fabricatr 1.0.0

draw_categorical() now returns a factor when category_labels is specified.
draw_ordered() now returns an ordered factor when break_labels is specified.
draw_ordered() performance is improved when returning a factor.

fabricatr 0.16.0

Replacement draw_likert function.
join function renamed to join_using to avoid confusion with other R functions and to more clearly indicate its purpose.
Bug fixes.

Added potential_outcomes (creates multiple columns of potential outcomes according to a formula)
Added reveal_outcomes (reveals observed outcomes on the basis of potential outcomes and a treatment assignment)
Added draw_multivariate (interface for adding correlated draws to fabricate calls)

Added split-modify-combine functionality to modify_level via the by argument.
Changed behavior of draw_ordered to ensure categories are data-independent and adding a strict option that replaces values with NA if values are outside all breaks.
Labeled correlate functionality as experimental.
Many bug fixes.

recycle() helper function for expanding and recycling data
Added correlate() function to allow users to generate arbitrary correlated random variables.
Added category_labels argument to draw_categorical
Added draw_quantile and split_quantile functions
Major updates to test harness
Changed ID label stapling behaviour in fabricate calls.
Allow users to pass ID_label in level creation calls to allow back-door unnamed level calls.
Bugfix for constants in certain modify_level contexts.
Changed specification of latent variables with links; users should now provide latent as an argument when using a link.
New syntax for using length(ID) from upper level to define N of lower level.
When passing a vector to draw_categorical prob argument, package now sends message instead of warning.
Major additions to documentation and vignettes, including vignettes for common social sciences variables, time series, and using other creation packages with fabricatr.
Bugfix in certain cases with single-level variables autocompleting to “data” and not working.
Added unique_labels argument to resample_data to allow block-level statistic calculation.

First CRAN submission for fabricatr