Package 'pkgstats'

Title: Metrics of R Packages
Description: Static code analyses for R packages using the external code-tagging libraries 'ctags' and 'gtags'. Static analyses enable packages to be analysed very quickly, generally a couple of seconds at most. The package also provides access to a database generating by applying the main function to the full 'CRAN' archive, enabling the statistical properties of any package to be compared with all other 'CRAN' packages.
Authors: Mark Padgham [aut, cre]
Maintainer: Mark Padgham <[email protected]>
License: GPL-3
Version: 0.2.0.048
Built: 2024-11-05 18:20:59 UTC
Source: https://github.com/ropensci-review-tools/pkgstats

Help Index

Install 'ctags' from a clone of the 'git' repository

Description

'ctags' is installed with this package on both Windows and macOS systems; this is an additional function to install from source on Unix systems.

Usage

ctags_install(bin_dir = NULL, sudo = TRUE)

Arguments

bin_dir

Prefix to pass to the autoconf configure command defining location to install the binary, with default of ⁠/usr/local⁠.
sudo

Set to FALSE if sudo is not available, in which case a value for bin_dir will also have to be explicitly specified, and be a location where a binary is able to be installed without sudo privileges.

Value

Nothing; the function will fail if installation fails, otherwise returns nothing.

See Also

Other tags: ctags_test(), tags_data()

Examples

## Not run: 
ctags_install (bin_dir = "/usr/local") # default

## End(Not run)

test a 'ctags' installation

Description

This uses the example from https://github.com/universal-ctags/ctags/blob/master/man/ctags-lang-r.7.rst.in and also checks the GNU global installation.

Usage

ctags_test(quiet = TRUE)

Arguments

quiet

If TRUE, display on screen whether or not 'ctags' is correctly installed.

Value

'TRUE' or 'FALSE' respectively indicating whether or not 'ctags' is correctly installed.

See Also

Other tags: ctags_install(), tags_data()

Examples

## Not run: 
ctags_test ()

## End(Not run)

Statistics from DESCRIPTION files

Description

Statistics from DESCRIPTION files

Usage

desc_stats(path)

Arguments

path

Directory to source code of package being analysed

Value

A data.frame with one row and 16 columns extracting various information from the 'DESCRIPTION' file, include websites, tallies of different kinds of authors and contributors, and package dependencies.

See Also

Other stats: loc_stats(), pkgstats(), pkgstats_summary(), rd_stats()

Examples

f <- system.file ("extdata", "pkgstats_9.9.tar.gz", package = "pkgstats")
# have to extract tarball to call function on source code:
path <- extract_tarball (f)
desc_stats (path)

Download latest version of 'pkgstats' data

Description

Download latest version of 'pkgstats' data

Usage

dl_pkgstats_data(current = TRUE, path = tempdir(), quiet = FALSE)

Arguments

current

If 'FALSE', download data for all CRAN packages ever released, otherwise (default) download data only for current CRAN packages.
path

Local path to download file.
quiet

If FALSE, display progress information on screen.

Value

(Invisibly) A data.frame of pkgstats results, one row for each package.

See Also

Other archive: pkgstats_cran_current_from_full(), pkgstats_fns_from_archive(), pkgstats_fns_update(), pkgstats_from_archive(), pkgstats_update()

Extract tarball of a package into temp directory and return path to extracted package

Description

Extract tarball of a package into temp directory and return path to extracted package

Usage

extract_tarball(tarball)

Arguments

tarball

Full path to local tarball of an R package.

Value

Path to extracted version of package (in tempdir()).

See Also

Other misc: pkgstats_fn_names()

Examples

f <- system.file ("extdata", "pkgstats_9.9.tar.gz", package = "pkgstats")
path <- extract_tarball (f)

Internal calculation of Lines-of-Code Statistics

Description

Internal calculation of Lines-of-Code Statistics

Usage

loc_stats(path)

Arguments

path

Directory to source code of package being analysed

Value

A list of statistics for each of three directories, 'R', 'src', and 'inst/include', each one having 5 statistics of total numbers of lines, numbers of empty lines, total numbers of white spaces, total numbers of characters, and indentation used in files in that directory.

Note

NA values are returned for directories which do not exist.

See Also

Other stats: desc_stats(), pkgstats(), pkgstats_summary(), rd_stats()

Examples

f <- system.file ("extdata", "pkgstats_9.9.tar.gz", package = "pkgstats")
# have to extract tarball to call function on source code:
path <- extract_tarball (f)
loc_stats (path)

Analyse statistics of one R package

Description

Analyse statistics of one R package

Usage

pkgstats(path = ".")

Arguments

path

Either a path to a local source repository, or a local '.tar.gz' file, containing code for an R package.

Value

List of statistics and data on function call networks (or object relationships in other languages). Includes the following components:

  1. loc: Summary of Lines-of-Code in all package directories

  2. vignettes: Numbers of vignettes and "demo" files

  3. data_stats: Statistics of numbers and sizes of package data files

  4. desc: Summary of contents of 'DESCRIPTION' file

  5. translations: List of translations into other (human) languages (where provides)

  6. objects: A data.frame of all functions in R, and all other objects (functions, classes, structures, global variables, and more) in all other languages

  7. network: A data.frame of object references within and between all languages; in R these are function calls, but may be more abstract in other languages.

  8. external_calls: A data.frame of all calls make to all functions from all other R packages, including base and recommended as well as contributed packages.

See Also

Other stats: desc_stats(), loc_stats(), pkgstats_summary(), rd_stats()

Examples

# 'path' can be path to a package tarball:
f <- system.file ("extdata", "pkgstats_9.9.tar.gz", package = "pkgstats")
## Not run: 
s <- pkgstats (f)

## End(Not run)
# or to a source directory:
path <- extract_tarball (f)
## Not run: 
s <- pkgstats (path)

## End(Not run)

Reduce data.frame of full CRAN archive data to current packages only.

Description

Reduce data.frame of full CRAN archive data to current packages only.

Usage

pkgstats_cran_current_from_full(prev_results, results_file = NULL)

Arguments

prev_results

Result of previous call to this function, if available. Submitting previous results will ensure that only newer packages not present in previous result will be analysed, with new results simply appended to previous results. This parameter can also specify a file to be read with readRDS().
results_file

Can be used to specify the name or full path of a .Rds file to which results should be saved once they have been generated. The '.Rds' extension will be automatically appended, and any other extensions will be ignored.

See Also

Other archive: dl_pkgstats_data(), pkgstats_fns_from_archive(), pkgstats_fns_update(), pkgstats_from_archive(), pkgstats_update()

Extract names of all functions for one R package

Description

Extract names of all functions for one R package

Usage

pkgstats_fn_names(path)

Arguments

path

Either a path to a local source repository, or a local '.tar.gz' file, containing code for an R package.

Value

A data.frame with three columns:

  • package: Name of package

  • version: Package version

  • fn_name: Name of function

See Also

Other misc: extract_tarball()

Examples

# 'path' can be path to a package tarball:
f <- system.file ("extdata", "pkgstats_9.9.tar.gz", package = "pkgstats")
path <- extract_tarball (f)
s <- pkgstats_fn_names (path)

Trawl a local CRAN archive to extract function names only from all packages

Description

Trawl a local CRAN archive to extract function names only from all packages

Usage

pkgstats_fns_from_archive(
  path,
  archive = FALSE,
  prev_results = NULL,
  results_file = NULL,
  chunk_size = 1000L,
  num_cores = 1L,
  results_path = fs::path_temp()
)

Arguments

path

Path to local archive of R packages, either as source directories, or '.tar.gz' files such as in a CRAN mirror.
archive

If TRUE, extract statistics for all packages in the ⁠/Archive⁠ sub-directory, otherwise only statistics for main directory (that is, current packages only).
prev_results

Result of previous call to this function, if available. Submitting previous results will ensure that only newer packages not present in previous result will be analysed, with new results simply appended to previous results. This parameter can also specify a file to be read with readRDS().
results_file

Can be used to specify the name or full path of a .Rds file to which results should be saved once they have been generated. The '.Rds' extension will be automatically appended, and any other extensions will be ignored.
chunk_size

Divide large archive trawl into chunks of this size, and save intermediate results to local files. These intermediate files can be combined to generate a single prev_results file, to enable jobs to be stopped and re-started without having to recalculate all results. These files will be named pkgstats-results-N.Rds, where "N" incrementally numbers each file.
num_cores

Number of machine cores to use in parallel, defaulting to single-core processing.
results_path

Path to save intermediate files generated by the chunk_size parameter described above.

Value

A data.frame object with one row for each function in each package and the following columns:

  • Package name

  • Package version

  • Function name

See Also

Other archive: dl_pkgstats_data(), pkgstats_cran_current_from_full(), pkgstats_fns_update(), pkgstats_from_archive(), pkgstats_update()

Update function names data from previous data and newly updated CRAN packages only.

Description

Update function names data from previous data and newly updated CRAN packages only.

Usage

pkgstats_fns_update(
  prev_results = NULL,
  results_file = NULL,
  chunk_size = 1000L,
  num_cores = 1L,
  results_path = tempdir()
)

Arguments

prev_results

Result of previous call to this function, if available. Submitting previous results will ensure that only newer packages not present in previous result will be analysed, with new results simply appended to previous results. This parameter can also specify a file to be read with readRDS().
results_file

Can be used to specify the name or full path of a .Rds file to which results should be saved once they have been generated. The '.Rds' extension will be automatically appended, and any other extensions will be ignored.
chunk_size

Divide large archive trawl into chunks of this size, and save intermediate results to local files. These intermediate files can be combined to generate a single prev_results file, to enable jobs to be stopped and re-started without having to recalculate all results. These files will be named pkgstats-results-N.Rds, where "N" incrementally numbers each file.
num_cores

Number of machine cores to use in parallel, defaulting to single-core processing.
results_path

Path to save intermediate files generated by the chunk_size parameter described above.

Value

A data.frame object with one row for each function in each package and the following columns:

  • Package name

  • Package version

  • Function name

See Also

Other archive: dl_pkgstats_data(), pkgstats_cran_current_from_full(), pkgstats_fns_from_archive(), pkgstats_from_archive(), pkgstats_update()

Trawl a local CRAN archive and extract statistics from all packages

Description

Trawl a local CRAN archive and extract statistics from all packages

Usage

pkgstats_from_archive(
  path,
  archive = TRUE,
  prev_results = NULL,
  results_file = NULL,
  chunk_size = 1000L,
  num_cores = 1L,
  save_full = FALSE,
  save_ex_calls = FALSE,
  results_path = fs::path_temp()
)

Arguments

path

Path to local archive of R packages, either as source directories, or '.tar.gz' files such as in a CRAN mirror.
archive

If TRUE, extract statistics for all packages in the ⁠/Archive⁠ sub-directory, otherwise only statistics for main directory (that is, current packages only).
prev_results

Result of previous call to this function, if available. Submitting previous results will ensure that only newer packages not present in previous result will be analysed, with new results simply appended to previous results. This parameter can also specify a file to be read with readRDS().
results_file

Can be used to specify the name or full path of a .Rds file to which results should be saved once they have been generated. The '.Rds' extension will be automatically appended, and any other extensions will be ignored.
chunk_size

Divide large archive trawl into chunks of this size, and save intermediate results to local files. These intermediate files can be combined to generate a single prev_results file, to enable jobs to be stopped and re-started without having to recalculate all results. These files will be named pkgstats-results-N.Rds, where "N" incrementally numbers each file.
num_cores

Number of machine cores to use in parallel, defaulting to single-core processing.
save_full

If TRUE, full pkgstats results are saved for each package to files in results_path.
save_ex_calls

If TRUE, the results of the external_calls component are saved for each package to files in results_path (only if save_full = FALSE).
results_path

Path to save intermediate files generated by the chunk_size parameter described above.

Value

A data.frame object with one row for each package containing summary statistics generated from the pkgstats_summary function.

See Also

Other archive: dl_pkgstats_data(), pkgstats_cran_current_from_full(), pkgstats_fns_from_archive(), pkgstats_fns_update(), pkgstats_update()

Examples

# Create fake archive directory with single tarball:
f <- system.file ("extdata", "pkgstats_9.9.tar.gz", package = "pkgstats")
tarball <- basename (f)

archive_path <- file.path (tempdir (), "archive")
if (!dir.exists (archive_path)) {
    dir.create (archive_path)
}
path <- file.path (archive_path, tarball)
file.copy (f, path)
tarball_path <- file.path (archive_path, "tarballs")
dir.create (tarball_path, recursive = TRUE)
file.copy (path, file.path (tarball_path, tarball))
## Not run: 
out <- pkgstats_from_archive (tarball_path)

## End(Not run)

Condense the output of pkgstats to summary statistics only

Description

Condense the output of pkgstats to summary statistics only

Usage

pkgstats_summary(s = NULL)

Arguments

s

Output of pkgstats, containing full statistical data on one package. Default of NULL returns a single row with NA values (used in pkgstats_from_archive).

Value

Summarised version of s, as a single row of a standardised data.frame object

Note

Variable names in the summary object use the following abbreviations:

  • "loc" = Lines-of-Code

  • "fn" = Function

  • "n_fns" = Number of functions

  • "npars" = Number of parameters

  • "doclines" = Number of documentation lines

  • "nedges" = Number of edges in function call network, as a count of unique edges, which may be less than the size of the network object returned by pkgstats, because that may include multiple calls between identical function pairs.

  • "n_clusters" = Number of connected clusters within the function call network.

  • "centrality" used as a prefix for several statistics, along with "dir" or "undir" for centrality calculated on networks respectively constructed with directed or undirected edges; "mn" or "md" for respective measures of mean or median centrality, and "no0" for measures excluding edges with zero centrality.

See Also

Other stats: desc_stats(), loc_stats(), pkgstats(), rd_stats()

Examples

f <- system.file ("extdata", "pkgstats_9.9.tar.gz", package = "pkgstats")
## Not run: 
p <- pkgstats (f)
s <- pkgstats_summary (p)

## End(Not run)

Update pkgstats' data on GitHub release

Description

This function is intended for internal rOpenSci use only. Usage by any unauthorized users will error and have no effect unless run with upload = FALSE, in which case updated data will be created in the sub-directory "pkgstats-results" of R's current temporary directory.

Usage

pkgstats_update(upload = TRUE)

Arguments

upload

If TRUE, upload updated results to GitHub release.

Value

Local path to directory containing updated results.

See Also

Other archive: dl_pkgstats_data(), pkgstats_cran_current_from_full(), pkgstats_fns_from_archive(), pkgstats_fns_update(), pkgstats_from_archive()

Plot interactive visNetwork visualisation of object-relationship network of package.

Description

Plot interactive visNetwork visualisation of object-relationship network of package.

Usage

plot_network(s, plot = TRUE, vis_save = NULL)

Arguments

s

Package statistics obtained from pkgstats function.
plot

If TRUE, plot the network using visNetwork which opens an interactive browser pane.
vis_save

Name of local file in which to save html file of network visualisation (will override plot to FALSE).

Value

(Invisibly) A visNetwork representation of the package network.

Note

Edge thicknesses are scaled to centrality within the package function call network. Node sizes are scaled to numbers of times each function is called from all other functions within a package.

Examples

f <- system.file ("extdata", "pkgstats_9.9.tar.gz", package = "pkgstats")
## Not run: 
p <- pkgstats (f)
plot_network (p)

## End(Not run)

Stats from '.Rd' files

Description

Stats from '.Rd' files

Usage

rd_stats(path)

Arguments

path

Directory to source code of package being analysed

Value

A data.frame of function names and numbers of parameters and lines of documentation for each, along with mean and median numbers of characters used to document each parameter.

See Also

Other stats: desc_stats(), loc_stats(), pkgstats(), pkgstats_summary()

Examples

f <- system.file ("extdata", "pkgstats_9.9.tar.gz", package = "pkgstats")
# have to extract tarball to call function on source code:
path <- extract_tarball (f)
rd_stats (path)

use ctags and gtags to parse call data

Description

use ctags and gtags to parse call data

Usage

tags_data(path, has_tabs = NULL, pkg_name = NULL)

Arguments

path

Path to local repository
has_tabs

A logical flag indicating whether or not the code contains any tab characters. This can be determined from loc_stats, which has a tabs column. If not given, that value will be extracted from internally calling that function.
pkg_name

Only used for external_call_network, to label package-internal calls.

Value

A list of three items:

  • "network" A data.frame of relationships between objects, generally as calls between functions in R, but other kinds of relationships in other source languages. This is effectively an edge-based network representation, and the data frame also include network metrics for each edge, calculated through representing the network in both directed (suffix "_dir") and undirected (suffix "_undir") forms.

  • "objects" A data.frame of statistics on each object (generally functions in R, and other kinds of objects in other source languages), including the kind of object, the language, numbers of lines-of-code, parameters, and lines of documentation, and a binary flag indicating whether or not R functions accept "three-dots" parameters (...).

  • "external_calls" A data.frame of every call from within every R function to any external R package, including base and recommended packages. The location of each calls is recorded, along with the external function and package being called.

See Also

Other tags: ctags_install(), ctags_test()

Examples

f <- system.file ("extdata", "pkgstats_9.9.tar.gz", package = "pkgstats")
# have to extract tarball to call function on source code:
path <- extract_tarball (f)
## Not run: 
tags <- tags_data (path)

## End(Not run)