Scratch

Scratch provides an alternative to inputs and outputs for passing information between tests. The table below shows the main differences between the two options:

Inputs are loaded from the database at the start of each test, so the major advantage of scratch is that its contents are stored in memory. Also, if tests require a large number of complex objects, scratch provides a way to avoid repeatedly loading, deserializing, and instantiating them. The major limitation of scratch, however, is that it only persists through a single test run (i.e., a single click of the Run Tests button). Inputs and outputs are therefore recommended over scratch when possible.

Example

scratch is a plain Hash that is passed to each test. It persists throughout a single test run, and changes made to it in one test will be visible to any subsequent tests in that run.

class MyGroup < Inferno::TestGroup
  # Since scratch is lost at the end of a test run, by making this group run
  # together, that ensures that the tests run together and the scratch will
  # remain present.
  run_as_group
  
  test do
    run do
      # retrieve a large number of FHIR resources
      # ...
      
      # store resources in scratch
      scratch[:observations] = a_large_number_of_fhir_observations
      scratch[:conditions] =  a_large_number_of_fhir_conditions
    end
  end
  
  test do
    run do
      # read resources from scratch
      perform_some_operation_on(scratch[:observations])
    end
  end
  
  test do
    run do
      # read resources from scratch
      perform_some_operation_on(scratch[:conditions])
    end
  end
end