speed up spec_is_possible() helper

[simonpcouch]

My old helper was slow🦥 Instead of joining, we can perform the filtering on the separate datasets, and with vctrs instead of dplyr. For a quick model fit on a small dataset, this cuts down fit time by ~22%!

With main dev:

library(parsnip)

bench::mark(
  total = fit(linear_reg(), mpg ~ ., mtcars),
  iterations = 25
)
#> # A tibble: 1 × 6
#>   expression      min   median `itr/sec` mem_alloc `gc/sec`
#>   <bch:expr> <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl>
#> 1 total        7.65ms   8.04ms      120.    7.19MB     67.6

With this PR:

#> # A tibble: 1 × 6
#>   expression      min   median `itr/sec` mem_alloc `gc/sec`
#>   <bch:expr> <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl>
#> 1 total        6.09ms   6.28ms      155.    7.12MB     60.3

^{Created on 2023-03-17 with reprex v2.0.2}

There are additional tests for this infrastructure in extratests—they are fine. :)

[EmilHvitfeldt]

looking clean!

[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.