Hello,
PLEASE DO NOT develop any new customizations based on Alfresco Explorer, especially if you are already on Alfresco 5.0.a. Alfresco Explorer has been removed starting with Alfresco 5.0.b and the more you add now the more you add to your pain afterwards.
I don't really understand what "path problem with document" is supposed to mean. When you construct a ZIP file via Java code, you are basically free to define the paths within the ZIP yourself and can just use the input stream from the content reader to fill the data for a ZIP entry. No need to access Alfresco documents as files where path might be an issue.
In terms of code / solution: There are already enough viable examples on the net, so I won't pollute it with yet another.
Regards
Axel