A set of Python scripts used to post-process CT datasets acquired at the 832 beamline at the Advanced Light Source (ALS).
Install using conda, create and activate an environment:
conda create -n tomoTools
conda activate tomoTools
Use conda to install mamba:
conda install -c conda-forge mamba python=3.7
Use mamba to install dependencies (faster than conda, but still takes some time):
mamba install -c conda-forge -c simpleitk -c anaconda tifffile numpy scipy matplotlib jupyter numexpr astropy h5py simpleitk tomopy jedi=0.17
Install PyTorch, make sure that a CUDA version is installed (check the official website for more instructions):
mamba install -c pytorch pytorch cudatoolkit=11.3
mamba install -c astra-toolbox/label/dev astra-toolbox
Then clone the scripts, and run them in a local notebook:
git clone https://github.com/eboigne/tomoTools_als832.git
jupyter-notebook
For graph-analysis of skeleton using package sknw from here:
mamba install -c conda-forge -c anaconda networkx numba
If the kernel dies every time the Astra algorithm is run, try uninstalling and re-installing Astra:
mamba remove astra-toolbox
mamba install -c astra-toolbox/label/dev astra-toolbox
AttributeError: module 'numpy.ctypeslib' has no attribute '_typecodes'
Appeared from Tomopy ring removal filter, requires numpy<=1.15:
mamba install -c conda-forge -c anaconda -c astra-toolbox/label/dev -c pytorch -c simpleitk numpy=1.15