Summary:
I find myself wanting something like this pretty often, so I tried my hand at adding it myself.

The `groupByAdjacent` operator creates groups bounded wherever the selector changes. It's especially useful for processing sources that've already been sorted on the selector, such as from a database query.

Given the following source sequence with keys `A` and `B`:

```
[A1, A2, A3, B1, B2, B3, A4, A5, B4, B5]
```

a regular `groupBy` would return something like:

```
[A:[A1, A2, A3, A4, A5], B:[B1, B2, B3, B4, B5]]
```

while this `groupByAdjacent` would return:

```
[A:[A1, A2, A3], B:[B1, B2, B3], A:[A4, A5], B:[B4, B5]]
```

Given a source where the items are presorted by selector, `groupByAdjacent` should behave identically to `groupBy`, except that `groupByAdjacent` supports infinite sources since it doesn't need to collect the entire source in memory before creating any output.

Reviewed By: yfeldblum

Differential Revision: D9475326

fbshipit-source-id: 1c8db3abadce5e68394e5fa38bf4bee0b413a03f