Improve error handling and logging

[MoritzWeber0]

Currently, the CLI seems to be very tolerant when it comes to errors. One example that I recently saw:

2024-07-25 18:11:20,962 - WARNING  Link creation for "<LogicalFunction '[REPLACED]' (b9068884-d922-48de-a79f-555b47900e6b)>" partially successful. Some links were not created:
	Requested attribute: realizes
	AttributeError 'LogicalFunction' object has no attribute 'realizes'

The reason for the error message comes from an invalid configuration:

la:
  LogicalFunction:
    links:
      - realizes

The error message got lost in the spam of log messages. The exit code of the run was 0, the Gitlab CI/CD pipeline was green. However, this is obviously a faulty configuration.

Here is my proposal:

• Add a validation step and try to fail fast. Stop before trying to execute a Capella2Polarion run with faulty configuration und unexpected outcome. Related to Config validation #63 and Use pydantic for config modeling and validation #66
• Improve severity of log messages. I'd consider an unknown attribute that was explicitly requested as ERROR or EXCEPTION, not just an WARNING.