Support datasets in CSV export

https://redmine.teklia.com/issues/11586

A new --with-datasets option should be added to arkindex export csv. This option causes the CSV to have three new columns: dataset_id, dataset_name and dataset_set_name.

Those columns should be repeated if any element belongs to multiple dataset sets, with dataset_id_1, dataset_id_2, etc.

If the --filter option is used to exclude all three columns, or if none of the elements to export belongs to any dataset, no additional query should be made to fetch datasets and a warning should be logged. This is the same behavior as with classifications, metadata and entities.

The CSV options documentation should be updated to include this new option. The CSV output section should include a new Datasets section with an example of the output.