[BUG]: "Run" in Edit model runs the code on the output model, not on the input model

[liunelson]

Describe the bug

Clicking "Run" in the Edit model operator multiple times runs the code on the model sequentially (i.e. on the output model), instead of always running it on the input model of the operator.

If the code adds a transition to a model, clicking "Run" n times adds the same template to the model n times.

For example, I clicked "Run" 3 times and 3 "Death" transitions/templates (H * μ) are added.

concepts_name_map = model.get_concepts_name_map()
subject_concept = concepts_name_map.get("H")

parameter_unit = Unit(expression = 1 / sympy.Symbol("day"))
parameters = {
    "μ": Parameter(name = "μ", value = 1.0, units = parameter_unit, description = "Rate of death due to infection")
}

model = model.add_template(
    template = NaturalDegradation(
        subject = subject_concept,
        rate_law = safe_parse_expr("μ * H", local_dict=_clash),
        name = "Death"
    ),
    parameter_mapping = parameters
)

Expected behavior

Clicking "Run" should always run the code on the original, input model.
This ensures that the code on the left is what's needed to reproduce the model on the right.
It allows the user to make edits incrementally, testing out the code on the model without having to start from scratch.