DBSTOP removal of overlaps refactoring needs

[paco-barreras]

• Go from components_by_parent (dict of lists of sets) to core_df pandas Series, sorted by timestamp or at least chronologically.
• cluster_df can be instantiated from core_df + the set of border points of the parent, and then mutated at this step
• We need to pipe through data to get the coords.

With these objects in place, we can iterate chronologically like:

for curr_time in core_df: that is, through the core points.

• Figure out the equivalent notion of "active cluster", considering that the natural order now is "parent cluster > child clusters" but, indeed, one of the children comes first.

[carolineychen8]

Will be resolved in PR #286