documentation improvements

[nhejazi]

More work on documentation improvements. 😭

Learners to document:

• CV_lrnr_sl: only requires an example (possibly wrapped in \dontrun{...} to avoid timing issues from running its examples). Maybe we can rename this learner (the capital "CV" is clunky and breaks the current naming pattern), perhaps Lrnr_sl_cv?
• Lrnr_base: the variety of methods it exposes should be documented, both for developers and for users. This might also be a good place to give basic explanations of the x$train() and x$predict() methods inherited by all learners as well as any internal methods that might impact users.
• Lrnr_bilstm: uses a documentation style that differs from that used by other learners and lacks an example. TODO: @imalenica
• Lrnr_caret: missing documentation updates and examples.
• Lrnr_cv: missing documentation updates and examples, as well as documentation for its helper functions.
• Lrnr_dbarts: missing documentation updates and examples.
• Lrnr_define_interactions: missing documentation updates and examples.
• Lrnr_density_discretize: missing documentation updates and examples. TODO: @nhejazi
• Lrnr_density_hse: missing documentation updates and examples. TODO: @nhejazi
• Lrnr_density_semiparametric: missing documentation updates and examples. TODO: @nhejazi
• Lrnr_grf: missing documentation updates and examples.
• Lrnr_gts: missing documentation updates and examples.
• Lrnr_h2o_glm: missing documentation updates and examples, as well as the same for the define_h2o_X model definition utility function.
• Lrnr_h2o_grid and helper functions in the same file: missing documentation updates and examples.
• Lrnr_hts: missing documentation updates and examples.
• Lrnr_independent_binomial: missing documentation updates and examples.
• Lrnr_multiple_ts: missing documentation updates and examples.
• Lrnr_multivariate: missing documentation updates and examples.
• Lrnr_nnet: missing documentation updates and examples.
• Lrnr_optim: missing documentation updates and examples.
• Lrnr_pca: missing documentation updates and examples. Maybe we just drop this?
• SuperLearner interface, including Lrnr_pkg_SuperLearner, Lrnr_pkg_SuperLearner_method, and Lrnr_pkg_SuperLearner_screener all require documentation updates and examples.
• Lrnr_polspline: missing documentation updates and examples.
• Lrnr_pooled_hazards: missing documentation updates and examples. TODO: @nhejazi
• Lrnr_randomForest: missing documentation updates and examples.
• Lrnr_revere_task: requires some documentation updates. Maybe this doesn't need to be exported?
• Lrnr_rpart: missing documentation updates and examples.
• Lrnr_screener_augment: missing documentation updates and examples.
• Lrnr_screener_coefs: missing documentation updates and examples.
• Lrnr_screener_correlation: missing documentation updates and examples.
• Lrnr_sl: missing documentation updates and examples. Could also point to the handbook chapter in the docs.
• Lrnr_solnp: missing documentation updates and examples.
• Lrnr_solnp_density: missing documentation updates and examples.
• Lrnr_stratified: missing documentation updates and examples.
• Lrnr_subset_covariates: requires some documentation updates. Maybe this doesn't need to be exported?
• Lrnr_tsDyn: missing documentation updates and examples. TODO: @imalenica
• Lrnr_ts_weights: missing documentation updates and examples.
• Pipeline: missing documentation updates and examples. Could also point to the handbook chapter in the docs.
• Shared_Data: requires some documentation updates. Maybe this doesn't need to be exported?
• Stack: missing documentation updates and examples. Could also point to the handbook chapter in the docs.
• Utility functions in debug_helpers.R. Examples wrapped in \dontrun{...} and some long-form description of how to use these tools would make them a lot more developer-friendly.
• Utility functions in density_utils.R, look mostly like those appearing in haldensify's data-reformatting function(s). TODO: @nhejazi
• importance and importance_plot: documentation looks good, need examples.
• Could we convert importance_plot to use ggplot2? TODO: @rachaelvp
• Utility functions in learner_helpers.R: documentation looks good but need examples if exported. Maybe these don't need to be exported? Some pointers to how/when to use these would still be useful.
• Loss function definitions in loss_functions.R: documentation looks good but need examples. Very easy to do.
• Metalearner definitions in metalearners.R: documentation looks ok, but would be helpful to have some description contrasting these for users to refer to. Also require examples if will remain exported.
• Utility functions in multinomial_helpers.R: requires some documentation updates. Maybe this doesn't need to be exported?
• prediction_plot: requires some minor documentation updates and examples.
• process_data: would be great to have some documentation describing what this is doing, especially for developer reference purposes.
• process_missing: I guess we can delete this file based on the comment at the top ("This file is deprecated")?
• Utility functions in reindex_folds.R: some documentation would be very helpful to make sure that we can all easily make sense of what these are doing. Example required for the exported subset_folds, but this probably doesn't need to be exported?
• sl3_revere_Task: requires some documentation updates. Maybe this doesn't need to be exported?
• inverse_sample (sampling.R): requires some minor documentation touch-ups; probably doesn't need an example. Maybe this doesn't need to be exported?
• sl3_Task: requires thorough documentation, including description of commonly used optional arguments (e.g., id, weights, folds); examples illustrating creation of different tasks, e.g., with folds respecting IDs, would be very helpful to users. Some description of less commonly used arguments (e.g., flag, row_index) would also be welcome.
• pooled_hazard_task (survival_utils.R): requires some minor documentation updates. Maybe this doesn't need to be exported? Needs an example if it is to remain exported.
• Utility functions in ts_utils.R: need the bare minimum of standard documentation.
• Utility functions in utils.R: documentation is mostly good, some functions need minor touch-ups. Exported functions need examples, but probably don't need to be exported.
• Variable_Type: explicit documentation would be very helpful in making clear how automatic variable type definitions are currently performed. Examples would help make this clear, both to users and current/potential developers. Questions routinely come up from new users about the option "sl3.pcontinuous".