Auto: predict to multiple grids?

[SimonDedman]

Now built.

1. Load samples & multiple grids files into environment
2. Set WD for grids1 e.g. autumn
3. Run BRT normally with samples & grids1
4. setWD for grids2 e.g. spring
5. Run BRT with loadgbm = "../autumn/SpeciesName/"
6. Repeat 4 & 5 for other folders

Should I instead have users load the model object with load() then just specify the name? No cos it'll get loaded in as Bin_Best_Model_Object and so they can only load 1 bin 1 gaus into the environment at any 1 time.

[SimonDedman]

Could instead have option to include multiple grids objects. grids = gridsspring length(grids) = 1 grids = c(gridsspring, gridsautumn) length(grids) = 2 if(length(grids) > 1) {}

Make the prediction section a subfunction or something so it can be respawned within the code? Label the outputs with a suffix e.g. based on name e.g. PredAbundMap_F.Atlantic.stingray_gridsspring.png

[SimonDedman]

1. Run gbm.auto as normal for first grids.
2. Setwd to next grids folder
3. Run second gbm.auto with loadgbm pointed to first gbm.auto run folder. L876 loadgbm inserted. Does all predicting & mapping, saving outputs in current wd's samples[i] subdirectory.

[SimonDedman]

Could as well allow multiple grids entries, so when gbm.auto gets to the predict stage, if there are >1 grids objects, it loops through them. Likely a ballache....