improve identification of filename in last URL segment

currently we "identify" if the last segment of a URL for a remote asset is a filename by checking if it contains a `.` which is not a robust way to do so. this should be improved.

Start by locating the logic that examines the last URL segment when loading a remote asset and review how it currently treats a segment containing a dot. Check whether existing tests cover filename identification; done means the classification no longer relies on that single heuristic and the relevant behavior is covered by tests.

Written by the indexing model from the issue text.