Pyomo.DoE bug fixes

[adowling2]

Fixes No issue

Summary/Motivation:

• Minor enhancements in Pyomo.DoE in preparation for a tutorial/workshop

Changes proposed in this PR:

• Fixed error about ambiguous a.any() versus a.all() for logical check of a list-like object
• Added a TypeError to prevent a downstream cryptic error message
• Update type checking per suggested best practices from @jsiirola and @mrmundt
• Improved initialization of the Jacobian and Cholesky factorization
• Added optionally passing theta parameter values into create_model when generating scenarios. This is especially helpful if create_model includes initialization such was using the simulator in pyomo.dae
• Added more comments
• Fixed degree of freedom issue by fixing the parameters on the nominal case model. This is important if create_model creates a full model in stage1 model. Our original design anticipated stage1 mode would be really lite, but that is inconvenient when trying to use the same create_model for parmest and pyomo.doe

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.

[adowling2]

@jsiirola @mrmundt Suggestions for type(lower_bounds) in [int, float]?

[adowling2]

Any reasons for not replacing if type(self.L_initial) != type(None) with if self.L_initial is not None?

[jsiirola]

For checking for numeric types, we generally use

from pyomo.common.numeric_types import native_numeric_types

if type(lower_bound) in native_numeric_types:
   pass
# or
if lower_bound.__class__ in native_numeric_types:
   pass

and yes: if self.L_initial is not None is the recommended idiom.

[mrmundt]

@adowling2 - Please run black -S -C on your changes to enforce PEP8 standards.

[adowling2]

More development plans:

• [x] Ensure discretize_model = None works on TCLab example (separate repo, really helpful for testing)
• [x] Use the jac_initial argument. Currently, it is stored in self and not used.
• [x] Automatically initialize Cholesky factorization
• [x] Confirm that the "square problem" has zero degrees of freedom. Correct if necessary.
• [x] Run black

[adowling2]

@adowling2 - Please run black -S -C on your changes to enforce PEP8 standards.

I got Error: no such option: -C

However, black -S worked fine. I have version 19.10b0 if that matters.

[adowling2]

Update: I reinstalled black and it works using the command provided by @mrmundt. Now using black, 24.4.2. I'm going to keep the comment above for my future reference.

[adowling2]

@dlakes94 @slilonfe5 @jialuw96 Check out these enhancements to Pyomo.DoE. We'll want to use this branch (which will hopefully be merged into main soon) when refreshing the Pyomo.DoE examples.

[mrmundt]

Update: I reinstalled black and it works using the command provided by @mrmundt. Now using black, 24.4.2. I'm going to keep the comment above for my future reference.

Future reference - yup, it matters! We use the newest black, unless there are bugs in it (e.g., 24.4.1 had a bug that was problematic).

[adowling2]

Anything else needed to get this reviewed and merged?