Tips for working on Sigma2#

Sharing files on NIRD through OpenDAP#

  • We have an openDAP server tha was set up as a custmoized service as described in the NIRIS/sigma2 documentataion

  • The local path on nird is:

    /projects/NS9081K/NORSTORE_OSL_DISK/NS9081K/shared

  • Files or symbolic links in that path will be visible online on (needs some testing of access right propagation)

    https://ns9081k.hyrax.sigma2.no/opendap/

  • The nird paths are also mounted on the computing clusters (/nird/projects/...), visible to login nodes, but not visible to compute nodes, as described in the NIRIS/sigma2 documentataion

Sigma2 data processing tools#

NIRD toolkit#

  • We have a projectspace connected to project NS9081K on the NIRD-toolkit. Here, fcarctic-ns9081k is accessible to members of Feide group
    fc:adhoc:dde5413a-7adc-4890-b900-b52093925959
    and mounting
    /nird/projects/NS9081K

  • manage groups through Feide Innsyn (needs interaction with sigma2 support)

  • active ocean analysis group jupyter hub (get in touch with Tore for access)

Betzy apps#

Creating local python environment and package installing#

Sigma2 python documentation
Stackoverflow tips

If you use JupyterLab: go to the terminal and type:
$ python -m venv my_new_pythonenv
Activate the environment:
$ source my_new_pythonenv/bin/activate
Install packages with pip. Here we install pandas.
$ python -m pip install pandas You then need to add my_new_venv to the iPython kernel:
$ ipython kernel install --user --name=my_new_venv
If this doesn’t work:
$ ipython kernel install --user --name=my_new_venv

If you use JupyterLab: go to the terminal and type:
$ python -m venv my_new_pythonenv
Activate the environment:
$ source my_new_pythonenv/bin/activate
Install packages with pip. Here we install pandas.
$ python -m pip install pandas also need to install ipykernel package. $ python -m pip install ipykernel You then need to add my_new_venv to the iPython kernel:
$ ipython kernel install --user --name=my_new_venv
Actually, need to make sure that the kernel is isntalled at the right place ~/my_new_pythonenv/bin/python -m ipykernel install --user --name=my_new_venv --display-name "Python (my_new_venv)"

Check: less /cluster/home/thinf/.local/share/jupyter/kernels/my_new_venv/kernel.json

If this doesn’t work:
$ ipython kernel install --user --name=my_new_venv

Then, to explore and reproduce a certain environment (e.g. of python packages), use:

pip freeze > requirements.txt

and

pip install -r requirements.txt