A Python Wrapper for SurveyGizmo's restful API service. Sort of - it's not really finished at the moment.
$ pip install SurveyGizmo
Start by instantiating the SurveyGizmo object and providing some configuration parameters. Options can also be set through the config
property.
from surveygizmo import SurveyGizmo
sg = SurveyGizmo(api_version='v3', response_type='json')
sg.config.auth_method = "user:pass"
sg.config.username = "username"
sg.config.password = "password"
Calls to the api are by object type then by function. For example,
sg.api.survey.list()
sg.api.survey.get('39501')
sg.api.survey.copy('39501', '39501 Copy')
sg.api.surveyresponse.list('39501')
Currently, user:pass
and user:md5
are the only supported authentication methods. There is the oauth_helper
module that is built on top of rauth, but it remains untested. You can try it I guess.
sg.config.auth_method = 'user:pass'
sg.config.username = 'username'
sg.config.password = 'password'
sg.config.auth_method = 'user:md5'
sg.config.username = 'username'
sg.config.md5_hash = '5f4dcc3b5aa765d61d8327deb882cf99'
or you can provide a password to config.password
and it will be automatically hashed.
sg.config.auth_method = 'oauth'
sg.config.consumer_key = 'f5e5f0fc0300519c7gef3f4cddec396cf3d28cf192f'
sg.config.consumer_secret = 'dfa3fb80a5be5a08c5b427dbs9c4vb1ad3'
sg.config.access_token = '12345'
sg.config.access_token_secret = '54321'
- api_version - 'v3', 'v4', 'head'. Defaults to 'head'
- auth_method - 'user:pass', 'user:md5', 'oauth'
- username
- password
- md5_hash
- consumer_key
- consumer_secret
- access_token
- access_token_secret
- response_type - None, 'json', 'pson', 'xml', 'debug'. If None, the response is returned as a python dictionary.
- requests_kwargs - Additional arguments passed to
requests.get
. Useful for setting timeouts and otherwise configuring the requests library. - prepare_url - Force the client to return the url after being prepared instead of executing the api call. This is useful in cases where you need to call the api asynchronously. Defaults to 'False'
- preserve_filters - Preserve filters between api calls. Defaults to 'False'. When 'True', use 'api.clear_filters' to manually clear the filters.
- handler52x - Handler for CloudFlare's 52x errors. Expects a callable (e.g.,
surveygizmo.default_52xhandler
). Defaults to 'None'.
Filters are currently untested (as is the whole library), but they should work. There is no magic here, just a convenience wrapper around the annoying filtering semantics. You could manually construct and pass these
TODO: Test filters
To filter, add filters to the query before making the api call. eg,
sg.api.add_filter('datesubmitted', '<=', '2013-07-01')
sg.api.surveyresponse.list('39501')
By default, filters are cleared out after calls to the api. If you want to keep filters for the next call, set the keep
parameter.
sg.api.surveyresponse.list('39501', keep=True)
sg.api.surveyresponse.list('39502')
After SurveyGizmo's move to CloudFlare, it isn't uncommon to see connectivity issues where the service is temporarily unreachable. These errors exist on the 52x range of HTTP status codes. To automatically handle 52x errors, set a callable for config.handler52x
. A basic handler is provided under surveygizmo.default_52xhandler
, which simply retries the request every second until a non-52x response is returned.
Only a small subset of the API is currently unimplemented.
TODO: Implement what's left.
0.2.0 is a forwards incompatible release, but only minorly so.
Forwards incompatible changes:
- Renamed the 'change' operations to 'update'. This is consistent with SurveyGizmo's API naming.
- Removed the 'keep' kwarg for preserving filters bettween api funcion calls. This is now configured with 'preserve_filters'. Filters are now cleared manually with
api.clear_filters()
- Removed the undocumented 'url_fetch' kwarg, which prevented api executioned and instead returned the prepared url.
Backwards incompatible changes:
- Modified 'api_version' to no longer has any effect on the client. SurveyGizmo provides no way to meaningfully differentiate between API versions, so this checking was unneeded and created code duplication
- Added 'prepare_url' as a replacement for 'url_fetch'. This forces the client to return the url after being prepared instead of executing the api call. This is useful in cases where you need to call the api asynchronously.
- Added 'requests_kwargs'. These are additional arguments passed to
requests.get
. Useful for setting timeouts and otherwise configuring the requests library. - Added handling for CloudFlare 52x errors
Copyright © 2013-2014 NC State University. See LICENSE for details.