Merge branch 'master' into pluginv2

datahub-graphql-core/src/main/java/com/linkedin/datahub/graphql/resolvers/chart/BrowseV2Resolver.java

@@ -49,6 +49,8 @@ public CompletableFuture<BrowseResultsV2> get(DataFetchingEnvironment environmen
     final int start = input.getStart() != null ? input.getStart() : DEFAULT_START;
     final int count = input.getCount() != null ? input.getCount() : DEFAULT_COUNT;
     final String query = input.getQuery() != null ? input.getQuery() : "*";
+    // escape forward slash since it is a reserved character in Elasticsearch
+    final String sanitizedQuery = ResolverUtils.escapeForwardSlash(query);
 
     return CompletableFuture.supplyAsync(() -> {
       try {
@@ -64,7 +66,7 @@ public CompletableFuture<BrowseResultsV2> get(DataFetchingEnvironment environmen
             maybeResolvedView != null
             ? SearchUtils.combineFilters(filter, maybeResolvedView.getDefinition().getFilter())
             : filter,
-            query,
+            sanitizedQuery,
             start,
             count,
             context.getAuthentication()

datahub-web-react/src/app/home/AcrylDemoBanner.tsx

@@ -33,13 +33,17 @@ const StyledLink = styled(Link)`
     font-weight: 700;
 `;
 
+const TextContent = styled.div`
+    max-width: 1025px;
+`;
+
 export default function AcrylDemoBanner() {
     return (
         <BannerWrapper>
             <Logo src={AcrylLogo} />
             <TextWrapper>
-                <Title>Schedule a Demo of Managed Datahub</Title>
-                <span>
+                <Title>Schedule a Demo of Managed DataHub</Title>
+                <TextContent>
                     DataHub is already the industry&apos;s #1 Open Source Data Catalog.{' '}
                     <StyledLink
                         href="https://www.acryldata.io/datahub-sign-up"
@@ -48,8 +52,17 @@ export default function AcrylDemoBanner() {
                     >
                         Schedule a demo
                     </StyledLink>{' '}
-                    of Acryl Cloud to see the advanced features that take it to the next level!
-                </span>
+                    of Acryl DataHub to see the advanced features that take it to the next level or purchase Acryl Cloud
+                    on{' '}
+                    <StyledLink
+                        href="https://aws.amazon.com/marketplace/pp/prodview-ratzv4k453pck?sr=0-1&ref_=beagle&applicationId=AWSMPContessa"
+                        target="_blank"
+                        rel="noopener noreferrer"
+                    >
+                        AWS Marketplace
+                    </StyledLink>
+                    !
+                </TextContent>
             </TextWrapper>
         </BannerWrapper>
     );

docs/how/updating-datahub.md

@@ -20,6 +20,8 @@ individually enable / disable desired field metrics.
 
 ### Deprecations
 
+- #8198: In the Python SDK, the `PlatformKey` class has been renamed to `ContainerKey`.
+
 ### Other notable Changes
 
 ## 0.10.4

metadata-ingestion/src/datahub/emitter/mce_builder.py

@@ -6,7 +6,17 @@
 import time
 from enum import Enum
 from hashlib import md5
-from typing import Any, List, Optional, Type, TypeVar, Union, cast, get_type_hints
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    List,
+    Optional,
+    Type,
+    TypeVar,
+    Union,
+    cast,
+    get_type_hints,
+)
 
 import typing_inspect
 
@@ -50,6 +60,9 @@
     os.getenv("DATAHUB_DATASET_URN_TO_LOWER", "false") == "true"
 )
 
+if TYPE_CHECKING:
+    from datahub.emitter.mcp_builder import DatahubKey
+
 
 # TODO: Delete this once lower-casing is the standard.
 def set_dataset_urn_to_lower(value: bool) -> None:
@@ -132,7 +145,11 @@ def dataset_key_to_urn(key: DatasetKeyClass) -> str:
     )
 
 
-def make_container_urn(guid: str) -> str:
+def make_container_urn(guid: Union[str, "DatahubKey"]) -> str:
+    from datahub.emitter.mcp_builder import DatahubKey
+
+    if isinstance(guid, DatahubKey):
+        guid = guid.guid()
     return f"urn:li:container:{guid}"
 
 

metadata-ingestion/src/datahub/emitter/mcp_builder.py

@@ -54,7 +54,9 @@ def guid(self) -> str:
         return _stable_guid_from_dict(bag)
 
 
-class PlatformKey(DatahubKey):
+class ContainerKey(DatahubKey):
+    """Base class for container guid keys. Most users should use one of the subclasses instead."""
+
     platform: str
     instance: Optional[str] = None
 
@@ -81,20 +83,27 @@ def guid_dict(self) -> Dict[str, str]:
     def property_dict(self) -> Dict[str, str]:
         return self.dict(by_alias=True, exclude_none=True)
 
+    def as_urn(self) -> str:
+        return make_container_urn(guid=self.guid())
+
+
+# DEPRECATION: Keeping the `PlatformKey` name around for backwards compatibility.
+PlatformKey = ContainerKey
+
 
-class DatabaseKey(PlatformKey):
+class DatabaseKey(ContainerKey):
     database: str
 
 
 class SchemaKey(DatabaseKey):
     db_schema: str = Field(alias="schema")
 
 
-class ProjectIdKey(PlatformKey):
+class ProjectIdKey(ContainerKey):
     project_id: str
 
 
-class MetastoreKey(PlatformKey):
+class MetastoreKey(ContainerKey):
     metastore: str
 
 
@@ -110,11 +119,11 @@ class BigQueryDatasetKey(ProjectIdKey):
     dataset_id: str
 
 
-class FolderKey(PlatformKey):
+class FolderKey(ContainerKey):
     folder_abs_path: str
 
 
-class BucketKey(PlatformKey):
+class BucketKey(ContainerKey):
     bucket_name: str
 
 
@@ -127,7 +136,7 @@ def default(self, obj: Any) -> Any:
         return json.JSONEncoder.default(self, obj)
 
 
-KeyType = TypeVar("KeyType", bound=PlatformKey)
+KeyType = TypeVar("KeyType", bound=ContainerKey)
 
 
 def add_domain_to_entity_wu(
@@ -188,7 +197,7 @@ def gen_containers(
     container_key: KeyType,
     name: str,
     sub_types: List[str],
-    parent_container_key: Optional[PlatformKey] = None,
+    parent_container_key: Optional[ContainerKey] = None,
     extra_properties: Optional[Dict[str, str]] = None,
     domain_urn: Optional[str] = None,
     description: Optional[str] = None,
@@ -199,9 +208,7 @@ def gen_containers(
     created: Optional[int] = None,
     last_modified: Optional[int] = None,
 ) -> Iterable[MetadataWorkUnit]:
-    container_urn = make_container_urn(
-        guid=container_key.guid(),
-    )
+    container_urn = container_key.as_urn()
     yield MetadataChangeProposalWrapper(
         entityUrn=f"{container_urn}",
         # entityKeyAspect=ContainerKeyClass(guid=parent_container_key.guid()),

metadata-ingestion/src/datahub/ingestion/graph/client.py

@@ -88,6 +88,12 @@ class RemovedStatusFilter(enum.Enum):
     """Search only soft-deleted entities."""
 
 
+@dataclass
+class RelatedEntity:
+    urn: str
+    relationship_type: str
+
+
 def _graphql_entity_type(entity_type: str) -> str:
     """Convert the entity types into GraphQL "EntityType" enum values."""
 
@@ -769,11 +775,6 @@ class RelationshipDirection(str, enum.Enum):
         INCOMING = "INCOMING"
         OUTGOING = "OUTGOING"
 
-    @dataclass
-    class RelatedEntity:
-        urn: str
-        relationship_type: str
-
     def get_related_entities(
         self,
         entity_urn: str,
@@ -794,7 +795,7 @@ def get_related_entities(
                 },
             )
             for related_entity in response.get("entities", []):
-                yield DataHubGraph.RelatedEntity(
+                yield RelatedEntity(
                     urn=related_entity["urn"],
                     relationship_type=related_entity["relationshipType"],
                 )

metadata-ingestion/src/datahub/ingestion/source/bigquery_v2/bigquery.py

@@ -20,7 +20,7 @@
     set_dataset_urn_to_lower,
 )
 from datahub.emitter.mcp import MetadataChangeProposalWrapper
-from datahub.emitter.mcp_builder import BigQueryDatasetKey, PlatformKey, ProjectIdKey
+from datahub.emitter.mcp_builder import BigQueryDatasetKey, ContainerKey, ProjectIdKey
 from datahub.ingestion.api.common import PipelineContext
 from datahub.ingestion.api.decorators import (
     SupportStatus,
@@ -434,7 +434,7 @@ def get_dataplatform_instance_aspect(
             entityUrn=dataset_urn, aspect=aspect
         ).as_workunit()
 
-    def gen_dataset_key(self, db_name: str, schema: str) -> PlatformKey:
+    def gen_dataset_key(self, db_name: str, schema: str) -> ContainerKey:
         return BigQueryDatasetKey(
             project_id=db_name,
             dataset_id=schema,
@@ -443,7 +443,7 @@ def gen_dataset_key(self, db_name: str, schema: str) -> PlatformKey:
             backcompat_env_as_instance=True,
         )
 
-    def gen_project_id_key(self, database: str) -> PlatformKey:
+    def gen_project_id_key(self, database: str) -> ContainerKey:
         return ProjectIdKey(
             project_id=database,
             platform=self.platform,

metadata-ingestion/src/datahub/ingestion/source/data_lake_common/data_lake_utils.py

@@ -3,9 +3,9 @@
 
 from datahub.emitter.mcp_builder import (
     BucketKey,
+    ContainerKey,
     FolderKey,
     KeyType,
-    PlatformKey,
     add_dataset_to_container,
     gen_containers,
 )
@@ -45,7 +45,7 @@ def create_emit_containers(
         container_key: KeyType,
         name: str,
         sub_types: List[str],
-        parent_container_key: Optional[PlatformKey] = None,
+        parent_container_key: Optional[ContainerKey] = None,
         domain_urn: Optional[str] = None,
     ) -> Iterable[MetadataWorkUnit]:
         if container_key.guid() not in self.processed_containers:

metadata-ingestion/src/datahub/ingestion/source/powerbi/powerbi.py

@@ -9,7 +9,7 @@
 import datahub.emitter.mce_builder as builder
 import datahub.ingestion.source.powerbi.rest_api_wrapper.data_classes as powerbi_data_classes
 from datahub.emitter.mcp import MetadataChangeProposalWrapper
-from datahub.emitter.mcp_builder import PlatformKey, gen_containers
+from datahub.emitter.mcp_builder import ContainerKey, gen_containers
 from datahub.ingestion.api.common import PipelineContext
 from datahub.ingestion.api.decorators import (
     SourceCapability,
@@ -104,7 +104,7 @@ def __init__(
         self.__reporter = reporter
         self.__dataplatform_instance_resolver = dataplatform_instance_resolver
         self.processed_datasets: Set[powerbi_data_classes.PowerBIDataset] = set()
-        self.workspace_key: PlatformKey
+        self.workspace_key: ContainerKey
 
     @staticmethod
     def urn_to_lowercase(value: str, flag: bool) -> str:
@@ -256,7 +256,6 @@ def to_datahub_schema(
         self,
         table: powerbi_data_classes.Table,
     ) -> SchemaMetadataClass:
-
         fields = []
         table_fields = (
             [self.to_datahub_schema_field(column) for column in table.columns]

metadata-ingestion/src/datahub/ingestion/source/powerbi/rest_api_wrapper/data_classes.py

@@ -2,7 +2,7 @@
 from enum import Enum
 from typing import Any, Dict, List, Optional, Union
 
-from datahub.emitter.mcp_builder import PlatformKey
+from datahub.emitter.mcp_builder import ContainerKey
 from datahub.metadata.schema_classes import (
     BooleanTypeClass,
     DateTypeClass,
@@ -28,11 +28,11 @@
 }
 
 
-class WorkspaceKey(PlatformKey):
+class WorkspaceKey(ContainerKey):
     workspace: str
 
 
-class DatasetKey(PlatformKey):
+class DatasetKey(ContainerKey):
     dataset: str
 
 
@@ -57,7 +57,7 @@ def get_workspace_key(
         platform_name: str,
         platform_instance: Optional[str] = None,
         workspace_id_as_urn_part: Optional[bool] = False,
-    ) -> PlatformKey:
+    ) -> ContainerKey:
         return WorkspaceKey(
             workspace=self.get_urn_part(workspace_id_as_urn_part),
             platform=platform_name,
@@ -150,7 +150,7 @@ def __eq__(self, instance):
     def __hash__(self):
         return hash(self.__members())
 
-    def get_dataset_key(self, platform_name: str) -> PlatformKey:
+    def get_dataset_key(self, platform_name: str) -> ContainerKey:
         return DatasetKey(
             dataset=self.id,
             platform=platform_name,

metadata-ingestion/src/datahub/ingestion/source/powerbi_report_server/constants.py

@@ -82,8 +82,6 @@ class Constant:
     Constant.LINKED_REPORTS: "{PBIRS_BASE_URL}/LinkedReports",
     Constant.LINKED_REPORT: "{PBIRS_BASE_URL}/LinkedReports({LINKED_REPORT_ID})",
     Constant.ME: "{PBIRS_BASE_URLL}/Me",
-    Constant.MOBILE_REPORTS: "{PBIRS_BASE_URL}/MobileReports",
-    Constant.MOBILE_REPORT: "{PBIRS_BASE_URL}/MobileReports({MOBILE_REPORT_ID})",
     Constant.POWERBI_REPORTS: "{PBIRS_BASE_URL}/PowerBiReports",
     Constant.POWERBI_REPORT: "{PBIRS_BASE_URL}/PowerBiReports({POWERBI_REPORT_ID})",
     Constant.POWERBI_REPORT_DATASOURCES: "{PBIRS_BASE_URL}/PowerBiReports({ID})/DataSources",