search

myousif9 / hippograd

GNU General Public License v3.0
0 stars 0 forks source link

reference gradient choices #3

Open jordandekraker opened 2 years ago

jordandekraker commented 2 years ago

https://github.com/myousif9/hippograd/blob/cd29d0777ac0d563b3ac54cd249c0ba6130dcfd0/hippograd/workflow/rules/post_hippunfold.smk#L21

use reference gradients from here: https://github.com/royhaast/smk-hcp-lifespan/blob/JDupdates/resources/AvgGradient_HCP-UR100_hemi-L.npy (could have this be an optional flag too though - eg. make new reference or manually specify a reference)

jordandekraker commented 2 years ago

you'll also need to resample the reference gradient wit this:

https://github.com/jordandekraker/hippunfold_toolbox/blob/caa60bec5ec28073acd3f70b299e882826911f97/Python/utils.py#L80

myousif9 commented 2 years ago

I've added the reference gradient to the latest commit and omitted the rule to use a dataset average gradient as the reference. I have yet to add in the ability to resample the reference gradient to different hippocampal surface densities since the density_interp function appears to require an argument called resource_dir to grab the starting and target unfolded surfaces. Would you be able to provide me with those surfaces and I could add them to the resources directory for this pipeline? Thanks!

jordandekraker commented 2 years ago

Ahh right, if you do this it will add hippunfold as a dependency, unfortunately. Its here: https://github.com/khanlab/hippunfold/tree/master/hippunfold/resources