Add trim_to_supervision_groups method

[desh2608]

This PR adds another way of trimming a cut set --- by trimming to "supervision groups". A supervision group (or an utterance group, see Fig. 2 in this paper) is defined as a group of contiguous supervisions that are either overlapping or separated by a small pause (shorter than some max_pause threshold. Such a way of trimming is useful for evaluating multi-talker ASR.

For example, the following cut::

                                        Cut
╔═════════════════════════════════════════════════════════════════════════════════╗
║┌──────────────────────┐                              ┌────────┐                 ║
║│ Hello this is John.  │                              │   Hi   │                 ║
║└──────────────────────┘                              └────────┘                 ║
║            ┌──────────────────────────────────┐            ┌───────────────────┐║
║            │     Hey, John. How are you?      │            │  What do you do?  │║
║            └──────────────────────────────────┘            └───────────────────┘║
╚═════════════════════════════════════════════════════════════════════════════════╝

is transformed into two cuts::

                    Cut 1                                       Cut 2
╔════════════════════════════════════════════════╗    ╔═══════════════════════════╗
║┌──────────────────────┐                        ║    ║┌────────┐                 ║
║│ Hello this is John.  │                        ║    ║│   Hi   │                 ║
║└──────────────────────┘                        ║    ║└────────┘                 ║
║            ┌──────────────────────────────────┐║    ║      ┌───────────────────┐║
║            │     Hey, John. How are you?      │║    ║      │  What do you do?  │║
║            └──────────────────────────────────┘║    ║      └───────────────────┘║
╚════════════════════════════════════════════════╝    ╚═══════════════════════════╝

[pzelasko]

Thanks!