add labels to prcomp object before tidying

[corybrunson]

This would fix #254 (see the discussion there for examples).

[alexpghayes]

If you pull in the most recent master the (AppVeyor) build should pass. It'd be nice to add a regression test and then I'd be happy to merge.

[corybrunson]

Thanks! I'm having some issues with the build (on a testing branch), but i'll push the merge and see what happens.

[alexpghayes]

Looks good. Would you add a quick test to make sure that the reprex from before:

library(broom)
test <- as.data.frame(matrix(1:9, ncol = 3) + rnorm(n = 9, sd = 0.25))
test
#>          V1       V2       V3
#> 1 0.7707063 3.870905 7.091106
#> 2 2.3385348 4.987070 7.835155
#> 3 2.6638938 6.006691 8.690749
pca <- prcomp(test)
tidy(pca, matrix = "u")
#> Error in data.frame(label = rep(labels, times = ncomp), samples): arguments imply differing number of rows: 0, 9
test2 

doesn't give an error?

[corybrunson]

It works for me!

[corybrunson]

Or, did you mean add a test to some part of the PR?

[alexpghayes]

Yeah I meant add a test to the PR so that if the implementation gets changed in the future the error doesn't return.

[alexpghayes]

The test current checks that the original reprex fails. We should instead make sure that it passes. I think you want expect_silent, not expect_error. It's weird that that test passed.

[corybrunson]

The check uses expect_error() with regexp = NA, indicating that no error should occur, so it should pass. Is there something else wrong?

[alexpghayes]

Woops, my bad. That's it then. Feel free to add yourself as a contributor in DESCRIPTION!