Added the build target that creates the downloadable zip with a corpus.xml and a primarySource.txt inside, the one being the complete corpus, the other being only the plain-text content of the text of primary source document transcriptions. That now seems to be working OK. In the process I discovered that there are still some issues with missing hashes in the @rendition attributes of <zone> elements in the image markup docs. Ideally I'll fix that in the original source files and then fix any fallout resulting from it in the static build.