josh-notes

Tests that load CSV data, which is better?
  Style 1:
    ```ruby
    def setup
      @dr = DistrictRepository.new
    end

    def test_find_by_name_returns_district
      assert_equal "ACADEMY 20", @dr.find_by_name("ACADEMY 20")[1].name
    end
    ```

  Style 2:
    ```ruby
    def test_it_can_find_an_enrollment_object
      er = EnrollmentRepository.new
      er.load_data({
        enrollment: {
          kindergarten: "./fixtures/Kindergartners in full-day program.csv"
        }
      })
      enrollment = er.find_by_name("ACADEMY 20")
      assert_kind_of(Enrollment, enrollment)
    end
    ```

  Immediate question for style 1:
    Where did it get the data from,
    and can I change that? eg for fixtures

  From this, we can determine that the second one is better
    And this is a specific example of a general truth:
    Design is usally better if you move integration points
    earlier in the callstack (here, the integration between
    the DistrictRepository data and the CSV location of the data)

  An example of taking this further
    would be to move the integration point of the persistence format (CSV)
    up higher, so that you could persist in other formats (JSON, XML, whatever)


Require:
  Don't start with ".." or ".", if you find yourself using that,
  there's a waypoint about it: https://github.com/turingschool/waypoints/blob/master/waypoints/require.md


Relative filepaths of the CSV data:
  You can use the same approach from editing the load path
  to make the path be the same from any directory.

  ```ruby
  def fixture_path
    File.expand_path("fixtures/Kindergartners in full-day program.csv", __dir__)
  end

  def test_something
    er = EnrollmentRepository.new
    er.load_data(enrollment: { kindergarten: fixture_path })
    # ...
  end
  ```