Improve testing from Python

[schmoelder]

Currently, our test cases are setup using createLWE (provided by CADET-Core) which creates an H5 file which in turn is read back to Python and then passed to a new instance of Cadet using the CLI or DLL runners respectively.

However, this has lead to several issues not being caught since the H5 file usually contains different types compared to setting up a configuration file directly with Python (e.g. lists are automatically cast to np.ndarrays).

To improve this, we should reconsider our testing setup. Here are some open questions:

• Add createLWE method to CADET-Python
• Use createLwE method by CADET-Process
• How does this integrate with CADET-Verification?