This PR begins the process of establishing CSVs for Zenodo upload by updating the tables/ directory to accommodate forthcoming changes, as follows:
run-tables.sh is now the name of the main script that runs the tables module.
output_tables.Rmd is renamed to write-manuscript-tables.Rmd, and tables are outputted to slightly different paths. Rather than sending all output to results/, output tables for main and supp tables respectively, go to manuscript-tables/ and manuscript-tables-supp/. results/ is now gone.
Set up a directory zenodo-upload/ to hold the forthcoming CSV files. It has a .gitkeep for now which will get removed once it is populated.
I created util/zenodo-utils.R to facilitate writing CSV files. We will always need to order by a given sample column and then write out, so this function can be used throughout OpenPBTA where needed.
README.md updates accordingly
Circle CI updates accordingly, including running the bash script run-tables.sh in CI, not only the notebook.
What to expect when you're expecting more PRs
Subsequent PRs will export figure data in one of two ways:
For figures created in figures/scripts/, CSVs will be written directly to tables/zenodo-upload/. Individual scripts in figures/scripts/ will get modified accordingly.
For figures creates in analysis/ modules, the CSVs will first be written out within those analysis modules (as described in #1692). Then, we will need to copy those CSVs into zenodo-upload/. A script copy-zenodo-tables.sh, which is already planned for in README.md, will be added here to copy those specific files (a lot like copying figure panels in figures/generate-figures.sh).
Note, we will not want to run copy-zenodo-tables.sh in CI, so it will eventually get added to run-tables.sh inside some appropriate logic
Goals for reviewing this PR
Now is the time for feedback on this overall organizational framework before I dive into writing CSVs!
@jashapiro I'm pretty sure I fixed CI here. Happily there was a data cache so these fresh checks should be done soon-ish!
Function has been purged as well.
Towards #1692
This PR begins the process of establishing CSVs for Zenodo upload by updating the
tables/
directory to accommodate forthcoming changes, as follows:run-tables.sh
is now the name of the main script that runs the tables module.output_tables.Rmd
is renamed towrite-manuscript-tables.Rmd
, and tables are outputted to slightly different paths. Rather than sending all output toresults/
, output tables for main and supp tables respectively, go tomanuscript-tables/
andmanuscript-tables-supp/
.results/
is now gone.zenodo-upload/
to hold the forthcoming CSV files. It has a.gitkeep
for now which will get removed once it is populated.util/zenodo-utils.R
to facilitate writing CSV files. We will always need to order by a given sample column and then write out, so this function can be used throughout OpenPBTA where needed.README.md
updates accordinglyrun-tables.sh
in CI, not only the notebook.What to expect when you're expecting more PRs
Subsequent PRs will export figure data in one of two ways:
figures/scripts/
, CSVs will be written directly totables/zenodo-upload/
. Individual scripts infigures/scripts/
will get modified accordingly.analysis/
modules, the CSVs will first be written out within those analysis modules (as described in #1692). Then, we will need to copy those CSVs intozenodo-upload/
. A scriptcopy-zenodo-tables.sh
, which is already planned for inREADME.md
, will be added here to copy those specific files (a lot like copying figure panels infigures/generate-figures.sh
).copy-zenodo-tables.sh
in CI, so it will eventually get added torun-tables.sh
inside some appropriate logicGoals for reviewing this PR
Now is the time for feedback on this overall organizational framework before I dive into writing CSVs!