This directory contains two scripts that can be used to perform compare testing.
The directory requests contain a number of XML/SOAP requests that can be sent to a running OpenSearch server. The directory responses contains a set of files matching the content of the requests directory, with expected responses from the requests.
Note, that responses are heavily dependent on the Datawell used, and other configuration for the OpenSearch server. This makes it hard to do a comparision between the actual and expected results. As an alternative, you can test two servers against eachother.
You can use the script compare_request_results.py to run requests from a folder against two different servers, and compare the results.
Most often you will use this to compare a feature branch against the master, for the same configuration.
Example of use:
./compare_request_results.py -d http://localhost:22222/5.2/ http://localhost:33333/5.2/ requestsOf, if started using one of the docker-compose files made to start servers locally:
./compare_request_results.py -d http://localhost:22222/ http://localhost:33333/ requests/issues/Run the script with the --help option to get more information.
Note: As of this writing (January 2020) the script below has not been migrated to Python 3, and the reponses may be outdated.
You can use the script search_webservice_tester.py as a test generator designed to test the opensearch web service.
It conceptually works the following way:
-
For each file found in the request folder, a matching file is identified in the response folder. Files are matched by name, and the request and response folders can contain subfolders.
-
For each request/response pair a new unittest is generated. The request is sent to the designated server and the returned actual response is compared the expected response. If the comparison produces a diff, the test fails.
example of usage:
python search_webservice_tester.py --url http://lakiseks.dbc.dk/opensearch/ requests/ responses/