Extracting plain text isn’t that much of a problem, relatively speaking. It’s interpreting more complex elements like nested lists, tables, side bars, footnotes/endnotes, cross-references, images and diagrams where things get challenging.