Dependency analysis on package tfdatasets

General information

CRAN link link
Package version2.7.0
Number of strong dependencies38
Number of all dependencies91
Number of parent packages8
Max heaviness from parent packages12
Total heaviness from parent packages13
Number of parent packages (including Suggests and Enhances)16
Max co-heaviness from parent packages ("tensorflow" and "reticulate")14

Dependency heatmap

In the following dependency heatmap, rows are the parent packages of tfdatasets and columns are the dependency packages that each parent package brings in. On the right side of the heatmap, there are three barplot annotations: 1. number of imported functions/S4 methods/S4 classes from parent packages; 2. number of dependency packages from each parent package; 3. heaviness of each parent package on tfdatasets.

Adjust heatmap size:
Base packages Other packages Imports Suggests utils methods grDevices stats graphics grid tools parallel glue rlang magrittr cli ellipsis vctrs R6 purrr jsonlite rappdirs withr rprojroot yaml tidyselect Matrix Rcpp RcppTOML here png lattice processx rstudioapi ps base64enc config reticulate tfruns tfautograph whisker backports crayon lifecycle fansi tibble pillar pkgconfig utf8 generics evaluate digest highr stringr xfun stringi dplyr tidyr cpp11 tensorflow bslib htmltools jquerylib knitr tinytex sass fastmap fs furrr slider future globals warp listenv parallelly codetools brio callr desc pkgload praise waldo diffobj rematch2 zeallot forge progress hms prettyunits 0 2 4 6 Imported methods 0 20 40 Required packages 0 10 20 Heaviness from parents magrittr stats rlang generics vctrs tidyselect reticulate tensorflow Metrics knitr dplyr rmarkdown rsample testthat keras tfestimators In total 38 packages are required directly or indirectly (91) when installing 'tfdatasets' (2.7.0) Imports Suggests Base packages that are required Recommended packages that are required Contributed packages that are required Packages that are required for installing 'tfdatasets' Imported functions The whole namespace is imported Package is listed in 'Imports' but namespace is not imported


Dependency table

"Import" information is from the NAMESPACE file of tfdatasets.

imports: number of imported functions/objects; importMethods: number of imported S4 methods; importClasses: number of imported S4 classes.

Required packages: number of strong dependency packages for each of the parent package (or in other words, number of dependency packages the parent package brings in).

Heaviness from parent on tfdatasets: number of required packages that can be reduced if moving parent package to Suggests of tfdatasets.

Parent package Field imports importMethods importClasses Required packages Heaviness from parent on tfdatasets
tensorflow Imports 4 0 0 36 12
generics Imports 1 0 0 1 1
magrittr Imports 1 0 0 0 0
stats Imports Parent package is listed in Imports of tfdatasets but no object from parent package is imported. 0 0
rlang Imports The whole set of functions/methods/classes from parent package is imported to the namespace of tfdatasets. 1 0
vctrs Imports Parent package is listed in Imports of tfdatasets but no object from parent package is imported. 5 0
tidyselect Imports 7 0 0 9 0
reticulate Imports 7 0 0 16 0
rsample Suggests No object is imported into the namespace of tfdatasets. 34 22
testthat Suggests No object is imported into the namespace of tfdatasets. 36 19
rmarkdown Suggests No object is imported into the namespace of tfdatasets. 27 16
tfestimators Suggests No object is imported into the namespace of tfdatasets. 52 15
dplyr Suggests No object is imported into the namespace of tfdatasets. 20 8
knitr Suggests No object is imported into the namespace of tfdatasets. 12 7
keras Suggests No object is imported into the namespace of tfdatasets. 39 2
Metrics Suggests No object is imported into the namespace of tfdatasets. 0 1

The following table lists pairs of parent packages with co-heaviness larger than 10. The co-heaviness measures the number of additional dependencies that two parent packages simultaneously import and are only imported by the two parents.

Note the co-heaviness from parents can be always easily observed in the dependency heatmap.

Parent 1 Parent 2 Co-heaviness on tfdatasets Co-heaviness as Jaccard coeffcient
reticulate tensorflow 14 0.538

Analysis was done with pkgndep.