openmetadata-ingestion 1.3.2.0rc1__py3-none-any.whl → 1.3.2.0rc3__py3-none-any.whl

metadata/ingestion/api/parser.py

@@ -69,6 +69,28 @@ from metadata.generated.schema.metadataIngestion.databaseServiceQueryUsagePipeli
     DatabaseServiceQueryUsagePipeline,
     DatabaseUsageConfigType,
 )
+from metadata.generated.schema.metadataIngestion.dbtconfig.dbtAzureConfig import (
+    DbtAzureConfig,
+)
+from metadata.generated.schema.metadataIngestion.dbtconfig.dbtCloudConfig import (
+    DbtCloudConfig,
+)
+from metadata.generated.schema.metadataIngestion.dbtconfig.dbtGCSConfig import (
+    DbtGcsConfig,
+)
+from metadata.generated.schema.metadataIngestion.dbtconfig.dbtHttpConfig import (
+    DbtHttpConfig,
+)
+from metadata.generated.schema.metadataIngestion.dbtconfig.dbtLocalConfig import (
+    DbtLocalConfig,
+)
+from metadata.generated.schema.metadataIngestion.dbtconfig.dbtS3Config import (
+    DbtS3Config,
+)
+from metadata.generated.schema.metadataIngestion.dbtPipeline import (
+    DbtConfigType,
+    DbtPipeline,
+)
 from metadata.generated.schema.metadataIngestion.messagingServiceMetadataPipeline import (
     MessagingMetadataConfigType,
     MessagingServiceMetadataPipeline,
@@ -125,6 +147,16 @@ SOURCE_CONFIG_CLASS_MAP = {
     DatabaseMetadataConfigType.DatabaseMetadata.value: DatabaseServiceMetadataPipeline,
     StorageMetadataConfigType.StorageMetadata.value: StorageServiceMetadataPipeline,
     SearchMetadataConfigType.SearchMetadata.value: SearchServiceMetadataPipeline,
+    DbtConfigType.DBT.value: DbtPipeline,
+}
+
+DBT_CONFIG_TYPE_MAP = {
+    "cloud": DbtCloudConfig,
+    "local": DbtLocalConfig,
+    "http": DbtHttpConfig,
+    "s3": DbtS3Config,
+    "gcs": DbtGcsConfig,
+    "azure": DbtAzureConfig,
 }
 
 
@@ -171,6 +203,7 @@ def get_source_config_class(
     Type[PipelineServiceMetadataPipeline],
     Type[MlModelServiceMetadataPipeline],
     Type[DatabaseServiceMetadataPipeline],
+    Type[DbtPipeline],
 ]:
     """
     Return the source config type for a source string
@@ -179,7 +212,7 @@ def get_source_config_class(
     """
     source_config_class = SOURCE_CONFIG_CLASS_MAP.get(source_config_type)
 
-    if source_config_type:
+    if source_config_class:
         return source_config_class
 
     raise ValueError(f"Cannot find the service type of {source_config_type}")
@@ -266,6 +299,27 @@ def _unsafe_parse_config(config: dict, cls: Type[T], message: str) -> None:
         raise err
 
 
+def _unsafe_parse_dbt_config(config: dict, cls: Type[T], message: str) -> None:
+    """
+    Given a config dictionary and the class it should match,
+    try to parse it or log the given message
+    """
+    logger.debug(f"Parsing message: [{message}]")
+    try:
+        # Parse the oneOf config types of dbt to check
+        dbt_config_type = config["dbtConfigSource"]["dbtConfigType"]
+        dbt_config_class = DBT_CONFIG_TYPE_MAP.get(dbt_config_type)
+        dbt_config_class.parse_obj(config["dbtConfigSource"])
+
+        # Parse the entire dbtPipeline object
+        cls.parse_obj(config)
+    except ValidationError as err:
+        logger.debug(
+            f"The supported properties for {cls.__name__} are {list(cls.__fields__.keys())}"
+        )
+        raise err
+
+
 def _parse_inner_connection(config_dict: dict, source_type: str) -> None:
     """
     Parse the inner connection of the flagged connectors
@@ -291,32 +345,35 @@ def parse_service_connection(config_dict: dict) -> None:
     :param config_dict: JSON configuration
     """
     # Unsafe access to the keys. Allow a KeyError if the config is not well formatted
-    source_type = config_dict["source"]["serviceConnection"]["config"].get("type")
-    if source_type is None:
-        raise InvalidWorkflowException("Missing type in the serviceConnection config")
+    if config_dict["source"].get("serviceConnection"):
+        source_type = config_dict["source"]["serviceConnection"]["config"].get("type")
+        if source_type is None:
+            raise InvalidWorkflowException(
+                "Missing type in the serviceConnection config"
+            )
 
-    logger.debug(
-        f"Error parsing the Workflow Configuration for {source_type} ingestion"
-    )
+        logger.debug(
+            f"Error parsing the Workflow Configuration for {source_type} ingestion"
+        )
 
-    service_type = get_service_type(source_type)
-    connection_class = get_connection_class(source_type, service_type)
+        service_type = get_service_type(source_type)
+        connection_class = get_connection_class(source_type, service_type)
 
-    if source_type in HAS_INNER_CONNECTION:
-        # We will first parse the inner `connection` configuration
-        _parse_inner_connection(
-            config_dict["source"]["serviceConnection"]["config"]["connection"][
-                "config"
-            ]["connection"],
-            source_type,
-        )
+        if source_type in HAS_INNER_CONNECTION:
+            # We will first parse the inner `connection` configuration
+            _parse_inner_connection(
+                config_dict["source"]["serviceConnection"]["config"]["connection"][
+                    "config"
+                ]["connection"],
+                source_type,
+            )
 
-    # Parse the service connection dictionary with the scoped class
-    _unsafe_parse_config(
-        config=config_dict["source"]["serviceConnection"]["config"],
-        cls=connection_class,
-        message="Error parsing the service connection",
-    )
+        # Parse the service connection dictionary with the scoped class
+        _unsafe_parse_config(
+            config=config_dict["source"]["serviceConnection"]["config"],
+            cls=connection_class,
+            message="Error parsing the service connection",
+        )
 
 
 def parse_source_config(config_dict: dict) -> None:
@@ -334,6 +391,13 @@ def parse_source_config(config_dict: dict) -> None:
 
     source_config_class = get_source_config_class(source_config_type)
 
+    if source_config_class == DbtPipeline:
+        _unsafe_parse_dbt_config(
+            config=config_dict["source"]["sourceConfig"]["config"],
+            cls=source_config_class,
+            message="Error parsing the dbt source config",
+        )
+
     _unsafe_parse_config(
         config=config_dict["source"]["sourceConfig"]["config"],
         cls=source_config_class,

metadata/ingestion/api/topology_runner.py

@@ -265,11 +265,12 @@ class TopologyRunnerMixin(Generic[C]):
             if entity:
                 same_fingerprint = True
 
-        create_entity_request_hash = generate_source_hash(
-            create_request=entity_request.right,
-        )
+        create_entity_request_hash = None
 
         if hasattr(entity_request.right, "sourceHash"):
+            create_entity_request_hash = generate_source_hash(
+                create_request=entity_request.right,
+            )
             entity_request.right.sourceHash = create_entity_request_hash
 
         if entity is None and stage.use_cache:

metadata/ingestion/lineage/parser.py

@@ -217,6 +217,11 @@ class LineageParser:
         """
         aliases = self.table_aliases
         values = identifier.value.split(".")
+
+        if len(values) > 4:
+            logger.debug(f"Invalid comparison element from identifier: {identifier}")
+            return None, None
+
         database_name, schema_name, table_or_alias, column_name = (
             [None] * (4 - len(values))
         ) + values
@@ -307,32 +312,39 @@ class LineageParser:
                 comparisons.append(sub)
 
         for comparison in comparisons:
-            if "." not in comparison.left.value or "." not in comparison.right.value:
-                logger.debug(f"Ignoring comparison {comparison}")
-                continue
-
-            table_left, column_left = self.get_comparison_elements(
-                identifier=comparison.left
-            )
-            table_right, column_right = self.get_comparison_elements(
-                identifier=comparison.right
-            )
-
-            if not table_left or not table_right:
-                logger.warning(
-                    f"Can't extract table names when parsing JOIN information from {comparison}"
+            try:
+                if (
+                    "." not in comparison.left.value
+                    or "." not in comparison.right.value
+                ):
+                    logger.debug(f"Ignoring comparison {comparison}")
+                    continue
+
+                table_left, column_left = self.get_comparison_elements(
+                    identifier=comparison.left
+                )
+                table_right, column_right = self.get_comparison_elements(
+                    identifier=comparison.right
                 )
-                logger.debug(f"Query: {sql_statement}")
-                continue
 
-            left_table_column = TableColumn(table=table_left, column=column_left)
-            right_table_column = TableColumn(table=table_right, column=column_right)
+                if not table_left or not table_right:
+                    logger.warning(
+                        f"Can't extract table names when parsing JOIN information from {comparison}"
+                    )
+                    logger.debug(f"Query: {sql_statement}")
+                    continue
+
+                left_table_column = TableColumn(table=table_left, column=column_left)
+                right_table_column = TableColumn(table=table_right, column=column_right)
 
-            # We just send the info once, from Left -> Right.
-            # The backend will prepare the symmetric information.
-            self.stateful_add_table_joins(
-                join_data, left_table_column, right_table_column
-            )
+                # We just send the info once, from Left -> Right.
+                # The backend will prepare the symmetric information.
+                self.stateful_add_table_joins(
+                    join_data, left_table_column, right_table_column
+                )
+            except Exception as exc:
+                logger.debug(f"Cannot process comparison {comparison}: {exc}")
+                logger.debug(traceback.format_exc())
 
     @cached_property
     def table_joins(self) -> Dict[str, List[TableColumnJoin]]:

metadata/ingestion/ometa/routes.py

@@ -90,6 +90,16 @@ from metadata.generated.schema.api.tests.createTestDefinition import (
 from metadata.generated.schema.api.tests.createTestSuite import CreateTestSuiteRequest
 from metadata.generated.schema.dataInsight.dataInsightChart import DataInsightChart
 from metadata.generated.schema.dataInsight.kpi.kpi import Kpi
+from metadata.generated.schema.entity.applications.app import App
+from metadata.generated.schema.entity.applications.createAppRequest import (
+    CreateAppRequest,
+)
+from metadata.generated.schema.entity.applications.marketplace.appMarketPlaceDefinition import (
+    AppMarketPlaceDefinition,
+)
+from metadata.generated.schema.entity.applications.marketplace.createAppMarketPlaceDefinitionReq import (
+    CreateAppMarketPlaceDefinitionRequest,
+)
 from metadata.generated.schema.entity.automations.workflow import Workflow
 from metadata.generated.schema.entity.bot import Bot
 from metadata.generated.schema.entity.classification.classification import (
@@ -232,4 +242,9 @@ ROUTES = {
     # Suggestions
     Suggestion.__name__: "/suggestions",
     CreateSuggestionRequest.__name__: "/suggestions",
+    # Apps
+    App.__name__: "/apps",
+    CreateAppRequest.__name__: "/apps",
+    AppMarketPlaceDefinition.__name__: "/apps/marketplace",
+    CreateAppMarketPlaceDefinitionRequest.__name__: "/apps/marketplace",
 }

metadata/ingestion/source/database/datalake/metadata.py

@@ -407,7 +407,7 @@ class DatalakeSource(DatabaseServiceSource):
         schema_name = self.context.database_schema
         try:
             table_constraints = None
-            data_frame = fetch_dataframe(
+            data_frame, raw_data = fetch_dataframe(
                 config_source=self.config_source,
                 client=self.client,
                 file_fqn=DatalakeTableSchemaWrapper(
@@ -415,10 +415,11 @@ class DatalakeSource(DatabaseServiceSource):
                     bucket_name=schema_name,
                     file_extension=table_extension,
                 ),
+                fetch_raw_data=True,
             )
             if data_frame:
                 column_parser = DataFrameColumnParser.create(
-                    data_frame[0], table_extension
+                    data_frame[0], table_extension, raw_data=raw_data
                 )
                 columns = column_parser.get_columns()
             else:

metadata/ingestion/source/database/dbt/dbt_config.py

@@ -173,7 +173,7 @@ def _(config: DbtCloudConfig):  # pylint: disable=too-many-locals
             params_data["job_definition_id"] = job_id
 
         response = client.get(f"/accounts/{account_id}/runs", data=params_data)
-        if not response and not response.get("data"):
+        if not response or not response.get("data"):
             raise DBTConfigException(
                 "Unable to get the dbt job runs information.\n"
                 "Please check if the auth token is correct and has the necessary scopes to fetch dbt runs"

metadata/ingestion/source/database/oracle/connection.py

@@ -38,6 +38,7 @@ from metadata.ingestion.connections.builders import (
 )
 from metadata.ingestion.connections.test_connections import test_connection_db_common
 from metadata.ingestion.ometa.ometa_api import OpenMetadata
+from metadata.ingestion.source.database.oracle.queries import CHECK_ACCESS_TO_DBA
 from metadata.utils.logger import ingestion_logger
 
 CX_ORACLE_LIB_VERSION = "8.3.0"
@@ -136,9 +137,13 @@ def test_connection(
     Test connection. This can be executed either as part
     of a metadata workflow or during an Automation Workflow
     """
+
+    test_conn_queries = {"CheckAccess": CHECK_ACCESS_TO_DBA}
+
     test_connection_db_common(
         metadata=metadata,
         engine=engine,
         service_connection=service_connection,
         automation_workflow=automation_workflow,
+        queries=test_conn_queries,
     )

metadata/ingestion/source/database/oracle/queries.py

@@ -88,7 +88,7 @@ WHERE
     type = 'PROCEDURE' and owner = '{schema}'
 """
 )
-
+CHECK_ACCESS_TO_DBA = "SELECT table_name FROM DBA_TABLES where ROWNUM < 2"
 ORACLE_GET_STORED_PROCEDURE_QUERIES = textwrap.dedent(
     """
 WITH SP_HISTORY AS (SELECT

metadata/ingestion/source/database/stored_procedures_mixin.py

@@ -138,7 +138,7 @@ class StoredProcedureMixin(ABC):
             return True
 
         if query_type == "INSERT" and re.search(
-            "^.*insert.*into.*select.*$", query_text, re.IGNORECASE
+            "^.*insert.*into.*select.*$", query_text.replace("\n", " "), re.IGNORECASE
         ):
             return True
 

metadata/ingestion/source/database/unitycatalog/connection.py

@@ -68,16 +68,20 @@ def test_connection(
             break
 
     def get_schemas(connection: WorkspaceClient, table_obj: DatabricksTable):
-        for schema in connection.schemas.list(catalog_name=table_obj.catalog_name):
-            table_obj.schema_name = schema.name
-            break
+        for catalog in connection.catalogs.list():
+            for schema in connection.schemas.list(catalog_name=catalog.name):
+                if schema.name:
+                    table_obj.schema_name = schema.name
+                    table_obj.catalog_name = catalog.name
+                    return
 
     def get_tables(connection: WorkspaceClient, table_obj: DatabricksTable):
-        for table in connection.tables.list(
-            catalog_name=table_obj.catalog_name, schema_name=table_obj.schema_name
-        ):
-            table_obj.name = table.name
-            break
+        if table_obj.catalog_name and table_obj.schema_name:
+            for table in connection.tables.list(
+                catalog_name=table_obj.catalog_name, schema_name=table_obj.schema_name
+            ):
+                table_obj.name = table.name
+                break
 
     test_fn = {
         "CheckAccess": connection.catalogs.list,

metadata/ingestion/source/pipeline/dagster/metadata.py

@@ -214,7 +214,9 @@ class DagsterSource(PipelineServiceSource):
             service_name=self.context.pipeline_service,
             pipeline_name=self.context.pipeline,
         )
-        pipeline_entity = self.metadata.get_by_name(entity=Pipeline, fqn=pipeline_fqn)
+        pipeline_entity = self.metadata.get_by_name(
+            entity=Pipeline, fqn=pipeline_fqn, fields=["tasks"]
+        )
         for task in pipeline_entity.tasks or []:
             try:
                 runs = self.client.get_task_runs(

metadata/ingestion/source/storage/storage_service.py

@@ -260,7 +260,7 @@ class StorageServiceSource(TopologyRunnerMixin, Source, ABC):
         metadata_entry: MetadataEntry,
     ) -> List[Column]:
         """Extract Column related metadata from s3"""
-        data_structure_details = fetch_dataframe(
+        data_structure_details, raw_data = fetch_dataframe(
             config_source=config_source,
             client=client,
             file_fqn=DatalakeTableSchemaWrapper(
@@ -269,10 +269,13 @@ class StorageServiceSource(TopologyRunnerMixin, Source, ABC):
                 file_extension=SupportedTypes(metadata_entry.structureFormat),
                 separator=metadata_entry.separator,
             ),
+            fetch_raw_data=True,
         )
         columns = []
         column_parser = DataFrameColumnParser.create(
-            data_structure_details, SupportedTypes(metadata_entry.structureFormat)
+            data_structure_details,
+            SupportedTypes(metadata_entry.structureFormat),
+            raw_data=raw_data,
         )
         columns = column_parser.get_columns()
         return columns

metadata/parsers/json_schema_parser.py

@@ -18,6 +18,8 @@ import traceback
 from enum import Enum
 from typing import List, Optional
 
+from pydantic.main import ModelMetaclass
+
 from metadata.generated.schema.type.schema import FieldModel
 from metadata.utils.logger import ingestion_logger
 
@@ -36,20 +38,25 @@ class JsonSchemaDataTypes(Enum):
     NULL = "null"
     RECORD = "object"
     ARRAY = "array"
+    UNKNOWN = "unknown"
 
 
-def parse_json_schema(schema_text: str) -> Optional[List[FieldModel]]:
+def parse_json_schema(
+    schema_text: str, cls: ModelMetaclass = FieldModel
+) -> Optional[List[FieldModel]]:
     """
     Method to parse the jsonschema
     """
     try:
         json_schema_data = json.loads(schema_text)
         field_models = [
-            FieldModel(
+            cls(
                 name=json_schema_data.get("title", "default"),
                 dataType=JsonSchemaDataTypes(json_schema_data.get("type")).name,
                 description=json_schema_data.get("description"),
-                children=get_json_schema_fields(json_schema_data.get("properties")),
+                children=get_json_schema_fields(
+                    json_schema_data.get("properties", {}), cls=cls
+                ),
             )
         ]
         return field_models
@@ -59,7 +66,9 @@ def parse_json_schema(schema_text: str) -> Optional[List[FieldModel]]:
     return None
 
 
-def get_json_schema_fields(properties) -> Optional[List[FieldModel]]:
+def get_json_schema_fields(
+    properties, cls: ModelMetaclass = FieldModel
+) -> Optional[List[FieldModel]]:
     """
     Recursively convert the parsed schema into required models
     """
@@ -67,9 +76,10 @@ def get_json_schema_fields(properties) -> Optional[List[FieldModel]]:
     for key, value in properties.items():
         try:
             field_models.append(
-                FieldModel(
-                    name=value.get("title", key),
-                    dataType=JsonSchemaDataTypes(value.get("type")).name,
+                cls(
+                    name=key,
+                    displayName=value.get("title"),
+                    dataType=JsonSchemaDataTypes(value.get("type", "unknown")).name,
                     description=value.get("description"),
                     children=get_json_schema_fields(value.get("properties"))
                     if value.get("type") == "object"

metadata/profiler/interface/profiler_interface.py

@@ -33,7 +33,7 @@ from metadata.generated.schema.entity.data.table import (
     TableData,
 )
 from metadata.generated.schema.entity.services.connections.connectionBasicType import (
-    SampleDataStorageConfig,
+    DataStorageConfig,
 )
 from metadata.generated.schema.entity.services.connections.database.datalakeConnection import (
     DatalakeConnection,
@@ -93,7 +93,7 @@ class ProfilerInterface(ABC):
         service_connection_config: Union[DatabaseConnection, DatalakeConnection],
         ometa_client: OpenMetadata,
         entity: Table,
-        storage_config: SampleDataStorageConfig,
+        storage_config: DataStorageConfig,
         profile_sample_config: Optional[ProfileSampleConfig],
         source_config: DatabaseServiceProfilerPipeline,
         sample_query: Optional[str],
@@ -248,7 +248,7 @@ class ProfilerInterface(ABC):
             DatabaseProfilerConfig,
             DatabaseAndSchemaConfig,
         ]
-    ):
+    ) -> Optional[DataStorageConfig]:
         if (
             config
             and config.sampleDataStorageConfig
@@ -264,7 +264,7 @@ class ProfilerInterface(ABC):
         database_profiler_config: Optional[DatabaseProfilerConfig],
         db_service: Optional[DatabaseService],
         profiler_config: ProfilerProcessorConfig,
-    ) -> Optional[SampleDataStorageConfig]:
+    ) -> Optional[DataStorageConfig]:
         """Get config for a specific entity
 
         Args:

metadata/profiler/processor/sample_data_handler.py

@@ -17,8 +17,13 @@ from datetime import datetime
 from functools import singledispatch
 from io import BytesIO
 
+from pydantic.json import ENCODERS_BY_TYPE
+
 from metadata.clients.aws_client import AWSClient
 from metadata.generated.schema.entity.data.table import Table, TableData
+from metadata.generated.schema.entity.services.connections.connectionBasicType import (
+    DataStorageConfig,
+)
 from metadata.generated.schema.security.credentials.awsCredentials import AWSCredentials
 from metadata.profiler.interface.profiler_interface import ProfilerInterface
 from metadata.utils.helpers import clean_uri
@@ -27,15 +32,45 @@ from metadata.utils.logger import profiler_logger
 logger = profiler_logger()
 
 
-def _get_object_key(table: Table, prefix: str, overwrite_data: bool) -> str:
+class PathPatternException(Exception):
+    """
+    Exception class need to validate the file path pattern
+    """
+
+
+def validate_path_pattern(file_path_format: str) -> None:
+    if not (
+        "{service_name}" in file_path_format
+        and "{database_name}" in file_path_format
+        and "{database_schema_name}" in file_path_format
+        and "{table_name}" in file_path_format
+        and file_path_format.endswith(".parquet")
+    ):
+        raise PathPatternException(
+            "Please provide a valid path pattern, "
+            "the pattern should include these components {service_name}, "
+            "{database_name}, {database_schema_name}, {table_name} and "
+            "it should end with extension .parquet"
+        )
+
+
+def _get_object_key(
+    table: Table, prefix: str, overwrite_data: bool, file_path_format: str
+) -> str:
+    validate_path_pattern(file_path_format)
+    file_name = file_path_format.format(
+        service_name=table.service.name,
+        database_name=table.database.name,
+        database_schema_name=table.databaseSchema.name,
+        table_name=table.name.__root__,
+    )
     if not overwrite_data:
-        file_name = f"sample_data_{datetime.now().strftime('%Y_%m_%d')}.parquet"
-    else:
-        file_name = "sample_data.parquet"
-    path = str(table.fullyQualifiedName.__root__).replace(".", "/")
+        file_name = file_name.replace(
+            ".parquet", f"_{datetime.now().strftime('%Y_%m_%d')}.parquet"
+        )
     if prefix:
-        return f"{clean_uri(prefix)}/{path}/{file_name}"
-    return f"{path}/{file_name}"
+        return f"{clean_uri(prefix)}/{file_name}"
+    return file_name
 
 
 def upload_sample_data(data: TableData, profiler_interface: ProfilerInterface) -> None:
@@ -45,9 +80,10 @@ def upload_sample_data(data: TableData, profiler_interface: ProfilerInterface) -
     import pandas as pd  # pylint: disable=import-outside-toplevel
 
     try:
-        sample_storage_config = profiler_interface.storage_config
+        sample_storage_config: DataStorageConfig = profiler_interface.storage_config
         if not sample_storage_config:
             return
+        ENCODERS_BY_TYPE[bytes] = lambda v: v.decode("utf-8", "ignore")
         deserialized_data = json.loads(data.json())
         df = pd.DataFrame(
             data=deserialized_data.get("rows", []),
@@ -59,6 +95,7 @@ def upload_sample_data(data: TableData, profiler_interface: ProfilerInterface) -
             table=profiler_interface.table_entity,
             prefix=sample_storage_config.prefix,
             overwrite_data=sample_storage_config.overwriteData,
+            file_path_format=sample_storage_config.filePathPattern,
         )
         upload_to_storage(
             sample_storage_config.storageConfig,

metadata/readers/dataframe/json.py

@@ -16,7 +16,7 @@ import gzip
 import io
 import json
 import zipfile
-from typing import List, Union
+from typing import Any, Dict, List, Optional, Tuple, Union
 
 from metadata.readers.dataframe.base import DataFrameReader
 from metadata.readers.dataframe.common import dataframe_to_chunks
@@ -47,7 +47,7 @@ class JSONDataFrameReader(DataFrameReader):
     @staticmethod
     def read_from_json(
         key: str, json_text: bytes, decode: bool = False, **__
-    ) -> List["DataFrame"]:
+    ) -> Tuple[List["DataFrame"], Optional[Dict[str, Any]]]:
         """
         Decompress a JSON file (if needed) and read its contents
         as a dataframe.
@@ -60,20 +60,25 @@ class JSONDataFrameReader(DataFrameReader):
         import pandas as pd
 
         json_text = _get_json_text(key=key, text=json_text, decode=decode)
+        raw_data = None
         try:
             data = json.loads(json_text)
+            if isinstance(data, dict) and data.get("$schema"):
+                raw_data = json_text
         except json.decoder.JSONDecodeError:
             logger.debug("Failed to read as JSON object. Trying to read as JSON Lines")
             data = [json.loads(json_obj) for json_obj in json_text.strip().split("\n")]
 
         # if we get a scalar value (e.g. {"a":"b"}) then we need to specify the index
         data = data if not isinstance(data, dict) else [data]
-        return dataframe_to_chunks(pd.DataFrame.from_records(data))
+        return dataframe_to_chunks(pd.DataFrame.from_records(data)), raw_data
 
     def _read(self, *, key: str, bucket_name: str, **kwargs) -> DatalakeColumnWrapper:
         text = self.reader.read(key, bucket_name=bucket_name)
+        dataframes, raw_data = self.read_from_json(
+            key=key, json_text=text, decode=True, **kwargs
+        )
         return DatalakeColumnWrapper(
-            dataframes=self.read_from_json(
-                key=key, json_text=text, decode=True, **kwargs
-            )
+            dataframes=dataframes,
+            raw_data=raw_data,
         )