represent unmarked write-ins in the original CVR snapshot

[adghayes]

The original CVR snapshot has mark thresholds for all marks. When unmarked write-ins are being considered, we should similarly track write-in thresholds for all write-in areas.

Currently, unmarked write-ins are represented in a special way in the modified snapshot. This isn't quite true to the CVR spec, in which the modified snapshot should only contain marks as interpreted for tabulation. At the point of export from VxScan or VxCentralScan, those unmarked write-ins are not being considered in tabulation.

On the original snapshot, the MarkMetricValue field is an array and the write-in threshold could be tucked in there.

On the VxAdmin import side, the import logic will have to be updated to consider the original snapshot when looking for unmarked write-ins.

[adghayes]

I'm very open to pushback if you think this is unnecessary, but it seems to me like the appropriate way to represent unmarked write-ins as we lean into actually using the original snapshots.