Easier installation with conda
I can’t figure out how to pull request this, so I’ll put it here. Could you please edit the installation instructions in the readme to include the following as an option for installation with conda. This worked at least on my ubuntu 18.04 system running miniconda.
conda create -n elsa01 python=2.7 numpy scipy gxx_linux-64 gfortran_linux-64 libgcc matplotlib
conda activate elsa01
Then navigate to the elsa directory
python setup.py install
And thats it. Also in this case elsa is contained inside of the elsa01 conda environment and doesn’t conflict with any other things floating around my system.
Comments (2)
-
-
I hope that this can help if someone is struggling with the eLSA installation.
Install Ubuntu
Ubuntu is available in Microsoft store (windows)
Set 2.7 python version
Python version has to be 2.7 . eLSA cannot be installed in other python versions.
Check all the versions existing in the systemls /usr/bin/python*
Check current version
python --version
If version 2.7 is not available in your system. Follow these steps:
sudo apt update sudo apt install python2
Change your current python version to 2.7
alias python='/usr/bin/python2.7'
Other environments
Install conda environment (e.g., Anaconda or Miniconda)
Then you need to install numpy (>=1.0) and scipy (>=0.6)conda install numpy conda install scipy
eLSA
Now you’re ready to install elsa.
Create elsa conda environment (I call it here elsa01).conda create -n elsa01 python=2.7 numpy scipy gxx_linux-64 gfortran_linux-64 libgcc matplotlib --yes
Activate the elsa01 environment
conda activate elsa01
Install eLSA from source
sudo apt-get update wget https://bitbucket.org/charade/elsa/get/master.tar.gz sudo tar xzvf master.tar.gz
Change directory (wherever your elsa directory is installed)
cd charade-elsa-c0bb9dd7d038 python2.7 setup.py install
The id number ‘c0bb9dd7d038’ may differ, be careful!
Then, change the directory and test eLSA to be sure that everything works fine (it takes some time)cd test . test.sh
You can run the following command to see the different arguments of the eLSA
lsa_compute -h
To deactivate the environment and return to ‘base’ you can use:
conda deactivate
When you want to run eLSA remember to activate the elsa01 environment.
- Log in to comment
Hi,
I confirm that the above worked for me too. It’s a pity that I didn’t see this comment earlier.
It would be great if you could edit the installation instructions indeed.