Update modularization work with the latest main (#661)

Merge latest changes from main into modularization-main It includes changes from #626, #630, #648, #649, and #651 By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license. --------- Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dlpzx <71252798+dlpzx@users.noreply.github.com> Co-authored-by: wolanlu <101870655+wolanlu@users.noreply.github.com> Co-authored-by: Amr Saber <amr.m.saber.mail@gmail.com> Co-authored-by: Noah Paige <69586985+noah-paige@users.noreply.github.com> Co-authored-by: kukushking <kukushkin.anton@gmail.com> Co-authored-by: Dariusz Osiennik <osiend@amazon.com> Co-authored-by: Dennis Goldner <107395339+degoldner@users.noreply.github.com> Co-authored-by: Abdulrahman Kaitoua <abdulrahman.kaitoua@polimi.it> Co-authored-by: akaitoua-sa <126820454+akaitoua-sa@users.noreply.github.com> Co-authored-by: Gezim Musliaj <102723839+gmuslia@users.noreply.github.com> Co-authored-by: Rick Bernotas <97474536+rbernotas@users.noreply.github.com> Co-authored-by: David Mutune Kimengu <57294718+kimengu-david@users.noreply.github.com> Co-authored-by: chamcca <40579012+chamcca@users.noreply.github.com> Co-authored-by: Dhruba <117375130+marjet26@users.noreply.github.com> Co-authored-by: dbalintx <132444646+dbalintx@users.noreply.github.com> Co-authored-by: Srinivas Reddy <srinivasreddych@outlook.com> Co-authored-by: mourya-33 <134511711+mourya-33@users.noreply.github.com> Co-authored-by: Noah Paige <noahpaig@amazon.com> Co-authored-by: dlpzx <dlpzx@amazon.com>
data-dot-all · Aug 16, 2023 · 90bc8e8 · 90bc8e8
1 parent f182ee2
commit 90bc8e8
Show file tree

Hide file tree

Showing 12 changed files with 208 additions and 100 deletions.
diff --git a/backend/dataall/modules/datasets/cdk/assets/gluedatabasecustomresource/index.py b/backend/dataall/modules/datasets/cdk/assets/gluedatabasecustomresource/index.py
@@ -49,6 +49,13 @@ def on_create(event):
     except ClientError as e:
         pass
 
+    default_db_exists = False
+    try:
+        glue_client.get_database(Name="default")
+        default_db_exists = True
+    except ClientError as e:
+        pass
+
     if not exists:
         try:
             db_input = props.get('DatabaseInput').copy()
@@ -63,7 +70,7 @@ def on_create(event):
             raise Exception(f"Could not create Glue Database {props['DatabaseInput']['Name']} in aws://{AWS_ACCOUNT}/{AWS_REGION}, received {str(e)}")
 
     Entries = []
-    for i, role_arn in enumerate(props.get('DatabaseAdministrators')):
+    for i, role_arn in enumerate(props.get('DatabaseAdministrators', [])):
         Entries.append(
             {
                 'Id': str(uuid.uuid4()),
@@ -103,6 +110,20 @@ def on_create(event):
                 'PermissionsWithGrantOption': ['SELECT', 'ALTER', 'DESCRIBE'],
             }
         )
+        if default_db_exists:
+            Entries.append(
+                {
+                    'Id': str(uuid.uuid4()),
+                    'Principal': {'DataLakePrincipalIdentifier': role_arn},
+                    'Resource': {
+                        'Database': {
+                            'Name': 'default'
+                        }
+                    },
+                    'Permissions': ['Describe'.upper()],
+                }
+            )
+
     lf_client.batch_grant_permissions(CatalogId=props['CatalogId'], Entries=Entries)
     physical_id = props['DatabaseInput']['Imported'] + props['DatabaseInput']['Name']
 

diff --git a/backend/dataall/modules/datasets/cdk/assets/glueprofilingjob/glue_script.py b/backend/dataall/modules/datasets/cdk/assets/glueprofilingjob/glue_script.py
@@ -1,4 +1,5 @@
 import json
+import os
 import logging
 import pprint
 import sys
@@ -8,7 +9,6 @@
 from awsglue.transforms import *
 from awsglue.utils import getResolvedOptions
 from pyspark.context import SparkContext
-from pydeequ.profiles import *
 
 sc = SparkContext.getOrCreate()
 sc._jsc.hadoopConfiguration().set('fs.s3.canned.acl', 'BucketOwnerFullControl')
@@ -32,6 +32,7 @@
     'environmentBucket',
     'dataallRegion',
     'table',
+    "SPARK_VERSION"
 ]
 try:
     args = getResolvedOptions(sys.argv, list_args)
@@ -43,6 +44,10 @@
     list_args.remove('table')
     args = getResolvedOptions(sys.argv, list_args)
 
+os.environ["SPARK_VERSION"] = args.get("SPARK_VERSION", "3.1")
+
+from pydeequ.profiles import *
+
 logger.info('Parsed Retrieved parameters')
 
 logger.info('Parsed Args = %s', pprint.pformat(args))

diff --git a/backend/dataall/modules/datasets/cdk/dataset_stack.py b/backend/dataall/modules/datasets/cdk/dataset_stack.py
@@ -300,24 +300,26 @@ def __init__(self, scope, id, target_uri: str = None, **kwargs):
                     ]
                 ),
                 iam.PolicyStatement(
-                    sid="CreateLoggingGlueCrawler",
+                    sid="CreateLoggingGlue",
                     actions=[
                         'logs:CreateLogGroup',
                         'logs:CreateLogStream',
                     ],
                     effect=iam.Effect.ALLOW,
                     resources=[
                         f'arn:aws:logs:{dataset.region}:{dataset.AwsAccountId}:log-group:/aws-glue/crawlers*',
+                        f'arn:aws:logs:{dataset.region}:{dataset.AwsAccountId}:log-group:/aws-glue/jobs/*',
                     ],
                 ),
                 iam.PolicyStatement(
-                    sid="LoggingGlueCrawler",
+                    sid="LoggingGlue",
                     actions=[
                         'logs:PutLogEvents',
                     ],
                     effect=iam.Effect.ALLOW,
                     resources=[
                         f'arn:aws:logs:{dataset.region}:{dataset.AwsAccountId}:log-group:/aws-glue/crawlers:log-stream:{dataset.GlueCrawlerName}',
+                        f'arn:aws:logs:{dataset.region}:{dataset.AwsAccountId}:log-group:/aws-glue/jobs/*',
                     ],
                 ),
                 iam.PolicyStatement(
@@ -443,7 +445,8 @@ def __init__(self, scope, id, target_uri: str = None, **kwargs):
                     'CreateTableDefaultPermissions': [],
                     'Imported': 'IMPORTED-' if dataset.imported else 'CREATED-'
                 },
-                'DatabaseAdministrators': dataset_admins
+                'DatabaseAdministrators': dataset_admins,
+                'TriggerUpdate': True
             },
         )
 
@@ -484,6 +487,7 @@ def __init__(self, scope, id, target_uri: str = None, **kwargs):
             '--enable-metrics': 'true',
             '--enable-continuous-cloudwatch-log': 'true',
             '--enable-glue-datacatalog': 'true',
+            '--SPARK_VERSION': '3.1',
         }
 
         job = glue.CfnJob(

diff --git a/backend/migrations/versions/b1cdc0dc987a_fix_template_column_in_table.py b/backend/migrations/versions/b1cdc0dc987a_fix_template_column_in_table.py
@@ -7,21 +7,61 @@
 """
 from alembic import op
 import sqlalchemy as sa
+from sqlalchemy import orm, Column, String
 from sqlalchemy.dialects import postgresql
+from sqlalchemy.ext.declarative import declarative_base
 
 # revision identifiers, used by Alembic.
 revision = 'b1cdc0dc987a'
 down_revision = '4392a0c9747f'
 branch_labels = None
 depends_on = None
 
+Base = declarative_base()
+
+
+class DataPipeline(Base):
+    __tablename__ = 'datapipeline'
+    DataPipelineUri = Column(
+        String, nullable=False, primary_key=True
+    )
+    devStrategy = Column(String, nullable=True)
+    devStages = Column(postgresql.ARRAY(String), nullable=True)
+
 
 def upgrade():
     # ### commands auto generated by Alembic - please adjust! ###
+    # Modify column types
+    print("Upgrade devStages and devStrategy column types. Updating nullable to True...")
     op.add_column(
         'datapipeline',
         sa.Column('template', sa.String(), nullable=True)
     )
+    op.alter_column(
+        'datapipeline',
+        'devStages',
+        existing_type=postgresql.ARRAY(sa.VARCHAR()),
+        nullable=True
+    )
+    op.alter_column(
+        'datapipeline',
+        'devStrategy',
+        existing_type=sa.VARCHAR(),
+        nullable=True
+    )
+    print("Backfilling values for devStages and devStrategy...")
+    # Backfill values
+    bind = op.get_bind()
+    session = orm.Session(bind=bind)
+    session.query(DataPipeline).filter(DataPipeline.devStrategy is None).update(
+        {DataPipeline.devStrategy: 'gitflowBlueprint'}, synchronize_session=False)
+
+    session.query(DataPipeline).filter(DataPipeline.devStages is None).update(
+        {DataPipeline.devStages: ['dev', 'test', 'prod']}, synchronize_session=False)
+    session.commit()
+
+    print("Backfilling values for devStages and devStrategy is done. Updating nullable to False...")
+    # Force nullable = False
     op.alter_column(
         'datapipeline',
         'devStages',

diff --git a/deploy/cdk_exec_policy/cdkExecPolicy.yaml b/deploy/cdk_exec_policy/cdkExecPolicy.yaml
@@ -1,9 +1,6 @@
 AWSTemplateFormatVersion: 2010-09-09
 Description: Custom least privilege IAM policy for linking environments to dataall
 Parameters:
-  AwsAccountId:
-    Description: AWS AccountId of the account that we wish to link.
-    Type: String
   PolicyName:
     Description: IAM policy name (The same name must be used during CDK bootstrapping. Default is DataAllCustomCDKPolicy.)
     Type: String
@@ -48,14 +45,14 @@ Resources:
             Effect: Allow
             Action: 'athena:CreateWorkGroup'
             Resource:
-              - !Sub 'arn:aws:athena:*:${AWS::AccountId}:workgroup/*'
+              - !Sub 'arn:${AWS::Partition}:athena:*:${AWS::AccountId}:workgroup/*'
           - Sid: IAM
             Action:
               - 'iam:CreatePolicy'
               - 'iam:GetPolicy'
             Effect: Allow
             Resource:
-              - !Sub 'arn:aws:iam::${AWS::AccountId}:policy/*'
+              - !Sub 'arn:${AWS::Partition}:iam::${AWS::AccountId}:policy/*'
           - Sid: IAMRole
             Action:
               - 'iam:AttachRolePolicy'
@@ -82,7 +79,7 @@ Resources:
               - 'iam:CreatePolicyVersion'
               - 'iam:DeletePolicyVersion'
             Resource:
-               - !Sub 'arn:aws:iam::${AWS::AccountId}:policy/service-role/AWSQuickSight*'
+               - !Sub 'arn:${AWS::Partition}:iam::${AWS::AccountId}:policy/service-role/AWSQuickSight*'
           - Sid: QuickSight
             Effect: Allow
             Action:
@@ -114,14 +111,14 @@ Resources:
               - 'kms:CreateAlias'
             Effect: Allow
             Resource:
-              - !Sub 'arn:aws:kms:*:${AWS::AccountId}:alias/*'
+              - !Sub 'arn:${AWS::Partition}:kms:*:${AWS::AccountId}:alias/*'
           - Sid: KMSKey
             Action:
               - 's3:PutBucketAcl'
               - 's3:PutBucketNotification'
             Effect: Allow
             Resource:
-              - !Sub 'arn:aws:s3:::${EnvironmentResourcePrefix}-logging-*'
+              - !Sub 'arn:${AWS::Partition}:s3:::${EnvironmentResourcePrefix}-logging-*'
           - Sid: ReadBuckets
             Action:
               - 'kms:CreateAlias'
@@ -136,7 +133,7 @@ Resources:
               - 'kms:PutKeyPolicy'
               - 'kms:TagResource'
             Effect: Allow
-            Resource: !Sub 'arn:aws:kms:*:${AWS::AccountId}:key/*'
+            Resource: !Sub 'arn:${AWS::Partition}:kms:*:${AWS::AccountId}:key/*'
           - Sid: Lambda
             Action:
               - 'lambda:AddPermission'
@@ -154,7 +151,7 @@ Resources:
             Action:
               - 'lambda:PublishLayerVersion'
             Resource:
-              - !Sub 'arn:aws:lambda:*:${AWS::AccountId}:layer:*'
+              - !Sub 'arn:${AWS::Partition}:lambda:*:${AWS::AccountId}:layer:*'
           - Sid: S3
             Action:
               - 's3:CreateBucket'
@@ -170,13 +167,13 @@ Resources:
               - 's3:DeleteBucketPolicy'
               - 's3:DeleteBucket'
             Effect: Allow
-            Resource: 'arn:aws:s3:::*'
+            Resource: !Sub 'arn:${AWS::Partition}:s3:::*'
           - Sid: SQS
             Effect: Allow
             Action:
               - 'sqs:CreateQueue'
               - 'sqs:SetQueueAttributes'
-            Resource: !Sub 'arn:aws:sqs:*:${AWS::AccountId}:*'
+            Resource: !Sub 'arn:${AWS::Partition}:sqs:*:${AWS::AccountId}:*'
           - Sid: SSM
             Effect: Allow
             Action:
@@ -190,18 +187,18 @@ Resources:
               - 'logs:CreateLogStream'
               - 'logs:PutLogEvents'
               - 'logs:DescribeLogStreams'
-            Resource: 'arn:aws:logs:*:*:*'
+            Resource: !Sub 'arn:${AWS::Partition}:logs:*:*:*'
           - Sid: STS
             Effect: Allow
             Action:
               - 'sts:AssumeRole'
               - 'iam:*Role*'
-            Resource: !Sub 'arn:aws:iam::${AWS::AccountId}:role/cdk-*'
+            Resource: !Sub 'arn:${AWS::Partition}:iam::${AWS::AccountId}:role/cdk-*'
           - Sid: CloudFormation
             Effect: Allow
             Action:
               - 'cloudformation:*'
-            Resource: !Sub 'arn:aws:cloudformation:*:${AWS::AccountId}:stack/CDKToolkit/*'
+            Resource: !Sub 'arn:${AWS::Partition}:cloudformation:*:${AWS::AccountId}:stack/CDKToolkit/*'
           - Sid: ECR
             Effect: Allow
             Action:
@@ -211,14 +208,14 @@ Resources:
               - 'ecr:DescribeRepositories'
               - 'ecr:CreateRepository'
               - 'ecr:DeleteRepository'
-            Resource: !Sub 'arn:aws:ecr:*:${AWS::AccountId}:repository/cdk-*'
+            Resource: !Sub 'arn:${AWS::Partition}:ecr:*:${AWS::AccountId}:repository/cdk-*'
           - Sid: SSMTwo
             Effect: Allow
             Action:
               - 'ssm:GetParameter'
               - 'ssm:PutParameter'
               - 'ssm:DeleteParameter'
-            Resource: !Sub 'arn:aws:ssm:*:${AWS::AccountId}:parameter/cdk-bootstrap/*'
+            Resource: !Sub 'arn:${AWS::Partition}:ssm:*:${AWS::AccountId}:parameter/cdk-bootstrap/*'
           - Sid: CloudformationTwo
             Effect: Allow
             Action:
@@ -232,7 +229,7 @@ Resources:
             Action:
               - 's3:*'
             Resource:
-              - !Sub 'arn:aws:s3:::cdktoolkit-stagingbucket-*'
+              - !Sub 'arn:${AWS::Partition}:s3:::cdk-hnb659fds-assets-${AWS::AccountId}-${AWS::Region}*'
           - Sid: Pipelines
             Effect: Allow
             Action:
@@ -261,15 +258,15 @@ Resources:
               - 's3:ListBucket'
               - 's3:GetBucketPolicy'
             Resource:
-              - 'arn:aws:s3::*:codepipeline-*'
+              - !Sub 'arn:${AWS::Partition}:s3::*:codepipeline-*'
           - Sid: CodeStarNotificationsReadOnly
             Effect: Allow
             Action:
               - 'codestar-notifications:DescribeNotificationRule'
             Resource: '*'
             Condition:
               'StringLike':
-                'codestar-notifications:NotificationsForResource': 'arn:aws:codepipeline:*'
+                'codestar-notifications:NotificationsForResource': !Sub 'arn:${AWS::Partition}:codepipeline:*'
           - Sid: Eventrules
             Effect: Allow
             Action:

diff --git a/deploy/stacks/backend_stack.py b/deploy/stacks/backend_stack.py
@@ -35,6 +35,7 @@ def __init__(
         image_tag=None,
         pipeline_bucket=None,
         vpc_id=None,
+        vpc_restricted_nacls=False,
         vpc_endpoints_sg=None,
         internet_facing=True,
         custom_domain=None,
@@ -64,6 +65,7 @@ def __init__(
             resource_prefix=resource_prefix,
             vpc_endpoints_sg=vpc_endpoints_sg,
             vpc_id=vpc_id,
+            restricted_nacl=vpc_restricted_nacls,
             **kwargs,
         )
         vpc = self.vpc_stack.vpc

diff --git a/deploy/stacks/backend_stage.py b/deploy/stacks/backend_stage.py
@@ -17,6 +17,7 @@ def __init__(
         tooling_account_id=None,
         pipeline_bucket=None,
         vpc_id=None,
+        vpc_restricted_nacls=False,
         vpc_endpoints_sg=None,
         internet_facing=True,
         custom_domain=None,
@@ -45,6 +46,7 @@ def __init__(
             pipeline_bucket=pipeline_bucket,
             image_tag=commit_id,
             vpc_id=vpc_id,
+            vpc_restricted_nacls=vpc_restricted_nacls,
             vpc_endpoints_sg=vpc_endpoints_sg,
             internet_facing=internet_facing,
             custom_domain=custom_domain,