Turbo controller center point issue

[roussel-ryan]

When using the Turbo controller the algorithm tab displays the following

We could use two changes to the center_x box:
(1) reduce the size of variable name boxes (when possible) such that we can see more of their numerical values
(2) Removing or adding to this list will break the controller, we should modify such that elements can't be removed or added (maybe make a new subclass for this?)

[MitchellAV]

@roussel-ryan I've addressed the formatting as best I can to prevent the second widget from being too small.

Could you explain in more detail your second issue you found? I can't replicate any error when adding or removing items from center_x

[roussel-ryan]

adding and removing items won't raise an error immediately, but if one tries to run more iterations it should raise a key error. Definitely if elements are removed, it might not if others are added, but this shouldn't be allowed since it would end up being ignored

[MitchellAV]

@roussel-ryan I am confused with some of you wording. Are you bringing up an issue or a desired function that you wish to be implemented? Your use of "should" is confusing in the context of item (2) from the original comment since you said "will break the controller". Any screenshot of the error message or explanation of the feature would clarify.

[roussel-ryan]

This error will be raised if you remove on of the elements of center_x and then try to run/re-run the optimization since its missing a center value key (in this case its missing `x2`)

[MitchellAV]

So the center_x requires x0, x1, x2 to be present and what you describing is when you remove x2 it does NOT show the red error message and will fail on a subsequent run. I can address the validation case not being translated correctly to the UI.

[roussel-ryan]

Yes, thats correct. I think we should remove the "Remove" button. Here is the stack trace:

Traceback (most recent call last):
  File "C:\Users\Ryan Roussel\Documents\GitHub\Badger\src\badger\core_subprocess.py", line 243, in run_routine_subprocess
    candidates = routine.generator.generate(1)[0]
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Ryan Roussel\Documents\GitHub\Xopt\xopt\generators\bayesian\bayesian_generator.py", line 370, in generate
    model = self.train_model(self.get_training_data(self.data))
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Ryan Roussel\Documents\GitHub\Xopt\xopt\generators\bayesian\bayesian_generator.py", line 559, in get_training_data
    data = self.turbo_controller.get_data_in_trust_region(data, self)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Ryan Roussel\Documents\GitHub\Xopt\xopt\generators\bayesian\turbo.py", line 255, in get_data_in_trust_region
    bounds = self.get_trust_region(generator)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Ryan Roussel\Documents\GitHub\Xopt\xopt\generators\bayesian\turbo.py", line 197, in get_trust_region
    [self.center_x[ele] for ele in self.vocs.variable_names],
     ~~~~~~~~~~~~~^^^^^
KeyError: 'x2'