We could catch if initialized labels are invalid very early on when they are created and before writing to a file. This would slow down initialization (how much depends on the % of labels checked), but give a higher confidence that the generated POS is valid and allow catching problems early.
It's essential that the problem of invalid labels is communicated to users in a way that is understandable so that they can take action by themself. We should not just panic with the message "labels are invalid" because users will be confused.
We could catch if initialized labels are invalid very early on when they are created and before writing to a file. This would slow down initialization (how much depends on the % of labels checked), but give a higher confidence that the generated POS is valid and allow catching problems early.
It's essential that the problem of invalid labels is communicated to users in a way that is understandable so that they can take action by themself. We should not just
panic
with the message "labels are invalid" because users will be confused.