You can clone with
HTTPS or Subversion.
Scrape ambiguous pages, too.
Don't scrape glossary pages as chapters.
Improve "defines" relations of dictionary pages.
Two more niceness facts about chapters.
Give a chapter page a dc:title with its h2, and make it type:chapter.
Add a README.
Can extract data from chapters, dictionaries.
This works to my satisfaction now. I can't wait to see the graph structure (-:
Improve scrape-glossary-page name.
Now can scrape glossary page inter-linking/names.