Load multiple datasets into a single view

[collinsr]

A typical DIVA use-case is to view multiple datasets on a single video, e.g.

• the ground-truth with a "_gt" suffix
• results from team A with a "_teamA" suffix
• results from team B with a "_teamB" suffix ...and so forth.

These results are generated independently, with possibly overlapping track IDs, and identical label domains (i.e. without the suffices.)

Describe the solution you'd like The ability to repeatedly add a dataset to a video (via GUI and API) such that:

• track IDs in the dataset are disambiguated from existing track IDs.
• an optional suffix is added to the labels as they're read in.

Describe alternatives you've considered Currently I use https://gitlab.kitware.com/meva/meva-data-repo/-/blob/master/tools/diva-dive-converter.py offline to incrementally build up the final CSV.

[subdavis]

Implementation is related to #942 -- solving one will probably solve the other.