# Access Sciserver datasets from datascope Technically we can access sciserver from datascope so you can also use your compute nodes in datascope to play with the sciserver data. All you need to do is change the catalog entry. Here are the steps to do it. * Go to Datascope ``` $ cd /export/scratch/asiddi24/ ``` * Make a fresh clone of oceanspy ``` $ git clone https://github.com/your_username_here/oceanspy.git $ cd oceanspy $ git remote add upstream https://github.com/hainegroup/oceanspy.git ``` * Make a new branch from upstream/main. This will help you to point a url for the catalog ``` $ git fetch upstream $ git checkout -b name_of_your_new_branch ``` * Create the Oceanography env ``` $ conda env create -f sciserver_catalogs/environment.yml $ conda activate Oceanography ``` * Now install oceanspy in dev mode ``` pip intall -e . ``` Now you can play around with oceanspy and push changes in your branch * Go to `catalog_xarray` and change the urlpath from `/home/idies/workspace/OceanCirculation/...` to `/sciserver/oceanography/...` in the required dataset. * Now use `git add .` and `git commit` to stage changes and then `git push -u origin name_of_branch` to push your branch online. * Now go to your fork of oceanspy and go to the branch that you just pushed. Go to the file `catalog_xarray.yaml`. Get the raw url for it. * Now, in datascope, in your local oceanspy folder, open `open_oceandataset.py` and change the catalog url. In my case this was `https://raw.githubusercontent.com/asiddi24/oceanspy/datascope_catalog/sciserver_catalogs/catalog_xarray.yaml` You're all set !