sbc rank histogram bug

[hyunjimoon]

Under the setting

precision ={
    "S": 1, # # of draws from prior
    "M": 500, # # of draws from posterior (# of chains * # of draws from each chain)
    "N": 30, # # of observation
    "R": 8 # # of subgroups for hierarchical Bayes
}

setting = {
    "est_param_names" : ["adj_frac1_loc","adj_frac2_loc", "adj_frac3_loc", "adj_frac1", "adj_frac2","adj_frac3"],# "ss2p_frac4"],# "ss2p_frac4"
    "hier_est_param_names": ["adj_frac1", "adj_frac2", "adj_frac3"],# "ss2p_frac4"],
    "target_simulated_vector_names" : ["ss", "s"],
    "driving_vector_names" : ["exog_demand", "process_noise_normal_driving"],
    "model_name": "2hier_s_lgnorm_.01"
}
init_order = 100

the following rank figure is observed; the total sum (2+2+1+3 = 8) which means each hierarchy is treated separately when calculating rank.