NYU-DiffusionMRI / DESIGNER-v2

Diffusion parameter EStImation with Gibbs and NoisE Removal pipeline Version 2
https://nyu-diffusionmri.github.io/DESIGNER-v2/
Other
16 stars 2 forks source link

Issue with b1correct #12

Closed RP-HAO closed 1 month ago

RP-HAO commented 1 month ago

Dear DESIGNER team,

Thank you very much for developing and maintaining this extremely useful processing pipeline.

My first (and main) question is about using option -b1correct for bias field correction. While running via docker I kept getting an error message as attached below:

Screen Shot 2024-07-26 at 22 54 48

I tried both on a Mac and a Windows and got the same error. May I ask if there's something wrong/missing with my installation, or could it be something else?

Another minor point is regarding printing the messages: when I try to input several commands (e.g., picture below used degibbs and _ants_motioncorrection), the outputs from run.command can always be printed in real time, whereas outputs from print function get delayed, with an example being:

Screen Shot 2024-07-26 at 23 02 48

While this doesn't affect any underlying image processings, having all output messages flushed in real time could be pretty informative when the pipeline takes very long to run. Would this be possible?

Best, Runpu

jchen33344 commented 1 month ago

Hello Runpu,

Thank you for your feedback. Since b1correct is not part of the default DESIGNER pipeline, we did not include ANTs as a dependency in docker. However, you are free to use the dev tools on Docker to add ANTs if you are comfortable.

I am not sure I am understanding the suggested improvement. When I run it on docker, the output from print seems to be in real time. Screenshot 2024-07-30 at 3 47 02 PM We also welcome you to send a pull request to suggest the change and we can test it.

Best, Jenny

RP-HAO commented 1 month ago

Hi Jenny,

Thanks for the info on b1correct!

My problem with not having output from print in real time seems to be related to the output buffering in python. I was able to solve it by running docker command with an additional environment variable -e PYTHONUNBUFFERED=1. :) Thanks again!

Best, Runpu