acryl-datahub 1.1.1rc3__py3-none-any.whl → 1.2.0__py3-none-any.whl

datahub/metadata/schemas/MLModelGroupKey.avsc

@@ -8,6 +8,7 @@
       "glossaryTerms",
       "editableMlModelGroupProperties",
       "domains",
+      "applications",
       "mlModelGroupProperties",
       "ownership",
       "status",
@@ -60,13 +61,17 @@
           "DEV": "Designates development fabrics",
           "EI": "Designates early-integration fabrics",
           "NON_PROD": "Designates non-production fabrics",
+          "PRD": "Alternative Prod spelling",
           "PRE": "Designates pre-production fabrics",
           "PROD": "Designates production fabrics",
           "QA": "Designates quality assurance fabrics",
           "RVW": "Designates review fabrics",
           "SANDBOX": "Designates sandbox fabrics",
+          "SBX": "Alternative spelling for sandbox",
+          "SIT": "System Integration Testing",
           "STG": "Designates staging fabrics",
           "TEST": "Designates testing fabrics",
+          "TST": "Alternative Test spelling",
           "UAT": "Designates user acceptance testing fabrics"
         },
         "name": "FabricType",
@@ -83,6 +88,10 @@
           "PROD",
           "CORP",
           "RVW",
+          "PRD",
+          "TST",
+          "SIT",
+          "SBX",
           "SANDBOX"
         ],
         "doc": "Fabric group type"

datahub/metadata/schemas/MLModelKey.avsc

@@ -8,6 +8,7 @@
       "glossaryTerms",
       "editableMlModelProperties",
       "domains",
+      "applications",
       "ownership",
       "mlModelProperties",
       "intendedUse",
@@ -74,13 +75,17 @@
           "DEV": "Designates development fabrics",
           "EI": "Designates early-integration fabrics",
           "NON_PROD": "Designates non-production fabrics",
+          "PRD": "Alternative Prod spelling",
           "PRE": "Designates pre-production fabrics",
           "PROD": "Designates production fabrics",
           "QA": "Designates quality assurance fabrics",
           "RVW": "Designates review fabrics",
           "SANDBOX": "Designates sandbox fabrics",
+          "SBX": "Alternative spelling for sandbox",
+          "SIT": "System Integration Testing",
           "STG": "Designates staging fabrics",
           "TEST": "Designates testing fabrics",
+          "TST": "Alternative Test spelling",
           "UAT": "Designates user acceptance testing fabrics"
         },
         "name": "FabricType",
@@ -97,6 +102,10 @@
           "PROD",
           "CORP",
           "RVW",
+          "PRD",
+          "TST",
+          "SIT",
+          "SBX",
           "SANDBOX"
         ],
         "doc": "Fabric group type"

datahub/metadata/schemas/MLPrimaryKeyKey.avsc

@@ -8,6 +8,7 @@
       "glossaryTerms",
       "editableMlPrimaryKeyProperties",
       "domains",
+      "applications",
       "mlPrimaryKeyProperties",
       "ownership",
       "institutionalMemory",

datahub/metadata/schemas/MetadataChangeEvent.avsc

@@ -2424,13 +2424,17 @@
                               "DEV": "Designates development fabrics",
                               "EI": "Designates early-integration fabrics",
                               "NON_PROD": "Designates non-production fabrics",
+                              "PRD": "Alternative Prod spelling",
                               "PRE": "Designates pre-production fabrics",
                               "PROD": "Designates production fabrics",
                               "QA": "Designates quality assurance fabrics",
                               "RVW": "Designates review fabrics",
                               "SANDBOX": "Designates sandbox fabrics",
+                              "SBX": "Alternative spelling for sandbox",
+                              "SIT": "System Integration Testing",
                               "STG": "Designates staging fabrics",
                               "TEST": "Designates testing fabrics",
+                              "TST": "Alternative Test spelling",
                               "UAT": "Designates user acceptance testing fabrics"
                             },
                             "name": "FabricType",
@@ -2447,6 +2451,10 @@
                               "PROD",
                               "CORP",
                               "RVW",
+                              "PRD",
+                              "TST",
+                              "SIT",
+                              "SBX",
                               "SANDBOX"
                             ],
                             "doc": "Fabric group type"
@@ -7733,13 +7741,15 @@
                                                   "type": "enum",
                                                   "symbolDocs": {
                                                     "EQUALS": "Whether the field matches the value",
+                                                    "NOT_EQUALS": "Whether the field does not match the value",
                                                     "STARTS_WITH": "Whether the field value starts with the value"
                                                   },
                                                   "name": "PolicyMatchCondition",
                                                   "namespace": "com.linkedin.pegasus2avro.policy",
                                                   "symbols": [
                                                     "EQUALS",
-                                                    "STARTS_WITH"
+                                                    "STARTS_WITH",
+                                                    "NOT_EQUALS"
                                                   ],
                                                   "doc": "The matching condition in a filter criterion"
                                                 },
@@ -7761,6 +7771,15 @@
                                 "name": "filter",
                                 "default": null,
                                 "doc": "Filter to apply privileges to"
+                              },
+                              {
+                                "type": [
+                                  "null",
+                                  "com.linkedin.pegasus2avro.policy.PolicyMatchFilter"
+                                ],
+                                "name": "privilegeConstraints",
+                                "default": null,
+                                "doc": "Constraints around what sub-resources operations are allowed to modify, i.e. NOT_EQUALS - cannot modify a particular defined tag, EQUALS - can only modify a particular defined tag, STARTS_WITH - can only modify a tag starting with xyz"
                               }
                             ],
                             "doc": "Information used to filter DataHub resource."

datahub/metadata/schemas/NotebookKey.avsc

@@ -15,6 +15,7 @@
       "browsePaths",
       "institutionalMemory",
       "domains",
+      "applications",
       "subTypes",
       "dataPlatformInstance",
       "browsePathsV2",

datahub/metadata/schemas/QuerySubjects.avsc

@@ -15,13 +15,6 @@
           "namespace": "com.linkedin.pegasus2avro.query",
           "fields": [
             {
-              "Relationship": {
-                "entityTypes": [
-                  "dataset",
-                  "schemaField"
-                ],
-                "name": "IsAssociatedWith"
-              },
               "Searchable": {
                 "fieldName": "entities",
                 "fieldType": "URN"
@@ -32,11 +25,7 @@
               "type": "string",
               "name": "entity",
               "doc": "An entity which is the subject of a query.",
-              "Urn": "Urn",
-              "entityTypes": [
-                "dataset",
-                "schemaField"
-              ]
+              "Urn": "Urn"
             }
           ],
           "doc": "A single subject of a particular query.\nIn the future, we may evolve this model to include richer details\nabout the Query Subject in relation to the query."

datahub/metadata/schemas/SchemaFieldKey.avsc

@@ -14,7 +14,8 @@
       "documentation",
       "testResults",
       "deprecation",
-      "subTypes"
+      "subTypes",
+      "logicalParent"
     ]
   },
   "name": "SchemaFieldKey",

datahub/sdk/__init__.py

@@ -18,9 +18,15 @@ from datahub.metadata.urns import (
     SchemaFieldUrn,
     TagUrn,
 )
+from datahub.sdk.chart import Chart
 from datahub.sdk.container import Container
+from datahub.sdk.dashboard import Dashboard
+from datahub.sdk.dataflow import DataFlow
+from datahub.sdk.datajob import DataJob
 from datahub.sdk.dataset import Dataset
 from datahub.sdk.main_client import DataHubClient
+from datahub.sdk.mlmodel import MLModel
+from datahub.sdk.mlmodelgroup import MLModelGroup
 from datahub.sdk.search_filters import Filter, FilterDsl
 
 # We want to print out the warning if people do `from datahub.sdk import X`.

datahub/sdk/_all_entities.py

@@ -1,19 +1,30 @@
 from typing import Dict, List, Type
 
+from datahub.sdk.chart import Chart
 from datahub.sdk.container import Container
+from datahub.sdk.dashboard import Dashboard
+from datahub.sdk.dataflow import DataFlow
+from datahub.sdk.datajob import DataJob
 from datahub.sdk.dataset import Dataset
 from datahub.sdk.entity import Entity
 from datahub.sdk.mlmodel import MLModel
 from datahub.sdk.mlmodelgroup import MLModelGroup
 
+# Base entity classes that don't have circular dependencies
+# Those that do are imported in the EntityClient where needed
 # TODO: Is there a better way to declare this?
 ENTITY_CLASSES_LIST: List[Type[Entity]] = [
     Container,
     Dataset,
     MLModel,
     MLModelGroup,
+    DataFlow,
+    DataJob,
+    Dashboard,
+    Chart,
 ]
 
+# Create the mapping of entity types to classes
 ENTITY_CLASSES: Dict[str, Type[Entity]] = {
     cls.get_urn_type().ENTITY_TYPE: cls for cls in ENTITY_CLASSES_LIST
 }

datahub/sdk/_shared.py

@@ -26,9 +26,12 @@ from datahub.emitter.mce_builder import (
 from datahub.emitter.mcp_builder import ContainerKey
 from datahub.errors import MultipleSubtypesWarning, SdkUsageError
 from datahub.metadata.urns import (
+    ChartUrn,
     ContainerUrn,
     CorpGroupUrn,
     CorpUserUrn,
+    DashboardUrn,
+    DataFlowUrn,
     DataJobUrn,
     DataPlatformInstanceUrn,
     DataPlatformUrn,
@@ -37,6 +40,7 @@ from datahub.metadata.urns import (
     DomainUrn,
     GlossaryTermUrn,
     OwnershipTypeUrn,
+    StructuredPropertyUrn,
     TagUrn,
     Urn,
     VersionSetUrn,
@@ -47,12 +51,21 @@ from datahub.utilities.urns.error import InvalidUrnError
 
 if TYPE_CHECKING:
     from datahub.sdk.container import Container
-
 UrnOrStr: TypeAlias = Union[Urn, str]
+ChartUrnOrStr: TypeAlias = Union[str, ChartUrn]
 DatasetUrnOrStr: TypeAlias = Union[str, DatasetUrn]
 DatajobUrnOrStr: TypeAlias = Union[str, DataJobUrn]
+DataflowUrnOrStr: TypeAlias = Union[str, DataFlowUrn]
+DashboardUrnOrStr: TypeAlias = Union[str, DashboardUrn]
+DataPlatformInstanceUrnOrStr: TypeAlias = Union[str, DataPlatformInstanceUrn]
+DataPlatformUrnOrStr: TypeAlias = Union[str, DataPlatformUrn]
 
 ActorUrn: TypeAlias = Union[CorpUserUrn, CorpGroupUrn]
+StructuredPropertyUrnOrStr: TypeAlias = Union[str, StructuredPropertyUrn]
+StructuredPropertyValueType: TypeAlias = Union[str, float, int]
+StructuredPropertyInputType: TypeAlias = Dict[
+    StructuredPropertyUrnOrStr, Sequence[StructuredPropertyValueType]
+]
 
 TrainingMetricsInputType: TypeAlias = Union[
     List[models.MLMetricClass], Dict[str, Optional[str]]
@@ -716,3 +729,107 @@ class HasVersion(Entity):
                 a for a in version_props.aliases if a.versionTag != alias
             ]
             self._set_aspect(version_props)
+
+
+class HasStructuredProperties(Entity):
+    """
+    Mixin for entities that support structured properties
+    """
+
+    __slots__ = ()
+
+    @property
+    def structured_properties(
+        self,
+    ) -> Optional[List[models.StructuredPropertyValueAssignmentClass]]:
+        """
+        Retrieve structured properties for the entity
+
+        Returns:
+            Optional list of structured property value assignments
+        """
+        sp_aspect = self._get_aspect(models.StructuredPropertiesClass)
+        return sp_aspect.properties if sp_aspect else None
+
+    def _ensure_structured_properties(self) -> models.StructuredPropertiesClass:
+        """
+        Ensure structured properties aspect exists, creating it if necessary
+
+        Returns:
+            StructuredPropertiesClass aspect
+        """
+        return self._setdefault_aspect(models.StructuredPropertiesClass(properties=[]))
+
+    def set_structured_property(
+        self,
+        property_urn: StructuredPropertyUrnOrStr,
+        values: Sequence[StructuredPropertyValueType],
+    ) -> None:
+        """
+        Update an existing structured property or add if it doesn't exist
+
+        Args:
+            property_urn: URN of the structured property
+            values: List of values for the property
+        """
+        # validate property_urn is a valid structured property urn
+        property_urn = StructuredPropertyUrn.from_string(property_urn)
+
+        properties = self._ensure_structured_properties()
+
+        # Find existing property assignment
+        existing_prop = next(
+            (
+                prop
+                for prop in properties.properties
+                if prop.propertyUrn == str(property_urn)
+            ),
+            None,
+        )
+        current_timestamp = make_ts_millis(datetime.now())
+
+        if existing_prop:
+            # Update existing property
+            existing_prop.values = list(values)
+            existing_prop.lastModified = models.AuditStampClass(
+                time=current_timestamp,
+                actor=DEFAULT_ACTOR_URN,
+            )
+        else:
+            # Create new property assignment
+            new_property = models.StructuredPropertyValueAssignmentClass(
+                propertyUrn=str(property_urn),
+                values=list(values),
+                created=models.AuditStampClass(
+                    time=current_timestamp,
+                    actor=DEFAULT_ACTOR_URN,
+                ),
+                lastModified=models.AuditStampClass(
+                    time=current_timestamp,
+                    actor=DEFAULT_ACTOR_URN,
+                ),
+            )
+            add_list_unique(
+                properties.properties,
+                key=lambda prop: prop.propertyUrn,
+                item=new_property,
+            )
+
+        self._set_aspect(properties)
+
+    def remove_structured_property(
+        self, property_urn: StructuredPropertyUrnOrStr
+    ) -> None:
+        """
+        Remove a structured property from the entity
+
+        Args:
+            property_urn: URN of the structured property to remove
+        """
+        remove_list_unique(
+            self._ensure_structured_properties().properties,
+            key=lambda prop: prop.propertyUrn,
+            item=models.StructuredPropertyValueAssignmentClass(
+                propertyUrn=str(property_urn), values=[]
+            ),
+        )