Extend Hecuba to support direct access by secondary indexes of Cassandra and Qbeast indexes

[yolandab]

Right now, hecuba supports the usage of multidimensional indexing through the iteration of a chunk of indexed elements. First, the programmer must filter the chunk of interest and, then, iterates across all the elements in the chunk.

We miss the interface to support direct access to an indexed element. This is a proposal of how to implement it:

class myclass (StorageObj):
'''
@ClassField mydict dict <<key:int>,x:float,y:float,z:float>
@index_on mydict x,y,z as index_name
'''
d=mydict("Somedata")
r=d.index_name[v1,v2,v3] 

After executing this code, r should contain the whole record. In this example (key,x,y,z)

Do you agree?