segmentation drops integrations from end

[caseyjlaw]

Creating segments currently assumes they must be on integration boundaries and are all the same size. That obligates us to drop some integrations at the end of a scan (up to the size of a segment).
Can we modify this scheme to search end integrations as a partial segment?

[caseyjlaw]

The state object has a search_ints method that could have a special case for partial segment at end.