[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Countrs of r-$FOO-packages, and Debian + Ubuntu package counts



I just got curious about how many R packages we would now have under the
different categories, and knowing that I could answer that via my RcppAPT
connector to the apt cache, I quickly run that.

On Debian testing, it is now almost a thousand packages. We can run the few
lines (in R, after installing libapt-pkg-dev and then R packages RcppAPT and
data.table)

  library(data.table)  # to filter and aggregate
  revs <- setDT(RcppAPT::reverseDepends("r-base-core"))
  revs[grepl("^r-([a-z]*)-*", package), c("r","where","name","other") := tstrsplit(package, "-")][
      "r"==r,.N,by=where][
      order(-N)]

we get (keeping the last compound data.table statement on one line)

  > revs[grepl("^r-([a-z]*)-*", package), c("r","where","name","other") := tstrsplit(package, "-")]["r"==r,.N,by=where][order(-N)]
           where   N
  1:        cran 974
  2:        bioc 134
  3:       other  10
  4:        base   4
  5:         doc   3
  6:    omegahat   1
  7: recommended   1
  8:     mathlib   1
  > 

where 'base', 'doc', 'recommended' and 'mathlib' all stem from my r-base
source package and can be ignored.  So for purposes of the INLA discussion we
have ten 'other' packages.

While this is a Debian list, it is still worth recalling that the Ubuntu PPA
effort by Michael Rutter is something to reckon with more than five times the
packages:

  > revs[grepl("^r-([a-z]*)-*", package), c("r","where","name","other") :=
         tstrsplit(package, "-")]["r"=  =r,.N,by=where][order(-N)]
           where    N
  1:        cran 5305
  2:        bioc  150
  3:        base   12
  4:         doc    9
  5:       other    8
  6: recommended    3
  7:     mathlib    3
  8:    omegahat    2
  > 

But before we jump up and down in joy, for Fedora and SUSE two volunteers
have assembled 15k+ binaries each covering essentially all of CRAN.  We need
a sponsor for more CPU power because running the build scripts is not that
hard given how well curated CRAN is.

Dirk

-- 
https://dirk.eddelbuettel.com | @eddelbuettel | edd@debian.org


Reply to: