[ESS] Best Practice for building R packages

Jeremie Juste jerem|eju@te @end|ng |rom gm@||@com
Thu Oct 8 15:14:42 CEST 2020


Hello Kasper,

Thanks for sharing.

> Personally, I'm a dinosaur so I'm still editing Rd files myself. I must
> say, I'm not convinced about the merits of roxygen2.
I admire your discipline. For my part, I feel that I would not document my
functions properly unless I constrain myself to do it on the fly. For
that I have created a [1] snippet (from [2] yassnippet package). I admit
that here is still room for improvement.

[2] https://github.com/joaotavora/yasnippet

[1] -- function -- 
# -*- mode: snippet -*-
# name: function
# key: fn
# --

##' .. content for \description{$3} (no empty lines) ..
##'
##' .. content for \details{} ..
##' @title $1 
##' @param $2
##' @return 
##' @author Jeremie Juste
$1 <- function($2){
    $0
}

So typing fn[tab] would expand the snippet and force me to write some
documentation.

I used to naively include the documentation in the function itself in
comments but that is not the R way of sharing code.

build..table <- function(obj){
### DD
    ## obj of class lm
### purpose
    ## extract coef name, coef, std and pval from estimation of lm
    
    res <- data.table(summary(obj)$coefficients[,c(1,2,4)],keep.rownames=TRUE)
    setnames(res,c("name","coef","std","pval"))
    res
}


Best regards,
-- 
Jeremie Juste



More information about the ESS-help mailing list