+    sfn.IntegrationPattern.RUN_JOB,
+  ];
+
+  protected readonly taskMetrics?: sfn.TaskMetricsConfig;
+  protected readonly taskPolicies?: iam.PolicyStatement[];
+
+  private readonly integrationPattern: sfn.IntegrationPattern;
+
+  /**
+   * @experimental
+   */
+  constructor(scope: Construct, id: string, private readonly props: GlueDataBrewStartJobRunProps) {
+    super(scope, id, props);
+    this.integrationPattern = props.integrationPattern ?? sfn.IntegrationPattern.REQUEST_RESPONSE;
+
+    validatePatternSupported(this.integrationPattern, GlueDataBrewStartJobRun.SUPPORTED_INTEGRATION_PATTERNS);
+
+    const actions = ['databrew:startJobRun'];
+
+    if (this.integrationPattern === sfn.IntegrationPattern.RUN_JOB) {
+      actions.push('databrew:stopJobRun', 'databrew:listJobRuns');
+    }
+
+    this.taskPolicies = [
+      new iam.PolicyStatement({
+        resources: [
+          cdk.Stack.of(this).formatArn({
+            service: 'databrew',
+            resource: 'job',
+            // If the name comes from input, we cannot target the policy to a particular ARN prefix reliably.
+            resourceName: sfn.JsonPath.isEncodedJsonPath(this.props.name) ? '*' : this.props.name,
+          }),
+        ],
+        actions: actions,
+      }),
+    ];
+  }
+
+  /**
+   * Provides the Glue DataBrew Start Job Run task configuration
+   * @internal
+   */
+  protected _renderTask(): any {
+    return {
+      Resource: integrationResourceArn('databrew', 'startJobRun', this.integrationPattern),
+      Parameters: sfn.FieldUtils.renderObject({
+        Name: this.props.name,
+      }),
+    };
+  }
+}
+
diff --git a/packages/@aws-cdk/aws-stepfunctions-tasks/lib/index.ts b/packages/@aws-cdk/aws-stepfunctions-tasks/lib/index.ts
@@ -43,3 +43,4 @@ export * from './athena/start-query-execution';
 export * from './athena/stop-query-execution';
 export * from './athena/get-query-execution';
 export * from './athena/get-query-results';
+export * from './databrew/start-job-run';
diff --git a/packages/@aws-cdk/aws-stepfunctions-tasks/package.json b/packages/@aws-cdk/aws-stepfunctions-tasks/package.json
@@ -75,6 +75,7 @@
     "@aws-cdk/aws-batch": "0.0.0",
     "@aws-cdk/aws-cloudwatch": "0.0.0",
     "@aws-cdk/aws-codebuild": "0.0.0",
+    "@aws-cdk/aws-databrew": "0.0.0",
     "@aws-cdk/aws-dynamodb": "0.0.0",
     "@aws-cdk/aws-ec2": "0.0.0",
     "@aws-cdk/aws-ecr": "0.0.0",
@@ -96,6 +97,7 @@
     "@aws-cdk/aws-batch": "0.0.0",
     "@aws-cdk/aws-cloudwatch": "0.0.0",
     "@aws-cdk/aws-codebuild": "0.0.0",
+    "@aws-cdk/aws-databrew": "0.0.0",
     "@aws-cdk/aws-dynamodb": "0.0.0",
     "@aws-cdk/aws-ec2": "0.0.0",
     "@aws-cdk/aws-ecr": "0.0.0",

diff --git a/packages/@aws-cdk/aws-stepfunctions-tasks/test/databrew/integ.start-job-run.expected.json b/packages/@aws-cdk/aws-stepfunctions-tasks/test/databrew/integ.start-job-run.expected.json
@@ -0,0 +1,255 @@
+{
+  "Resources": {
+    "JobOutputBucketACE3BC7B": {
+      "Type": "AWS::S3::Bucket",
+      "UpdateReplacePolicy": "Delete",
+      "DeletionPolicy": "Delete"
+    },
+    "DataBrewRole7E60F80D": {
+      "Type": "AWS::IAM::Role",
+      "Properties": {
+        "AssumeRolePolicyDocument": {
+          "Statement": [
+            {
+              "Action": "sts:AssumeRole",
+              "Effect": "Allow",
+              "Principal": {
+                "Service": "databrew.amazonaws.com"
+              }
+            }
+          ],
+          "Version": "2012-10-17"
+        },
+        "ManagedPolicyArns": [
+          "arn:aws:iam::aws:policy/service-role/AWSGlueDataBrewServiceRole"
+        ],
+        "Path": "/",
+        "Policies": [
+          {
+            "PolicyDocument": {
+              "Statement": [
+                {
+                  "Action": [
+                    "s3:GetObject",
+                    "s3:PutObject",
+                    "s3:DeleteObject",
+                    "s3:ListBucket"
+                  ],
+                  "Effect": "Allow",
+                  "Resource": [
+                    "arn:aws:s3:::databrew-public-datasets-test-region/*",
+                    "arn:aws:s3:::databrew-public-datasets-test-region",
+                    {
+                      "Fn::Join": [
+                        "",
+                        [
+                          {
+                            "Fn::GetAtt": [
+                              "JobOutputBucketACE3BC7B",
+                              "Arn"
+                            ]
+                          },
+                          "/*"
+                        ]
+                      ]
+                    },
+                    {
+                      "Fn::GetAtt": [
+                        "JobOutputBucketACE3BC7B",
+                        "Arn"
+                      ]
+                    }
+                  ]
+                }
+              ],
+              "Version": "2012-10-17"
+            },
+            "PolicyName": "DataBrewPolicy"
+          }
+        ]
+      }
+    },
+    "DataBrewRecipe": {
+      "Type": "AWS::DataBrew::Recipe",
+      "Properties": {
+        "Name": "recipe-1",
+        "Steps": [
+          {
+            "Action": {
+              "Operation": "UPPER_CASE",
+              "Parameters": {
+                "sourceColumn": "description"
+              }
+            }
+          },
+          {
+            "Action": {
+              "Operation": "DELETE",
+              "Parameters": {
+                "sourceColumn": "doc_id"
+              }
+            }
+          }
+        ]
+      }
+    },
+    "DataBrewDataset": {
+      "Type": "AWS::DataBrew::Dataset",
+      "Properties": {
+        "Input": {
+          "S3InputDefinition": {
+            "Bucket": "databrew-public-datasets-test-region",
+            "Key": "votes.csv"
+          }
+        },
+        "Name": "dataset-1"
+      }
+    },
+    "DataBrewProject": {
+      "Type": "AWS::DataBrew::Project",
+      "Properties": {
+        "DatasetName": "dataset-1",
+        "Name": "project-1",
+        "RecipeName": "recipe-1",
+        "RoleArn": {
+          "Fn::GetAtt": [
+            "DataBrewRole7E60F80D",
+            "Arn"
+          ]
+        }
+      },
+      "DependsOn": [
+        "DataBrewDataset",
+        "DataBrewRecipe"
+      ]
+    },
+    "DataBrewJob": {
+      "Type": "AWS::DataBrew::Job",
+      "Properties": {
+        "Name": "job-1",
+        "RoleArn": {
+          "Fn::GetAtt": [
+            "DataBrewRole7E60F80D",
+            "Arn"
+          ]
+        },
+        "Type": "RECIPE",
+        "Outputs": [
+          {
+            "Location": {
+              "Bucket": {
+                "Ref": "JobOutputBucketACE3BC7B"
+              }
+            }
+          }
+        ],
+        "ProjectName": "project-1"
+      },
+      "DependsOn": [
+        "DataBrewProject"
+      ]
+    },
+    "StateMachineRoleB840431D": {
+      "Type": "AWS::IAM::Role",
+      "Properties": {
+        "AssumeRolePolicyDocument": {
+          "Statement": [
+            {
+              "Action": "sts:AssumeRole",
+              "Effect": "Allow",
+              "Principal": {
+                "Service": {
+                  "Fn::Join": [
+                    "",
+                    [
+                      "states.",
+                      {
+                        "Ref": "AWS::Region"
+                      },
+                      ".amazonaws.com"
+                    ]
+                  ]
+                }
+              }
+            }
+          ],
+          "Version": "2012-10-17"
+        }
+      }
+    },
+    "StateMachineRoleDefaultPolicyDF1E6607": {
+      "Type": "AWS::IAM::Policy",
+      "Properties": {
+        "PolicyDocument": {
+          "Statement": [
+            {
+              "Action": "databrew:startJobRun",
+              "Effect": "Allow",
+              "Resource": {
+                "Fn::Join": [
+                  "",
+                  [
+                    "arn:",
+                    {
+                      "Ref": "AWS::Partition"
+                    },
+                    ":databrew:",
+                    {
+                      "Ref": "AWS::Region"
+                    },
+                    ":",
+                    {
+                      "Ref": "AWS::AccountId"
+                    },
+                    ":job/job-1"
+                  ]
+                ]
+              }
+            }
+          ],
+          "Version": "2012-10-17"
+        },
+        "PolicyName": "StateMachineRoleDefaultPolicyDF1E6607",
+        "Roles": [
+          {
+            "Ref": "StateMachineRoleB840431D"
+          }
+        ]
+      }
+    },
+    "StateMachine2E01A3A5": {
+      "Type": "AWS::StepFunctions::StateMachine",
+      "Properties": {
+        "RoleArn": {
+          "Fn::GetAtt": [
+            "StateMachineRoleB840431D",
+            "Arn"
+          ]
+        },
+        "DefinitionString": {
+          "Fn::Join": [
+            "",
+            [
+              "{\"StartAt\":\"Start DataBrew Job run\",\"States\":{\"Start DataBrew Job run\":{\"End\":true,\"Type\":\"Task\",\"Resource\":\"arn:",
+              {
+                "Ref": "AWS::Partition"
+              },
+              ":states:::databrew:startJobRun\",\"Parameters\":{\"Name\":\"job-1\"}}},\"TimeoutSeconds\":30}"
+            ]
+          ]
+        }
+      },
+      "DependsOn": [
+        "StateMachineRoleDefaultPolicyDF1E6607",
+        "StateMachineRoleB840431D"
+      ]
+    }
+  },
+  "Outputs": {
+    "stateMachineArn": {
+      "Value": {
+        "Ref": "StateMachine2E01A3A5"
+      }
+    }
+  }
+}