Here is the information from their OData Site:
they mention you can curl it, but apparently RDF-a (
Rendering the XHTML source for a given concept or heading through a
RDFa-aware processing tool, such as the RDFa Distiller offered by the
W3C. For example, "Bahia grass", when processed by the RDFa distiller,
yields this result.
) is an option, traditional crawling/scraping (ugh), and they also mention wget, and Perl: LWP::UserAgent, Java: HttpClient, Python: httplib.
Also from that site, if you use the breadcrumbs under Technical Center, it provides format types, metadata types, Search/query options (you can apparently also use OpenQuery on their site when on a modern browser, [I did not try this] (If you don't know about this, look it up OpenSearch.org, it is pretty sweet) and serialization options.
Supported RDF MIME types and serialization formats
RDF Data returned from the Linked Data Service can be serialized in
several formats. •RDF/XML (MIME type: application/rdf+xml; Extension:
.rdf) Example •N-Triples (MIME type: text/plain; Extension: .nt)
Extension: .json) Example •JSON-P (JSON) (Extension: .jsonp; accepts
"callback" query string parameter) Example •XHTML + RDFa 1.0 (MIME
type: application/xhtml+xml; Extension: .html) Example
Supported Search serialization formats
Search results from the Linked Data Service can be serialized in
several formats. •XHTML (Default setting) •ATOM (XML)( (add query
string parameter: "format=atom") Example •JSONML (JSON) (add query
string parameter: "format=json") Example •JSON-P (JSON) (add query
string parameter: "format=jsonp"; also acccepts "callback" query
string parameter) Example