Skip to content

The pytest plugin including a set of common tools for ADCM testing

License

Notifications You must be signed in to change notification settings

arenadata/adcm_pytest_plugin

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ADCM Pytest Plugin

Overview

The pytest plugin which includes a set of common tools for ADCM tests.

Requirements

  • python (3.7+)
  • pip

Installation

pip install adcm_pytest_plugin

Fixtures

A word about naming convention

The most of the fixture names introduced by this plugin has a suffix which indicates fixture scope. The following list of suffixes are in use:

  • _fs - function scope
  • _ms - module scope
  • _ss - session scope

Here in after the name of any fixture used without scope suffix unless stated.

E.g. adcm which expands to:

  • adcm_fs
  • adcm_ms
  • adcm_ss

List of fixtures

  • imagesession scope only - creates initialized ADCM image for further usage in tests
  • cmd_optssession scope only - fixture aimed to access values of cmd_line options
  • adcm - returns instance of ADCM wrapper (ADCM API and Docker container)
  • sdk_client - returns ADCMClient instance bounded to ADCM instance
  • adcm_api_credentials - returns dict with default ADCM credentials

Functions and methods

utils.py contains a lot of methods useful for testing. See docstrings for more info.

Basic usage

Assume running from adcm_test.

conftest.py

import os

import pytest

from adcm_pytest_plugin.utils import random_string

from adcm_client.objects import Bundle, Cluster, ADCMClient


@pytest.fixture()
def dummy_bundle(sdk_client_fs: ADCMClient) -> Bundle:
    """
    Uploads bundle from dummy_bundle folder 
    """
    bundle = sdk_client_fs.upload_from_fs(
        os.path.dirname(os.path.abspath(__file__)) + "/dummy_bundle"
    )

    return bundle


@pytest.fixture()
def dummy_cluster(dummy_bundle: Bundle) -> Cluster:
    """
    Initialize cluster (based on the cluster prototype)
    """

    cluster = dummy_bundle.cluster_prototype().cluster_create(
        name=f"test_cluster_{random_string()}"
    )

    return cluster

dummy_bundle/config.yaml see ADCM docs for details

---
- name: dummy_cluster
  type: cluster
  version: '1.0'
  config:
    - name: some_boolean_param
      type: boolean
      required: true
      default: false
  actions:
    dummy_job:
      description: "Will fail if config param is false"
      script: cluster_action.yaml
      script_type: ansible
      type: job
      states:
        available:
          - created

dummy_bundle/cluster_action.yaml

---
- name: fail_if_some_boolean_param_is_false
  hosts: localhost
  tasks:
    - name: Assert bool value
      assert:
        that:
          - cluster.config.some_boolean_param == true

test_cluster_action.py

from adcm_pytest_plugin.steps.actions import (
    run_cluster_action_and_assert_result,
)
from adcm_client.objects import Cluster


def test_cluster_action(dummy_cluster: Cluster):
    """Test cluster action run and result"""

    run_cluster_action_and_assert_result(
        cluster=dummy_cluster, action="dummy_job", status="failed"
    )

    dummy_cluster.config_set_diff({"some_boolean_param": True})

    run_cluster_action_and_assert_result(
        cluster=dummy_cluster, action="dummy_job", status="success"
    )

Then run pytest with command line arguments described below.

Command line options

List of available options:


ADCM image options

--staticimage

Use single ADCM docker image instead of initializing new one at the test session start

Property Value
value any valid docker image name
default none
example --staticimage arenadata/adcm:test or --staticimage some_repo/some_image:some_tag

--dontstop

If passed then ADCM containers will remain running after tests

Property Value
value none
default false

--adcm-image

Exact name of ADCM docker image to run tests on

Incompatible with --adcm-images and --adcm-min-version

Property Value
value valid image name:tag
default arenadata/adcm:latest

--adcm-images

Names of ADCM docker images to run tests on. Each image name should be passed as individual arg

Incompatible with --adcm-image and --adcm-min-version

Property Value
value valid image name:tag
default none
example --adcm-images arenadata/adcm:2020.01.30.15 arenadata/adcm:2020.10.15.28

--adcm-min-version

If passed then tests will be executed on all ADCM release images newer than version passed

Incompatible with --adcm-images and --adcm-image

Property Value
value string of ADCM version format
default none
example --adcm-min-version 2020.01.30.15

--nopull

If passed then no pull action will be performed on docker run

Property Value
value none
default false

Mics

--remote-executor-host

If passed then ADCM API will be initialized with external IP to allow incoming connections from any remote executor (ex. Selenoid) Tests will fail if remote host is unreachable. This option will be ignored if --remote-docker option is passed

Property Value
value string with fqdn
default none

--remote-docker

If passed then ADCM instances will be created on a remote host. Docker daemon should be running and be available with provided host:port

Property Value
value string of host:port format
default none
example --remote-docker '10.92.7.14:2375'

--verbose-actions

If passed then ADCM actions will be started with 'verbose' checkbox selected. Applied only to action calls over adcm_client. Does not affect UI action calls in tests.

Property Value
value none
default false

--actions-report-dir

If passed then plugin will collect information about all actions calls made with run_action_*** wrappers and create summary report in the provided directory

Property Value
value string with absolute or elative path to dir where actions report will be stored
default none
example --actions-report-dir relative/path/to/dir or --actions-report-dir /absolute/path/to/dir

Writing tests for plugin

At the moment, the project has a set of tests, which is located in the ./tests directory. Current tests and tests that will be written should be based on usage of the tesdir fixture from the pytester plugin. This approach allow tests to be run inside tests. Nested launch is important for us, since many functions and fixtures of the plugin depends on launch parameters (pytest command-line options)

Source code of nested tests can be stored as:

  1. .py files in ./tests/_test_files directory
  2. multiline string inside code of the test that will run nested test

Example 1.

Given that the code for a nested test is stored in the test_some.py file. To run such a test, you need to run the following code:

def test_some_fixture(testdir):
    testdir.copy_example("test_some.py")
    result = testdir.runpytest()
    # This number determines the number of successful tests in the nested test
    result.assert_outcomes(passed=1)

Example 2.

Given that the code for a nested test is stored in multiline string. To run such a test, you need to run the following code:

def test_some_fixture(testdir):
    testdir.makepyfile(
        """
        def test_some():
          assert True
        """
    )
    result = testdir.runpytest()
    # This number determines the number of successful tests in the nested test
    result.assert_outcomes(passed=1)

You can read more about writing tests for pytest plugins here

How to run unit tests for plugin

To run plugin tests, you can execute this from your project root:

pip install -e .
pip install -r tests/requirements.txt
cd tests
pytest ./plugin --alluredir allure-result

To open allure report, you can execute this:

allure serve allure-result

Pre-commit hook

We are using black, pylint and pre-commit to care about code formating and linting.

So you have to install pre-commit hook before you do something with code.

pip install pre-commit # Or do it with your preffered way to install pip packages
pre-commit install

After this you will see invocation of black and pylint on every commit.