search

raamana / graynet

Subject-wise networks from structural MRI, both vertex- and voxel-wise features (thickness, GM density, curvature, gyrification)
https://raamana.github.io/graynet/
MIT License
35 stars 7 forks source link

Is current volumetric build functional? #38

Closed jamesruffle closed 4 years ago

jamesruffle commented 4 years ago

Hello,

I am interested in using your toolkit for gray matter volumetric connectivity using a custom ROI.

I have just downloaded the package and tried it on the example data provided, using the example code provided, however it returns an error. Is the current build functional for volumetric analyses? Output as follows below.

Thanks for your advice! James

user:~/Desktop/graynet-master/example_data$ graynet -s volumetric_CAT12/sub_id_list.txt -f spm_cat_gmdensity -i volumetric_CAT12/ -w manhattan -a cat_aal -o volumetric_CAT12/graynet_volumetric/

Data resampled to cat_aal atlas, smoothed at 10 with node size None Computing single edge ...

Processing spm_cat_gmdensity features Reading spm_cat_gmdensity for CAM_0002_01 ... Done.

Processing id CAM_0002_01 -- weight manhattan (1/1) :Reading spm_cat_gmdensity for UBC_0018_01 ... Done.

Processing id UBC_0018_01 -- weight manhattan (1/1) :........................ Unable to extract manhattan features for UBC_0018_01 Traceback (most recent call last): File "/usr/local/lib/python3.6/dist-packages/graynet/volumetric.py", line 107, in extract_per_subject_volumetric graph.node[roi]['x'] = float(centroids[roi][0]) AttributeError: 'Graph' object has no attribute 'node' ........................ Unable to extract manhattan features for CAM_0002_01 Traceback (most recent call last): File "/usr/local/lib/python3.6/dist-packages/graynet/volumetric.py", line 107, in extract_per_subject_volumetric graph.node[roi]['x'] = float(centroids[roi][0]) AttributeError: 'Graph' object has no attribute 'node' Done.Reading spm_cat_gmdensity for TGH_0070_01 ... Done.

Processing id TGH_0070_01 -- weight manhattan (1/1) :........................ Unable to extract manhattan features for TGH_0070_01 Traceback (most recent call last): File "/usr/local/lib/python3.6/dist-packages/graynet/volumetric.py", line 107, in extract_per_subject_volumetric graph.node[roi]['x'] = float(centroids[roi][0]) AttributeError: 'Graph' object has no attribute 'node' Done. graynet computation done.

raamana commented 4 years ago

Thanks James for the bug report. The CI tests seem to have passed, let me try again, and see what the problem is.

raamana commented 4 years ago

the tests pass locally for me. Can you try updating your networkx package (and others) and try again with the following commands:

pip uninstall networkx graynet
pip install -U networkx
pip install -U graynet
raamana commented 4 years ago

Do NOT download the package from this repo btw. So delete any directly downloaded files as they may have not been ready for release. It must be installed via pip as noted in the documentation.

raamana commented 4 years ago

So there seems to have been a problem with API for latest networkx v2.4.. Fixed it and released a new graynet v0.4.8. So if you try this, your bug shouldn't exist anymore with graynet v0.4.8:

pip uninstall graynet
pip install -U graynet
jamesruffle commented 4 years ago

Hi Raamana,

This seems to be functioning now as intended. I'm impressed how speedy it is... 500 subjects with an ROI atlas of ~350 regions in 20 mins. Great work!

raamana commented 4 years ago

Very glad to hear that James :)

Let me know if you have any issues in using the results - read this first: https://raamana.github.io/graynet/results_howto.html

You seem to be processing Canadian/OBI datasets, are you? Also, If you can, would you mind sharing the focus of your project?