Integrate testing into DataClass

[seabbs]

Currently testing is standardised across all objects that inherit from data class. This standardised testing is then used to check the integrity of datasets with a daily check and also used with a data snapshot for unit testing.

Using our data sources I have found myself wanting to check the validity of our data sources. This makes me think the incorporating testing into the package either as a function or a class method (this would be nice as it would allow classes to inherit a standard set of whilst adding more to them.

Some things to work on with this:

• [ ] testing should be a suggests and not a core part of the package.
• [ ] testing should probably only work once the download method has been called
• [ ] we probably do not want it to impact the state of the class when it is called (this has tricky implications for example with filter)
• [ ] We want to be able to easily test on both real data and a snapshot (perhaps a secondary issue)

[joseph-palmer]

So have a test method defined in the generic class (DataClass or CountryDataClass) and then the test functions will just call the test from each country (Italy$run_tests()). Then if a country has some specific tests to add it can define its own test to overwtire the parent (whilst still calling it so it can add extra), so for UK it might be

run_tests <- function() {
  super$run_tests()
  ....
}

[seabbs]

Yes that is exactly what I was thinking (though test as the method name 😄 )