Fix use of multiple parameter files

[m-philipps]

When reading in parameter tables using parameters.get_parameter_df:

• check all parameter tables (not just those resulting from several parameter files) for contradicting parameter definitions
• remove the removal of duplicate parameters, will instead raise ValueError
• ensure that parameters that only differ by their parameterId are recognized as distinct and not dropped (fix Parameters dropped when using subset parameter files #154 )
• adjust/extend test_parameters.py accordingly

[dilpath]

👍

Thanks for fixing this, and the nice tests!

[dilpath]

As the drop_duplicates call is removed, the duplicated parameter definitions may be consistent (e.g. your test raises an error for redundance too), so this check is now just about duplicated parameter IDs.

The logic for this should now be equivalent to petab.lint.assert_unique_parameter_ids. Would be good if that could be reused here, but fine as is, since the circular dependency would need to be resolved.