Use drop = FALSE for glmnet predictions on single observation

[juliasilge]

Closes #392

This PR adds drop = FALSE for the linear, logistic, and multinomial glmnet models.

library(parsnip)

fitted_parsnip <- linear_reg(
  penalty = 0.1 # set for simiplicity
) %>%
  set_mode('regression') %>%
  set_engine('glmnet') %>%
  fit(Sepal.Length ~ Sepal.Width + Petal.Length, iris)

predict(fitted_parsnip, head(iris))
#> # A tibble: 6 x 1
#>   .pred
#>   <dbl>
#> 1  5.05
#> 2  4.95
#> 3  4.96
#> 4  5.01
#> 5  5.07
#> 6  5.24
predict(fitted_parsnip, head(iris, 1))
#> # A tibble: 1 x 1
#>   .pred
#>   <dbl>
#> 1  5.05

^{Created on 2020-11-17 by the reprex package (v0.3.0.9001)}

[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.