Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

No matlab #64

Merged
merged 36 commits into from
May 28, 2021
Merged

No matlab #64

merged 36 commits into from
May 28, 2021

Conversation

akhanf
Copy link
Member

@akhanf akhanf commented May 19, 2021

@jordandekraker making this draft PR so I can bring it up to date with master, and translate the warps/gifti code too. I've integrated everything else so far and tested only with dry run, will try running for real soon..

jordandekraker and others added 30 commits March 29, 2021 14:35
…e directory temporarily, but will integrate later.
Had a couple free hours today so I integrated the new code into the
snakemake workflow. Dry-run with T2w works at this point, but just has a
placeholder rule to generate the warps/gifti files, as I still need to
translat that matlab code. I'll give this a whirl and see what happens,
at the least will give me some files so I can write the warps/gifti
generation code outside of matlab
new shape injection and laplace. components tested, and workflow dry
run, but running a full end to end test now..

TODO:
 - finish reorganizing some files (e.g. warps are still in
autotop dir)
 - end-to-end test
 - add greedy to container
midsurf vis had to be disabled for now, since without extrapolate we have nans in
the surf vertices, and so nilearn plotting fails
- adds new labels for PD sink based on dilating DG into SRLM and BG,
  fixes problem where GM domain for PD wqs different from other coords
  resulting in warp ambiguity
- adds new template laplace coords, from running with 1e-8 stopping
  criterion, above fix, and multires init (included snakefile for this)
- incorporate shape inject init of laplace and the new PD sink heuristic
  in main workflow, cuts run time to less than 15 min total (around a
  minute for AP laplace)
@akhanf akhanf marked this pull request as ready for review May 28, 2021 01:03
@akhanf
Copy link
Member Author

akhanf commented May 28, 2021

This should be ready to merge now:

  • replaces all matlab/mcr with python
  • shape injection uses greedy (much faster)
  • laplace coords initialized with template
  • default behaviour will tar up each subject's work folder after completion (--keep_work to prevent)

Some remaining things to fix in a later PR:

  • midsurf visualization disabled right now, failure could be due to fov of surfaces (?) need to investigate
  • documentation / install instructions could be updated

@akhanf akhanf merged commit da316ca into master May 28, 2021
@akhanf akhanf deleted the nomatlab branch May 28, 2021 01:25
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants