Make better use of {cli}

[EmilHvitfeldt]

This is a collection issue I'll use to keep track of all the {cli} based changes that is going to happen over the next couple of months.

Infrastructure

• check_new_data() New data checks #1014 (old)
• check_type() switch check_type() to use cli package #1235
• recipes_eval_select() switch to cli abort in recipes_eval_select() #1238
• check_name() fix typos in check_name() #1240
• get_keep_original_cols() use cli in get_keep_original_cols() #1241
• case weights functions update case weights functions to use cli #1242
• check_nominal_type() use cli in check_nominal_type() #1245
• smaller functions in formula.R and misc.R switch to cli in various functions #1244
• roles functions in roles.R use cli in roles functions #1246
• everything in recipe.R use cli for all recipe.R functions #1251
• everything non-abort() Cli last edits #1264

Steps

• move from rlang::abort to cli::cli_abort in steps - first batch #1255
• Use cli::cli_abort() in all steps #1262

[EmilHvitfeldt]

I have been using .var to denote variables in the data set, but I'm wondering if .field is better?

[simonpcouch]

I feel conflicted about .var vs. .field, too. In many cases I prefer the appearance of .field but also try to be faithful to the type of thing over the way the thing appears, since ultimately cli could change that mapping whenever.