Simplify the ID schema

[gongcastro]

Currently, the package works with three different ID systems: id (unique to BVQ), id_exp (unique to the experiment the participant was recruited to participate in, if any), and id_db (unique to the BabyLab participant database.

In theory, id_db shold be sufficient: the three remain constant and unique for the same participant. If anything, id_exp might be useful for adding BVQ data to experiment-specific datasets, although 1) such dataset shoould already contain id_db, and id_exp can be extracted from bvq_participant.