Managing package dependencies

• Some CRAN packages with names pkg\({}^{*}\)

• packrat, deepdep, pak, etc.

Introduction to crandep

• To obtain “all” the dependencies of CRAN packages
  • Depends, Imports, Suggests, LinkingTo, Enhances
  • their reverse counterparts
• Readily available formats:
  • As a data frame
  • As an igraph object
• Some useful functions

All types, all CRAN packages

df1 <- get_dep_all_packages() # utilises tools::CRAN_package_db()
head(df1)

##    from             to    type reverse
## 2 aaSEA             DT imports   FALSE
## 3 aaSEA      networkD3 imports   FALSE
## 4 aaSEA          shiny imports   FALSE
## 5 aaSEA shinydashboard imports   FALSE
## 6 aaSEA       magrittr imports   FALSE
## 7 aaSEA       Bios2cor imports   FALSE

tail(df1)

##         from         to     type reverse
## 411665 Zelig  stargazer enhances    TRUE
## 411666 Zelig     texreg enhances    TRUE
## 411697   zoo      chron enhances    TRUE
## 411698   zoo  lubridate enhances    TRUE
## 411699   zoo RcisTarget enhances    TRUE
## 411700   zoo        tis enhances    TRUE

dplyr::count(df1, type, reverse)

##          type reverse     n
## 1     depends   FALSE 11392
## 2     depends    TRUE  9900
## 3    enhances   FALSE   513
## 4    enhances    TRUE   533
## 5     imports   FALSE 72582
## 6     imports    TRUE 66248
## 7  linking to   FALSE  4319
## 8  linking to    TRUE  4647
## 9    suggests   FALSE 45738
## 10   suggests    TRUE 50728

length(unique(df1$from)) # number of CRAN packages

## [1] 16708

nrow(df1) # number of dependencies

## [1] 266600

Growth of dependencies

• Total number of forward & reverse dependencies

Network of dependencies

(g0 <- get_graph_all_packages(type = "depends"))

## IGRAPH 625b0ca DN-- 4940 8281 -- 
## + attr: name (v/c)
## + edges from 625b0ca (vertex names):
##  [1] A3         ->xtable     A3         ->pbapply    abc        ->abc.data  
##  [4] abc        ->nnet       abc        ->quantreg   abc        ->MASS      
##  [7] abc        ->locfit     abcdeFBA   ->Rglpk      abcdeFBA   ->rgl       
## [10] abcdeFBA   ->corrplot   abcdeFBA   ->lattice    ABCp2      ->MASS      
## [13] abctools   ->abc        abctools   ->abind      abctools   ->plyr      
## [16] abctools   ->Hmisc      abd        ->nlme       abd        ->lattice   
## [19] abd        ->mosaic     abn        ->nnet       abn        ->lme4      
## [22] abodOutlier->cluster    Ac3net     ->data.table acc        ->mhsmm     
## + ... omitted several edges

igraph::is.dag(g0) # whether it is a directed acyclic graph

## [1] TRUE

visNetwork plot

Degree distribution - power law?

Future work

• Modelling the dynamic growth of network
• Grouping the strong dependencies
  • Depends + Imports (+ LinkingTo)
• Does the preferential attachment model fit the data?
  • It leads to the power law degree distribution