sparqlhttp 1.14

The classes are a mess. Here's how they should look:

def queryd(self, query, initBindings=None, format='python'):
def countQuery(self, query, initBindings={}):
def add(self, triples, ctx): # changed to a list, not *args
def contains(self, stmt):
def label(self, subj, default=''):
def value(self, subj, pred, default=None):
def subgraphLength(self, ctx):
def subgraphClear(self, ctx):
def dumpAllStatements(self):
def remove(self, triples, ctx): # changed to a list, not *args
def save(self, context): # for certain remote graphs

cache = None
cache = PredicateCache(timeout...) # expires according to the predicate
cache = SubjectCache() # invalidates everything about a subject whenever you write to it

target can be an in-process rdflib berkeleydb path too

graph = SyncGraph(protocol='sesame', target='http://...', cache)
graph = AsyncGraph(protocol='sesame', target='http://...', cache)

inside, each real method calls prep, request, and processResult
pieces, which are connected together differently in the case of the
deferred api.

request(method, rootUrl, path, queryparams, headers, payload)
# streaming support for restkit?

# cache api
class Cache(object):
def __init__(self):
#
def get(self, methodName, *args):
# return result or Miss
def set(self, result, methodName, *args):
# None

todo:
- remove 'remote' prefix everywhere?
- not makeConnection, but SyncGraph and DeferredGraph?