Ferret (Apptainer)
Modules and program start
An installation of PyFerret is provided in /sw/container/jupyterhub/ferret.sif
The container can be used on Jupyter HPC for usage inside of notebooks:
%load_ext ferretmagic
%%ferret
cancel data/all
use levitus_climatology
show data
%%ferret -q
shade temp[k=1]; go land
It is also possible to connect to the HPC cluster with X11 forwarding enabled (ssh -X) and start ferret directly:
Note that the $WORK directory currently differs between partitions, for which you can find details here and in the cpu partitions table.
In this example the large96s:shared partition is used.
srun --partition large96s:shared -c 1 --x11 --pty bash -c "module load apptainer squashfuse; apptainer run --bind $WORK,$TMPDIR /sw/container/jupyterhub/ferret.sif"
If you want to execute ferret with particular command line arguments you can also first start a shell inside the container:
srun --partition large96s:shared -c 1 --x11 --pty bash -c "module load apptainer squashfuse; apptainer exec --bind $WORK,$TMPDIR /sw/container/jupyterhub/ferret.sif /bin/bash"
Documentation
Ferret is an interactive computer visualization and analysis environment for oceanographers and meteorologists analyzing large and complex gridded data sets. A detailed description is found at the Ferret home page, where an online-manual for the latest version can be found.
Specific features related to pyferret, i.e., new graphic capabilities and the integration of ferret in python are described on the pyferret home page.
Special features of the installation
pyferret
Usage of the pyferret-package with your own python - environment will be enabled after installation of a stable python - version. The wrapper startup-script ferret starts pyferret.
Special go-Files, additional data sets
All additional files are in the directory /sw/viz/ferret/ferret_iow.
| Tool | scripts (use go/help <script> for help) |
|---|---|
| Median filter | median_l , median_test.jnl |
| Baltic Sea landmask | modified fland |
| Baltic Sea coastlines | land_balt_100, land_balt_200 |
| West-African coastlines | land_south |
| comfortable aggregation | aggregate |
The median filter is written by E. D. Cokelet. There is a test-script median_test.jnl.
For models of the Baltic Sea regional, topographic data sets are compiled and formatted like the etopo data, but with 1 n.m. oder 2 n.m. resolution. For this reason go fland 2 or go fland 1 works too for the Baltic Sea area.
Specific palettes
Palettes adcplus.spk, adcpmin.spk and adcp.spk are suitable for visualisation of current fields oscillating around 0 and are suitable for centered fill or shade levels.
Data set aggregation
Ferret allows to aggregate data distributed over many files into one logical data set.
- Aggregation related manual section
- Simplified aggregation over the time-dimension
For convenience, the go-script aggregate can be used.