• I don't want to take the namespace data (scikit-image) or datasets (scikit-learn), might need these later
• I prefer the librosa.example approach
• but I don't see a need to hide it in a sub-package like librosa.util.files right now
• I want to include a handful of files in the package, but
• Probably we should just go ahead and use pooch and be able to grab entire datasets later

so:

• add vocalpy.examples module and import function example from it at top level, to allow something like vocalpy.example('gy6or6-song.cbin') or vocalpy.example('birdsong-rec.xml')`
• adapt from https://github.com/librosa/librosa/blob/main/librosa/util/files.py
• use https://github.com/fatiando/pooch
• add example zebra finch songs + examples from warbleR, mouse ... what else?