brain-life / workflow_ui Goto Github PK

I need to put instance status maker back to let user know which one is running, finished, failed, etc..

At the moment, there is no instance status that aggregates all tasks running inside the instance which makes this a bit difficult. I believe I need to do following.

1. When a task status changes (not when task gets updated), update instance status[] field.
2. Above implies adding task status[] under instance (again..)
3. Emit event when instance record gets updated
4. Update UI to subscribe and receive the instance update event
5. Update UI to calculate the current instance status based on task status.

When job is running, allow users to stop all tasks currently running.

The content on this page is wrong / outdated.

http://www.brainlife.io/tracking

Right now, the UI doesn't normalize / single-shell input data. Sample data I am using are already normalized / single shell-ed (2000) so it works, but user may upload data that's not normalized.

I could do either...

1. [easy] check to see if the bvals are normalized / single shell-ed at the validation stage.
2. [hard] add normalize / single-shell (at user configured b-value) for all workflows.

Problem is though.. I might also need to flip bvecs .. so I could also do..

3. [harder] add Lindsey's bvec analyzer and run normalize/bvec flip/single-shell for all workflows.

Uploading file requires resource access. Don't let user submit any job if there is no resource available to upload - or let user use aux resource?

This processing portal is amazing!! Thank you so much for this effort. Is it possible to transfer processed data between modules. For example, if FREESURFER is run on a data set in the LIFE module, it would be helpful to be able to transfer those outputs to the connectome module. Rather than having to repeat the process for both.

Hi!
congratulations for this great project!

I was testing some of the pipelines, such as LiFE, Streamline Tracking, O3D and White Matter Tracking, and I realized that after several days all of them are stuck in some part of the process. Well, 3 are stuck in the FS phase and one is in the tracking step.

All of the processes were fed with the same data I downloaded from your HCP 3T database
101915_dwi.bvec 101915_dwi_corrected.bval 101915_dwi.nii.gz t1.nii.gz

What should I do know? Are there any steps I should take to check, reboot, evaluate the situation? Why do you think this happened?

Thanks again for this project and best regards,
Gari

soichih/sca-service-conneval-validate currently runs on Karst, so if Karst is down user can't upload data. Also, sample data ..

           files: {
                t1: "/N/u/hayashis/Karst/testdata/demo_data_encode/STN/sub-FP/anatomy/t1.nii.gz",
                dwi: "/N/u/hayashis/Karst/testdata/demo_data_encode/STN/sub-FP/dwi/run01_fliprot_aligned_trilin.nii.gz",
                bvecs: "/N/u/hayashis/Karst/testdata/demo_data_encode/STN/sub-FP/dwi/run01_fliprot_aligned_trilin.bvecs",
                bvals: "/N/u/hayashis/Karst/testdata/demo_data_encode/STN/sub-FP/dwi/run01_fliprot_aligned_trilin.bvals",
            }

are currently stored in IU HPC and default resource set to use Karst to load this.

I should move the sample data to brain-life/dev1 and make validator run on brain-life server.

Currently, http://www.brain-life.org/tracking is using "brain-life/app-tracking" which just run mrtrix 3 times to yield SD_STREAM, SD_PROB, and DT_STREAM output. This is fine for all other apps, but for "Ensemble Tracking" app, we should use "brain-life/app-ensembletracking".

Further detials about the required file formats for the DWI, T1, bvec and bval files would be helpful. In addition, to clarify on the errors for why files fail to load. Also, individual validation checks for each file would help not have to load all 4 each time.