Multi-resolution search

[JulienPeloton]

Currently, the methods to perform the cross-match between 2 data sets are at fixed grid resolution (the resolution of the grid within a partition - we would always keep the Spark partitioning the same). The idea would be to first perform a quick search of match with a coarse resolution to reduce the number of candidates, and then use a finer resolution with the resulting small set of objects to identify true match.