Smooth Package Updates - Test Failing

[mdancho84]

Hi Matt,

I'm preparing an update for the package smooth, where es() and msarima() functions will be wrappers for adam(). This means that they will return the same set of objects as adam(). I don't know what specifically your functions use and how, but one of your tests failed with the following error:

── Error ('test-algo-exp_smoothing-ets.R:290'): (code run outside of test_that()) ──
<tibble_error_incompatible_size/tibble_error/rlang_error/error/condition>
Error: Tibble columns must have compatible sizes.
• Size 275: Existing data.
• Size 0: Column .actual.
ℹ Only values of size one are recycled.
Backtrace:
▆
1. ├─model_spec %>% ... at test-algo-exp_smoothing-ets.R:290:0
2. ├─generics::fit(., log(value) ~ date, data = training(splits))
3. ├─parsnip::fit.model_spec(., log(value) ~ date, data = training(splits))
4. │ └─parsnip:::form_xy(...)
5. │ └─parsnip:::xy_xy(...)
6. │ └─parsnip:::eval_mod(...)
7. │ └─rlang::eval_tidy(e, env = envir, ...)
8. └─modeltime::smooth_fit_impl(x = maybe_data_frame(x), y = y)
9. ├─modeltime::new_modeltime_bridge(...)
10. │ └─base::is.data.frame(data)
11. └─tibble::tibble(...)
12. └─tibble:::tibble_quos(xs, .rows, .name_repair)
13. └─tibble:::vectbl_recycle_rows(res, first_size, j, given_col_names[[j]])
14. └─rlang::cnd_signal(error_incompatible_size(n, name, size, "Existing data"))

It might be that you are extracting actual values, and do that via a specific object. If you need the in-sample actuals, there is a function actuals() in greybox package that does that.

Anyway, can you please check and amend? The newest version of smooth that I want to submit to CRAN is here: https://github.com/config-i1/smooth/ I actually already submitted on CRAN, but it returned that error.

Kind regards,
Ivan Svetunkov