Generate a dataframe that gather infos about the sound files in a LibriSpeech subdirectory
(librispeech_dir, subdir, speakers_md)
source not stored for this graph (policy: none)
no outgoing calls
no test coverage detected