/* 

  Click graph in Picat.

  From PSI I paper.
  Where the objective is to determine how similar the 
  clicks of A and B are.

  Cf my Gamble model gamble_click_graph.rkt

  This program was created by Hakan Kjellerstrand, hakank@gmail.com
  See also my Picat page: http://www.hakank.org/picat/

*/

import ppl_distributions, ppl_utils.
import util.
% import ordset.

main => go.

/*
  clicks = [[true,true,true,false,false],[true,true,true,false,false]]
  var : sim all
  Probabilities (truncated):
  0.998639848082624: 0.0010000000000000
  0.998059035743614: 0.0010000000000000
  0.997746356761896: 0.0010000000000000
  0.997715513220856: 0.0010000000000000
  .........
  0.011753603821506: 0.0010000000000000
  0.010036382363195: 0.0010000000000000
  0.004697684200805: 0.0010000000000000
  0.001868278254694: 0.0010000000000000
  mean = 0.599464

  var : sim 1
  Probabilities:
  true: 0.6550000000000000
  false: 0.3450000000000000
  mean = [true = 0.655,false = 0.345]


  clicks = [[true,true,true,false,true,true],[false,false,false,true,false,false]]
  var : sim all
  Probabilities (truncated):
  0.986526463174506: 0.0010000000000000
  0.984096966676459: 0.0010000000000000
  0.977998407547361: 0.0010000000000000
  0.977298801288613: 0.0010000000000000
  .........
  0.001832617447634: 0.0010000000000000
  0.00168651528735: 0.0010000000000000
  0.001626677811903: 0.0010000000000000
  0.000806179363656: 0.0010000000000000
  mean = 0.317108

  var : sim 1
  Probabilities:
  false: 0.7190000000000000
  true: 0.2810000000000000
  mean = [false = 0.719,true = 0.281]

*/
go ?=>
  % Very similar clicks
  SimilarClicks = [[true,true,true,false,false],
                   [true,true,true,false,false]],
  % Very different clicks
  DifferentClicks = [[true,true,true,false,true,true],% true,false],
                     [false,false,false,true,false,false] % ,false,true]
                    ],

  member(Clicks,[SimilarClicks,DifferentClicks]),
  println(clicks=Clicks),
  reset_store,
  run_model(130_000,$model(Clicks),[show_probs_trunc,mean,
                           % show_percentiles,show_histogram,
                           % show_hpd_intervals,hpd_intervals=[0.84],
                           min_accepted_samples=1000,show_accepted_samples=false % true
                          ]),
  nl,
  % show_store_lengths,nl,
  fail,
  nl.
go => true.
  
model([As,Bs]) =>
  Len = As.len,

  SimAll = uniform(0,1),
  Sim = flip_n(SimAll,Len),

  PA = uniform_dist_n(0,1,Len),
  PB = [ cond(Sim[I] == true, PA[I], uniform(0,1)) : I in 1..Len],

  ClicksA = [flip(PA[I]) : I in 1..Len],
  ClicksB = [flip(PB[I]) : I in 1..Len],
  
  observe(As==ClicksA),
  observe(Bs==ClicksB),  
  
  if observed_ok then
    add("sim all",SimAll),
    add("sim 1",Sim[1]),
  end.