log_info() in generate_signals() was the final statement, so the logging was the return value, instead of the dataset itself. I didn't think I could find anywhere else where that error existed.
I think test being unset should always be TRUE. Runs should only be done intentionally. Are we sure we don't want test to be internal to the function, so we just use the env variable and remove the duplicated calls in each update_{indicator}.R script?
I have changed it so we only throw errors on warnings when not interactive. This should allow for a lot more flexibility when interactively developing. What do you think?
Some other small changes I notice when doing some runs.
Fixes to some issues found in the code.
log_info()
ingenerate_signals()
was the final statement, so the logging was the return value, instead of the dataset itself. I didn't think I could find anywhere else where that error existed.test
being unset should always beTRUE
. Runs should only be done intentionally. Are we sure we don't wanttest
to be internal to the function, so we just use the env variable and remove the duplicated calls in eachupdate_{indicator}.R
script?