TODO.txt

DONE - Change references to "toc" to "nav" or "navigation"
	cache dir
	rake tasks
	etc.

DONE - Implement richer navigation model
	each node should know about its parent
		parse json into composite object?
	remove first_child code?
		allow id to not be present in arg list, opts default to last if hash
		instead, if nav node doesn't have id, get first child's id etc.?
		-- handle this logic in view helper?


Improved multi-valued faceting?
  https://issues.apache.org/jira/browse/SOLR-475

Check out - http://www.railsinside.com/plugins/164-7-tree-data-structure-plugins-for-rails.html
	
	DocMeta.find(:document_id=>'swinburne').each{|d|d.destroy}
	
	xml = Nokogiri::XML open('swinburne.xml')
	
	dm = DocMeta.new(:document_id=>'swinburne', :label=>'Swinburne')
	dm.item 'Info', xml.at('teiHeader').to_xml, :type=>:xml
	dm.item 'Poems' do |poems_meta|
		xml.search('//text').each do |poem_xml|
			poems_meta.item poem_xml['n'] do |poem_meta|
				Raven::XML::Fragmenter.fragment(poem_xml, 'pb') do |fragment|
					poem_meta.item fragment['n'], fragment.to_xml, :type=>:xml
				end
			end
		end
	end
	
	dm.flatten.select{|i|i.type==:xml}.each do |item|
		solr_doc = {
			:id=>"#{item.document_id}-#{item.id}",
			:text=>Nokogiri::XML(item.value).text,
			:author_s => xml.at('//author').text
		}
	end
	
Add serialization and fetching to Raven::Navigation ->
	Raven::Navigation.store!
	root = Raven::Navigation.fetch('swinburne-CW')
	navigation = root.locate(params[:fragment_id])
	@document = solr.find_by_id navigation.opts['solr_id']

ID's
	Should solr docs know about nav node ids?
	Should nav nodes know about solr doc ids?
	Should nav node ids and solr doc ids always be the same?

Allow blank variant_id (copy of a book, for example), navigation id's are getting "swinburne--poems" instead of "sunwburne-poems"
	This is a bug in NavBuilder (check for empty values before joining or compact the array)

Solidify generic Solr schema -- don't use dynamic fields?

Break-up Raven module into standard Ruby gem directory structure

Various export formats for document details view
	juxta, pdf, text-only etc.

Bring back JS text diff plugin

DONE - Bug in view code -- navigation is now hash (single node root) instead of array

DONE - xslt transform failing in details view
	switched to saxon (temporary)
		Nokogiri doing some weird stuff