Need precomputed database of crossovers and overlaps

[karlmsmith]

Need to add something to SOCAT when adding or updating datasets that will automatically check for crossovers (good things - close matches between two unrelated datasets) as well as overlaps (bad things - duplications in a dataset or a series of datasets). These crossovers and overlaps need to be precomputed and stored in a database, and LAS then needs to report these values in the database instead of trying to do an on-the-fly calculation when the user presses a button.

The current on-the-fly crossover checker IN LAS has a known bug in that it uses the decimated data; otherwise it would be a lot slower. But this is missing a good number of crossovers. It may be using decimated data for one dataset and retrieves all the data for the other dataset which could explain the "one-way" crossover of 33GG20150708 with 316420141024. Another possibility is the code is not ignoring samples that are out-of-order in time and should have a WOCE-4 flag, so the code to limit the number of comparisons being made is jumping past the match in one of the directions.