kedro-pytest 0.1.0

Helps testing Kedro plugin.

kedro-pytest

Helps you testing Kedro plugins with pytest.

from kedro_pytest import TestKedro

def test_count_cli(tkedro: TestKedro):
    tkedro.new('project')
    # ^ creates a new project in a temporary directory
    tkedro.create_pipeline('pipe')
    # ^ ceates a pipeline and two csv entries in catalog by default
    tkedro.update_yml('conf/base/catalog.yml',
        {'new_dset': {'type': 'pandas.JSONDataSet', 'filepath': 'a.json'}})
    # ^ adds a new dataset to the catalog
    result = tkedro.cli(['count', 'catalog'], ['--type', 'pandas.CSVDataSet'])
    # ^ returns a click.testing.Result object
    assert int(result.output) == 2
    tkedro.stop()
    # ^ cleans the temporary directory (only required for doctests)

def test_helloworld_hook(tkedro: TestKedro):
    with tkedro.new('project') as tk:
      tk.create_pipeline('pipe')
      assert 'Hello world!' in tk.run('pipe').output

This plugin focuses on generating minimal Kedro structures, such as small pipelines, and a small project, in order to simplify testing of Kedro plugins in a real Kedro environment. Also, to make it safe for testing, the plugin relies on pytest-tmpfs to create a temporary filesystem for the project.

Install

For installing this package, just type the code below in your terminal:

pip install kedro-pytest

Usage

As a fixture

You can use the KedroPytest as a fixture by passing tkedro as an argument of your test function.

from kedro_pytest import TestKedro

def test_fixture(tkedro: TestKedro):
    ...

In doctests

You can always use fixtures in doctests. To do so, you have to inject it into the doctest namespace in the conftest file.

# conftest.py
@pytest.fixture(autouse=True)
def add_to_doctests(doctest_namespace: dict, tkedro: TestKedro):
    """Adds the tkedro fixture to the doctest namespace."""
    doctest_namespace["kedro"] = tkedro

# code.py
def function():
  """
  Example:
      >>> kedro.new('project')
      >>> kedro.create_pipeline('pipe')
      >>> kedro.stop()
  """
  ...

Your own instance

If you seek to use this plugin without pointing to tmp_path it is possible to create your own instance and manipulate it the way you want.

from kedro_pytest import TestKedro
from pathlib import Path

def test_my_path():
    tkedro = TestKedro(path=Path.cwd())
    ...

Methods

The following table contains all methods that TestKedro implements and its descriptions:

Method	Description	Parameters	Return
new	Creates a new project in a temporary directory.	name: str	TestKedro
create_pipeline	Creates a minimal dummy pipeline with two csv entries in catalog by default and a parameter.	name: str, *content: str	~
write_yml	Writes a yaml file in the project.	path: str, content: dict	str: abs path to the file
read_yml	Reads a yaml file in the project.	path: str	dict: content of the file
update_yml	Updates a yaml file in the project. Can be nested, it will replace recursively.	path: str, content: dict	str: abs path to the file
cli	Runs a kedro cli command in the project.	command: list[str], args: list[str]	click.testing.Result
run	Runs a pipeline in the project.	pipeline: str, run_command: list[str]	click.testing.Result
stop	Cleans the temporary directory.	~	~

Note: The create_pipeline method differs from the cli command because it doesn't need to have a downloaded starter to work.

Warning: cwd switches between temporary and the last cwd when using new and stop methods of TestKedro. A new instance of the fixture is created for each test, so you don't have to worry about it in tests, but in doctests it may be useful. Also, the use of stop can be avoided by using with statements.

Advanced usage

If the methods mentioned above are not enough for your needs, you can always use the fs attribute of the TestKedro instance to manipulate the filesystem.

import json

def test_advanced(tkedro: TestKedro):
    tkedro.new('project')
    tkedro.cli(['json', 'catalog'])
    text = tkedro.fs.read_text('conf/base/catalog.json')
    ...

Contributing

Feel free to contribute with this project, just remember to install and use pre-commit and to write unit tests for it.