Type:	Package
Title:	Testing Infrastructure for Broom Model Generics
Version:	0.1.6
Description:	Provides a number of testthat tests that can be used to verify that tidy(), glance() and augment() methods meet consistent specifications. This allows methods for the same generic to be spread across multiple packages, since all of those packages can make the same guarantees to users about returned objects.
License:	MIT + file LICENSE
URL:	https://github.com/alexpghayes/modeltests
BugReports:	https://github.com/alexpghayes/modeltests/issues
Depends:	R (≥ 3.1)
Imports:	dplyr (≥ 0.7.6), generics, purrr (≥ 0.2.5), testthat (≥ 2.0.0), tibble (≥ 3.0.0)
Suggests:	covr
Encoding:	UTF-8
LazyData:	true
RoxygenNote:	7.2.3
NeedsCompilation:	no
Packaged:	2024-05-03 17:35:53 UTC; alex
Author:	Alex Hayes [aut, cre], Simon Couch [aut]
Maintainer:	Alex Hayes <alexpghayes@gmail.com>
Repository:	CRAN
Date/Publication:	2024-05-03 17:50:02 UTC

modeltests: Testing Infrastructure for Broom Model Generics

Description

Provides a number of testthat tests that can be used to verify that tidy(), glance() and augment() methods meet consistent specifications. This allows methods for the same generic to be spread across multiple packages, since all of those packages can make the same guarantees to users about returned objects.

Author(s)

Maintainer: Alex Hayes alexpghayes@gmail.com (ORCID)

Authors:

• Simon Couch simonpatrickcouch@gmail.com

See Also

Useful links:

• https://github.com/alexpghayes/modeltests

• Report bugs at https://github.com/alexpghayes/modeltests/issues

Determine acceptable names for augment output

Description

Given a data frame (or tibble), and a model object, makes a character vector of acceptable columns names for augment output. This includes:

• Any column names of the passed dataset

• Any syntactically correct column names generated by calling stats::model.frame() on the object in question.

Usage

acceptable_augment_colnames(object, passed_data)

Arguments

Value

A vector of colnames that are acceptable in augment output.

Allowed argument names in tidiers

Description

Allowed argument names in tidiers

Usage

argument_glossary

Format

A tibble with 3 variables:

method

One of "glance", "augment" or "tidy".

argument

Character name of allowed argument name.

description

Character description of argument use.

Examples

argument_glossary

Get copies of a dataset with various rowname behaviors

Description

Helper function for check_augment_data_specification(). There should be no need to ever use this directly in tests. Takes a dataset and returns a list with three copies of the dataset. Optionally introduces NA values into the dataset. Useful for checking that tibbles, data frames, and data frames with rownames are treated equivalently.

Usage

augment_data_helper(data, add_missing)

Arguments

Value

A list with three copies of data:

• tibble: the data in a tibble::tibble().

• no_row: the data in a data.frame() without row names.

• row_nm: the data in a data.frame, with row names.

See Also

.row_names_info(), rownames(), tibble::rownames_to_column()

Check that tidying methods use allowed argument names

Description

Call this function to perform tests. If a tests fails, an informative error will be thrown. Otherwise silent.

Tests when strict = FALSE:

• None

Tests when strict = TRUE:

• tidy_method has a conf.int argument if it has a conf.level argument.

• tidy_method has a conf.level argument if it has a conf.int argument.

• conf.int defaults to FALSE when present.

• conf.level defaults to '0.95“ when present.

• exponentiate defaults to FALSE when present.

• All arguments to tidy_method are listed in the argument_glossary.

Usage

check_arguments(tidy_method, strict = TRUE)

Arguments

Value

An invisible NULL. This function should be called for side effects, not return values.

See Also

testthat, testthat::expect_true()

Check that augment behavior is consistent for dataframes and tibbles

Description

Call this function to perform tests. If a tests fails, an informative error will be thrown. Otherwise silent.

Uses augment_data_helper() to create copies of the same dataset as a tibble, data frame and dataframe with rownames. When add_missing = TRUE these datasets have missing values along the diagonal, and one row of entirely missing values. Once the datasets have been generated, tests that:

• augment(fit, data = generated_dataset) passes check_tibble() for each generated dataset.

• Output of augment(fit, data = generated_dataset) is the same for all three generated datasets, except the data frame with rownames should also generate a .rownames column that the tibble and nameless data frame do not.

Additional tests when test_newdata = TRUE:

• head(aug(model, newdata = data)) equals aug(head(model, newdata = data)). This commutativity check catches issues where the output of predict changes for the same data point depending on the rest of the dataset.

Usage

check_augment_data_specification(aug, model, data, add_missing, test_newdata)

Arguments

Value

An invisible NULL. This function should be called for side effects, not return values.

Check an augment method

Description

Call this function to perform tests. If a tests fails, an informative error will be thrown. Otherwise silent.

Test when strict = FALSE:

• aug(model, data = data) passes check_tibble()

• aug(model, newdata = newdata) passes check_tibble()

Additional tests when strict = TRUE:

• aug(model, data = data) passes check_augment_data_specification().

• aug(model, newdata = newdata) passes check_augment_data_specification().

• aug(model, newdata = newdata) passes check_augment_data_specification() with add_missing = TRUE.

• If aug has a newdata argument, the newdata argument takes precedence over a data argument, i.e. calls check_augment_newdata_precedence().

• aug either gives an informative error or produces a reasonable tibble, i.e. calls check_augment_no_data().

Note that it doesn't make sense to test that aug(model, data = data) passes check_augment_data_specification() with add_missing = TRUE. This is because the user is already guaranteeing that data is the original dataset used to create model.

Usage

check_augment_function(aug, model, data = NULL, newdata = NULL, strict = TRUE)

Arguments

Value

An invisible NULL. This function should be called for side effects, not return values.

Check that newdata argument has higher precedence than data argument

Description

Call this function to perform tests. If a tests fails, an informative error will be thrown. Otherwise silent.

Usage

check_augment_newdata_precedence(aug, model, data, strict = TRUE)

Arguments

Value

An invisible NULL. This function should be called for side effects, not return values.

Check an augment method when no data or newdata is passed

Description

Call this function to perform tests. If a tests fails, an informative error will be thrown. Otherwise silent.

Test when strict = FALSE:

• None

Additional tests when strict = TRUE:

• aug(model) either returns an informative error or produces output that passes check_tibble().

• If the output passes check_tibble, will issue warning when:

  • Augmented data is missing rows from original data.

  • Augmented data is missing columns from original data.

  • Original data has rownames but ugmented data is missing .rownames column.

Usage

check_augment_no_data(aug, model, passed_data, strict = TRUE)

Arguments

Value

An invisible NULL. This function should be called for side effects, not return values.

Check that tibble has expected dimensions.

Description

Check that tibble has expected dimensions.

Usage

check_dims(data, expected_rows = NULL, expected_cols = NULL)

Arguments

Examples

check_dims(iris, expected_rows = 150)

Check the output of a glance method

Description

Call this function to perform tests. If a tests fails, an informative error will be thrown. Otherwise silent.

Tests when strict = FALSE:

• Each item passed to ... passes check_tibble()

• Each item passed to ... has exactly 1 row.

Additional tests when strict = TRUE:

• Column names and order agree across all elements of ....

Usage

check_glance_outputs(..., strict = TRUE)

Arguments

Value

An invisible NULL. This function should be called for side effects, not return values.

See Also

check_tibble()

Check the output of an augment method

Description

Call this function to perform tests. If a tests fails, an informative error will be thrown. Otherwise silent.

Test when strict = FALSE:

• au passes check_tibble().

• All column names present in passed_data are also present in au.

Additional tests when strict = TRUE:

• If passed_data has rownames other than simple row numbers (i.e. paste(1:5)), au contains a column called .rownames.

Usage

check_single_augment_output(au, passed_data, model = NULL, strict = TRUE)

Arguments

Value

An invisible NULL. This function should be called for side effects, not return values.

Check the output of a tidying method

Description

Call this function to perform tests. If a tests fails, an informative error will be thrown. Otherwise silent.

Tests when strict = FALSE:

• output is a tibble.

Additional tests when strict = TRUE:

• columns are listed in the column_glossary.

Usage

check_tibble(output, method, columns = colnames(output), strict = TRUE)

Arguments

Details

Do not call directly. Helper function used by check_tidy_output(), check_glance_outputs() and check_augment_function().

Value

An invisible NULL. This function should be called for side effects, not return values.

Check the output of a tidy method

Description

Call this function to perform tests. If a tests fails, an informative error will be thrown. Otherwise silent.

A thin wrapper around check_tibble().

Usage

check_tidy_output(td, strict = TRUE)

Arguments

Value

An invisible NULL. This function should be called for side effects, not return values.

Allowed column names in tidied tibbles

Description

Allowed column names in tidied tibbles

Usage

column_glossary

Format

A tibble with 4 variables:

method

One of "glance", "augment" or "tidy".

column

Character name of allowed output column.

description

Character description of expected column contents.

Examples

column_glossary

Check whether or not a data-frame-like object has rownames

Description

Check whether or not a data-frame-like object has rownames

Usage

has_rownames(df)

Arguments

Value

Logical indicating if df has rownames. If df is a tibble, returns FALSE. If df is a data.frame, return FALSE if the rownames are simply row numbers. If the rownames are anything other than the return row numbers, returns TRUE.