Make tests more reproducible

At the moment some of the tests in [openalex/examples](https://github.com/benedekh/ktalex/tree/main/src/test/kotlin/ktalex/openalex/examples) group depends on the "randomness" of the OpenAlex API, i.e. it's not sure that a certain ID will be returned. Make such tests more robust in a way that the test case remains meaningful and fulfills its original purpose.