Adding a Command-Line Interface¶

Prepare¶

Activate the Conda workshop enviroment:

$ source activate workshop

We are going to write a command line interface (CLI) for our Python plot function.

Objectives:

Go to the plotter tutorial source:

$ cd ~/birdhouse-workshop/tutorials/03_plotter_cli

See the command line options of our plotter:

$ python plotter.py -h
usage: plotter.py [-h] [-V [VARIABLE]] dataset

Plot our well-know image:

$ python plotter.py --variable air ../../data/air.mon.ltm.nc

Play a little bit with the command-line options. Try some other options (-V), use invalid input (water) and skip some arguments.

Use external data from a Thredds service, like NOAA:

See access methods for a dataset, note OpenDAP:

Use OpenDAP URLs directly as dataset input:

$ python plotter.py --variable air \
      http://www.esrl.noaa.gov/psd/thredds/dodsC/Datasets/ncep.reanalysis.derived/surface/air.mon.ltm.nc

Extend the command line and with an optional parameter for the timestep to be plotted.

Open your editor on plotter.py … and happy hacking.

Don’t forget to test often:

$ pytest plotter.py

The output name of the plot is always plot.png. Add an optional output parameter to set an output filename.