Extend Hecuba to support secondary indexes of Cassandra

[yolandab]

We need to support the creation of secondary indexes from Hecuba. We can use the same keyword that we use to specify multidimensional indexes implemented by Qbeast:

class miobj <StorageObj>:
'''
@ClassField dictname dict <key:int>,x:float,y:float,z:float>
@index_on dictname x,y,z  #this creates a Qbeast index on x,y,z
'''

With this class specification we are using Qbeast to create a multidimensional index on dictname using attributes x,y and z as dimensions.

If after the keyword "index_on" we find only one attribute we can assume that we want to add on that attribute a secondary index implemented by Cassandra

class miobj <StorageObj>:
'''
@ClassField dictname dict <key:int>,x:float,y:float,z:float>
@index_on dictname x  #this creates a Cassandra secondary index on x
'''