Local virtual environment
Note: the following information is to set up an environment that works with the RFI simulations only. Other scripts may need more packages to be added to requirements.txt.
Differences based on Operating System
Linux: you may create a virtual environment with conda, virtualenvwrapper, or other python-based virtual environment tool. Installing the requirements via pip should work in all.
MacOS: you will need to create the environemnt with conda. That is because python-casacore
does not currently behave well, when trying to install it via pip
into a standard python environment.
Create a virtual environment
virtualenvwrapper
To install and set up virtualenvwrapper follow this guide.
Create an environment: replace my-environment with the name you prefer, and replace python3.7 with
the path to your python3 installation. If the PYTHONPATH used by virtualenvwrapper is the python3 version
you want to use, then you can omit the -p
option.
mkvirtualenv -p python3.7 my-environment
Start environment:
workon my-environemnt
Deactivate environment:
deactivate
conda
To install and set up conda follow the conda guide.
Create an environemnt (replace my-environment with the name you prefer):
conda create --name my-environment python=3.7
Start environment:
conda activate my-environemnt
Deactivate environment:
conda deactivate
Install requirements
Linux
Once you have activated your environment and navigated into the ska-sim-low directory (i.e. the root directory of the git repository), run the following:
pip install -r requirements.txt --pre
--pre
will allow you to download the latest beta versions of dependencies.
This is necessary to get the latest RASCIL version from PyPi.
Depending on what python version you used to create the environment, the pip within that will be tied to that python version. This command should install all of the necessary requirements.
In addition, you will have to obtain RASCIL data. RASCIL will be installed via pip as part of the above
command, however the additional setup described at RASCIL Installation
is required. If you encounter with a Tuple index out of range
error while running RASCIL-dependent code,
you may also need to go through the Git LFS
steps on the same page under “Installation via git clone”.
You will also need OSKAR set up. On Linux, you may use the Singularity image.
MacOS
On MacOS, python-casacore
, a dependency of RASCIL, does not behave well with pip
,
so you will need conda
to install it.
Install python-casacore
with conda:
conda install -c conda-forge python-casacore
Install the rest of the requirements using pip
:
pip install -r requirements.txt --pre
--pre
will allow you to download the latest beta versions of dependencies.
This is necessary to get the latest RASCIL version from PyPi.
Depending on what python version you used to create the environment, the pip within that will be tied to that python version. This command should install all of the necessary requirements.
In addition, you will have to obtain RASCIL data. RASCIL will be installed via pip as part of the above
command, however the additional setup described at RASCIL Installation
is required. If you encounter with a Tuple index out of range
error while running RASCIL-dependent code,
you may also need to go through the Git LFS
steps on the same page under “Installation via git clone”.
You will also need OSKAR set up, which you can do via installing the binary version.