Just had Ryan N run into the uninformative error "should be a data.frame, matrix, etc." when giving an epi_recipe and epi_archive, so he just fed in a tibble and then epi_recipe just silently returned a recipe object, which cryptically rejected any attempt to add an epi_lag. We should be explicit about expecting an epi_df here and error otherwise. In the future, we can add epi_archive support.
Just had Ryan N run into the uninformative error "should be a data.frame, matrix, etc." when giving an epi_recipe and epi_archive, so he just fed in a tibble and then epi_recipe just silently returned a recipe object, which cryptically rejected any attempt to add an
epi_lag
. We should be explicit about expecting anepi_df
here and error otherwise. In the future, we can add epi_archive support.