Populate scan_type in live db

[rfdougherty]

There is a scan_type field on the Epoch table that is currently not used for a real scan type (it is currently populated with the psd name). The nimsdata refactor will populate this with one of the following values: shim, localizer, calibration, fieldmap, spectroscopy, diffusion, functional, 3D anatomy, 2D anatomy, or (as a last resort) anatomy. To use scan_type in search, we need a script to loop over all epochs, open the primary dataset, and update the scan_type field.

[gsfr]

That's a nice list. A couple of suggestions/questions:

• functional -> function? (makes every option a noun, but may be too weird)
• To make searching for any type of anatomy efficient by using a database index, the 2D and 3D qualifiers should be in the back, as in anatomy-2D.
• Are 2D and 3D anatomies really different enough to warrant 3 top-level types?
• I also remember hearing perfusion. Should we add that to the list?

[rfdougherty]

• I guess function is ok, but does sound a little odd to my ear. Not sure if the part of speech uniformity trumps common parlance here.
• Yea, I'm ambivalent on this. I considered creating a sub-type for all of them, but only putting something meaningful in the anatomy type. But decided that was stupid. The 2D/3D distinction is useful, as it can generally (though not perfectly) allow someone to find all the high-res T1's (i.e., anatomy-3d will get those with >95% accuracy). Let's go with anatomy-2d, anatomy-3d, anatomy for now. In Nims 2, the 2d/3d thing mich become part of the subtype.
• I was lazy, and just did the low-hanging fruit. But yea, perfusion should be easy. I'll find some and see what heuristics I can craft.

----- Original Message -----

From: "Gunnar Schaefer" notifications@github.com To: "cni/nims" nims@noreply.github.com Cc: "Bob Dougherty" bobd@stanford.edu Sent: Saturday, August 17, 2013 9:17:07 PM Subject: Re: [nims] Populate scan_type in live db (#45)

That's a nice list. A couple of suggestions/questions:

* functional -> function? (makes every option a noun, but may be too weird) 
* To make searching for any type of anatomy efficient by using a database index, the 2D and 3D qualifiers should be in the back, as in anatomy-2D. 
* Are 2D and 3D anatomies really different enough to warrant 3 top-level types? 
* I also remember hearing perfusion. Should we add that to the list? 

— Reply to this email directly or view it on GitHub .

[sbenito]

Now working on beta version. Thanks Robert and Gunnar!