Detect overlapping windows in datasets

[yetinam]

In sufficiently complete data sets, waveform windows could overlap. As it might sometimes be of interest, to get the metadata information from the overlapping traces as well (e.g., overlapping events), we could add a routine to detect those overlapping windows. This can probably be done by grouping the metadata by station and ordering the traces in time.

Side note: Our format does not actually prohibit to use the same trace name in multiple metadata rows. This might actually prove helpful, when writing data set conversion scripts with low temporal spacing between the events.