Add linear scaling for batch recipes

[ehughsbaird]

Summary

Features "Add another form of scaling for batch recipes, enabling creation of bulk recipes"

Purpose of change

Resolves #82744 by allowing creation of bulk threshing recipes.
As discussed in #82795

Describe the solution

Allow specifying recipe scaling as either logistic (the existing mode) or linear with a max batch size.

For linear scaling with offset T, and a max batch size of M, the time taken for batch of n is (ceil(n/M) * T) + (n * (time - T))).

Move recipe batch information to it's own struct, with a variant of substructures specifying the data for whichever mode is selected.

Testing

Examine the batch time taken for the bulk and non-bulk threshed oats recipes, see that they match (e.g. 1 bulk oats is 10 oats, and so on until 5/50).

[PatrikLundell]

ITEM_CRAFT_AND_DISASSEMBLY.md:

• I would like the batch_craft_factors numbers only format to be described as legacy, to be phased out, to encourage new users to use the full format.
• Is the "at" valure description really correct? I thought it was "savings_after". Otherwise [25, 1] would mean you'd get the reduction on everything, since you're always going to produce the "at" number. Thus, I had thought the savings in the example would be applied to the 5 beyond the limit. The savings_string code also prints '>'.
• I think "offset" is harder to understand than "setup" or "preparation". I'd also say that 'M' defaults to 1 when not specified (you can figure that out by comparing the formulae, or assume it by using common sense, but I think being explicit is useful).
• I'd like to see it clarified that "max" isn't the maximum size of the UI specified batches, but the maximum at which the setup cost applies before it has to be spent again.

Note that I'm usually in favor of verbose descriptions covering as much as is reasonable, and others might be less amused (and have a lower threshold for "reasonable")...

recipe.cpp:

• batch_savings::apply():
  • I continue to dislike assignments within "if" statements...
• batch_savings::deserialize():
  • I'd add a comment to the first section indicating it should be removed once all JSON has been converted to make it easier to find once the job has been completed.
  • I'd consider the addition of validation code rejecting illegal values (rsize < 1, rscale < 0, >= 100, offset > time (but equal could be OK: boiling 3 eggs taking the same as 1), max < 2)).

[ehughsbaird]

I'd also say that 'M' defaults to 1 when not specified

M defaults to n when not specified.