Package Data

Package data is feature of val.meter - it is the framework by which we calculate package qualities and derive metrics. Package data is any derived information about a package. Fundamentally, new data is implemented using the methods pkg_data_derive() and pkg_data_info() used for calculating and providing metadata respectively.

Usage

impl_data(
  name,
  fn,
  for_resource = resource,
  ...,
  overwrite = FALSE,
  quiet = FALSE
)

impl_data_info(
  name,
  class = class_any,
  title = character(0L),
  description = character(0L),
  tags = class_tags(c()),
  permissions = class_permissions(character(0L)),
  suggests = class_suggests(character(0L)),
  metric = FALSE,
  overwrite = FALSE,
  quiet = FALSE
)

impl_data_derive(name, fn, for_resource, overwrite = FALSE, quiet = FALSE)

Arguments

name

character(1L) the data field name to implement.

fn

function to use when deriving the data for a pkg.

for_resource

resource classes that can use this method to derive data. Classes can also be S7::new_union()s or abstract classes to allow multiple resources to use the same method.

...

Additional arguments passed from impl_data to impl_data_info and are provided as metadata.

overwrite

logical(1L) flag indicating that the method should overwrite an existing method that uses the same dispatch arguments.

quiet

logical(1L) flag indicating that overwriting should be quiet, suppressing messages emitted during overwriting.

class

S7::S7_class or character(n). A return type that is type checked after method evaluation. character values will be coerced into S7 representations of S3 classes.

title

character(1L) data title, used for user-facing communication about data derivation or purpose.

description

character(n) or Rd object used as longer-form documentation about the data.

tags

tags() associated with the data.

permissions

permissions() required to compute the data.

suggests

character(n) packages which must be installed in order to derive the data.

metric

logical(1L) flag indicating whether the data should be user-facing as a metric. When TRUE, class must be atomic.

Details

However, implementing these functions directly requires a bit of knowledge of how the internals of the package are structured. Instead, it is recommended to use impl_data(), which provides a high-level interface that helps make this process as simple as possible.

Functions

• impl_data(): Helper for implementing all the necessary methods for package data. Internally this is a wrapper for [impl_data_meta()] (associate metadata with the data field), [impl_data_derive()] (associate a derivation function for a combination of data field and package resource) and [impl_metric()] (declare a piece of data to be a metric).

• impl_data_info(): Associate metadata with the data field

• impl_data_derive(): Register a derivation function for a data field and package resource.

Implementing a new metric

To implement some new data, you can use impl_data(), providing, at a minimum, the field name for the data and the way it should be calculated.

impl_data("name_character_count", function(pkg, ...) nchar(pkg$name))

Examples

p <- random_pkg()
impl_data("name_character_count", function(pkg, ...) nchar(pkg$name))
p$name_character_count
#> [1] 32