azure-devopsazure-pipelinespytest

Send arguments/params to pytest via command without conftest.py


I have a yml template with a job that triggers pytest class.

The file with allt he test classes is (test_invoke_pipelines.py:

import logging
import os

import pytest
from azure.identity import DefaultAzureCredential
from azure.mgmt.datafactory import DataFactoryManagementClient

logging.basicConfig(level=logging.INFO)

LOAD_DATA_ENHANCED_PIPELINE = "my-enhanced-pipeline"
PREPROCESS_PIPELINE = "my-preprocess-pipeline"
TRAIN_MONTHLY_PIPELINE = "my-train-pipeline"
SCORE_MONTHLY_PIPELINE = "my-score-pipeline"


class TestDataLoad:
    """Data Load pipeline test cases."""

    @pytest.mark.order(1)
    def test_load_data_pipeline_succeeded(self, adf_pipeline_run):
        """Test that pipeline has data in SQL."""
        this_run = adf_pipeline_run(LOAD_DATA_ENHANCED_PIPELINE, run_inputs={"pl_full_load": True})
        assert this_run.status == "Succeeded"  # noqa: S101

    @pytest.mark.order(2)
    def test_preprocess_pipeline_succeeded(self, adf_pipeline_run):
        """Test that preprocess pipeline works."""
        this_run = adf_pipeline_run(PREPROCESS_PIPELINE, run_inputs={})
        assert this_run.status == "Succeeded"  # noqa: S101


class TestTrainingPipeline:
    """Train pipeline test cases - Used for CT."""

    subscription_id = os.environ.get("SUBSCRIPTION_ID")
    resource_group_name = os.environ.get("DATA_RESOURCE_GROUP")
    factory_name = os.environ.get("ADF_NAME")
    pipeline_name = os.environ.get("PIPELINE_NAME")

    print(f"Subscription ID: {subscription_id}")
    print(f"RG Name: {resource_group_name}")
    print(f"Factory Name: {factory_name}")
    print(f"Pipeline Name: {pipeline_name}")

    def test_pipeline_runs(self):
        """Test train pipeline case."""
        credential = DefaultAzureCredential()
        adf_client = DataFactoryManagementClient(credential, self.subscription_id)

        # Trigger the pipeline run
        pipeline_run = adf_client.pipelines.create_run(self.resource_group_name, self.factory_name, self.pipeline_name)

        # Assert that the pipeline run was successfully initiated
        print(f"Pipeline Run ID: {pipeline_run.run_id}")
        assert pipeline_run.run_id is not None  # noqa: S101

        # Check the pipeline status at the start
        pipeline_status = adf_client.pipeline_runs.get(
            self.resource_group_name, self.factory_name, pipeline_run.run_id
        ).status
        assert pipeline_status in ["InProgress", "Queued", "Canceling"]  # noqa: S101
        print(f"##vso[task.setvariable variable=ADF_PIPELINE_RUN_ID]{pipeline_run.run_id}")

for the TestDataLoad, it uses adf test plugin and I call the test using the template (integration-template)below:

- task: AzureCLI@2
  displayName: '3 Run Integration Test Data Load'
   inputs:
       azureSubscription: 'datanalytics-${{ parameters.ENV }}'
       scriptType: bash
       scriptLocation: inlineScript
       inlineScript:
           pytest test_invoke_pipelines.py::TestDataLoad --sub_id ${{ parameters.SUBSCRIPTION_ID }} --rg_name ${{ parameters.DATA_RESOURCE_GROUP }} --adf_name ${{ parameters.ADF_NAME }} --sp_tenant_id ${{ parameters.TENANT_ID }} --doctest-modules --junitxml=junit/test-integration-results.xml
       addSpnToEnvironment: true
       workingDirectory: tests

and pytest automatically picks the arguments like subscription_id, resource_group_name when passed with above command as TestDataLoad uses adf test plugin/extension in pytest.

But for the TestTrainingPipeline, I dnt use the adf plugin and instead use adf sdk to trigger an adf pipeline and write the adf pipeline run id to a variable to it can be used for subsequent steps.

The problem is:

How do I pass the arguments like subscription_id and others to the TestTrainingPipeline class. I have another pipeline that calls this test calss as:

 - task: AzureCLI@2
      name: CTTraining
      displayName: '4 Run Integration Tests Train'
      inputs:
        azureSubscription: 'datanalytics-${{ parameters.ENV }}'
        scriptType: bash
        scriptLocation: inlineScript
        inlineScript: |
          echo "##vso[task.setvariable variable=SUBSCRIPTION_ID]${{ parameters.SUBSCRIPTION_ID }}"
          echo "##vso[task.setvariable variable=DATA_RESOURCE_GROUP]${{ parameters.DATA_RESOURCE_GROUP }}"
          echo "##vso[task.setvariable variable=ADF_NAME]${{ parameters.ADF_NAME }}"
          echo "##vso[task.setvariable variable=PIPELINE_NAME]${{ parameters.TRAIN_PIPELINE_NAME }}"
          
          pytest test_invoke_pipelines.py::TestTrainingPipeline --verbose  --capture=no --doctest-modules --junitxml=junit/test-ct-results.xml
        addSpnToEnvironment: true
        workingDirectory: tests

but it throws the error that arguments are unrecognized.

I do not want to use the conftest.py file. I have also tried writing the parameters to os vars before calling pytest:

echo "##vso[task.setvariable variable=SUBSCRIPTION_ID]${{ parameters.SUBSCRIPTION_ID }}"
echo "##vso[task.setvariable variable=DATA_RESOURCE_GROUP]${{ parameters.DATA_RESOURCE_GROUP }}"
echo "##vso[task.setvariable variable=ADF_NAME]${{ parameters.ADF_NAME }}"
echo "##vso[task.setvariable variable=PIPELINE_NAME]${{ parameters.TRAIN_PIPELINE_NAME }}"

but in TestTrainingPipeline class I get None for all these variables:

subscription_id = os.environ.get("SUBSCRIPTION_ID")
resource_group_name = os.environ.get("DATA_RESOURCE_GROUP")
factory_name = os.environ.get("ADF_NAME")
pipeline_name = os.environ.get("PIPELINE_NAME", "pl_train_monthly")

Is there a way to pass these arguments from a yml pipeline to the pytestclass?

The trace I get from the failed test is:

=================================== FAILURES =================================== ___________________ TestTrainingPipeline.test_pipeline_runs ____________________

self = <test_invoke_pipelines.TestTrainingPipeline object at 0x7f8b5bd82800>

def test_pipeline_runs(self):
    """Test train pipeline case."""
    credential = DefaultAzureCredential()
  adf_client = DataFactoryManagementClient(credential, self.subscription_id)

test_invoke_pipelines.py:55: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ../../../_tool/Python/3.10.12/x64/lib/python3.10/site-packages/azure/mgmt/datafactory/_data_factory_management_client.py:132: in init self._config = DataFactoryManagementClientConfiguration(


self = <azure.mgmt.datafactory._configuration.DataFactoryManagementClientConfiguration object at 0x7f8b5bb7e590> credential = <azure.identity._credentials.default.DefaultAzureCredential object at 0x7f8b5bd83d00> subscription_id = None, kwargs = {}, api_version = '2018-06-01'

def __init__(self, credential: "TokenCredential", subscription_id: str, **kwargs: Any) -> None:
    api_version: str = kwargs.pop("api_version", "2018-06-01")

    if credential is None:
        raise ValueError("Parameter 'credential' must not be None.")
    if subscription_id is None:
      raise ValueError("Parameter 'subscription_id' must not be None.") E           ValueError: Parameter 'subscription_id' must not

be None.

../../../_tool/Python/3.10.12/x64/lib/python3.10/site-packages/azure/mgmt/datafactory/_configuration.py:42: ValueError -------- generated Nunit xml file: /azp/_work/3/s/tests/test-output.xml -------- ------ generated xml file: /azp/_work/3/s/tests/junit/test-ct-results.xml ------ =========================== short test summary info ============================ FAILED test_invoke_pipelines.py::TestTrainingPipeline::test_pipeline_runs - V... ============================== 1 failed in 0.61s ===============================

It fails becasue the subscription_id is None


Solution

  • Setting up variable sin the bash worked.

    SUBSCRIPTION_ID="${{ parameters.SUBSCRIPTION_ID }}"
    DATA_RESOURCE_GROUP="${{ parameters.DATA_RESOURCE_GROUP }}"
    ADF_NAME="${{ parameters.ADF_NAME }}"
    PIPELINE_NAME="${{ parameters.TRAIN_PIPELINE_NAME }}"
    
    export SUBSCRIPTION_ID DATA_RESOURCE_GROUP ADF_NAME PIPELINE_NAME
    

    And in the script access as:

    subscription_id = os.environ.get("SUBSCRIPTION_ID")
    resource_group_name = os.environ.get("DATA_RESOURCE_GROUP")
    factory_name = os.environ.get("ADF_NAME")
    pipeline_name = os.environ.get("PIPELINE_NAME")