Harmonize errors for set_mode() and set_engine()

[juliasilge]

Closes #494

This PR harmonizes the errors (previously a mix of warnings + errors) for set_mode() and set_engine(). I made little functions for the error messages and removed the functionality to choose a default engine, as we discussed. Folks will now get results like this for the engine:

library(parsnip)
linear_reg() %>% set_engine()
#> Error: Available engines are: 'lm', 'glmnet', 'stan', 'spark', 'keras'
linear_reg() %>% set_engine(NULL)
#> Error: Available engines are: 'lm', 'glmnet', 'stan', 'spark', 'keras'
linear_reg() %>% set_engine(3)
#> Error: Available engines are: 'lm', 'glmnet', 'stan', 'spark', 'keras'
linear_reg() %>% set_engine("ranger")
#> Error: Available engines are: 'lm', 'glmnet', 'stan', 'spark', 'keras'

^{Created on 2021-05-21 by the reprex package (v2.0.0)}

And this for mode:

library(parsnip)
logistic_reg() %>% set_mode()
#> Error: Available modes are: 'unknown', 'classification'
logistic_reg() %>% set_mode(NULL)
#> Error: Available modes are: 'unknown', 'classification'
logistic_reg() %>% set_mode(42)
#> Error: Available modes are: 'unknown', 'classification'
logistic_reg() %>% set_mode("regression")
#> Error: Available modes are: 'unknown', 'classification'

rand_forest() %>% set_mode()
#> Error: Available modes are: 'unknown', 'classification', 'regression'
rand_forest() %>% set_mode(NULL)
#> Error: Available modes are: 'unknown', 'classification', 'regression'
rand_forest() %>% set_mode("ranger")
#> Error: Available modes are: 'unknown', 'classification', 'regression'

^{Created on 2021-05-21 by the reprex package (v2.0.0)}

[github-actions]

This pull request has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.