normalise query fields

[jefferis]

There are different forms of query for different datasets. For example '/DA2_ORN' searches the type field for hemibrain but not flywire (which expects '/type:DA2_ORN'). Furthermore malecns uses superclass where flywire uses super_class and manc uses class. We should give the option (in cf_ids()) to handle these vagaries – as well as the option not to do so if the end user wants complete control. For new users this would be more convenient (I think) so it should be on by default. Therefore there should probably also be a global option for this behaviour in case some scripts currently rely on it. We may want to display a message to alert the user if the query is modified in this way

• Add implicit type for all queries introduced with a /
• Give cf_ids a normalise_query arg (defaults to TRUE)
• Make a coconatfly.normalise_query option (implicitly TRUE)
• Display a message if the query is modified
  • likely also requires a function (e.g. idspec_type()) that figures out if the id specification is one of:
    • URL (^http)
    • numeric ids (possibly space/comma delimited)
    • keys (is_key() provides a pattern)
    • query (possibly divided into full blown regex query and an exact type specification)
    • note that is_key() has a compound argument that seems to indicate if it is a vector of one or more keys or potentially a single string listing multiple comma/space separated keys.
• should probably be part of a v0.3 coconatfly release that also changes the cf_meta output fields to superclass, class, subclass.