A lot of the native code in Calibre (not nearly all of it, more of a nontrivial chunk) is dedicated to parsing Compressed HTML Help and Microsoft Reader files, since the two formats share a lot in common — namely that both are variants of the *.cab format.

There's a couple good libraries out there for handling the compression formats supported by CHM and LIT files, but those libraries haven't been updated in a while and seem to have had some serious security issues in the past.