Merge branch 'master' into feature/cus-3546-fix-tableau-authentication

Author: sgomezvillamor

datahub-graphql-core/src/main/resources/entity.graphql

@@ -262,8 +262,16 @@ type Query {
     Fetch all Business Attributes
     """
     listBusinessAttributes(input: ListBusinessAttributesInput!): ListBusinessAttributesResult
+
+    """
+    Fetch a Data Process Instance by primary key (urn)
+    """
+    dataProcessInstance(urn: String!): DataProcessInstance
+
+
 }
 
+
 """
 An ERModelRelationship is a high-level abstraction that dictates what datasets fields are erModelRelationshiped.
 """
@@ -9832,15 +9840,45 @@ type MLModelGroup implements EntityWithRelationships & Entity & BrowsableEntity
     privileges: EntityPrivileges
 }
 
+"""
+Properties describing a group of related ML models
+"""
 type MLModelGroupProperties {
+    """
+    Display name of the model group
+    """
+    name: String
 
+    """
+    Detailed description of the model group's purpose and contents
+    """
     description: String
 
-    createdAt: Long
+    """
+    When this model group was created
+    """
+    created: AuditStamp
 
+    """
+    When this model group was last modified
+    """
+    lastModified: AuditStamp
+
+    """
+    Version identifier for this model group
+    """
     version: VersionTag
 
+    """
+    Custom key-value properties for the model group
+    """
     customProperties: [CustomPropertiesEntry!]
+
+    """
+    Deprecated creation timestamp
+    @deprecated Use the 'created' field instead
+    """
+    createdAt: Long @deprecated(reason: "Use `created` instead")
 }
 
 """
@@ -9990,40 +10028,103 @@ description: String
 }
 
 type MLMetric {
+    """
+    Name of the metric (e.g. accuracy, precision, recall)
+    """
     name: String
 
+    """
+    Description of what this metric measures
+    """
     description: String
 
+    """
+    The computed value of the metric
+    """
     value: String
 
+    """
+    Timestamp when this metric was recorded
+    """
     createdAt: Long
 }
 
 type MLModelProperties {
+    """
+    The display name of the model used in the UI
+    """
+    name: String!
 
+    """
+    Detailed description of the model's purpose and characteristics
+    """
     description: String
 
-    date: Long
+    """
+    When the model was last modified
+    """
+    lastModified: AuditStamp
 
+    """
+    Version identifier for this model
+    """
     version: String
 
+    """
+    The type/category of ML model (e.g. classification, regression)
+    """
     type: String
 
+    """
+    Mapping of hyperparameter configurations
+    """
     hyperParameters: HyperParameterMap
 
-    hyperParams:  [MLHyperParam]
+    """
+    List of hyperparameter settings used to train this model
+    """
+    hyperParams: [MLHyperParam]
 
+    """
+    Performance metrics from model training
+    """
     trainingMetrics: [MLMetric]
 
+    """
+    Names of ML features used by this model
+    """
     mlFeatures: [String!]
 
+    """
+    Tags for categorizing and searching models
+    """
     tags: [String!]
 
+    """
+    Model groups this model belongs to
+    """
     groups: [MLModelGroup]
 
+    """
+    Additional custom properties specific to this model
+    """
     customProperties: [CustomPropertiesEntry!]
 
+    """
+    URL to view this model in external system
+    """
     externalUrl: String
+
+    """
+    When this model was created
+    """
+    created: AuditStamp
+
+    """
+    Deprecated timestamp for model creation
+    @deprecated Use 'created' field instead
+    """
+    date: Long @deprecated(reason: "Use `created` instead")
 }
 
 type MLFeatureProperties {
@@ -12804,3 +12905,92 @@ type CronSchedule {
     """
     timezone: String!
 }
+
+
+"""
+Properties describing a data process instance's execution metadata
+"""
+type DataProcessInstanceProperties {
+    """
+    The display name of this process instance
+    """
+    name: String!
+
+    """
+    URL to view this process instance in the external system
+    """
+    externalUrl: String
+
+    """
+    When this process instance was created
+    """
+    created: AuditStamp
+
+    """
+    Additional custom properties specific to this process instance
+    """
+    customProperties: [CustomPropertiesEntry!]
+}
+
+"""
+Properties specific to an ML model training run instance
+"""
+type MLTrainingRunProperties {
+    """
+    Unique identifier for this training run
+    """
+    id: String
+
+    """
+    List of URLs to access training run outputs (e.g. model artifacts, logs)
+    """
+    outputUrls: [String]
+    
+    """
+    Hyperparameters used in this training run
+    """
+    hyperParams: [MLHyperParam]
+
+    """
+    Performance metrics recorded during this training run
+    """
+    trainingMetrics: [MLMetric]
+}
+
+extend type DataProcessInstance {
+
+    """
+    Additional read only properties associated with the Data Job
+    """
+    properties: DataProcessInstanceProperties
+
+    """
+    The specific instance of the data platform that this entity belongs to
+    """
+    dataPlatformInstance: DataPlatformInstance
+
+    """
+    Sub Types that this entity implements
+    """
+    subTypes: SubTypes
+
+    """
+    The parent container in which the entity resides
+    """
+    container: Container
+
+    """
+    Standardized platform urn where the data process instance is defined
+    """
+    platform: DataPlatform!
+
+    """
+    Recursively get the lineage of containers for this entity
+    """
+    parentContainers: ParentContainersResult
+
+    """
+    Additional properties when subtype is Training Run
+    """
+    mlTrainingRunProperties: MLTrainingRunProperties
+}

metadata-ingestion/src/datahub/ingestion/source/looker/looker_common.py

@@ -1408,6 +1408,15 @@ class LookerDashboardSourceReport(StaleEntityRemovalSourceReport):
     dashboards_with_activity: LossySet[str] = dataclasses_field(
         default_factory=LossySet
     )
+
+    # Entities that don't seem to exist, so we don't emit usage aspects for them despite having usage data
+    dashboards_skipped_for_usage: LossySet[str] = dataclasses_field(
+        default_factory=LossySet
+    )
+    charts_skipped_for_usage: LossySet[str] = dataclasses_field(
+        default_factory=LossySet
+    )
+
     stage_latency: List[StageLatency] = dataclasses_field(default_factory=list)
     _looker_explore_registry: Optional[LookerExploreRegistry] = None
     total_explores: int = 0

metadata-ingestion/src/datahub/ingestion/source/looker/looker_source.py

@@ -68,6 +68,7 @@
     ViewField,
     ViewFieldType,
     gen_model_key,
+    get_urn_looker_element_id,
 )
 from datahub.ingestion.source.looker.looker_config import LookerDashboardSourceConfig
 from datahub.ingestion.source.looker.looker_lib_wrapper import LookerAPI
@@ -165,6 +166,9 @@ def __init__(self, config: LookerDashboardSourceConfig, ctx: PipelineContext):
         # Required, as we do not ingest all folders but only those that have dashboards/looks
         self.processed_folders: List[str] = []
 
+        # Keep track of ingested chart urns, to omit usage for non-ingested entities
+        self.chart_urns: Set[str] = set()
+
     @staticmethod
     def test_connection(config_dict: dict) -> TestConnectionReport:
         test_report = TestConnectionReport()
@@ -642,6 +646,7 @@ def _make_chart_metadata_events(
         chart_urn = self._make_chart_urn(
             element_id=dashboard_element.get_urn_element_id()
         )
+        self.chart_urns.add(chart_urn)
         chart_snapshot = ChartSnapshot(
             urn=chart_urn,
             aspects=[Status(removed=False)],
@@ -1380,7 +1385,9 @@ def _get_folder_and_ancestors_workunits(
         yield from self._emit_folder_as_container(folder)
 
     def extract_usage_stat(
-        self, looker_dashboards: List[looker_usage.LookerDashboardForUsage]
+        self,
+        looker_dashboards: List[looker_usage.LookerDashboardForUsage],
+        ingested_chart_urns: Set[str],
     ) -> List[MetadataChangeProposalWrapper]:
         looks: List[looker_usage.LookerChartForUsage] = []
         # filter out look from all dashboard
@@ -1391,6 +1398,15 @@ def extract_usage_stat(
 
         # dedup looks
         looks = list({str(look.id): look for look in looks}.values())
+        filtered_looks = []
+        for look in looks:
+            if not look.id:
+                continue
+            chart_urn = self._make_chart_urn(get_urn_looker_element_id(look.id))
+            if chart_urn in ingested_chart_urns:
+                filtered_looks.append(look)
+            else:
+                self.reporter.charts_skipped_for_usage.add(look.id)
 
         # Keep stat generators to generate entity stat aspect later
         stat_generator_config: looker_usage.StatGeneratorConfig = (
@@ -1414,7 +1430,7 @@ def extract_usage_stat(
             stat_generator_config,
             self.reporter,
             self._make_chart_urn,
-            looks,
+            filtered_looks,
         )
 
         mcps: List[MetadataChangeProposalWrapper] = []
@@ -1669,7 +1685,7 @@ def get_workunits_internal(self) -> Iterable[MetadataWorkUnit]:
         if self.source_config.extract_usage_history:
             self.reporter.report_stage_start("usage_extraction")
             usage_mcps: List[MetadataChangeProposalWrapper] = self.extract_usage_stat(
-                looker_dashboards_for_usage
+                looker_dashboards_for_usage, self.chart_urns
             )
             for usage_mcp in usage_mcps:
                 yield usage_mcp.as_workunit()