Consider lazydata = FALSE and removing raster from imports

[Nowosad]

No description provided.

[Nowosad]

@Robinlovelace @jannes-m What do you think about the idea of disabling lazydata and removing raster from imports in spDataLarge?

[Robinlovelace]

Good plan, 👍 needed by end of 2023, right?

[Nowosad]

At latest. I think we can do that when (a) we avoid using this feature (lazydata from spDataLarge) in geocompr2, and (b) geocompr2 is published.

[jannes-m]

Should I convert the raster files in the landslides, the mongon and the ep_eco datasets from raster to terra? Hopefully, the corresponding code adjustments will be not too labour intensive...

[Nowosad]

@jannes-m Correct me if I am wrong, but you cannot save terra objects directly to rda (you would need to use wrap to save it, and rast to read it). My current approach is to use .tif files saved in inst/raster and documented in the R folder (https://github.com/Nowosad/spDataLarge/blob/master/R/rasters.R).

[jannes-m]

Ah ok, no, I didn't know that one can not save terra objects as rds files. I didn't know either that you have already put most of the raster files into a raster folder. So thanks for letting me know. Some terrain attributes of the raster bricks ta and ep are still missing in the raster folder (catchment area, catchment slope, etc.) . Should we add them to the raster folder as multi-layer files or do you prefer to keep ta and ep?

[Nowosad]

I think we should use terra objects (or raster files) consistently through the book (and avoid raster objects). If you plan to use ta and ep in the book, then yes -- save them in the raster folder and add/update the documentation. Let me know if you need any help. J

[jannes-m]

Ok, then I will update these files and use terra files throughout the book.

[Nowosad]

This is not a priority now, but I made a version of the package that does not depend on {raster} -- #31. We can marge this PR whenever we want...

[Nowosad]

Merge completed.

[Robinlovelace]

Great work Jakub!