AGU 2018 Demo

This notebook shows how to use birdy’s high-level interface to WPS processes.

Here we access a test server called Emu offering a dozen or so dummy processes.

The shell interface

%%bash
export WPS_SERVICE="http://localhost:5000/wps?Service=WPS&Request=GetCapabilities&Version=1.0.0"
birdy -h

%%bash
export WPS_SERVICE="http://localhost:5000/wps?Service=WPS&Request=GetCapabilities&Version=1.0.0"
birdy hello -h

%%bash
export WPS_SERVICE="http://localhost:5000/wps?Service=WPS&Request=GetCapabilities&Version=1.0.0"
birdy hello --name stranger

The python interface

The WPSClient function creates a mock python module whose functions actually call a remote WPS process. The docstring and signature of the function are dynamically created from the remote’s process description. If you type wps. and then press Tab, you should see a drop-down list of available processes. Simply call help on each process of type ? after the process to print the docstring for that process.

from birdy import WPSClient

url = "http://localhost:5000/wps?Service=WPS&Request=GetCapabilities&Version=1.0.0"
wps = WPSClient(url, verify=False)
help(wps.binaryoperatorfornumbers)

Type wps. and the press Tab, you should see a drop-down list of available processes.

# wps.

Process execution

Processes are executed by calling the function. Each process instantaneoulsy returns a WPSExecute object. The actual output values of the process are obtained by calling the get method. This get method returns a namedtuple storing the process outputs as native python objects.

resp = wps.binaryoperatorfornumbers(1, 2, operator="add")
print(resp)
resp.get()

For instance, the inout function returns a wide variety of data types (float, integers, dates, etc) all of which are converted into a corresponding python type.

wps.inout().get()

Retrieving outputs by references

For ComplexData objects, WPS servers often return a reference to the output (an http link) instead of the actual data. This is useful if that output is to serve as an input to another process, so as to avoid passing back and forth large files for nothing.

With birdy, the outputs are by default return values are the references themselves, but it’s also possible to download these references in the background and convert them into python objects. To trigger this automatic conversion, set asobj to True when calling the get method. In the example below, we’re using a dummy process called output_formats, whose first output is a netCDF file, and second output is a json file. With asobj=True, the netCDF file is opened and returned as a netcdf4.Dataset instance, and the json file into a dictionary.

# NBVAL_SKIP
# This cell is failing due to an unautheticated SSL certificate
out = wps.output_formats()
nc, json = out.get()
print(out.get())
ds, json = out.get(asobj=True)
print(json)
ds

Progress bar

It’s possible to display a progress bar when calling a process. The interface to do so at the moment goes like this. Note that the cancel button does not do much here, as the WPS server does not support interruption requests.

wps = WPSClient("http://localhost:5000/wps", progress=True)
resp = wps.sleep()