mold$clean()
and mold$process()
This is a purely developer facing function, that is only used if you are
creating a completely new blueprint inheriting only from new_blueprint()
, and
not from one of the more common: new_xy_blueprint()
, new_recipe_blueprint()
,
new_formula_blueprint()
.
run_mold(blueprint, ...)
A preprocessing blueprint.
Not used. Required for extensibility.
The preprocessed result, as a named list.
Because mold()
has different interfaces (like XY and formula),
which require different arguments (x
and y
vs data
), their
corresponding blueprints also have different arguments for the
blueprint$mold$clean()
and blueprint$mold$process()
functions. The sole
job of run_mold()
is simply to call these two functions with the right
arguments.
The only time you need to implement a method for run_mold()
is if you
are creating a new_blueprint()
that does not follow one of the three core
blueprint types. In that special case, create a method for run_mold()
with
your blueprint type, and pass through whatever arguments are necessary to call
your blueprint specific clean()
and process()
functions.
If you go this route, you will also need to create a mold()
method if x
is not a data frame / matrix, recipe, or formula. If x
is one of
those types, then run_mold()
will be called for you by the
existing mold()
method, you just have to supply the run_mold()
method
for your blueprint.