This reverts commit 781e331.

<https://cran.r-project.org/web/packages/progressr/vignettes/progressr-intro.html>)

also introduce global option `projpred.use_progressr` so that all users (even
those with `progressr` installed) may refrain from using `progressr`;
also handle the export of `use_progressr`

to avoid potential problems when exporting to parallel workers (the export is
handled here explicitly as well)

fix a bug in `kfold_varsel()` where object `inds` does not exist

…used

…f `foreach::foreach()`

…ure`

backend for parallelization (this issue does not get resolved when installing
projpred (instead of using `devtools::load_all()`))

projpred (instead of using `devtools::load_all()`))

…tent may be a list, for example

…t know why

both `object` and `refmodel` exist)

argument `ndraws` explains that it is `nclusters` which decides the
clustering/thinning mechanism by being `NULL` or not)

the clustering/thinning used

…ies_fast` is always `NULL`)

…ed to

…V for subsampled PSIS-LOO-CV

`getOption("projpred.verbose", interactive())`

…varsel.vsel()`)

argument `ndraws` explains that it is `nclusters` which decides the
clustering/thinning mechanism by being `NULL` or not; the same holds for
`ndraws_pred` / `nclusters_pred`)

… the

new column behavior of `summary.vsel()`, and the omittance of option `"best"` of
argument `baseline`

…vsel()`, the"

This reverts commit ea18e9a.

Reason for the revert: It's easier if we first add everything that should be in
the final (separate) pull request that incorporates all changes from this pull
request (including the new "mixed deltas" plot, the new column behavior of
`summary.vsel()` etc.). Afterwards, we can simplify the current pull request and
leave things like the new "mixed deltas" plot and the new column behavior of
`summary.vsel()` for the separate pull request.

avoid that `\n` accidentally remains after removing the "NOTE" in the future)

…ries_fast_sub <- NULL` to one line

the intentioned usage was probably in `get_stat()` (because `loo_inds <-
which(!is.na(lppd))` was error-prone; e.g., in the augmented-data or latent
projection, `lppd` might consist of only `NA`s)

also perform some cleaning in terms of whitespace and object names (in R,
calling objects `n` is not convenient because it complicates debugging [there,
`n` means "next"]; also, such a generic name makes it hard to search for these
objects in the code; for `w`, the problem is that there are several different
kinds of weights within projpred, so we should point out what they refer to)

(in R, calling objects `n` is not convenient because it complicates debugging
[there, `n` means "next"]; also, such a generic name makes it hard to search for
these objects in the code; for `w`, the problem is that there are several
different kinds of weights within projpred, so we should point out what they
refer to)

…mong the empty initialized objects, so move it back

…then

`correct_baseline` is `0` anyway)

…diffe$hat_v_y)` into `.srs_diff_est_w()`

…nloo < refmodel[["nobs"]]`

…t + srs_diffe$hat_v_y)` into `.srs_diff_est_w()`"

This reverts commit 2727cff.

Reason for the revert: See <2727cff#commitcomment-145785150>.

`NULL$mu` is `NULL` and `rep(NULL$mu, c(4, 5))` (for example) as well)

new column behavior of `summary.vsel()`, and the omittance of option `"best"` of
argument `baseline`

subsampled LOO (`n_loo < n_full`) and everything else (`n_loo == n_full`)

does not change from definition of `var_mse_e` to use of `var_mse_e`, the extra
definition of `var_mse_e` can be avoided

`is.null(summaries_baseline)`); one reason is that I'm not sure whether it was
supposed to read `mu_baseline <- y` in that case (instead of `mu_baseline <- 0`)

…of `cv_varsel()`

…)` calls

This fixes commit 0d73c8e. However, before
commit 0d73c8e, `is.null(mu_baseline)` should
have never occurred because if `summaries_baseline` was `NULL`, then
`mu_baseline` was set to `0` (and if `summaries_baseline` was not `NULL`, then
`mu_baseline` was set to `summaries_baseline$mu` which should not be `NULL`
either). Hence, this fixup here does not only fix commit
0d73c8e, but also the incorrect behavior which
existed before it.