Since you're working with XML data, the apoc.load.xml and apoc.xml.parse procedures can load XML from a file, however they will parse and transform the XML into a map/dictionary/object strucutre. Not sure what your usecase is but you could use apoc.xml.parse to load the XML file and then pass the result to apoc.util.compress.
thanks.. I will dig into these again.. I didn't see how to do the connection between these mechanisms.. Basically we have these xml files, which don't seem to be in a standard format (yuck) and we want to understand them, so I proposed using a neo4j graphs to at least visualize them. the non-standard format is what makes this hard as a non-programmer/non graph person would have to manually build all those "with" statements..