Error early when trying to deploy a model from a temporary board

[juliasilge]

Related to rstudio/vetiver-python#193

When calling vetiver_deploy_rsconnect() or the functions for Docker, we should check that the board isn't a board_temp() and error early if it is.

The SageMaker functions already check that the board uses S3: https://github.com/rstudio/vetiver-r/blob/5d3fd53d91148f8115a3b3d56339f9ae765a991d/R/sagemaker.R#L66-L68

[juliasilge]

Oh, slight hiccup here! Our board_temp() doesn't have a separate class from a board_folder():

library(pins)
class(board_temp())
#> [1] "pins_board_folder" "pins_board"

^{Created on 2023-10-11 with reprex v2.0.2}