I've been wondering lately if the xml that Solr returns from a search isn't actually "true" xml, and by true, I mean searchable in the the sense that you can normally pull specific nodes from an xml document and return only that information.
I've been trying to figure out a way to make my responses from solr be more human-friendly. Solr will index my pdfs and dump the output from the pdfs into a _content node, and I can return that node on the screen which is all fine and well.
Even using highlighting in my queries though, it's still pretty unreadable by any useful standards.
Consider the following response to a query.
<response> <lst name="responseHeader"> <int name="status">0</int> <int name="QTime">57</int> <lst name="params"> <str name="q">_content:[search term]</str> <str name="hl">on</str> <str name="fl">_content</str> <str name="hl.fl">_content</str> </lst></lst> <result name="response" numFound="5" start="0"> <doc> <arr name="_content"> <str>[block of content here with zero formatting]</str> </arr> </doc> </result>
While I'm glad that solr returned a response, it's really not very useful. I would have to have it return something that I could style with html and look presentable to the user.
I'm not currently using any custom search indexes...is that would I would need to do to be able to return something that I can mold and shape?