Load MSSQL source with no system tables

[tommy1x]

DLT version 1.6.0
I'm trying to load data from a Microsoft Dataverse TDS endpoint.
I'm setting up the connection by providing a SQLAlchemy engine instance, so I can use a Microsoft Entra ID token for authentication.
I'm getting the following error:
[Microsoft][ODBC Driver 18 for SQL Server][SQL Server]{"Message":"Invalid table name COLUMNS for schema INFORMATION_SCHEMA.

The TDS endpoint seems to have sys tables (e.g. sys.tables, sys.columns) as well as INFORMATION_SCHEMA.TABLES, but it doesn't seem to have INFORMATION_SCHEMA.COLUMNS.
I can get around it by specifying metadata for the load but that isn't ideal. Is there any way to load it without it trying to query INFORMATION_SCHEMA.COLUMNS, even without provided metadata?

This is roughly the code I've tried which works when providing metadata:

metadata = MetaData()
Table(
    "my_table",
    metadata,
    Column("id", String, primary_key=True),
    Column("other_column", String)
)
source = sql_table(
    my_sqlalchemy_engine,
    metadata=metadata,
    table="my_table",
    included_columns=["id", "other_column"]
)