Jeremy Coyle

It's useful for learners to have "names" for things like model stacking. Right now this is being done as a haphazard combination of the class name and any specified params....

Currently we handle clustering by making folds for the clusters, and then expanding into folds for the individuals. This can result in unequal fold sizes if the clusters are of...

* Add sampling support to `Lf_known` * Add a node property detailing whether or not this node is observed. This helps generate a sample of observable data for estimation, vs...

If a user manually generates a delta_Y variable, and uses it with a spec, it will not be recognized. It should be.

Maybe it's not needed anymore, maybe it just needs to have different defaults

SequentialJob and FutureJob don't handle seeds the same. See an sl3 example here: https://github.com/tlverse/sl3/issues/402 It seems we should make SequentialJob act more like FutureJob: The help on `future` says: >To...

* Adds (optional) functionality to track timing and memory demands of various subtasks * Adds (optional) debugging functionality to better diagnose when tasks are failing, even in a parallel environment

In the logic that determines which of the "ready" tasks gets assigned to the next available worker, we currently make no distinction between sequential and parallel tasks. Sequential tasks run...

Required for compatibility with Shiny async. Implementation of similar functionality for future here: https://github.com/rstudio/promises/blob/b7060b3987e5f6c0e442dd99e7bcb7ba8315889c/R/promise.R#L360

Covariates passed to learner should match the order of the covariate param, not the order of the covariates in the task. This protects against prediction data with a different covariate...