rajarshi/rest-ws

… then it accepts two form elements called query (a SMARTS pattern string) and target (a comma separate string of one or more SMILES). Return is a plain/text doc with N rows for N input SMILES, each row being true, false or fail if the pattern matched the SMILES, didn't match or failed to parse respectively

…spatch to actually get the prediction

…uires base 64 encoded smiles"

This reverts commit c7e6337.

…se 64 encoded smiles

…to get predictions from arbitrary models. Added example manifest file -allows us to serialize model details. It's the first version, so expect changes to manifast schema

…DME describing the services

…it simplified the code significantly. Looks a bit uglier though

client specifically sets the Accept heder to just text/plain

matching content type request, rathr than carrying on through the loop

Rather than just looking at the first content type, we loop over all
supplied content types and take the first one that matches the supported
types. This allows clients to suggest preferred types and if not supported
get some other type of return page