Closed colinpmillar closed 1 year ago
from @arnima-github: I've created a basic pathway to run a custom bootstrap data script. See current version of icesTAF on GitHub (master branch is currently same as develop). The corresponding test analysis is at https://github.com/ices-taf-dev/bootscript.
some potential ways forward:
source(local=TRUE)
, sys.source(envir=new.env())
, local(source())
, etcsource = {script}
, forcing script to be the same name as the data output. All output could be put in a folder of the same name as the script possiblyWe could then screen the bootstrap/data folder for files or folders that are not reference in the DATA.bib
file and warn or delete them
Based on discussions it seems a combo of 1. and 2. from above would be easiest. For example, the given the following record:
@Misc{datras,
originator = {ICES},
year = {2019},
title = {Survey indices: IBTS1, IBTS3, Unknown, and UKBTS3},
period = {1977-2019},
source = {script},
}
The rules would be:
datras.R
.bootstrap/data/datras/
folder.This should ensure that all files and folders in the bootstrap/data/
folder have a name corresponding to a DATA.bib
entry while still being relatively easy to implement in icesTAF::process.bib
.
brainstorming:
process.bib
will set the working directory accordinglyinstruction to user: "please do not specify target directory"
Summary
this would work like:
the
name
allows us to trace the directory or file that the record is referencing.Tasks required
Links to other issues?