search

seung-lab / kimimaro

Skeletonize densely labeled 3D image segmentations with TEASAR. (Medial Axis Transform)
GNU General Public License v3.0
135 stars 23 forks source link

Verbose flag #77

Closed 0xJustin closed 1 year ago

0xJustin commented 1 year ago

Hi - I am somewhat of a power user of both this and meshparty skeletonize. Is there any chance you could add a verbose flag to control whether or not the for loop here uses tqdm: https://github.com/seung-lab/kimimaro/blob/master/kimimaro/intake.py#L377 Thanks :)

william-silversmith commented 1 year ago

Hi Justin,

Thanks for writing in! Which interface are you using, Python or CLI? It seems the tqdm instance is already instrumented with a progress flag. Do you mean you want to disable that progress bar specifically?

I tried running kimimaro forge --progress -p 4 connectomics.npy --no-fill-holes and there seemed to be only one progress bar. To disable progress, just don't add --progress. In the python interface, you can set:

kimimaro.skeletonize(..., progress=False)

Thanks! Will

william-silversmith commented 1 year ago

By the way, in case you weren't aware, I believe there are a set of mesh to skeleton algorithms including the meshparty algorithm included in https://github.com/navis-org/skeletor though I haven't used it.

0xJustin commented 1 year ago

Wow I'm a complete goofball I can't believe I missed that flag! Thanks for your nice response- also thanks for that link! I've been doing quite a bit of mesh skeletonization as part of a pipeline for finding EM error locations using the mesh, and the skeletonization part of the approach is certainly the computational bottleneck. I'm keen to see if any of the other algorithms have potential for speedup