Parsing the XML with an XPATH parser is the easiest way.
A more tedious way would be to import the XML schema into a domain model and then retrieve the attribute containing the file path.
Import XML Documents | Mendix Documentation
Do those files exist in your resources file? If they do you can use GetFileContentsFromResource Java Action from the Community Commons module in the Marketplace to get the file into a FileDocument. Once it’s in a FileDocument you can use it in an Import from XML action, or however you want to handle the document.
Hope this helps.