to research: processing speed when multiple thematic_dicts are calculated together.

[dieuska]

brdr-processing seems to get slower when there are more reference_items in the reference_data, compared when you load reference data per thematic item. Is there something to do about that?
Some (architectal) change, likt f.e keeping a sepeterate reference data dict for each thematic feature seperately?

To research?
Or better to start a new Aligner for each thematic_object?