intubate (version 1.0.0)

intubate: 1) Interfaces "on demand" and 2) calling of non-pipe-aware functions directly.

Description

intubate is a helper function used to implement "on demand" interfaces to functions you want to use in pipelines implemented by magrittr. Suppose intubate does not implement an interface to a fantasy function (such as fantasy(formula, data, ...)) that is non-pipe-aware (because data is not its first parameter). To create an interface you only need the line of code ntbt_fantasy <- intubate, after which ntbt_fantasy(data, ...) can be used in a data science pipeline. See examples for details and discussion. ntbt lets you interface the non-pipe-aware function directly without creating an interface.

Usage

## Helper function to define interfaces
intubate(data, ...)
## Function to call non-pipe-aware functions directly
ntbt(data, fti, ...)
## Deprecated helper functions
## (for compatibility with 0.99.2. Will be removed at some point)
ntbt_function_formula_data(data, ...)
ntbt_function_model_data(data, ...)
ntbt_function_object_data(data, ...)
ntbt_function_x_data(data, ...)

Arguments

data

data frame, tibble, list, ...

fti

function to interface.

...

Other arguments passed to interfaced function.

Value

Object returned by interfaced function. If you call intubate directly it will fail.

Examples

## Not run: 
# library(intubate)
# 
# ## NOTE: intubate implements an interface to
# ##       *xyplot* (in package lattice), called *ntbt_xyplot*.
# ##       For the sake of argument, let's suppose the
# ##       interface does not exist, and you want to implement
# ##       it "on demand" to use it in a pipeline.
# 
# ## Original function you would like to interface
# library(lattice)
# xyplot(Sepal.Length + Sepal.Width ~ Petal.Length + Petal.Width | Species,
#        iris, scales = "free", layout = c(2, 2),
#        auto.key = list(x = .6, y = .7, corner = c(0, 0)))
# 
# ## If you try to use *xyplot* directly in a data pipeline
# ## it will raise an error
# library(magrittr)
# try(iris %>%
#     xyplot(Sepal.Length + Sepal.Width ~ Petal.Length + Petal.Width | Species,
#            scales = "free", layout = c(2, 2),
#            auto.key = list(x = .6, y = .7, corner = c(0, 0))),
#     silent = TRUE)
# geterrmessage()
# 
# ## The error disappears if you create an interface to *xyplot*.
# 
# ## Step needed to create an interface to *xyplot*.
# 
# ntbt_xyplot <- intubate
# 
# ## NOTE: names of interfaces must start with
# ##       *ntbt_* followed by the name of the function
# ##       (*xyplot* in this case) you want to interface.
# 
# ## Now you can use the interface instead of the original
# ## function. Just remember to switch the order of
# ## *data* and *x* (there is no need to name the parameters).
# ntbt_xyplot(iris, 
#             Sepal.Length + Sepal.Width ~ Petal.Length + Petal.Width | Species,
#             scales = "free", layout = c(2, 2),
#             auto.key = list(x = .6, y = .7, corner = c(0, 0)))
# 
# ## The newly created interface can be used easily in a pipeline.
# library(magrittr)
# iris %>%
#   ntbt_xyplot(Sepal.Length + Sepal.Width ~ Petal.Length + Petal.Width | Species,
#               scales = "free", layout = c(2, 2),
#               auto.key = list(x = .6, y = .7, corner = c(0, 0)))
# 
# ## Alternative: call non-pipe-aware function directly.
# ## You can also avoid creating an interface, by calling ntbt with the name of
# ## the function to interface.
# ntbt(iris, xyplot,
#      Sepal.Length + Sepal.Width ~ Petal.Length + Petal.Width | Species,
#      scales = "free", layout = c(2, 2),
#      auto.key = list(x = .6, y = .7, corner = c(0, 0)))
# 
# ## In a pipeline
# iris %>%
#   ntbt(xyplot, Sepal.Length + Sepal.Width ~ Petal.Length + Petal.Width | Species,
#        scales = "free", layout = c(2, 2),
#        auto.key = list(x = .6, y = .7, corner = c(0, 0)))
# ## End(Not run)