acryl-datahub 0.15.0rc25__py3-none-any.whl → 0.15.0.1rc2__py3-none-any.whl

datahub/ingestion/source/looker/looker_source.py

@@ -145,7 +145,9 @@ class LookerDashboardSource(TestableSource, StatefulIngestionSourceBase):
         self.source_config: LookerDashboardSourceConfig = config
         self.reporter: LookerDashboardSourceReport = LookerDashboardSourceReport()
         self.looker_api: LookerAPI = LookerAPI(self.source_config)
-        self.user_registry: LookerUserRegistry = LookerUserRegistry(self.looker_api)
+        self.user_registry: LookerUserRegistry = LookerUserRegistry(
+            self.looker_api, self.reporter
+        )
         self.explore_registry: LookerExploreRegistry = LookerExploreRegistry(
             self.looker_api, self.reporter, self.source_config
         )
@@ -1673,5 +1675,14 @@ class LookerDashboardSource(TestableSource, StatefulIngestionSourceBase):
                 yield usage_mcp.as_workunit()
             self.reporter.report_stage_end("usage_extraction")
 
+        # Dump looker user resource mappings.
+        logger.info("Ingesting looker user resource mapping workunits")
+        self.reporter.report_stage_start("user_resource_extraction")
+        yield from auto_workunit(
+            self.user_registry.to_platform_resource(
+                self.source_config.platform_instance
+            )
+        )
+
     def get_report(self) -> SourceReport:
         return self.reporter

datahub/ingestion/source/mlflow.py

@@ -38,16 +38,30 @@ T = TypeVar("T")
 class MLflowConfig(EnvConfigMixin):
     tracking_uri: Optional[str] = Field(
         default=None,
-        description="Tracking server URI. If not set, an MLflow default tracking_uri is used (local `mlruns/` directory or `MLFLOW_TRACKING_URI` environment variable)",
+        description=(
+            "Tracking server URI. If not set, an MLflow default tracking_uri is used"
+            " (local `mlruns/` directory or `MLFLOW_TRACKING_URI` environment variable)"
+        ),
     )
     registry_uri: Optional[str] = Field(
         default=None,
-        description="Registry server URI. If not set, an MLflow default registry_uri is used (value of tracking_uri or `MLFLOW_REGISTRY_URI` environment variable)",
+        description=(
+            "Registry server URI. If not set, an MLflow default registry_uri is used"
+            " (value of tracking_uri or `MLFLOW_REGISTRY_URI` environment variable)"
+        ),
     )
     model_name_separator: str = Field(
         default="_",
         description="A string which separates model name from its version (e.g. model_1 or model-1)",
     )
+    base_external_url: Optional[str] = Field(
+        default=None,
+        description=(
+            "Base URL to use when constructing external URLs to MLflow."
+            " If not set, tracking_uri is used if it's an HTTP URL."
+            " If neither is set, external URLs are not generated."
+        ),
+    )
 
 
 @dataclass
@@ -279,12 +293,23 @@ class MLflowSource(Source):
         )
         return urn
 
-    def _make_external_url(self, model_version: ModelVersion) -> Union[None, str]:
+    def _get_base_external_url_from_tracking_uri(self) -> Optional[str]:
+        if isinstance(
+            self.client.tracking_uri, str
+        ) and self.client.tracking_uri.startswith("http"):
+            return self.client.tracking_uri
+        else:
+            return None
+
+    def _make_external_url(self, model_version: ModelVersion) -> Optional[str]:
         """
         Generate URL for a Model Version to MLflow UI.
         """
-        base_uri = self.client.tracking_uri
-        if base_uri.startswith("http"):
+        base_uri = (
+            self.config.base_external_url
+            or self._get_base_external_url_from_tracking_uri()
+        )
+        if base_uri:
             return f"{base_uri.rstrip('/')}/#/models/{model_version.name}/versions/{model_version.version}"
         else:
             return None

datahub/ingestion/source/powerbi/config.py

@@ -9,7 +9,7 @@ from pydantic.class_validators import root_validator
 
 import datahub.emitter.mce_builder as builder
 from datahub.configuration.common import AllowDenyPattern, ConfigModel
-from datahub.configuration.source_common import DatasetSourceConfigMixin
+from datahub.configuration.source_common import DatasetSourceConfigMixin, PlatformDetail
 from datahub.configuration.validate_field_deprecation import pydantic_field_deprecated
 from datahub.ingestion.source.common.subtypes import BIAssetSubTypes
 from datahub.ingestion.source.state.stale_entity_removal_handler import (
@@ -232,19 +232,6 @@ def default_for_dataset_type_mapping() -> Dict[str, str]:
     return dict_
 
 
-class PlatformDetail(ConfigModel):
-    platform_instance: Optional[str] = pydantic.Field(
-        default=None,
-        description="DataHub platform instance name. To generate correct urn for upstream dataset, this should match "
-        "with platform instance name used in ingestion "
-        "recipe of other datahub sources.",
-    )
-    env: str = pydantic.Field(
-        default=builder.DEFAULT_ENV,
-        description="The environment that all assets produced by DataHub platform ingestion source belong to",
-    )
-
-
 class DataBricksPlatformDetail(PlatformDetail):
     """
     metastore is an additional field used in Databricks connector to generate the dataset urn

datahub/ingestion/source/powerbi/dataplatform_instance_resolver.py

@@ -2,8 +2,8 @@ import logging
 from abc import ABC, abstractmethod
 from typing import Union
 
+from datahub.configuration.source_common import PlatformDetail
 from datahub.ingestion.source.powerbi.config import (
-    PlatformDetail,
     PowerBiDashboardSourceConfig,
     PowerBIPlatformDetail,
 )

datahub/ingestion/source/powerbi/m_query/pattern_handler.py

@@ -5,13 +5,13 @@ from typing import Dict, List, Optional, Tuple, Type, cast
 
 from lark import Tree
 
+from datahub.configuration.source_common import PlatformDetail
 from datahub.emitter import mce_builder as builder
 from datahub.ingestion.api.common import PipelineContext
 from datahub.ingestion.source.powerbi.config import (
     Constant,
     DataBricksPlatformDetail,
     DataPlatformPair,
-    PlatformDetail,
     PowerBiDashboardSourceConfig,
     PowerBiDashboardSourceReport,
     PowerBIPlatformDetail,

datahub/ingestion/source/snowflake/snowflake_lineage_v2.py

@@ -265,64 +265,17 @@ class SnowflakeLineageExtractor(SnowflakeCommonMixin, Closeable):
         with PerfTimer() as timer:
             self.report.num_external_table_edges_scanned = 0
 
-            for (
-                known_lineage_mapping
-            ) in self._populate_external_lineage_from_copy_history(discovered_tables):
-                self.sql_aggregator.add(known_lineage_mapping)
-            logger.info(
-                "Done populating external lineage from copy history. "
-                f"Found {self.report.num_external_table_edges_scanned} external lineage edges so far."
-            )
-
-            for (
-                known_lineage_mapping
-            ) in self._populate_external_lineage_from_show_query(discovered_tables):
-                self.sql_aggregator.add(known_lineage_mapping)
-
-            logger.info(
-                "Done populating external lineage from show external tables. "
-                f"Found {self.report.num_external_table_edges_scanned} external lineage edges so far."
-            )
+            for entry in self._get_copy_history_lineage(discovered_tables):
+                self.sql_aggregator.add(entry)
+            logger.info("Done populating external lineage from copy history. ")
 
             self.report.external_lineage_queries_secs = timer.elapsed_seconds()
 
-    # Handles the case for explicitly created external tables.
-    # NOTE: Snowflake does not log this information to the access_history table.
-    def _populate_external_lineage_from_show_query(
-        self, discovered_tables: List[str]
-    ) -> Iterable[KnownLineageMapping]:
-        external_tables_query: str = SnowflakeQuery.show_external_tables()
-        try:
-            for db_row in self.connection.query(external_tables_query):
-                key = self.identifiers.get_dataset_identifier(
-                    db_row["name"], db_row["schema_name"], db_row["database_name"]
-                )
-
-                if key not in discovered_tables:
-                    continue
-                if db_row["location"].startswith("s3://"):
-                    yield KnownLineageMapping(
-                        upstream_urn=make_s3_urn_for_lineage(
-                            db_row["location"], self.config.env
-                        ),
-                        downstream_urn=self.identifiers.gen_dataset_urn(key),
-                    )
-                    self.report.num_external_table_edges_scanned += 1
-
-                self.report.num_external_table_edges_scanned += 1
-        except Exception as e:
-            logger.debug(e, exc_info=e)
-            self.structured_reporter.warning(
-                "Error populating external table lineage from Snowflake",
-                exc=e,
-            )
-            self.report_status(EXTERNAL_LINEAGE, False)
-
     # Handles the case where a table is populated from an external stage/s3 location via copy.
     # Eg: copy into category_english from @external_s3_stage;
     # Eg: copy into category_english from 's3://acryl-snow-demo-olist/olist_raw_data/category_english'credentials=(aws_key_id='...' aws_secret_key='...')  pattern='.*.csv';
     # NOTE: Snowflake does not log this information to the access_history table.
-    def _populate_external_lineage_from_copy_history(
+    def _get_copy_history_lineage(
         self, discovered_tables: List[str]
     ) -> Iterable[KnownLineageMapping]:
         query: str = SnowflakeQuery.copy_lineage_history(

datahub/ingestion/source/snowflake/snowflake_queries.py

@@ -247,9 +247,6 @@ class SnowflakeQueriesExtractor(SnowflakeStructuredReportMixin, Closeable):
                 for entry in self.fetch_copy_history():
                     queries.append(entry)
 
-            # TODO: Add "show external tables" lineage to the main schema extractor.
-            # Because it's not a time-based thing, it doesn't really make sense in the snowflake-queries extractor.
-
             with self.report.query_log_fetch_timer:
                 for entry in self.fetch_query_log():
                     queries.append(entry)

datahub/ingestion/source/snowflake/snowflake_schema_gen.py

@@ -16,6 +16,7 @@ from datahub.ingestion.glossary.classification_mixin import (
     ClassificationHandler,
     classification_workunit_processor,
 )
+from datahub.ingestion.source.aws.s3_util import make_s3_urn_for_lineage
 from datahub.ingestion.source.common.subtypes import (
     DatasetContainerSubTypes,
     DatasetSubTypes,
@@ -35,6 +36,7 @@ from datahub.ingestion.source.snowflake.snowflake_connection import (
 )
 from datahub.ingestion.source.snowflake.snowflake_data_reader import SnowflakeDataReader
 from datahub.ingestion.source.snowflake.snowflake_profiler import SnowflakeProfiler
+from datahub.ingestion.source.snowflake.snowflake_query import SnowflakeQuery
 from datahub.ingestion.source.snowflake.snowflake_report import SnowflakeV2Report
 from datahub.ingestion.source.snowflake.snowflake_schema import (
     SCHEMA_PARALLELISM,
@@ -65,6 +67,7 @@ from datahub.ingestion.source.sql.sql_utils import (
     get_domain_wu,
 )
 from datahub.ingestion.source_report.ingestion_stage import (
+    EXTERNAL_TABLE_DDL_LINEAGE,
     METADATA_EXTRACTION,
     PROFILING,
 )
@@ -96,7 +99,10 @@ from datahub.metadata.com.linkedin.pegasus2avro.schema import (
     TimeType,
 )
 from datahub.metadata.com.linkedin.pegasus2avro.tag import TagProperties
-from datahub.sql_parsing.sql_parsing_aggregator import SqlParsingAggregator
+from datahub.sql_parsing.sql_parsing_aggregator import (
+    KnownLineageMapping,
+    SqlParsingAggregator,
+)
 from datahub.utilities.registries.domain_registry import DomainRegistry
 from datahub.utilities.threaded_iterator_executor import ThreadedIteratorExecutor
 
@@ -180,7 +186,8 @@ class SnowflakeSchemaGenerator(SnowflakeStructuredReportMixin):
 
         # These are populated as side-effects of get_workunits_internal.
         self.databases: List[SnowflakeDatabase] = []
-        self.aggregator: Optional[SqlParsingAggregator] = aggregator
+
+        self.aggregator = aggregator
 
     def get_connection(self) -> SnowflakeConnection:
         return self.connection
@@ -212,6 +219,19 @@ class SnowflakeSchemaGenerator(SnowflakeStructuredReportMixin):
                 self.report.set_ingestion_stage(snowflake_db.name, METADATA_EXTRACTION)
                 yield from self._process_database(snowflake_db)
 
+            self.report.set_ingestion_stage("*", EXTERNAL_TABLE_DDL_LINEAGE)
+            discovered_tables: List[str] = [
+                self.identifiers.get_dataset_identifier(
+                    table_name, schema.name, db.name
+                )
+                for db in self.databases
+                for schema in db.schemas
+                for table_name in schema.tables
+            ]
+            if self.aggregator:
+                for entry in self._external_tables_ddl_lineage(discovered_tables):
+                    self.aggregator.add(entry)
+
         except SnowflakePermissionError as e:
             self.structured_reporter.failure(
                 GENERIC_PERMISSION_ERROR_KEY,
@@ -1082,3 +1102,33 @@ class SnowflakeSchemaGenerator(SnowflakeStructuredReportMixin):
 
         # Access to table but none of its constraints - is this possible ?
         return constraints.get(table_name, [])
+
+    # Handles the case for explicitly created external tables.
+    # NOTE: Snowflake does not log this information to the access_history table.
+    def _external_tables_ddl_lineage(
+        self, discovered_tables: List[str]
+    ) -> Iterable[KnownLineageMapping]:
+        external_tables_query: str = SnowflakeQuery.show_external_tables()
+        try:
+            for db_row in self.connection.query(external_tables_query):
+                key = self.identifiers.get_dataset_identifier(
+                    db_row["name"], db_row["schema_name"], db_row["database_name"]
+                )
+
+                if key not in discovered_tables:
+                    continue
+                if db_row["location"].startswith("s3://"):
+                    yield KnownLineageMapping(
+                        upstream_urn=make_s3_urn_for_lineage(
+                            db_row["location"], self.config.env
+                        ),
+                        downstream_urn=self.identifiers.gen_dataset_urn(key),
+                    )
+                    self.report.num_external_table_edges_scanned += 1
+
+                self.report.num_external_table_edges_scanned += 1
+        except Exception as e:
+            self.structured_reporter.warning(
+                "External table ddl lineage extraction failed",
+                exc=e,
+            )

datahub/ingestion/source/snowflake/snowflake_v2.py

@@ -161,35 +161,32 @@ class SnowflakeV2Source(
         # For database, schema, tables, views, etc
         self.data_dictionary = SnowflakeDataDictionary(connection=self.connection)
         self.lineage_extractor: Optional[SnowflakeLineageExtractor] = None
-        self.aggregator: Optional[SqlParsingAggregator] = None
-
-        if self.config.use_queries_v2 or self.config.include_table_lineage:
-            self.aggregator = self._exit_stack.enter_context(
-                SqlParsingAggregator(
-                    platform=self.identifiers.platform,
-                    platform_instance=self.config.platform_instance,
-                    env=self.config.env,
-                    graph=self.ctx.graph,
-                    eager_graph_load=(
-                        # If we're ingestion schema metadata for tables/views, then we will populate
-                        # schemas into the resolver as we go. We only need to do a bulk fetch
-                        # if we're not ingesting schema metadata as part of ingestion.
-                        not (
-                            self.config.include_technical_schema
-                            and self.config.include_tables
-                            and self.config.include_views
-                        )
-                        and not self.config.lazy_schema_resolver
-                    ),
-                    generate_usage_statistics=False,
-                    generate_operations=False,
-                    format_queries=self.config.format_sql_queries,
-                )
+
+        self.aggregator: SqlParsingAggregator = self._exit_stack.enter_context(
+            SqlParsingAggregator(
+                platform=self.identifiers.platform,
+                platform_instance=self.config.platform_instance,
+                env=self.config.env,
+                graph=self.ctx.graph,
+                eager_graph_load=(
+                    # If we're ingestion schema metadata for tables/views, then we will populate
+                    # schemas into the resolver as we go. We only need to do a bulk fetch
+                    # if we're not ingesting schema metadata as part of ingestion.
+                    not (
+                        self.config.include_technical_schema
+                        and self.config.include_tables
+                        and self.config.include_views
+                    )
+                    and not self.config.lazy_schema_resolver
+                ),
+                generate_usage_statistics=False,
+                generate_operations=False,
+                format_queries=self.config.format_sql_queries,
             )
-            self.report.sql_aggregator = self.aggregator.report
+        )
+        self.report.sql_aggregator = self.aggregator.report
 
         if self.config.include_table_lineage:
-            assert self.aggregator is not None
             redundant_lineage_run_skip_handler: Optional[
                 RedundantLineageRunSkipHandler
             ] = None
@@ -487,8 +484,6 @@ class SnowflakeV2Source(
 
         databases = schema_extractor.databases
 
-        # TODO: The checkpoint state for stale entity detection can be committed here.
-
         if self.config.shares:
             yield from SnowflakeSharesHandler(
                 self.config, self.report
@@ -540,6 +535,7 @@ class SnowflakeV2Source(
                 identifiers=self.identifiers,
                 schema_resolver=schema_resolver,
                 discovered_tables=discovered_datasets,
+                graph=self.ctx.graph,
             )
 
             # TODO: This is slightly suboptimal because we create two SqlParsingAggregator instances with different configs

datahub/ingestion/source/sql/mssql/job_models.py

@@ -1,11 +1,17 @@
 from dataclasses import dataclass, field
 from typing import Dict, List, Optional, Union
 
-from datahub.emitter.mce_builder import make_data_flow_urn, make_data_job_urn
+from datahub.emitter.mce_builder import (
+    make_data_flow_urn,
+    make_data_job_urn,
+    make_data_platform_urn,
+    make_dataplatform_instance_urn,
+)
 from datahub.metadata.schema_classes import (
     DataFlowInfoClass,
     DataJobInfoClass,
     DataJobInputOutputClass,
+    DataPlatformInstanceClass,
 )
 
 
@@ -204,6 +210,18 @@ class MSSQLDataJob:
             status=self.status,
         )
 
+    @property
+    def as_maybe_platform_instance_aspect(self) -> Optional[DataPlatformInstanceClass]:
+        if self.entity.flow.platform_instance:
+            return DataPlatformInstanceClass(
+                platform=make_data_platform_urn(self.entity.flow.orchestrator),
+                instance=make_dataplatform_instance_urn(
+                    platform=self.entity.flow.orchestrator,
+                    instance=self.entity.flow.platform_instance,
+                ),
+            )
+        return None
+
 
 @dataclass
 class MSSQLDataFlow:
@@ -238,3 +256,14 @@ class MSSQLDataFlow:
             customProperties=self.flow_properties,
             externalUrl=self.external_url,
         )
+
+    @property
+    def as_maybe_platform_instance_aspect(self) -> Optional[DataPlatformInstanceClass]:
+        if self.entity.platform_instance:
+            return DataPlatformInstanceClass(
+                platform=make_data_platform_urn(self.entity.orchestrator),
+                instance=make_dataplatform_instance_urn(
+                    self.entity.orchestrator, self.entity.platform_instance
+                ),
+            )
+        return None

datahub/ingestion/source/sql/mssql/source.py

@@ -639,6 +639,13 @@ class SQLServerSource(SQLAlchemySource):
             aspect=data_job.as_datajob_info_aspect,
         ).as_workunit()
 
+        data_platform_instance_aspect = data_job.as_maybe_platform_instance_aspect
+        if data_platform_instance_aspect:
+            yield MetadataChangeProposalWrapper(
+                entityUrn=data_job.urn,
+                aspect=data_platform_instance_aspect,
+            ).as_workunit()
+
         if include_lineage:
             yield MetadataChangeProposalWrapper(
                 entityUrn=data_job.urn,
@@ -654,6 +661,13 @@ class SQLServerSource(SQLAlchemySource):
             entityUrn=data_flow.urn,
             aspect=data_flow.as_dataflow_info_aspect,
         ).as_workunit()
+
+        data_platform_instance_aspect = data_flow.as_maybe_platform_instance_aspect
+        if data_platform_instance_aspect:
+            yield MetadataChangeProposalWrapper(
+                entityUrn=data_flow.urn,
+                aspect=data_platform_instance_aspect,
+            ).as_workunit()
         # TODO: Add SubType when it appear
 
     def get_inspectors(self) -> Iterable[Inspector]:

datahub/ingestion/source/tableau/tableau.py

@@ -645,7 +645,7 @@ def report_user_role(report: TableauSourceReport, server: Server) -> None:
         # the site-role might be different on another site
         logged_in_user: UserInfo = UserInfo.from_server(server=server)
 
-        if not logged_in_user.is_site_administrator_explorer():
+        if not logged_in_user.has_site_administrator_explorer_privileges():
             report.warning(
                 title=title,
                 message=message,
@@ -896,10 +896,9 @@ class TableauSiteSource:
         return f"/{self.config.env.lower()}{self.no_env_browse_prefix}"
 
     def _re_authenticate(self):
-        tableau_auth: Union[
-            TableauAuth, PersonalAccessTokenAuth
-        ] = self.config.get_tableau_auth(self.site_id)
-        self.server.auth.sign_in(tableau_auth)
+        # Sign-in again may not be enough because Tableau sometimes caches invalid sessions
+        # so we need to recreate the Tableau Server object
+        self.server = self.config.make_tableau_client(self.site_id)
 
     @property
     def site_content_url(self) -> Optional[str]:

datahub/ingestion/source/tableau/tableau_constant.py

@@ -82,4 +82,6 @@ PROJECT = "Project"
 SITE = "Site"
 IS_UNSUPPORTED_CUSTOM_SQL = "isUnsupportedCustomSql"
 SITE_PERMISSION = "sitePermission"
-SITE_ROLE = "SiteAdministratorExplorer"
+ROLE_SITE_ADMIN_EXPLORER = "SiteAdministratorExplorer"
+ROLE_SITE_ADMIN_CREATOR = "SiteAdministratorCreator"
+ROLE_SERVER_ADMIN = "ServerAdministrator"

datahub/ingestion/source/tableau/tableau_server_wrapper.py

@@ -11,8 +11,12 @@ class UserInfo:
     site_role: str
     site_id: str
 
-    def is_site_administrator_explorer(self):
-        return self.site_role == c.SITE_ROLE
+    def has_site_administrator_explorer_privileges(self):
+        return self.site_role in [
+            c.ROLE_SITE_ADMIN_EXPLORER,
+            c.ROLE_SITE_ADMIN_CREATOR,
+            c.ROLE_SERVER_ADMIN,
+        ]
 
     @staticmethod
     def from_server(server: Server) -> "UserInfo":

datahub/ingestion/source/tableau/tableau_validation.py

@@ -28,7 +28,7 @@ def check_user_role(
 
     try:
         # TODO: Add check for `Enable Derived Permissions`
-        if not logged_in_user.is_site_administrator_explorer():
+        if not logged_in_user.has_site_administrator_explorer_privileges():
             capability_dict[c.SITE_PERMISSION] = CapabilityReport(
                 capable=False,
                 failure_reason=f"{failure_reason} Their current role is {logged_in_user.site_role}.",

datahub/ingestion/source/unity/source.py

@@ -26,6 +26,9 @@ from datahub.emitter.mcp_builder import (
     gen_containers,
 )
 from datahub.emitter.sql_parsing_builder import SqlParsingBuilder
+from datahub.ingestion.api.auto_work_units.auto_ensure_aspect_size import (
+    EnsureAspectSizeProcessor,
+)
 from datahub.ingestion.api.common import PipelineContext
 from datahub.ingestion.api.decorators import (
     SupportStatus,
@@ -260,6 +263,7 @@ class UnityCatalogSource(StatefulIngestionSourceBase, TestableSource):
             StaleEntityRemovalHandler.create(
                 self, self.config, self.ctx
             ).workunit_processor,
+            EnsureAspectSizeProcessor(self.get_report()).ensure_aspect_size,
         ]
 
     def get_workunits_internal(self) -> Iterable[MetadataWorkUnit]:

datahub/ingestion/source_report/ingestion_stage.py

@@ -14,6 +14,7 @@ LINEAGE_EXTRACTION = "Lineage Extraction"
 USAGE_EXTRACTION_INGESTION = "Usage Extraction Ingestion"
 USAGE_EXTRACTION_OPERATIONAL_STATS = "Usage Extraction Operational Stats"
 USAGE_EXTRACTION_USAGE_AGGREGATION = "Usage Extraction Usage Aggregation"
+EXTERNAL_TABLE_DDL_LINEAGE = "External table DDL Lineage"
 QUERIES_EXTRACTION = "Queries Extraction"
 PROFILING = "Profiling"
 

datahub/sql_parsing/sql_parsing_aggregator.py

@@ -490,7 +490,7 @@ class SqlParsingAggregator(Closeable):
             self._exit_stack.push(self._query_usage_counts)
 
         # Tool Extractor
-        self._tool_meta_extractor = ToolMetaExtractor()
+        self._tool_meta_extractor = ToolMetaExtractor.create(graph)
         self.report.tool_meta_report = self._tool_meta_extractor.report
 
     def close(self) -> None: