Add conda env for pytest in aml-benchmark

[iamrk04]

There are a couple of issues (listed below) with the default asset-test workflow which we want to avoid and hence we are creating our own workflow to make sure new changes do not break the existing functionality:

• default asset-test workflow uses latest python version i.e. python==3.12, and some of the packages (azureml-dataprep and azureml-dataprep-native) that we depend upon do not have support for python==3.12, which simply breaks the workflow and we are unable to run our tests.
• If we change code in multiple components, it triggers the test multiple times (for multiple components), which is unnecessary. We just want the test to be triggered once as all the tests will be executed for any change anywhere. (Please note we want to run all the tests for any change in the code anywhere, not just for change in component but also for change in test)

Update: To address the issue, we are now leveraging the conda_environment property for pytest.

[github-actions]

Test Results for aml-benchmark-test

220 tests   207 ✔️  33m 48s ⏱️
    1 suites      1 💤
    1 files      10 ❌  2 🔥

For more details on these failures and errors, see this check.

Results for commit 749535bd.

♻️ This comment has been updated with latest results.

[vizhur]

lets not duplicate CIs and make improvements to the current one

[github-actions]

Test Results for assets-test

233 tests   225 ✔️  1h 45m 58s ⏱️
    1 suites      1 💤
    1 files        7 ❌

For more details on these failures, see this check.

Results for commit bbe8ff2.

♻️ This comment has been updated with latest results.

addressed