Easier public apis for Conditional statments + Marginalization mask

[khosravipasha]

1. Easy CON API
   Currently user has to manually call MAR twice to get conditionals. Have some api like this:

conditional(circuit::ProbCircuit,  data::DataFrame,  q_indices)

Where q_indices give indexes of variables in q for each datapoint (all variables mentioned in q[i] should NOT be missing in the correspoding data[i]). In the end we are computing p( q | x^o) for each data point where x^o is all observed variables except the ones in q.

2. MAR with Marginalization mask
   Makes sense to add another API to marginalize even over observed varibles. (for example above we need p^o but x^o, q is observed). This helps to avoid copying data for each query.

initial attempt here: master...cond