Ideas for classes of objects saved in package

[andrew-edwards]

I was thinking of the classes just being pacea_t for a time series, pacea_s for spatial, and pacea_st for spatiotemporal, and then having functions like plot.pacea_t() to plot an object in the correct style.
Currently implement (not fully working) for climate indices, so that
plot(oni) does a sensible plot.

Currently implementing estimated annual recruitment and biomass from hake assessment, for which different types of plot are needed. So now thinking to have classes:

• pacea_index - climate and oceanographic indices
• pacea_biomass - estimated biomass
• pacea_recruitment - estimated recruitment
• pacea_s - spatial (to add to a map)
• pacea_st - spatiotemporal, such as ROMS output, but may want to be more specific.

The current plot.pacea_t() is getting a bit big with different switches depending on what's in the object, so makes sense just to add more classes with sensible default plotting options.

[andrew-edwards]

Looks like classes can be documented like functions:
https://adv-r.hadley.nz/s3.html#s3-constructor (a few links on from https://r-pkgs.org/man.html )

[andrew-edwards]

We're ending up with more classes. So go through them and just briefly document them as per above, to help for adding future data objects (so we'll know if we can just call them an existing class).

[andrew-edwards]

Be nice to document the classes, but not completely necessary for Friday.

[travistai2]

Should documenting the classes just be an object .R documentation? Or what would the other options be?

[andrew-edwards]

There was something in the link above (May 2 comment). I took the label off as didn't think it was 100% necessary for tomorrow (kind of helps us more than users).