On Mon, 15 Jun 2020, Aditya Mahajan wrote:

> On Mon, 15 Jun 2020, Aditya Mahajan wrote:
>
>> On Sun, 14 Jun 2020, Fabrice L wrote:
>> 
>>> Hi,
>>> 
>>> Following a question about the broken R module (R the statistical 
>>> software), Aditya suggested me the filter module (thanks to him). I’m 
>>> about to distribute a set of course notes to colleagues, and I need to 
>>> deal with details now. I have one problem and two questions.
>>> 
>>> 1) the filter collect R code between \startR / \stopR and submit this code 
>>> to R. As it can be seen in the following minimal example, when a label 
>>> contains an accented character (« Fréquence » in my example), the pdf 
>>> graphic does not contains the « é ». Strangely, the snippet of code (which 
>>> is saved locally as « test2-temp-R-0.tmp » (for test2.tex)) when submitted 
>>> with the same command as the filtercommand, works correctly; that means 
>>> the same file works when submitted to R outside of ConTeXt. I have no idea 
>>> how to solve this.
>> 
>> Are you running the exact same command as the `filtercommand`. If so, I 
>> don't know why running the `filtercommand` through context vs directly 
>> typing it on the terminal should behave differently. The filter module 
>> effectively just runs os.execute("filtercommand"). I don't use R so I am 
>> unable to debug further.
>
> I can reproduce this output side the filter module. If I directly run 
> os.execute(...) from a lua session, the output is correct but running it 
> through context gives the wrong output. I am guessing this is something to do 
> with locale. I notice that the output of os.execute("locale") from context is 
> different from that from my shell. But I don't know why locale should affect 
> UTF characters in R.

It is indeed related to locale. Here is a minimal example that works. I use a beforebuffer to set the locale so that you don't have to do it in each snippet and then filter out the first line in the typeset code listing and filter out the first two lines in the output.

\usemodule[filter]
\usemodule[vim]

\definecolor  [darkblue]   [r=0.1,  g=0.11, b=0.84]
\definecolor  [darkgreen]  [g=0.7]
\definecolor  [lightblue]   [h=87cefa]
\definecolor  [lightgreen]  [g=0.95,b=0.8,r=0.8]

\defineframedtext
   [leftbartext]
   [
     width=local,
     frame=off,
     framecolor=darkblue,
     leftframe=on,
     rulethickness=1ex,
     offset=0.25ex,
     loffset=3ex,
     background=color,
     backgroundcolor=lightblue,
   ]

\definevimtyping
   [RCODE]
   [
     vimcommand=nvim,
     syntax=r,
     directory=output,
     strip=yes,
     escape=yes,
     before={\blank[quarterline]\startleftbartext},
     after={\stopleftbartext},
   ]

\define[1]\ShowOutput
     {\typeRCODEfile[start=2]{\externalfilterinputfile}
      \startleftbartext[framecolor=darkgreen, backgroundcolor=lightgreen]
        \setupalign[flushleft, verytolerant]
        \typefile[range={3,-1}]{#1}
      \stopleftbartext}

\startbuffer[locale]
    Sys.setlocale(category="LC_ALL", locale="en_US.UTF-8")
\stopbuffer

\defineexternalfilter
   [R]
   [ filtercommand={R CMD BATCH -q --no-timing --save --restore
\externalfilterinputfile\space \externalfilteroutputfile},
      output=\externalfilterbasefile.out,
      directory=output,
      readcommand=\ShowOutput,
      read=no,
      bufferbefore=locale,
      cache=yes,
      purge=yes,
    ]



\starttext

Some text...

\startR[read=yes]
    pdf("RPlots/MyHistogram.pdf",5,5)
    X <- rnorm(200,mean=10,sd=2)
    hist(X, col =  "red3" ,  xlab="Score QI" , main="", ylab="fréquence")
\stopR

\externalfigure[RPlots/MyHistogram.pdf][width=.5\textwidth]

\stoptext

Aditya