heaviness_on_downstream: Heaviness on all downstream packages

View source: R/heaviness.R

heaviness_on_downstreamR Documentation

Heaviness on all downstream packages

Description

Heaviness on all downstream packages

Usage

heaviness_on_downstream(package, add_values_attr = FALSE, via = NULL,
    total = FALSE, internal = FALSE)

Arguments

package

A package name.

add_values_attr

Whether to include "values" attribute? Internally used.

via

Whether to only consider downstream packages via a intermediate package?

total

Whether to return the total heaviness?

internal

Whether to use internally calculated heaviness?

Details

It is calculated based on a specific CRAN/Bioconductor snapshot. The version is set via pkgndep_opt$heaviness_db_version.

Value

The value is the mean heaviness of the package on all its downstream packages. Denote n as the number of all its downstream packages, k_i as the number of required packages for package i, v_1 as the total number of required packages for all downstream packages, i.e. v_1 = sum_i^n {k_i}. Denote p_i as the number of required packages if moving package to Suggests, and v_2 as the total number of required packages, i.e. v_1 = sum_i^n {p_i}. The final heaviniss on downstream packages is (v_1 - v_2)/n.

Note since the interaction from package to its downstream packages may go through several intermediate packages, which means, the reduction of required packages for a downstream package might be joint effects from all its upstream packages, thus, to properly calculate the heaviness of a package to its downstream packages, we first make a copy of the package database and move package to Suggests for all packages which depends on package. Then for all downstream packages of package, dependency analysis by pkgndep is redone with the modified package database. Finally, the heaviness on downstream packages is collected and the mean heaviness is calculated.

Examples

## Not run: 
heaviness_on_downstream("ComplexHeatmap")

## End(Not run)

pkgndep documentation built on Oct. 13, 2023, 9:08 a.m.