[CV] Refactor Examinador

What is the issue?
Currently Examinador (our integration testing tool) has a number of major issues:

1. It is missing a lot of core functionality (generic validation of documents' contents, convenient data export from a local cluster, etc.) that makes it hard to write new tests.
2. In addition to the point above, all functions and test suites handle the exported or generated documents in different ways, there is no standardised way of representing a JSON document in Examinador.
3. There is no clear guidance as to when Robot keywords and when Python functions (that can be turned into Robot keywords) should be used (there is a number of keywords and Python functions that basically only do CLI calls so it is confusing which of those should be used when a new CLI call functionality needs to be added).
4.  A lot of keywords or functions that are meant to be used in tests are not easy to use because they do not provide a neat way of handling expectations (for example, a keyword that checks that something executed successfully should not return a 200 code that has to be explicitly checked in every tests that uses that function, it should be done inside of the keyword and managed by an 'expected execution result' argument).