Changelog • crossmap

cross_fit() gains the argument clusters, allowing mapping along cluster specifications for functions that support it, like estimatr::lm_robust().
cross_fit_robust() is a wrapper for cross_fit(fn = estimatr::lm_robust).
Functions now use rlang::check_installed() for suggested packages, giving the user the option to install the package interactively.
Move stats from suggested to imported packages.

New features

Added cross_fit_glm(), which works like cross_fit() but allows you to also specify a crossing of glm() model families.
Added tidy_glance(), which returns a tibble with information from both broom::tidy() and broom::glance().
- tidy_glance() is now the default tidier in cross_fit().
Added future_xmap_raw() and future_xwalk().

Patches

cross_join(), cross_list(), cross_tbl() and cross_df() now silently ignore NULL inputs.
future_*() functions now prompt the user to select a future plan if R is not set up for parallelization.

Miscellaneous

Promoted broom and dplyr from suggested to imported packages.

New features

Added weights argument to cross_fit().
- You can now cross model specifications in three dimensions: formulas, subsets, and weights.
- Weights are specified as a list of column names, or NULL or NA for an unweighted model.

Miscellaneous

Added tibble as a suggested package.
Added a NEWS.md file to track changes to the package.
Added a URL to pkgdown YAML.

Initial CRAN release.