The queries for FileModels for app scanning needs to limit to given app's files.

Currently, the app rules queries simply query for all files of certain qualities.

Once we start supporting server dir scans too, this won't work - we really don't want to unzip and decompile and analyze the jars from servers.

Therefore, a node representing each user provided directory/archive must be created and all files linked to it (if it isn't the case already).

And then, the queries in rules need to take these edges into account, and they will only query for the files that belong to the respective application.