Pass a reference to the sequencer when calling TestClass.test

[clint-lawrence]

To maintain backward compatibility, we wouldn't do this directly. Instead, we can add a new internal method to TestClass which will be called by the sequencer, and that new method will then call the original TestClass.test. And instead of directly passing a reference to the sequencer, I think we could create a SequencerProxy object. The methods on this object would represent the public interface that is O.K. for test scripts rely on.

e.g.

class SequencerProxy:
    def __init__(self, sequencer: Sequencer):
      self._sequencer = sequencer
  
    @property
    def context(self):
        return list(self.sequencer.context)
  
    @property
    def script_params(self):
        return self.sequencer.context_data

class TestClass:
    # the sequencer would call this method instead of calling test directly.
    def test_internal(self, sequencer_proxy):
        # Returns a dict of test data that this class needs from containing TestLists
        kwargs = get_test_data(sequencer_proxy.context)  
        self.test(**kwargs)

    def test(self):
        # overridden by the test implementation.
        raise NotImplementedError

   ... etc

This idea was motivated by thinking about #126 At the moment, we can't easily share data from the sequencer to tests. We added the context_data dict to the sequencer and some scripts use that as a way to pass some data from the sequencer into the script. But this requires the TestClass.test method to reach into RESOURCES["SEQUENCER"], accessing the global sequencer instance.

With this change, we could implement Ryan's idea "5" from here without having to add much extra complexity to the sequencer and provide a more general method to make the use of context_data more explicit where it's used.

If the need arises, we could also do the same treatment to setup/teardown/enter/exit.