Since we have no interest in the actual spectrum data of each scan, we could have etree.parse use a collector that drops data to only capture the hierarchy and the attributes of each terms but not their data content (avoiding loading binary scan data in memory).
Since we have no interest in the actual spectrum data of each scan, we could have etree.parse use a collector that drops data to only capture the hierarchy and the attributes of each terms but not their data content (avoiding loading binary scan data in memory).
See here for details.