Conda install Google colab + Pip installation for CIL

[epapoutsellis]

I have tried a lot of times to install CIL on colab or kaggle.

I did !conda install -c conda-forge -c ccpi cil and I got the following error :

---------------------------------------------------------------------------
OSError                                   Traceback (most recent call last)
<ipython-input-40-6b83ac0934ef> in <module>()
      1 # Import libraries
      2 
----> 3 from cil.framework import  AcquisitionGeometry
      4 
      5 from cil.optimisation.functions import L2NormSquared, BlockFunction, MixedL21Norm, IndicatorBox

3 frames
/usr/lib/python3.7/ctypes/__init__.py in __init__(self, name, mode, handle, use_errno, use_last_error)
    362 
    363         if handle is None:
--> 364             self._handle = _dlopen(self._name, mode)
    365         else:
    366             self._handle = handle

OSError: libcilacc.so: cannot open shared object file: No such file or directory

One year ago, a conda install it was sufficient and importing CIL was working. I thought that the problem was with the new version of CIL, but I tried older versions and did not work. I tried many things to fix this error but I failed miserably. Reading online, I understood that conda and colab are not very good friends. Maybe that's why there is no Miniconda installation in google colab by default. In order to use libraries that are not pre-installed in the default conda environment of colab, the best is to use pip install.

[paskino]

I think you have written some instructions by email @epapoutsellis . these could go in an appropriate doc file.

[paskino]

Check TomographicImaging/CIL-Demos#89

[gfardell]

TomographicImaging/CIL-Demos#111