Closed dtm2451 closed 1 year ago
Just curious, have you tried not filling it with a mock path? I usually just assign no directory if I'm not doing scaffold analysis, eg: gated_fcs_dir: ""
But in general I think the PR allows for more flexibility depending on what inputs users might give!
I haven't tried that, but I suspect it would bug out the same way nowadays cuz you'd still end up with FALSE at the
if( dir.exists(gated_fcs_dir) )
logic point, and thus skip past where make_scaffold_map
is changed from NULL.
> dir.exists("")
[1] FALSE
Checking the string version is needed... I hadn't tested at the time of the first commit or else this would have all been in one, =p, because the first commit didn't work. exists()
takes in a string and checks for the existence of an object with that name.
Is there a specific reason you like exists(gated_fcss)
over exists("gated_fcss")
?
Oh, sorry I confused exists() with dir.exists() that I use above! :| P.S. sorry about closing and opening the PR. I am just learning this aspect of git. :D
haha, no wories, no harm done!
I think a recent update introduced a bug where the config file's 'gated_fcs_dir' must be a real directory or else the pipeline errors out in step 6.
The bug:
A cyclone run where the config file 'gated_fcs_dir' is not a real directory (because I don't have gated fcs files to power SCAFFoLD analysis; derp value I filled it in with was "/boop/boop/boop") fails after completing step 5 with this at the end of the log file.
Fix here:
This PR adjusts the logic around setting the 'make_scaffold_map' logical value in order to ensure it still gets set.