Working with Web Processing Service with Python and OWSLib

Basic WPS interaction is described in the OWSLib documentation

We suggest using OWSLib 0.17.1 and up.

Getting a list of processes:

from owslib.wps import WebProcessingService
wps = WebProcessingService('http://localhost:8081/pywps')
wps.getcapabilities()
processes = [x.identifier for x in wps.processes]

Inputs identifiers:

process = wps.describeprocess('some_process')
inputs = [x.identifier for x in process.dataInputs]

Execute process (synchronous):

myinputs = [('some_parameter', 'some_value')]
execution = wps.execute('some_process', myinputs, async=False)
execution.getStatus()
execution.statusLocation  # useful for accessing the xml status file
execution.processOutputs[0].data

Execute process with file inputs/outputs (asynchronous):

from owslib.wps import ComplexDataInput
myinputs = [
    ('some_file_input', ComplexDataInput('http://localhost/file.nc')),
    ('some_string_parameter', 'some_value'),
]
execution = wps.execute('some_process', myinputs, output='OUTPUT')
while execution.status != 'ProcessSucceeded':
    execution.checkStatus(sleepSecs=1)
    if execution.status == 'ProcessFailed':
        break
execution.processOutputs[0].reference

On a local flyingpigeon, the results can also be found in:

~/birdhouse/var/lib/pywps/outputs/

If the WPS is protected behind magpie:

import requests
credentials = dict(provider_name='ziggurat',
                   user_name='magpie_username',
                   password='magpie_password')
s = requests.Session()
response = s.post('{0}/signin'.format(https://localhost/magpie), data=credentials, verify=False)
auth_tkt = response.cookies.get('auth_tkt', domain='localhost.local')
headers = dict(Cookie='auth_tkt={0}'.format(auth_tkt))
wps = WebProcessingService('https://localhost/twitcher/ows/proxy/wpsandbox/pywps', headers=headers, verify=False)