Good questions there.
1. I am afraid the application needs a fully qualified path to work with. We will look at this in the next release. For now, there can be 2 approaches possible.
a. Wrap the OpenXMLViewer.exe with another application layer that will be able to determine the actual physical path and call the exe
b. The C++ source code is available in the source branch. The code can be modified, built etc to get the relative path.
2. The file generated is <document name>.html / <document name>.xhtml. This can be used to get the respective file name. However for the images, a change in code may be needed. The current application uses the file names as stored in the docx.
Hence it is possible that the image file may be overwritten in case multiple documents use the same file name and the files are extracted to the same output folder.
The C# code branch is not supported and is legacy code. Hence it may not function as expected.