Interferogram network from compressed slcs needs logic for mid-stack reference index

[scottstanie]

If we had comressed SLC for day 1_3_6 and real slcs 7,8,9, the outputs of phase linking will be

(1, 7)
(1, 8)
(1, 9)

as of now, we can handle a reference of 1 either using a single-reference-ifg-network , or a max-bandwidth network.

We need to figure out the exact logic for handling a mid-stack reference switch. there's a reference_index- should that apply to the whole SLC stack? just the real SLCs? should we require a date instead?