update to include max datarow operations

Author: jtsodapop

labelbox/schema/dataset.py

@@ -15,12 +15,16 @@
 from labelbox.exceptions import InvalidQueryError, LabelboxError, ResourceNotFoundError, InvalidAttributeError
 from labelbox.orm.db_object import DbObject, Updateable, Deletable
 from labelbox.orm.model import Entity, Field, Relationship
+from labelbox.exceptions import MalformedQueryException
 
 if TYPE_CHECKING:
     from labelbox import Task, User, DataRow
 
 logger = logging.getLogger(__name__)
 
+MAX_DATAROW_PER_API_OPERATION = 150000
+MAX_DATAROW_WITH_METADATA = 30000
+
 
 class Dataset(DbObject, Updateable, Deletable):
     """ A Dataset is a collection of DataRows.
@@ -393,14 +397,18 @@ def convert_item(item):
 
         # TODO: If any datarows contain metadata, we're limiting max # of datarows
         # until we address performance issues with datarow create with metadata
-        max_datarow_with_metadata = 30_000
-        if (len(items) > max_datarow_with_metadata):
+        if (len(items) > MAX_DATAROW_WITH_METADATA):
             for row in items:
                 if 'metadata_fields' in row:
-                    raise ValueError(
-                        f"Cannot create more than {max_datarow_with_metadata} DataRows, if any DataRows contain metadata"
+                    raise MalformedQueryException(
+                        f"Cannot create more than {MAX_DATAROW_WITH_METADATA} DataRows, if any DataRows contain metadata"
                     )
 
+        if len(items) > MAX_DATAROW_PER_API_OPERATION:
+            raise MalformedQueryException(
+                f"Cannot create more than {MAX_DATAROW_PER_API_OPERATION} DataRows per function call."
+            )
+
         with ThreadPoolExecutor(file_upload_thread_count) as executor:
             futures = [executor.submit(convert_item, item) for item in items]
             items = [future.result() for future in as_completed(futures)]