suggested changes to allocate

[dylanbeaudette]

I can't figure out how to document all of the arguments to allocate such that arguments specific to sub-functions are clear in the manual page.

Ideas:

• application of droplevels should probably happen after sub-functions are done, fewer arguments to pass around
• sub-functions should be exported with their own documentation, manual page for allocate is generic with links
• consider splitting fao_lev into two vectors (salinity and sodicity), matching by absolute position within fao_lev is tricky and breaks if classes change
• consider optional output with saline / sodic classes in their own columns (result is a matrix or data.frame) so that cross-classification is possible e.g. "slightly saline / moderately sodic".
• allocate(EC = NA, pH = 5, ESP = 1, to = 'FAO Salt Severity', droplevels = FALSE) results in an error

[dylanbeaudette]

Function details moved to Details section for now.

[brownag]

I intend to open a PR linked to this issue with my suggestions and comments in it. I feel this will be easier to provide specific feedback on instances/lines of code. I am not going to do that before 1.27 goes to CRAN though -- and probably not before upcoming stats class -- so as to not detract from anything we need to do before that.

[brownag]

Good work on fixing it up for this release though #179

[dylanbeaudette]

Hi @smroecker any thoughts on the above commentary / ideas for allocate? This won't prevent aqp 2.0 from being released, but now is a convenient time to make large changes.