refactor: add get_base_grid() to states

[goccert25]

Pull Request doc

Purpose

What is the larger goal of this change?
To pay down some tech debt and hopefully make PowerSimData codebase more maintainable

What the code is doing

How is the purpose executed?
Removing many similar/copy and pasted calls to initialize the original Grid that a scenario was derived from via information from the scenario object

Testing

How did you test this change (unit/functional testing, manual testing, etc.)?
Existing unit tests

Time estimate

How long will it take for reviewers and observers to understand this code change?
20 minutes

[jenhagg]

Nice, the implementation looks good to me. One other thing I would consider is updating the usages of scenario.state.get_base_grid() to be simply scenario.get_base_grid().

[goccert25]

Nice, the implementation looks good to me. One other thing I would consider is updating the usages of scenario.state.get_base_grid() to be simply scenario.get_base_grid().

@jon-hagg I used scenario.state.get_base_grid() because in the code there was already scenario.state.get_grid() throughout. I'm happy with switching over to scenario.get_base_grid() but if so I would ask if I could also do scenario.get_grid() to be consistent. Let me know!

[danielolsen]

The code is clear, and it simplifies the logic within the powersimdata.design module. I support a general change from scenario.state.foo to scenario.foo, which is available since #503. Previously we simplified the user-facing example code but kept the more verbose calls within the internal code so that the tests would still pass with the mock objects, but now we should be able to simplify everywhere.

One comment is that with the variable renaming from grid_new to grid within the investment cost functions, we should be sure to remember that what's in this object named grid is not really a grid in the usual sense of the word, but a grid difference for one table within the grid (e.g. the branch table or the plant table or the dcline table), which we use a Grid object for because it's convenient to have those other Grid attributes around (i.e. the bus table, sub table, other non-table attributes, etc.) within the investment cost calculation functions. One day we might decide that we want a cleaner implementation, e.g. maybe with a new GridDelta class that could have one or more tables that are really tables of element differences, not of elements themselves, but I think that's outside of the scope of this PR.

[jenhagg]

@goccert25 good call - I would also support switching to scenario.get_grid().

[goccert25]

@jon-hagg done

@danielolsen renamed the grid variable to grid_differences to be more appropriate

Feel free to approve and merge whenever, I rebased on develop as of the time of this comment

[jenhagg]

LGTM

[BainanXia]

Thanks!

[danielolsen]

Looks good, thanks!