For the examples in this vignette glue will be used as an example. glue version 1.6.2.9000 is included in the system files of PaRe and is thus accessible even if these examples are ran offline.

PaRe does fetch some online resources through the package pak. And by default online stored csv-files in the PaRe::whiteList data.frame. If no connection can be made, functions using these methods to reference these online resources will return NULL.

Whitelist Data Frame

PaRe includes a data frame which contains links to csv-files to be used in the PaRe::checkDependencies and PaRe::getDefaultPermittedPackages functions.

By default the data frame contains the following information.

PaRe::whiteList
#> # A tibble: 3 × 4
#>   source    link                                                 package version
#>   <chr>     <chr>                                                <chr>   <chr>  
#> 1 tidyverse https://raw.githubusercontent.com/mvankessel-EMC/De… package version
#> 2 darwin    https://raw.githubusercontent.com/mvankessel-EMC/De… package version
#> 3 hades     https://raw.githubusercontent.com/mvankessel-EMC/De… package version

The data frame contains 4 columns:

  1. source: Source name.
  2. link: Link or path to the csv-file.
  3. package: Column name in the referenced csv-file that contains the package names.
  4. version: Column name in the referenced csv-file that contains the package versions.

If you wish to alter the sources in just your R-session, you can either add, remove, or replace individual rows in the whiteList data frame.

sessionWhiteList <- rbind(
  whiteList,
  list(
    source = "dummySession",
    link = "some/file.csv",
    package = "package",
    version = "version"
  )
)

sessionWhiteList
#> # A tibble: 4 × 4
#>   source       link                                              package version
#>   <chr>        <chr>                                             <chr>   <chr>  
#> 1 tidyverse    https://raw.githubusercontent.com/mvankessel-EMC… package version
#> 2 darwin       https://raw.githubusercontent.com/mvankessel-EMC… package version
#> 3 hades        https://raw.githubusercontent.com/mvankessel-EMC… package version
#> 4 dummySession some/file.csv                                     package version

If you wish to make more permanent alterations to the whiteList data frame, you can edit the whiteList.csv file in the PaRe system files.

fileWhiteList <- rbind(
  read.csv(
    system.file(
      package = "PaRe",
      "whiteList.csv"
    )
  ),
  list(
    source = "dummyFile",
    link = "some/file.csv",
    package = "package",
    version = "version"
  )
)

fileWhiteList
#>      source
#> 1 tidyverse
#> 2    darwin
#> 3     hades
#> 4 dummyFile
#>                                                                                               link
#> 1 https://raw.githubusercontent.com/mvankessel-EMC/DependencyReviewerWhitelists/main/tidyverse.csv
#> 2    https://raw.githubusercontent.com/mvankessel-EMC/DependencyReviewerWhitelists/main/darwin.csv
#> 3     https://raw.githubusercontent.com/mvankessel-EMC/DependencyReviewerWhitelists/main/hades.csv
#> 4                                                                                    some/file.csv
#>   package version
#> 1 package version
#> 2 package version
#> 3 package version
#> 4 package version
write.csv(
  fileWhiteList,
  system.file(
    package = "PaRe",
    "whiteList.csv"
  )
)

Dependency Review

Before we start diving into the dependency usage of glue we should first establish what our dependency white list even looks like. We can retrieve our full list of whitelisted dependencies buy calling the getDefaultPermittedPackages function.

getDefaultPermittedPackages

PaRe::getDefaultPermittedPackages(base = TRUE)

getDefaultPermittedPackages takes one parameter:

  1. base which is set to TRUE by default. Packages that listed as base packages will be included in the white list.

Setting up a Repository object

# Temp dir to clone repo to
tempDir <- tempdir()
pathToRepo <- file.path(tempDir, "glue")

# Clone IncidencePrevalence to temp dir
git2r::clone(
  url = "https://github.com/tidyverse/glue.git",
  local_path = pathToRepo
)

repo <- PaRe::Repository$new(path = pathToRepo)
#> cloning into 'C:\Users\MVANKE~1\AppData\Local\Temp\RtmpQRpwDq/glue'...
#> Receiving objects:   1% (50/4925),   63 kb
#> Receiving objects:  11% (542/4925),  352 kb
#> Receiving objects:  21% (1035/4925), 1768 kb
#> Receiving objects:  31% (1527/4925), 2345 kb
#> Receiving objects:  41% (2020/4925), 2737 kb
#> Receiving objects:  51% (2512/4925), 2849 kb
#> Receiving objects:  61% (3005/4925), 3241 kb
#> Receiving objects:  71% (3497/4925), 3746 kb
#> Receiving objects:  81% (3990/4925), 3858 kb
#> Receiving objects:  91% (4482/4925), 4026 kb
#> Receiving objects: 100% (4925/4925), 4892 kb, done.

checkDependencies

Now that we know what is included in the white list, we can make our first step into reviewing glue, which is to ensure the (suggested) dependencies glue uses are in our white list.

PaRe::checkDependencies(repo = repo)
→ The following are not permitted: covr, microbenchmark, R.utils, rprintf, testthat                  
→ Please open an issue here: https://github.com/mvankessel-EMC/DependencyReviewerWhitelists/issues
package version
covr *
microbenchmark *
R.utils *
rprintf *
testthat 3.0.0

Not all suggested dependencies are approved. The function prints a message and returns a data frame, containing all packages that are not listed in our white list.

checkDependecies takes two parameters:

  1. pkgPath which specifies the path to the pacakge.
  2. dependencyType a vector of character items which specify kinds of imports to look at.

getGraphData

glue depends on (suggested) dependencies. These dependencies in turn import other dependencies, and so on. We can investigate how these recursive dependencies depend on one another, by investigating it as a graph.

graphData <- PaRe::getGraphData(
  repo = repo,
  packageTypes = c("imports", "suggests")
)

We can compute several statistics about our dependency graph

data.frame(
  countVertices = length(igraph::V(graphData)),
  countEdges = length(igraph::E(graphData)),
  meanDegree = round(mean(igraph::degree(graphData)), 2),
  meanDistance = round(mean(igraph::distances(graphData)), 2)
)
  • countVertices resembles the amount of recursive dependencies glue depends on.
  • countEdges: are the total amount of imports of all dependencies.
  • meanDegree: is the average amount of imports per dependency.
  • meanDistance: is the average amount of dependencies between glue and all other recursive dependencies.

We can then plot the graph.

plot(graphData)
glue graph
glue graph

Package wide function use

PaRe allows you to get insight in the function usage in a package.

summariseFunctionUse

funsUsed <- PaRe::getFunctionUse(repo = repo)
funsUsed
#> # A tibble: 426 × 4
#>    file     line pkg     fun         
#>    <chr>   <int> <chr>   <chr>       
#>  1 color.R    59 base    function    
#>  2 color.R    59 base    parent.frame
#>  3 color.R    60 unknown glue        
#>  4 color.R    65 base    function    
#>  5 color.R    65 base    parent.frame
#>  6 color.R    66 unknown glue_data   
#>  7 color.R    69 base    function    
#>  8 color.R    70 base    function    
#>  9 color.R    70 base    parse       
#> 10 color.R    70 base    tryCatch    
#> # ℹ 416 more rows

getDefinedFunctions

defFuns <- PaRe::getDefinedFunctions(repo = repo)
head(defFuns)
#>                name lineStart lineEnd nArgs cycloComp             fileName
#> 1          glue_col        59      61     3         1              color.R
#> 2     glue_data_col        65      67     4         1              color.R
#> 3 color_transformer        69      99     2         8              color.R
#> 4           .onLoad        47      50     1         1 compat-s3-register.R
#> 5       s3_register        53     122     4         8 compat-s3-register.R
#> 6    get_method_env        64      71     1         2 compat-s3-register.R

Besides the location of each function being displayed, the number of arguments for each function, and the cyclometic complexity is also included in the result.

PaRe::pkgDiagram(repo = repo) %>%
  DiagrammeRsvg::export_svg() %>%
  magick::image_read_svg()

Lines of code

PaRe::countPackageLines(repo = repo)
#> # A tibble: 1 × 6
#>       R   cpp     o     h  java   sql
#>   <int> <int> <int> <int> <int> <int>
#> 1  1118     0     0     0     0     0

glue contains 1056 lines of R-code.

Linting

messages <- PaRe::lintRepo(repo = repo)
PaRe::lintScore(repo = repo, messages = messages)
#> Warning: Returning more (or less) than 1 row per `summarise()` group was deprecated in
#> dplyr 1.1.0.
#>  Please use `reframe()` instead.
#>  When switching from `summarise()` to `reframe()`, remember that `reframe()`
#>   always returns an ungrouped data frame and adjust accordingly.
#>  The deprecated feature was likely used in the PaRe package.
#>   Please report the issue at <https://github.com/darwin-eu-dev/PaRe/issues>.
#> This warning is displayed once every 8 hours.
#> Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
#> generated.
#> # A tibble: 2 × 2
#>   type      pct
#>   <chr>   <dbl>
#> 1 style   10.2 
#> 2 warning  3.94
head(messages)
#>   filename line_number column_number    type
#> 1  color.R           8            81   style
#> 2  color.R          14            81   style
#> 3  color.R          59             1   style
#> 4  color.R          59            81   style
#> 5  color.R          60             3 warning
#> 6  color.R          60            81   style
#>                                                 message
#> 1          Lines should not be more than 80 characters.
#> 2          Lines should not be more than 80 characters.
#> 3 Variable and function name style should be camelCase.
#> 4          Lines should not be more than 80 characters.
#> 5      no visible global function definition for 'glue'
#> 6          Lines should not be more than 80 characters.
#>                                                                                             line
#> 1  #' Using the following syntax will apply the function [crayon::blue()] to the text 'foo bar'.
#> 2              #' If you want an expression to be evaluated, simply place that in a normal brace
#> 3             glue_col <- function(..., .envir = parent.frame(), .na = "NA", .literal = FALSE) {
#> 4             glue_col <- function(..., .envir = parent.frame(), .na = "NA", .literal = FALSE) {
#> 5   glue(..., .envir = .envir, .na = .na, .literal = .literal, .transformer = color_transformer)
#> 6   glue(..., .envir = .envir, .na = .na, .literal = .literal, .transformer = color_transformer)
#>                linter
#> 1  line_length_linter
#> 2  line_length_linter
#> 3  object_name_linter
#> 4  line_length_linter
#> 5 object_usage_linter
#> 6  line_length_linter