porespy
porespy copied to clipboard
A set of tools for characterizing and analying 3D images of porous materials
This is an api breaking change, so no rush.
My recent descent into dependency hell has me thinking about making truly 'archived' installations. I'm opening this issue to capture the ideas and options.
Currently, voxel size is passed as an argument to SNOW, meaning that once the network is extracted, it's not possible to easily rescale the output based on a different voxel...
Similar to PR #463, but for unit tests. Not urgent though, as our unit tests are still manageable.
We currently use the marching_cubes method to get more accurate surface area values. It got me thinking that throat perimeter could/should be done more accurately too. * Find the 3D...
https://github.com/PMEAL/porespy/runs/4590324027?check_suite_focus=true
I removed Dask progress bars for now (in PR #414). A more permanent solution would be to wait for Dask to fix the progress bar upstream. See dask/distributed#3414.
The distance transform is useful in network extraction algs and therefore often needs to be computed anyway. Why not pass dt into _estimate_overlap instead of calculating it twice? I propose...
To reproduce: ```python import porespy as ps im = ps.generators.blobs([100, 100, 100], blobiness=0.75, divs=2) ```
Currently, it shows the log generated by `pip install tensorflow`, etc. Should be turned off.