[Bioc-devel] Error when building in command line but not in RStudio

[AJ Kinstlick]

Hi all,
I'm trying to upgrade a package that uses R and python together with temp
files generated by the system. I encounter an error when building via the
command line, but can build the package with no errors using RStudio.


*Do I need to worry about any command line errors if the package can be
built correctly using RStudio?*

More detail about my error:
I'm currently updating the package 'seq2pathway'.
Saving the 'seq2pathway' directory as an R package and building it in
RStudio gives me a couple warnings but no errors. However, building it in
console gives this error:

> R CMD build seq2pathway

* checking for file ‘seq2pathway/DESCRIPTION’ ... OK

* preparing ‘seq2pathway’:

* checking DESCRIPTION meta-information ... OK

* checking whether ‘INDEX’ is up-to-date ... OK

* installing the package to build vignettes

* creating vignettes ... ERROR

--- re-building ‘seq2pathwaypackage.Rnw’ using Sweave


Warning: replacing previous import ‘WGCNA::cor’ by ‘stats::cor’ when
loading ‘seq2pathway’

Warning in system2(mypython, args = script_path, stdout = TRUE, stderr =
TRUE) :

  running command
''/Library/Frameworks/Python.framework/Versions/3.8/bin/python3'
/var/folders/9r/bgcsznps6sz6kcn851gqf47m0000gp/T//Rtmp4ADNbj/inputfile_20210323_13_33_30_seq2gene_log.py
2>&1' had status 1


Error: processing vignette 'seq2pathwaypackage.Rnw' failed with diagnostics:

 chunk 7 (label = runseq2gene)

Error in read.table(file = tmpoutfile_UTR3, header = TRUE, sep = "\t") :

  no lines available in input


--- failed re-building ‘seq2pathwaypackage.Rnw’


SUMMARY: processing the following file failed:

  ‘seq2pathwaypackage.Rnw’


Error: Vignette re-building failed.

Execution halted


I've encountered and debugged this error before, and it usually means that
python encountered an error when trying to read the input file from the
temp folder, and exits before writing anything to the output temp file.
However, I have tested the same R function with the same command outside of
building the package, and am sure that it can be run without error if I
directly source the .R file containing the function.
Any help on this would be greatly appreciated, but once again I'm not sure
it's even necessary if I can just build the package in RStudio and submit
it to Bioconductor.

Thank you,
AJ Kinstlick

	[[alternative HTML version deleted]]