schul-cloud-search-tests 1.0.41

Schul-Cloud Search API

These are the tests for the Schul-Cloud search API. You can read motivation blog post about

• common tests for all search engines

• testing during live deployment

Installation

You can install the tests by installing Python version 3 and pip. Then, you run the installation from the command line:

pip install --user schul-cloud-search-tests

Installation for Development

If you would like to contribute code to this repository, you can clone it first.

git clone https://github.com/schul-cloud/schul_cloud_search_tests.git
cd schul_cloud_search_tests

Install the required packages for Python:

pip3 install --user -r requirements.txt pip-tools==1.6.5

Note that if you would like to change the requirements, please edit the requirements.in file and run this command to update the dependencies:

pip-compile --output-file requirements.txt requirements.in

Specification

The idea is that these tests run in between your search client and your search engine or search adapter. You use the test interface instead of your search engine interface to run checks on every request the client makes:

• Is the client request formulated correctly?

• Is the server response folmulated correctly?

These cases can be defined and are defined:

• When the client issues a successful request and the server responds correctly, the request is forwarded, optionally including a note that the tests passed.

• When the client issues a malformed request, HTTP error 400 is returned including the information which tests did not pass. The request is forwarded to the server and the response is expected to be 400, too.

• When the client issues a correct request, and the server response is malformed, then HTTP error 409 is returned including a list of error descriptions of the mistakes made by the server.

To make the decisions transparent, the client request and the server response are included in the error reponses.

Usage

Suppose you have a server running on http://localhost:1234/v1/search/. You can tun the search engine tests as a proxy on port 8080 like this:

python3 -m schul_cloud_search_tests.proxy 8080 /endpoint/ http://localhost:1234/v1/search/

Now, all your requests to http://localhost:8080/endpoint/ will be forwarded to http://localhost:1234/v1/search/.

When you are done, you can visit http://localhost:8080/stop to stop the server or run this command:

python3 -m schul_cloud_search_tests.stop 8080

The return code is zero (success) if all tests of all requests passed. If one test fails, it is a number greater than zero.

Development Process

The idea is stated in the motivation blog post. We can use the tests to test the search engines. However, the tests can become complex and must be tested themselves. Therefore, the following development process is proposed.

1. Have a look at the specification:

   • The Search API describes how to request a search.

   • The Response Schema describes what to expect as a response.

   • The Error Schema describes what an error should look like.

   The specification is the most important document. It determines what needs to be tested.

2. Implement tests according to examples of the specification. These tests are located in the schul_cloud_search_tests/tests folder. They test how you would like to have the search proxy respond to the different valid and invalid requests.

3. Make the tests run.

   • If it is a new condition under which a proxy request succeeds or fails, you should implement these as tests in the schul_cloud_search_tests/search_tests folder. These tests are executed when a search request goes through the proxy.

   • If this is a communication feature of the proxy, it must be described in the Specification section. The code in the schul_cloud_search_tests/proxy.py should be touched.

Further Reading

• Readme Driven Development

---

You can edit this document on Github and check it with this editor.