Implement Query Method for Python Client

[crasanders]

The Python client can currently only send setup, ask, tell, and resume messages to the server. We would also like a query method that sends "query" messages to the server. This method would take in query_type, probability_space, and other arguments that get passed to the server's query method. See server.query, server.handle_query, and test_server.test_strat_query to see examples of "query" messages and what types of arguments can be sent.

Examples of use:

client.query(type="max") # return the x value with the highest y value in the form of a d-length vector, where d is the dimensionality. See the output of AEPsychModel.get_max.

client.query(type="min", locked_dims={0: [0, 1]}) # return the x value with the lowest y value in the form of a d-length vector; the first element must be between 0 and 1. See the output of AEPsychModel.get_min.

client.query(type="prediction", x=torch.tensor([[0, 1], [2,3]])) # return the model predictions at x=[0,1] and x=[2,3], in the form of a 2 n-length vectors, where n is the number of x values given. The first vector is the means, the second is the variances. See the output of BinaryClassificationGP.predict.

client.query(type="inverse", y=0.0) # # return the x value where y=0.0. Returns the actual y value of this point as a float as well as x in the form of a d-length vector, where d is the dimensionality. See the output of AEPsychModel.inv_query.