Skip to content

Latest commit

 

History

History
49 lines (36 loc) · 3.13 KB

README.md

File metadata and controls

49 lines (36 loc) · 3.13 KB

Tests

Which Connections are compatible with Visual Edit?

As explained in the deployment guide, we strongly recommend using an SQL connection for the editlog and edits datasets; it is not necessary to change the connection of the Original Dataset nor of the edited dataset.

You may use other connections for demo/evaluation purposes (e.g. looking at the webapp UX and the integration with the Flow).

Here is an overview of what was explicitly tested and what is expected to work.

  • Persisting edits (i.e. appending rows to the editlog dataset):
    • ✅ We run automated tests on PostgreSQL and FileSystem.
    • ✅ We successfully tested manually on Snowflake.
    • ⚠️ We successfully tested manually on S3 but each edit resulted in a new file, which is not ideal.
    • ❓ Persisting edits may not work with other connections, including SQL ones. For instance, it fails on BigQuery.
    • This is because the implementation is based on the write_dataframe method of the Dataiku API, which may fail when the dataset is in "append only" mode (as is the case with the editlog).
  • Building edits and edited datasets from an editlog, and showing edited data in the webapp, is expected to work on all connections as it is based on the get_dataframe and write_dataframe methods of the Dataiku API.
  • Initializing a secure editlog: successful manual tests on PostgreSQL (see deployment guide).

Automated Integration Tests

These tests are written using the Gherkin specification language and a Dataiku library containing generic Gherkin steps. See the parent folder's README for more context.

Contents

We test the following:

  • Editing of columns of type int, float, and string.
  • Editing datasets with a single primary key and with several primary key columns.
  • Editing of a validation column (which should have a special behavior)
  • Impact of the labels and lookup_columns configuration variables of Linked Records
  • Impact of the authorized_users configuration variable on the ability to make edits.

All tests are performed both on a FileSystem and on a PostgreSQL connection. We test that edits are stored in the editlog dataset as expected; we also build the edits and edited datasets and we test their contents as well.

How to set up an environment to run tests locally

  • Create a Pyhton environment with test requirements found in this directory.
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
  • VS Code:
    • Install the Cucumber extension.
    • In tests/.vscode/, duplicate settings.json.example to settings.json and launch.json.example to launch.json.
    • Fill in the environment variables in settings.json.
  • Configure .dataiku/config.json to select the instance where the tests will run.
  • Open a .feature file and press F5 to run the tests.