Disallow calling del_component with ComponentData arguments

[emma58]

Fixes #3435

Summary/Motivation:

I think silently deleting the containing component when we call del_component on a ComponentData is rude, and the post-processing in Fourier-Motzkin elimination was falling victim to this mistake. This PR adds a ValueError to del_component if it is given a ComponentData and revises the FME transformation post-processing routine.

Changes proposed in this PR:

• Sneaks in a documentation fix for ComponentData
• Adds error to del_component if the arguments is a ComponentData
• Revises FME post-processing routine to not misuse del_component and, in fact, to not add and delete as many components in general.

Legal Acknowledgement

By contributing to this software project, I have read the contribution guide and agree to the following terms and conditions for my contribution:

1. I agree my contributions are submitted under the BSD license.
2. I represent I am authorized to make the contributions and grant the license. If my employer has rights to intellectual property that includes these contributions, I represent that I have received permission to make contributions and grant the required license on behalf of that employer.

[jsiirola]

I am pretty sure that the change to del_component won't actually catch attempts to delete a component data.

[jsiirola]

This doesn't do what you think it does: line 1133 (obj = self.component(name_or_object)) maps the incoming object into its component, so you shouldn't be able to trigger this exception. If we are going to change the behavior of del_component to only admit Components (either by name or reference), we should

• change (possibly inline the logic from) line 1133
• document this in the docstring
• possibly add a deprecation path (because we are changing behavior that has been in Pyomo for over a decade)??

[michaelbynum]

I would argue that this is a bug and does not need a deprecation path.

[michaelbynum]

I think block.component() only works with strings if the string refers to a component and not a component data, so we can probably just check if name_or_object is a string?

[emma58]

I agree with @michaelbynum that this is a bug and shouldn't have a deprecation path. Especially after I saw where it got hit in the codebase already... Assuming tests pass (I think they finally will?), this is ready for review again.

[mrmundt]

We are probably not going to aim to merge this before the patch release next week (because it's fairly intricate / could break some folks). We don't want to sneak this in without sufficient planning/warning.

[jsiirola]

One minor question, but overall this looks good.

[emma58]

This is ready to merge when tests pass.