Better handle speakers at chunk boundaries in pyannote.diarization

[hbredin]

I am wondering whether it would be a good idea to actually overlap chunks a bit. Instead of showing [10, 20] -> [20, 30], we could actually show [9, 21] -> [19, 31] but only save [10, 20] -> [20, 30] in the database.

As a matter of fact, it is sometimes tricky for a human to annotate the very beginning of a chunk without this additional context.

We already add some context for the pipeline -- so it could be reused for the human :)

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.