Add generation of scene/snapshots for fMRIQC

[mharms]

Still need to revise and add the scene template itself, but wanted to get the code up for initial review.

Of note, the bias corrected, normalized, but not brain masked version of the scout is named with a "_notmasked" suffix currently. "unmasked" was a bit too cryptic in my opinion, although open to other suggestions.

[coalsont]

_nomask could be a shorter but equivalent specifier, I like shorter when it doesn't sacrifice clarity. For brainstorming, "with background" is a different way to think about it, but _withbg doesn't look as clear, and background is fairly long to put in a filename.

I think a few files use _brain for the masked version, and while _nobrain might be amusing, it is unlikely to be helpful.

[mharms]

_nomask seems like a good alternative.

[glasserm]

The code looks good to me and thanks for making a no mask SBRef. I would like to look over some test results prior to approving.

[mharms]

With the above additions, this should now be operational. I'm running another test overnight (denovo execution of fMRIVolume and fMRISurface) and can generate a -zip-scene-file for sharing tomorrow.

[mharms]

The code as it exists in this PR works. I'll email you the location of the scene file.

[mharms]

@glasserm : It would be great if you can review the scene file. Would like to get this PR into the master, tag a new 4.3rc, and then have QuNex folks build a new container using that.

[glasserm]

@glasserm : It would be great if you can review the scene file. Would like to get this PR into the master, tag a new 4.3rc, and then have QuNex folks build a new container using that.

I looked but did not find an e-mail from you showing the location of the scene file.

[mharms]

@glasserm I sent the email again.

[mharms]

@coalsont: Here's the git diff on the approach that @hodgem took for the set -e issue:

 main()
 {
+
        log_Msg "Starting main functionality"
+       set -e
 
        # Retrieve positional parameters
        local StudyFolder="${1}"
@@ -1045,7 +1047,9 @@ main()
         fi
     done
 
+       set +e
        log_Msg "Completing main functionality"
+
 }

He confirmed that this does indeed catch the "error" when wb_command got killed due to insufficient memory. Although your coupling its placement to the debug.shlib line seems cleaner, if it doesn't need to be inside main()

[mharms]

Just for public documentation purposes, the set -e issue mentioned above it completely unrelated to this PR, and relates to something we need to tweak in DeDriftAndResample

[coalsont]

The set -e should be done immediately before sourcing debug.shlib, so that it can disable it for interactive shells, substantially reducing a considerable annoyance when doing bash debugging (unexpectedly losing the entire shell, with all carefully set up script arguments, when set -e trips).