UPDATED August 2023: The basic recipe as outlined below is pretty much the same for the current version (2.6.3), except that you don't need to avoid the newest versions of Python and numpy anymore! So step 2 is even easier!
...
A quick one here! The ISCE developer team at JPL and Caltech continue to make the software easy-peasy to install. Now we can use the conda-forge to get the job done, via conda.
As of ISCE version 2.6.1 (August 2022), there are a couple of little tweaks to the setup that enable it to play nice with dependencies. So if you are having trouble, and are running some form of Ubuntu Linux (including, wouldn't you know it, WSL), maybe this will help?
1) Make sure you have conda installed. I usually go for miniconda, as all of my conda-ing is from the command line.
2) Make a new conda environment and install ISCE. At this point (December 2022), ISCE does not play nice with the latest versions of python3 or numpy, so make sure to use older versions:
conda create --name isce2.6.1 python=3.8 numpy=1.19
conda activate isce2.6.1
conda install -c conda-forge isce2
This installs isce2 in a conda environment called isce2.6.1. The 'conda activate' command is necessary to activate the conda environment in future.
UPDATE, August 2023: This part is even easier now, under the latest version of ISCE (2.6.3), which has been modified to deal with the latest versions of Python and numpy, so you don't need to specify older ones:
conda create --name isce2.6.3
conda activate isce2.6.3
conda install -c conda-forge isce2
3) Make a script that adds another critical directory to the path. In your favorite text editor make a script that contains the following line:
export PATH=$ISCE_HOME/applications:$PATH
Assuming that you called it something like 'ISCE_config.sh', then you should run it before you so any processing:
source ISCE_config.sh
(make sure to give the correct path and name to your script). And that should be that. Happy ISCEing!
No comments:
Post a Comment