fixes cpu-only environment and also auxilary data condition for cupy

conda/environment-cpu.yml

@@ -4,14 +4,18 @@ channels:
   - anaconda
   - httomo
 dependencies:
-  - conda-forge::click
+  - conda-forge::tomopy==1.15
+  - httomo::httomolib==2.0
+  - conda-forge::click>=8.1.3
   - conda-forge::mpi4py>=3.1
-  - conda-forge::h5py==3.8.0[build=*openmpi*]
+  - conda-forge::h5py=*=*mpi_openmpi*
   - conda-forge::pyyaml
-  - conda-forge::numpy<1.24
-  - conda-forge::python
-  - conda-forge::tomopy
+  - conda-forge::numpy<=1.25
+  - conda-forge::python 
   - conda-forge::numexpr>=2.8.4
   - anaconda::ipython
-  - anaconda::loguru
-  - httomo::httomolib
+  - anaconda::loguru  
+  - conda-forge::graypy
+  - conda-forge::plumbum
+  - conda-forge::tqdm
+  - conda-forge::typing_extensions

docs/source/howto/installation.rst

@@ -15,12 +15,13 @@ Install as a Python module
 ===========================
 
 If installation above for some reason is not working for you, then the best way to install HTTomo is to create conda environment first and then 
-`pip install` HTTomo into it. You will need to `git clone` HTTomo repository to your disk first. 
+`pip install` HTTomo into it. You will need to `git clone` HTTomo repository to your disk first.  Use `environment.yml` file to install 
+the GPU-supported HTTomo. For CPU-only version, please use `environment-cpu.yml` instead.
 
 .. code-block:: console
     
    $ git clone [email protected]:DiamondLightSource/HTTomo.git # clone the repo
-   $ conda env create --name httomo --file conda/environment.yml # install dependencies
+   $ conda env create --name httomo --file conda/environment.yml # install dependencies for GPU version   
    $ conda activate httomo # activate environment
    $ pip install . # Install the module
 

httomo/runner/auxiliary_data.py

@@ -81,7 +81,8 @@ def _get_field(self, field: str, gpu=False) -> generic_array:
         if gpu:
             array = xp.asarray(array)
         else:
-            array = xp.asnumpy(array)
+            if xp.__name__ == 'cupy':
+                array = xp.asnumpy(array)
 
         setattr(self, f"_{field}", array)