Package 'orbital'

Description

augment() will add column(s) for predictions to the given data.

Usage

## S3 method for class 'orbital_class'
augment(x, new_data, ...)

Arguments

Details

This function is a shorthand for the following code

dplyr::bind_cols(
  predict(orbital_obj, new_data),
  new_data
)

Note that augment() works better and safer than above as it also works on data set in data bases.

This function is confirmed to not work work in spark data bases or arrow tables.

Value

A modified data frame or remote database table.

Examples

library(workflows)
library(recipes)
library(parsnip)

rec_spec <- recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_numeric_predictors())

lm_spec <- linear_reg()

wf_spec <- workflow(rec_spec, lm_spec)

wf_fit <- fit(wf_spec, mtcars)

orbital_obj <- orbital(wf_fit)

augment(orbital_obj, mtcars)

Description

Fitted workflows, parsnip objects, and recipes objects can be turned into an orbital object that contain all the information needed to perform predictions.

Usage

orbital(x, ..., prefix = ".pred")

Arguments

Details

An orbital object contains all the information that is needed to perform predictions. This makes the objects substantially smaller than the original objects. The main downside with this object is that all the input checking has been removed, and it is thus up to the user to make sure the data is correct.

The printing of orbital objects reduce the number of significant digits for easy viewing, the can be changes by using the digits argument of print() like so print(orbital_object, digits = 10). The printing likewise truncates each equation to fit on one line. This can be turned off using the truncate argument like so print(orbital_object, truncate = FALSE).

Full list of supported models and recipes steps can be found here: vignette("supported-models").

These objects will not be useful by themselves. They can be used to predict() with, or to generate code using functions such as orbital_sql() or orbital_dt().

Value

An orbital object.

Examples

library(workflows)
library(recipes)
library(parsnip)

rec_spec <- recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_numeric_predictors())

lm_spec <- linear_reg()

wf_spec <- workflow(rec_spec, lm_spec)

wf_fit <- fit(wf_spec, mtcars)

orbital(wf_fit)

# Also works on parsnip object by itself
fit(lm_spec, mpg ~ disp, data = mtcars) %>%
  orbital()

# And prepped recipes
prep(rec_spec) %>%
  orbital()

Description

Returns data.table code that is equivilant to prediction code.

Usage

orbital_dt(x)

Arguments

Value

data.table code.

Examples

library(workflows)
library(recipes)
library(parsnip)

rec_spec <- recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_numeric_predictors())

lm_spec <- linear_reg()

wf_spec <- workflow(rec_spec, lm_spec)

wf_fit <- fit(wf_spec, mtcars)

orbital_obj <- orbital(wf_fit)

orbital_dt(orbital_obj)

Description

Use orbital object splicing function to apply orbital prediction in a quosure aware function such as dplyr::mutate().

Usage

orbital_inline(x)

Arguments

Details

This function is mostly going to be used for Dots Injection. This function is used internally in predict(), but is also exported for user flexibility. Should be used with ⁠!!!⁠ as seen in the examples.

Note should be taken that using this function modifies existing variables and creates new variables, unlike predict() which only returns predictions.

Value

a list of quosures.

Examples

library(workflows)
library(recipes)
library(parsnip)

rec_spec <- recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_numeric_predictors())

lm_spec <- linear_reg()

wf_spec <- workflow(rec_spec, lm_spec)

wf_fit <- fit(wf_spec, mtcars)

orbital_obj <- orbital(wf_fit)

orbital_inline(orbital_obj)

library(dplyr)

mtcars %>%
  mutate(!!!orbital_inline(orbital_obj))

Description

Reading an orbital object from disk

Usage

orbital_json_read(path)

Arguments

Details

This function is aware of the version field of the orbital object, and will read it in correctly, according to its specification.

Value

An orbital object.

See Also

orbital_json_write()

Examples

library(workflows)
library(recipes)
library(parsnip)

rec_spec <- recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_numeric_predictors())

lm_spec <- linear_reg()

wf_spec <- workflow(rec_spec, lm_spec)

wf_fit <- fit(wf_spec, mtcars)

orbital_obj <- orbital(wf_fit)

tmp_file <- tempfile()

orbital_json_write(orbital_obj, tmp_file)

orbital_json_read(tmp_file)

Description

Saving an orbital object to disk in a human and machine readable way.

Usage

orbital_json_write(x, path)

Arguments

Details

The structure of the resulting JSON file allows for easy reading, both by orbital itself with orbital_json_read(), but potentially by other packages and langauges. The file is versioned by the version field to allow for changes why being backwards combatible with older file versions.

Value

Nothing.

See Also

orbital_json_read()

Examples

library(workflows)
library(recipes)
library(parsnip)

rec_spec <- recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_numeric_predictors())

lm_spec <- linear_reg()

wf_spec <- workflow(rec_spec, lm_spec)

wf_fit <- fit(wf_spec, mtcars)

orbital_obj <- orbital(wf_fit)

tmp_file <- tempfile()

orbital_json_write(orbital_obj, tmp_file)

readLines(tmp_file)

Description

Returns a R file that contains a function that output predictions when applied to data frames.

Usage

orbital_r_fun(x, name = "orbital_predict", file)

Arguments

Details

The generated function is only expected to work on data frame objects. The generated function doesn't require the orbital package to be loaded. Depending on what models and steps are used, other packages such as dplyr will need to be loaded as well.

Value

Nothing.

Examples

library(workflows)
library(recipes)
library(parsnip)

rec_spec <- recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_numeric_predictors())

lm_spec <- linear_reg()

wf_spec <- workflow(rec_spec, lm_spec)

wf_fit <- fit(wf_spec, mtcars)

orbital_obj <- orbital(wf_fit)

file_name <- tempfile()

orbital_r_fun(orbital_obj, file = file_name)

readLines(file_name)

Description

Returns SQL code that is equivilant to prediction code.

Usage

orbital_sql(x, con)

Arguments

Details

This function requires a database connection object, as the resulting code SQL code can differ depending on the type of database.

Value

SQL code.

Examples

library(workflows)
library(recipes)
library(parsnip)

rec_spec <- recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_numeric_predictors())

lm_spec <- linear_reg()

wf_spec <- workflow(rec_spec, lm_spec)

wf_fit <- fit(wf_spec, mtcars)

orbital_obj <- orbital(wf_fit)

library(dbplyr)
con <- simulate_dbi()

orbital_sql(orbital_obj, con)

Description

Running prediction on data frame of remote database table, without needing to load original packages used to fit model.

Usage

## S3 method for class 'orbital_class'
predict(object, new_data, ...)

Arguments

Details

Using this function should give identical results to running predict() or bake() on the orginal object.

The prediction done will only return prediction colunms, a opposed to returning all modified functions as done with orbital_inline().

Value

A modified data frame or remote database table.

Examples

library(workflows)
library(recipes)
library(parsnip)

rec_spec <- recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_numeric_predictors())

lm_spec <- linear_reg()

wf_spec <- workflow(rec_spec, lm_spec)

wf_fit <- fit(wf_spec, mtcars)

orbital_obj <- orbital(wf_fit)

predict(orbital_obj, mtcars)