Candidate Aggregate Metrics Setup + sessId Focus

This sets up the design for how candidate aggregate metrics will be
collected, and runs through the whole flow up to emits. Currently, many
values are default values, and will be filled in as the design is
extended out. It does handle the split session id logic, however,
between the two tracks. Also, after deliberation, we now use the
hashCode of a random UUID to get closer to ~ (1/2^(32))
collision rate - their hashcode is pretty random, utilizing an int
conversion from a bitwise xor operation of the most and least
significant bits from the already high entropy uuid. This will provide
more security than the binder implementation, especially under time and
device id buckets, against collisions.

Bug: 271135048
Test: Build and Won't Submit without E2E Test
Change-Id: I8e7455dd6eaf8adb15035f6a9262f34a0c20531f