Sam Bayer identified this issue and shared the attached triggering document. We should extract images from odt.
A bit of a hack, but we're pulling everything under "/Pictures" and "/Thumbnails" for now.
I couldn't figure out how to embed 1) a pdf without it being just the raw bytes or 2) a doc file without it being converted to an embedded odt object...maybe we want to add handling for embedded objects (odt within an odt).
For anyone who knows more about embedded objects in odt, please open another issue.
Thank you, Sam. Cheers, Tim.
triggering file with very small image.