Side note: My script was not extracting all of the sample file paths because not all samples are listed as such. It seems that any sample that gets sliced is then stored in Song.xml as an Instrument. In that case the file path of the source wav is not included.
This is not (probably) a big deal for me since I've been giving my wav samples unique names. The Song.xml has the base name of the source wav (e.g. "0301_00102_116bpm") which will be good enough for me. My concern is to know if I've inadvertently reused the same sample in multiple songs, and to help avoid reusing a sample as assemble new songs. Or at least know that I'm reusing something and need to make sure there's some sufficient additional transformation to avoid having songs sound too much like one another.