Add merge/reduce FLORIS objects

[paulf81]

Just a rough draft for now @misi9170 and @Bartdoekemeijer of sketching in merging/reducing FLORIS objects (taking these functions out of FLASC and into FLORIS), we'll look a mess until we merge refactor...

Then also, first decision, is this the right place? Do we prefer a static member function?

[misi9170]

I think this way (as a static function in floris_model.py) is reasonable. Will we want to do the same for uncertain models and parallel models? perhaps we can figure out a convenient function that could handle any of them? In that case, we may want this function on a separate file.

I'm more into the merge function than the reduce function. For reducing, we already have the following options:

• Updating layout_x and layout_y in set()
• Using disable_turbines for "temporarily" reducing (although the computation will still be "large").
  I guess I feel that things could easily go wrong with the reduction function

[paulf81]

I think this way (as a static function in floris_model.py) is reasonable. Will we want to do the same for uncertain models and parallel models? perhaps we can figure out a convenient function that could handle any of them? In that case, we may want this function on a separate file.

I'm more into the merge function than the reduce function. For reducing, we already have the following options:

• Updating layout_x and layout_y in set()
• Using disable_turbines for "temporarily" reducing (although the computation will still be "large").
  I guess I feel that things could easily go wrong with the reduction function

Ok sounds good, I'll set it as a static member of the FlorisModel and remove the reduce function. I think for uncertain/parallel these should be defined specifically for those, rather than trying to make a catch-all for now, which we could do later, but that moment catching all the special cases could be something. At any rate I think I will just make the change in FlorisModel and then pause and wait for refactor merge just to simplify a little

[paulf81]

nevermind... hit the wrong button

[paulf81]

Ok @misi9170 , I think this one is ready, I think for now just adding the function is good. I tested it out a bit with some test code, but I think for now we won't call it in the examples and can add that later (or just make greater use in flasc). This is ready from my perspective

[misi9170]

Should these UserWarnings be errors?

[misi9170]

Same comment as above

[misi9170]

fmodel.reset_operation()

[misi9170]

@paulf81 I made a few changes in d127c70

• Changed UserWarning to ValueError in a few places (please feel free to change back if you intended to raise UserWarnings)
• Changed fi and FlorisInterface to fmodel and FlorisModel throughout, and made warning about not providing FlorisModels more general
• Included a reset_opeartion() call so that any control setpoints, which are specified for a specific number of turbines, are dropped.

If you're happy with those changes, feel free to merge this in.

[rafmudaf]

@misi9170 @paulf81 I just came across this and I think it's not validating fmodel_list correctly. If I understand correctly that it's a 1-D list, then it's only validating the first element. You can check all elements with something like:

# does `isinstance` on each element of the list through `map`
# then checks whether all the elements are True.
# If any elements are not True, then it goes into the if-statement
if not all( list( map(isinstance, fmodel_list, len(fmodel_list)*[FlorisModel]) ) ):

# or the same thing but more simply
if not all( type(fm) == FlorisModel for fm in fmodel_list ):

It may or may not be worth fixing right now, but I wanted to give you a heads up in case you see anything funny in the future with this function.