This project aims to provide an easy way to validate a DBT project by enabling the unit-testing capability.
This project is based on the pytest-postgresql library, so we can validate our DBT postgres project without relying on an existing database.
The initial idea resides on pytest-postgres providing us an empty and isolated postgres where we can validate different use-cases with our DBT code.
If your project uses external sources, you will need to make sure they are properly defined in the yml files, since the project generates the DDLs from there.
import unittest
from pytest_dbt_postgres import auto_inject_fixtures
from pytest_dbt_postgres.dbt_executor import DbtExecutor
from pytest_dbt_postgres.dbt_validator import DbtValidator
from pytest_dbt_postgres.postgres.connector import PostgresConnector
from pytest_dbt_postgres.postgres.sql_methods import export_credentials_to_env, get_credentials
dbt_project_dir = '...'
resources_folder = '...'
@auto_inject_fixtures("postgresql")
class TestDbtValidator(unittest.TestCase):
def setUp(self) -> None:
credentials = get_credentials(self.postgresql)
self.connector = PostgresConnector(credentials)
# SETUP
export_credentials_to_env(credentials)
executor = DbtExecutor(dbt_project_dir=dbt_project_dir, profiles_dir=resources_folder)
self.validator = DbtValidator(connector=self.connector, executor=executor, resources_folder=resources_folder)
def test_orders(self) -> None:
# GIVEN
sources_to_load = [
("external.customers", "csv/input/customers.csv"),
("external.orders", "csv/input/orders.csv"),
]
selector = "orders"
outputs_to_validate = [
("dbt_unittest.orders", "csv/output/orders.csv"),
]
# THEN
self.validator.validate(sources_to_load, selector, outputs_to_validate)