Replace required_field property for CSV corpora

The XML corpus supports a `required` property for corpus fields - skip the document if that field is empty.

For historic reasons, the `CSVCorpus` has a property `required_field` which refers to a column/field in the CSV file (_not_ a field in the corpus), i.e. skip all rows where that column is empty. It does not have an implementation for the `required` property for corpus fields.

This is unnecessarily confusing. It should be possible to remove `required_field` and instead use the `required` property on corpus fields. `required_field` is used in the following places:

- https://github.com/UUDigitalHumanitieslab/I-analyzer/blob/5603be785b8687ac436110fd560e06b6da60bec1/backend/corpora/parliament/germany-old.py#L31
- https://github.com/UUDigitalHumanitieslab/I-analyzer/blob/5603be785b8687ac436110fd560e06b6da60bec1/backend/corpora/parliament/canada.py#L25
- https://github.com/UUDigitalHumanitieslab/I-analyzer/blob/5603be785b8687ac436110fd560e06b6da60bec1/backend/corpora/parliament/germany-new.py#L26
- https://github.com/UUDigitalHumanitieslab/I-analyzer/blob/5603be785b8687ac436110fd560e06b6da60bec1/backend/corpora/parliament/denmark.py#L46

Start by locating the CSVCorpus implementation and compare its required_field handling with the XML corpus's required corpus-field behavior. Review backend/corpora/parliament/germany-old.py, canada.py, germany-new.py, and denmark.py, then update their configurations to use required on corpus fields. Done means required_field is no longer needed and empty required corpus fields cause the corresponding document to be skipped.

Written by the indexing model from the issue text.