acryl-datahub 0.15.0.6rc1__py3-none-any.whl → 0.15.0.6rc2__py3-none-any.whl

{acryl_datahub-0.15.0.6rc1.dist-info → acryl_datahub-0.15.0.6rc2.dist-info}/RECORD

@@ -1,7 +1,8 @@
 datahub/__init__.py,sha256=aq_i5lVREmoLfYIqcx_pEQicO855YlhD19tWc1eZZNI,59
 datahub/__main__.py,sha256=pegIvQ9hzK7IhqVeUi1MeADSZ2QlP-D3K0OQdEg55RU,106
-datahub/_version.py,sha256=U9IGE-YR9bmigwAxXVjXLxWIGHYH0FW2G6D5UE_-ZIg,324
+datahub/_version.py,sha256=BhSRdcjgmWwrve2zKjSYh9Z1fk_7CHpbmijLhhWMrJM,324
 datahub/entrypoints.py,sha256=2TYgHhs3sCxJlojIHjqfxzt3_ImPwPzq4vBtsUuMqu4,8885
+datahub/errors.py,sha256=w6h8b27j9XlmPbTwqpu7-wgiTrXlHzcnUOnJ_iOrwzo,520
 datahub/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datahub/_codegen/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datahub/_codegen/aspect.py,sha256=PJRa-Z4ouXHq3OkulfyWhwZn-fFUBDK_UPvmqaWdbWk,1063
@@ -203,7 +204,7 @@ datahub/ingestion/source/ldap.py,sha256=CNr3foofIpoCXu_GGqfcajlQE2qkHr5isYwVcDut
 datahub/ingestion/source/metabase.py,sha256=m9Gfhrs8F1z23ci8CIxdE5cW--25stgxg_IQTKwkFrk,31532
 datahub/ingestion/source/mlflow.py,sha256=pmIkmsfidi7dOGdQ61rab7m8AnKZhIRE2IA9in9HGFU,12144
 datahub/ingestion/source/mode.py,sha256=HVxhzMIY4HjkAG_T6y00Po2B9XwjALP6i5XQThuyYM4,63488
-datahub/ingestion/source/mongodb.py,sha256=0P3PHVvMSXFkFimGvQzOQZF7APjsFOyzQAVQjVlVbuk,21172
+datahub/ingestion/source/mongodb.py,sha256=2C2Cxn8DXL53IbNiywIuKt8UT_EMcPg9f8su-OPSNGU,21237
 datahub/ingestion/source/nifi.py,sha256=FgIbZSCu-mcdnbIpqwvmADnUIxptogUq1sSEkrkwtrc,56089
 datahub/ingestion/source/openapi.py,sha256=MGsRLseZompW10UVMN_tU1GZgqPgTAM4lnqCJ8eVRoY,17386
 datahub/ingestion/source/openapi_parser.py,sha256=1_68wHWe_SzWYEyC1YVDw9vxoadKjW1yv8DecvyIhwY,13606
@@ -410,7 +411,7 @@ datahub/ingestion/source/s3/config.py,sha256=Zs1nrBZKLImteZreIcSMMRLj8vBGgxakNDs
 datahub/ingestion/source/s3/datalake_profiler_config.py,sha256=FfrcgK-JEF94vw-l3q6pN6FENXb-wZzW2w1VUZVkwW8,3620
 datahub/ingestion/source/s3/profiling.py,sha256=yKNCKpr6w7qpCH-baeSkNE9VjkN6eBot_weD-2_Jxzk,17579
 datahub/ingestion/source/s3/report.py,sha256=fzkTdTewYlWrTk4f2Cyl-e8RV4qw9wEVtm0cdKD-Xgw,542
-datahub/ingestion/source/s3/source.py,sha256=IE_K_HE_S7w8fpGPT8OptU5-VmwapntsI5PePv_wUQA,47412
+datahub/ingestion/source/s3/source.py,sha256=JwEmVWDEFtPt6iMo82n5DQava8QAKXk_xYl01KAfdHk,47614
 datahub/ingestion/source/sac/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datahub/ingestion/source/sac/sac.py,sha256=zPSO9ukuyhvNaaVzeAYpA-_sFma_XMcCQMPaGvDWuTk,30226
 datahub/ingestion/source/sac/sac_common.py,sha256=-xQTDBtgH56AnpRXWGDnlmQqUuLRx-7wF1U1kQFWtX8,998
@@ -444,8 +445,8 @@ datahub/ingestion/source/snowflake/snowflake_profiler.py,sha256=0DJiSwII6FY34url
 datahub/ingestion/source/snowflake/snowflake_queries.py,sha256=M-FBoYeiW91-g3gOUpCTj8cKWHH-wqyFtD5UcewfI2k,28121
 datahub/ingestion/source/snowflake/snowflake_query.py,sha256=Ex9FZZzz02cQis4bV3tzd53Pmf8p3AreuWnv9w95pJ0,39642
 datahub/ingestion/source/snowflake/snowflake_report.py,sha256=ahea-bwpW6T0iDehGo0Qq_J7wKxPkV61aYHm8bGwDqo,6651
-datahub/ingestion/source/snowflake/snowflake_schema.py,sha256=fdDN7jO5aHsmTFYC8cahXRT9BSAoDY72heM_WrkSxXo,25648
-datahub/ingestion/source/snowflake/snowflake_schema_gen.py,sha256=5DUP_uTmbLU01NegFXClqPkpB8LTruIkyIaGUBOTCQw,54718
+datahub/ingestion/source/snowflake/snowflake_schema.py,sha256=GFgcKV5T6VHyNwPBzzw_f8cWA9YFlWug0m6nkLoGXus,25979
+datahub/ingestion/source/snowflake/snowflake_schema_gen.py,sha256=Yba6OIWYtukAFMoNtEtX2BXWwJee17Dl58DUyK0myho,54530
 datahub/ingestion/source/snowflake/snowflake_shares.py,sha256=maZyFkfrbVogEFM0tTKRiNp9c_1muv6YfleSd3q0umI,6341
 datahub/ingestion/source/snowflake/snowflake_summary.py,sha256=kTmuCtRnvHqM8WBYhWeK4XafJq3ssFL9kcS03jEeWT4,5506
 datahub/ingestion/source/snowflake/snowflake_tag.py,sha256=1eLYTcgmfzDs9xktMTTE74L5SeNP48Qg3uLr9y-Ez3Y,8733
@@ -869,6 +870,16 @@ datahub/metadata/schemas/VersionSetKey.avsc,sha256=psjGNNcFua3Zs9Xlh4HnUHNmBEU74
 datahub/metadata/schemas/VersionSetProperties.avsc,sha256=yrhhVNioD11nFlDO7IfUbxAQjhA9Tr-4wnAYH5I9W74,1172
 datahub/metadata/schemas/ViewProperties.avsc,sha256=3HhcbH5493dJUnEUtFMYMVfbYQ52aDedm5L4j77Nym4,1032
 datahub/metadata/schemas/__init__.py,sha256=uvLNC3VyCkWA_v8e9FdA1leFf46NFKDD0AajCfihepI,581
+datahub/sdk/__init__.py,sha256=fYD-f338EW5WPFW2NSiirMAsHkNgZfolIvneM7yxgBk,977
+datahub/sdk/_all_entities.py,sha256=0XFtmgeEtrWOXy_oWcwqrtvfvzo8obPIq3Z1fEr5-34,400
+datahub/sdk/_attribution.py,sha256=05iNVT_IDO7aU3vU6dJKCF8EoDAI1mwh4mAg_EBf3RY,1121
+datahub/sdk/_entity.py,sha256=A_AWqNjaV6Y1YvmIiPNm0UyC0aO52G_L-iwgp9XEtYs,3225
+datahub/sdk/_shared.py,sha256=FIEcYGm3LgRKLqn_i1FDePaqvMljkXIA9f3RMqaF1kY,11137
+datahub/sdk/container.py,sha256=6Y19pQLN5vXGM86TEKW7ItTBwE3S0U4BsAMilWhyXqw,6511
+datahub/sdk/dataset.py,sha256=BydV2papBEYL2de72UMNjSkAbsDNpVlXqiEkMlXJiLY,21514
+datahub/sdk/entity_client.py,sha256=DcHytfCM8X9J6mm_QXzFR-2vDQa88I9Q2ktSNC2oSUI,4277
+datahub/sdk/main_client.py,sha256=IKKzBMiKtT5zIMEHrvaMyxJ7DYBVNpqua70J0Ydl74Y,2068
+datahub/sdk/resolver_client.py,sha256=UxI0bwg0ITm9dWPmAAKxkTvpLiGACtSCuEDNxLJipjs,3395
 datahub/secret/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datahub/secret/datahub_secret_store.py,sha256=9u9S87-15jwhj4h0EsAVIMdQLgvstKc8voQux2slxgU,2477
 datahub/secret/datahub_secrets_client.py,sha256=nDmhziKdvseJHlaDVUcAwK8Fv8maeAaG-ktZtWG2b70,1316
@@ -1001,9 +1012,9 @@ datahub_provider/operators/datahub_assertion_operator.py,sha256=uvTQ-jk2F0sbqqxp
 datahub_provider/operators/datahub_assertion_sensor.py,sha256=lCBj_3x1cf5GMNpHdfkpHuyHfVxsm6ff5x2Z5iizcAo,140
 datahub_provider/operators/datahub_operation_operator.py,sha256=aevDp2FzX7FxGlXrR0khoHNbxbhKR2qPEX5e8O2Jyzw,174
 datahub_provider/operators/datahub_operation_sensor.py,sha256=8fcdVBCEPgqy1etTXgLoiHoJrRt_nzFZQMdSzHqSG7M,168
-acryl_datahub-0.15.0.6rc1.dist-info/LICENSE,sha256=9xNHpsD0uYF5ONzXsKDCuHHB-xbiCrSbueWXqrTNsxk,11365
-acryl_datahub-0.15.0.6rc1.dist-info/METADATA,sha256=YPtgGtRPaltE8LkNq8PlWsrlzhbE-FQV0wWYkkuPqDc,175375
-acryl_datahub-0.15.0.6rc1.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
-acryl_datahub-0.15.0.6rc1.dist-info/entry_points.txt,sha256=U1e5ZwqPX1OaIbvGrwvozcdB8SbzFYXQM7plpdLKKeo,9592
-acryl_datahub-0.15.0.6rc1.dist-info/top_level.txt,sha256=iLjSrLK5ox1YVYcglRUkcvfZPvKlobBWx7CTUXx8_GI,25
-acryl_datahub-0.15.0.6rc1.dist-info/RECORD,,
+acryl_datahub-0.15.0.6rc2.dist-info/LICENSE,sha256=9xNHpsD0uYF5ONzXsKDCuHHB-xbiCrSbueWXqrTNsxk,11365
+acryl_datahub-0.15.0.6rc2.dist-info/METADATA,sha256=Bv5h1BaxWarbwfDdiiaD1wwyonRclpbgYtingIwN4B4,175375
+acryl_datahub-0.15.0.6rc2.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+acryl_datahub-0.15.0.6rc2.dist-info/entry_points.txt,sha256=U1e5ZwqPX1OaIbvGrwvozcdB8SbzFYXQM7plpdLKKeo,9592
+acryl_datahub-0.15.0.6rc2.dist-info/top_level.txt,sha256=iLjSrLK5ox1YVYcglRUkcvfZPvKlobBWx7CTUXx8_GI,25
+acryl_datahub-0.15.0.6rc2.dist-info/RECORD,,

datahub/_version.py

@@ -1,6 +1,6 @@
 # Published at https://pypi.org/project/acryl-datahub/.
 __package_name__ = "acryl-datahub"
-__version__ = "0.15.0.6rc1"
+__version__ = "0.15.0.6rc2"
 
 
 def is_dev_mode() -> bool:

datahub/errors.py

@@ -0,0 +1,35 @@
+from datahub.configuration.common import MetaError
+
+# TODO: Move all other error types to this file.
+
+
+class SdkUsageError(MetaError):
+    pass
+
+
+class AlreadyExistsError(SdkUsageError):
+    pass
+
+
+class ItemNotFoundError(SdkUsageError):
+    pass
+
+
+class MultipleItemsFoundError(SdkUsageError):
+    pass
+
+
+class SchemaFieldKeyError(SdkUsageError, KeyError):
+    pass
+
+
+class IngestionAttributionWarning(Warning):
+    pass
+
+
+class MultipleSubtypesWarning(Warning):
+    pass
+
+
+class ExperimentalWarning(Warning):
+    pass

datahub/ingestion/source/mongodb.py

@@ -219,26 +219,27 @@ def construct_schema_pymongo(
     """
 
     aggregations: List[Dict] = []
+
+    # The order of the aggregations impacts execution time. By setting the sample/limit aggregation first,
+    # the subsequent aggregations process a much smaller dataset, improving performance.
+    if sample_size:
+        if use_random_sampling:
+            aggregations.append({"$sample": {"size": sample_size}})
+        else:
+            aggregations.append({"$limit": sample_size})
+
     if should_add_document_size_filter:
         doc_size_field = "temporary_doc_size_field"
         # create a temporary field to store the size of the document. filter on it and then remove it.
-        aggregations = [
-            {"$addFields": {doc_size_field: {"$bsonSize": "$$ROOT"}}},
-            {"$match": {doc_size_field: {"$lt": max_document_size}}},
-            {"$project": {doc_size_field: 0}},
-        ]
-    if use_random_sampling:
-        # get sample documents in collection
-        if sample_size:
-            aggregations.append({"$sample": {"size": sample_size}})
-        documents = collection.aggregate(
-            aggregations,
-            allowDiskUse=True,
+        aggregations.extend(
+            [
+                {"$addFields": {doc_size_field: {"$bsonSize": "$$ROOT"}}},
+                {"$match": {doc_size_field: {"$lt": max_document_size}}},
+                {"$project": {doc_size_field: 0}},
+            ]
         )
-    else:
-        if sample_size:
-            aggregations.append({"$limit": sample_size})
-        documents = collection.aggregate(aggregations, allowDiskUse=True)
+
+    documents = collection.aggregate(aggregations, allowDiskUse=True)
 
     return construct_schema(list(documents), delimiter)
 

datahub/ingestion/source/s3/source.py

@@ -866,8 +866,21 @@ class S3Source(StatefulIngestionSourceBase):
         Returns:
         List[Folder]: A list of Folder objects representing the partitions found.
         """
+
+        def _is_allowed_path(path_spec_: PathSpec, s3_uri: str) -> bool:
+            allowed = path_spec_.allowed(s3_uri)
+            if not allowed:
+                logger.debug(f"File {s3_uri} not allowed and skipping")
+                self.report.report_file_dropped(s3_uri)
+            return allowed
+
+        s3_objects = (
+            obj
+            for obj in bucket.objects.filter(Prefix=prefix).page_size(PAGE_SIZE)
+            if _is_allowed_path(path_spec, f"s3://{obj.bucket_name}/{obj.key}")
+        )
+
         partitions: List[Folder] = []
-        s3_objects = bucket.objects.filter(Prefix=prefix).page_size(PAGE_SIZE)
         grouped_s3_objects_by_dirname = groupby_unsorted(
             s3_objects,
             key=lambda obj: obj.key.rsplit("/", 1)[0],
@@ -878,10 +891,6 @@ class S3Source(StatefulIngestionSourceBase):
             modification_time = None
 
             for item in group:
-                file_path = self.create_s3_path(item.bucket_name, item.key)
-                if not path_spec.allowed(file_path):
-                    logger.debug(f"File {file_path} not allowed and skipping")
-                    continue
                 file_size += item.size
                 if creation_time is None or item.last_modified < creation_time:
                     creation_time = item.last_modified

datahub/ingestion/source/snowflake/snowflake_schema.py

@@ -6,6 +6,7 @@ from datetime import datetime
 from typing import Callable, Dict, Iterable, List, MutableMapping, Optional
 
 from datahub.ingestion.api.report import SupportsAsObj
+from datahub.ingestion.source.common.subtypes import DatasetSubTypes
 from datahub.ingestion.source.snowflake.constants import SnowflakeObjectDomain
 from datahub.ingestion.source.snowflake.snowflake_connection import SnowflakeConnection
 from datahub.ingestion.source.snowflake.snowflake_query import (
@@ -100,6 +101,9 @@ class SnowflakeTable(BaseTable):
     def is_hybrid(self) -> bool:
         return self.type is not None and self.type == "HYBRID TABLE"
 
+    def get_subtype(self) -> DatasetSubTypes:
+        return DatasetSubTypes.TABLE
+
 
 @dataclass
 class SnowflakeView(BaseView):
@@ -109,6 +113,9 @@ class SnowflakeView(BaseView):
     column_tags: Dict[str, List[SnowflakeTag]] = field(default_factory=dict)
     is_secure: bool = False
 
+    def get_subtype(self) -> DatasetSubTypes:
+        return DatasetSubTypes.VIEW
+
 
 @dataclass
 class SnowflakeSchema:
@@ -154,6 +161,9 @@ class SnowflakeStream:
     column_tags: Dict[str, List[SnowflakeTag]] = field(default_factory=dict)
     last_altered: Optional[datetime] = None
 
+    def get_subtype(self) -> DatasetSubTypes:
+        return DatasetSubTypes.SNOWFLAKE_STREAM
+
 
 class _SnowflakeTagCache:
     def __init__(self) -> None:

datahub/ingestion/source/snowflake/snowflake_schema_gen.py

@@ -21,7 +21,6 @@ from datahub.ingestion.glossary.classification_mixin import (
 from datahub.ingestion.source.aws.s3_util import make_s3_urn_for_lineage
 from datahub.ingestion.source.common.subtypes import (
     DatasetContainerSubTypes,
-    DatasetSubTypes,
 )
 from datahub.ingestion.source.snowflake.constants import (
     GENERIC_PERMISSION_ERROR_KEY,
@@ -467,7 +466,13 @@ class SnowflakeSchemaGenerator(SnowflakeStructuredReportMixin):
                 context=f"{db_name}.{schema_name}",
             )
 
-    def _process_tags(self, snowflake_schema, schema_name, db_name, domain):
+    def _process_tags(
+        self,
+        snowflake_schema: SnowflakeSchema,
+        schema_name: str,
+        db_name: str,
+        domain: str,
+    ) -> None:
         snowflake_schema.tags = self.tag_extractor.get_tags_on_object(
             schema_name=schema_name, db_name=db_name, domain=domain
         )
@@ -837,15 +842,7 @@ class SnowflakeSchemaGenerator(SnowflakeStructuredReportMixin):
         if dpi_aspect:
             yield dpi_aspect
 
-        subTypes = SubTypes(
-            typeNames=(
-                [DatasetSubTypes.SNOWFLAKE_STREAM]
-                if isinstance(table, SnowflakeStream)
-                else [DatasetSubTypes.VIEW]
-                if isinstance(table, SnowflakeView)
-                else [DatasetSubTypes.TABLE]
-            )
-        )
+        subTypes = SubTypes(typeNames=[table.get_subtype()])
 
         yield MetadataChangeProposalWrapper(
             entityUrn=dataset_urn, aspect=subTypes
@@ -932,9 +929,9 @@ class SnowflakeSchemaGenerator(SnowflakeStructuredReportMixin):
                         "OWNER_ROLE_TYPE": table.owner_role_type,
                         "TABLE_NAME": table.table_name,
                         "BASE_TABLES": table.base_tables,
-                        "STALE_AFTER": table.stale_after.isoformat()
-                        if table.stale_after
-                        else None,
+                        "STALE_AFTER": (
+                            table.stale_after.isoformat() if table.stale_after else None
+                        ),
                     }.items()
                     if v
                 }

datahub/sdk/__init__.py

@@ -0,0 +1,33 @@
+import warnings
+
+import datahub.metadata.schema_classes as models
+from datahub.errors import ExperimentalWarning, SdkUsageError
+from datahub.ingestion.graph.config import DatahubClientConfig
+from datahub.metadata.urns import (
+    ChartUrn,
+    ContainerUrn,
+    CorpGroupUrn,
+    CorpUserUrn,
+    DashboardUrn,
+    DataPlatformInstanceUrn,
+    DataPlatformUrn,
+    DatasetUrn,
+    DomainUrn,
+    GlossaryTermUrn,
+    SchemaFieldUrn,
+    TagUrn,
+)
+from datahub.sdk.container import Container
+from datahub.sdk.dataset import Dataset
+from datahub.sdk.main_client import DataHubClient
+
+warnings.warn(
+    "The new datahub SDK (e.g. datahub.sdk.*) is experimental. "
+    "Our typical backwards-compatibility and stability guarantees do not apply to this code. "
+    "When it's promoted to stable, the import path will change "
+    "from `from datahub.sdk import ...` to `from datahub import ...`.",
+    ExperimentalWarning,
+    stacklevel=2,
+)
+del warnings
+del ExperimentalWarning

datahub/sdk/_all_entities.py

@@ -0,0 +1,15 @@
+from typing import Dict, List, Type
+
+from datahub.sdk._entity import Entity
+from datahub.sdk.container import Container
+from datahub.sdk.dataset import Dataset
+
+# TODO: Is there a better way to declare this?
+ENTITY_CLASSES_LIST: List[Type[Entity]] = [
+    Container,
+    Dataset,
+]
+
+ENTITY_CLASSES: Dict[str, Type[Entity]] = {
+    cls.get_urn_type().ENTITY_TYPE: cls for cls in ENTITY_CLASSES_LIST
+}

datahub/sdk/_attribution.py

@@ -0,0 +1,48 @@
+from __future__ import annotations
+
+import contextlib
+from typing import Iterator
+
+from datahub.utilities.str_enum import StrEnum
+
+
+class KnownAttribution(StrEnum):
+    INGESTION = "INGESTION"
+    INGESTION_ALTERNATE = "INGESTION_ALTERNATE"
+
+    UI = "UI"
+    SDK = "SDK"
+
+    PROPAGATION = "PROPAGATION"
+
+    def is_ingestion(self) -> bool:
+        return self in (
+            KnownAttribution.INGESTION,
+            KnownAttribution.INGESTION_ALTERNATE,
+        )
+
+
+_default_attribution = KnownAttribution.SDK
+
+
+def get_default_attribution() -> KnownAttribution:
+    return _default_attribution
+
+
+def set_default_attribution(attribution: KnownAttribution) -> None:
+    global _default_attribution
+    _default_attribution = attribution
+
+
+@contextlib.contextmanager
+def change_default_attribution(attribution: KnownAttribution) -> Iterator[None]:
+    old_attribution = get_default_attribution()
+    try:
+        set_default_attribution(attribution)
+        yield
+    finally:
+        set_default_attribution(old_attribution)
+
+
+def is_ingestion_attribution() -> bool:
+    return get_default_attribution().is_ingestion()

datahub/sdk/_entity.py

@@ -0,0 +1,89 @@
+import abc
+from typing import List, Optional, Type, Union
+
+from typing_extensions import Self
+
+import datahub.metadata.schema_classes as models
+from datahub.emitter.mce_builder import Aspect as AspectTypeVar
+from datahub.emitter.mcp import MetadataChangeProposalWrapper
+from datahub.errors import SdkUsageError
+from datahub.metadata.urns import Urn
+from datahub.utilities.urns._urn_base import _SpecificUrn
+
+
+class Entity:
+    __slots__ = ("_urn", "_prev_aspects", "_aspects")
+
+    def __init__(self, /, urn: Urn):
+        # This method is not meant for direct usage.
+        if type(self) is Entity:
+            raise SdkUsageError(f"{Entity.__name__} cannot be instantiated directly.")
+
+        assert isinstance(urn, self.get_urn_type())
+        self._urn: _SpecificUrn = urn
+
+        # prev_aspects is None means this was created from scratch
+        self._prev_aspects: Optional[models.AspectBag] = None
+        self._aspects: models.AspectBag = {}
+
+    @classmethod
+    def _new_from_graph(cls, urn: Urn, current_aspects: models.AspectBag) -> Self:
+        # If an init method from a subclass adds required fields, it also needs to override this method.
+        # An alternative approach would call cls.__new__() to bypass the init method, but it's a bit
+        # too hacky for my taste.
+        entity = cls(urn=urn)
+        return entity._init_from_graph(current_aspects)
+
+    def _init_from_graph(self, current_aspects: models.AspectBag) -> Self:
+        self._prev_aspects = current_aspects
+        aspect: models._Aspect
+        for aspect_name, aspect in (current_aspects or {}).items():  # type: ignore
+            aspect_copy = type(aspect).from_obj(aspect.to_obj())
+            self._aspects[aspect_name] = aspect_copy  # type: ignore
+        return self
+
+    @classmethod
+    @abc.abstractmethod
+    def get_urn_type(cls) -> Type[_SpecificUrn]: ...
+
+    @property
+    def urn(self) -> _SpecificUrn:
+        return self._urn
+
+    def _get_aspect(
+        self,
+        aspect_type: Type[AspectTypeVar],
+        /,
+    ) -> Optional[AspectTypeVar]:
+        return self._aspects.get(aspect_type.ASPECT_NAME)  # type: ignore
+
+    def _set_aspect(self, value: AspectTypeVar, /) -> None:
+        self._aspects[value.ASPECT_NAME] = value  # type: ignore
+
+    def _setdefault_aspect(self, default_aspect: AspectTypeVar, /) -> AspectTypeVar:
+        # Similar semantics to dict.setdefault.
+        if existing_aspect := self._get_aspect(type(default_aspect)):
+            return existing_aspect
+        self._set_aspect(default_aspect)
+        return default_aspect
+
+    def _as_mcps(
+        self,
+        change_type: Union[str, models.ChangeTypeClass] = models.ChangeTypeClass.UPSERT,
+    ) -> List[MetadataChangeProposalWrapper]:
+        urn_str = str(self.urn)
+
+        mcps = []
+        for aspect in self._aspects.values():
+            assert isinstance(aspect, models._Aspect)
+            mcps.append(
+                MetadataChangeProposalWrapper(
+                    entityUrn=urn_str,
+                    aspect=aspect,
+                    changeType=change_type,
+                )
+            )
+        return mcps
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}('{self.urn}')"