Connect learner ORSF from package obliqueRSF

[RaphaelS1]

Checklist before requesting a review

• [x] Run styler::style_pkg(style = styler::mlr_style) (install pat-s/styler@mlr-style if not yet done)

• [x] Run lintr::lint_package() and fix all issues.

• [x] Run usethis::use_tidy_description() to format the DESCRIPTION file.

• [x] Check that the learner package name is all lower case, e.g. mlr3learners.partykit.

• [x] Ensure that there are not leftover of <package>, <algorithm> or <type> within the learner repo.

• [x] Ensure that the "Parameter Check" passed in the CI (both for the train and predict functions)

• [x] Ensure that "R CMD check" passed in the CI.

• [x] Check that your learners upstream package is not listed in the "Imports" but in the "Suggests" section within the DESCRIPTION file.

• [x] If you changed any parameter defaults: Did you document the change (reason and new default) in the help page of the respective learner?

• [x] Open a Pull Request in the mlr3learners repo to add your learner to the list of "In Progress" learners. Once approved, it will be moved to the "Approved" section.

[RaphaelS1]

Ready for review!

[pat-s]

looks good 👍

[RaphaelS1]

deployed

[RaphaelS1]

Although, I can see it's been added to the bin folder but can't see it in src/contrib, not sure if there's a time delay on this or I've done something wrong

[pat-s]

src/contrib is for Linux. If the runner errored during the deployment run, nothing will be deployed.

Also the folder structure changed for R 4.0 on macOS. I think this needs to be accounted upstream in drat though.

We need to find a logic to have some kind of CRON auto-deploy of all existing learners in mlr3learners.drat to avoid missing sources for specific OS of some learners. This is not trivial though because we are essentially doing a small CRAN setup here.

[RaphaelS1]

Okay, I'll leave this to you then!