fix intermittent failure for flaky test

[shanbady]

What are the relevant tickets?

Closes #1772

Description (What does it do?)

Fixes the intermittent test failure in test_similar_resources_endpoint_only_returns_published

How can this be tested?

1. checkout main
2. decorate learning_resources/views_test.py::test_similar_resources_endpoint_only_returns_published with @pytest.mark.parametrize('execution_number', range(100)) and run it via pytest learning_resources/views_test.py::test_similar_resources_endpoint_only_returns_published -s -vvv -x to see it fail
3. checkout this branch and try the same and note it succeeds every time

[ChristopherChudzicki]

Approving because this works, but:

1. What was causing the failures before?
2. Mentioning: I see that get_similar_resources does

           return LearningResource.objects.for_search_serialization().filter(
           id__in=[
               resource.id
               for resource in response.hits
               if resource.id != value_doc["id"] and resource.published
           ]
       )
   

   So I guess this tests the last part of that filter. I think we also exclude unpublished stuff from the search indexes (do we?) . But if I'm correct, no harm in doing it here, too.

[shanbady]

Approving because this works, but:

1. What was causing the failures before?

2. Mentioning: I see that get_similar_resources does

           return LearningResource.objects.for_search_serialization().filter(
           id__in=[
               resource.id
               for resource in response.hits
               if resource.id != value_doc["id"] and resource.published
           ]
       )
   

   So I guess this tests the last part of that filter. I think we also exclude unpublished stuff from the search indexes (do we?) . But if I'm correct, no harm in doing it here, too.

The initial issue happened when the document id we pass into the view happens to also be the published=True resource we test in the response. The fix was to ensure that the document id we pass in is not in the list of mocked resources for the response