add functions to manage more than 1 query variables

bn_functions.py

@@ -60,7 +60,7 @@ def average_prob(ref_cpds_table, current_cpds_table):
             res_cpds_table[elem1][elem2] = (ref_cpds_table[elem1][elem2]+current_cpds_table[elem1][elem2])/2
     return res_cpds_table
 
-def update_cpds_tables(bn_model,variables_tables):
+def update_cpds_tables(bn_model, variables_tables):
     '''
     This function updates the bn model with the variables_tables provided in input
     Args:
@@ -144,6 +144,7 @@ def get_stochastic_action(actions_distr_prob):
                     index = i
         return index
 
+
     actions_distr_prob_scaled = [0]*len(actions_distr_prob)
     accum = 0
     for i in range(len(actions_distr_prob)):
@@ -155,27 +156,60 @@ def get_stochastic_action(actions_distr_prob):
 
     return action_id
 
-def interpret_user_output(action_id):
+def flat_action_probs(action_probs):
+    flat_array_user_action_prob = None
+    column = 0
+    if len(action_probs.values.shape)==2:
+        flat_array_user_action_prob = [action_probs.values[j][i] for j in range(action_probs.values.shape[0]) for i in range(action_probs.values.shape[1])]
+        column = action_probs.values.shape[0]
+        row = action_probs.values.shape[1]
+    else:
+        assert "Did you forget to add the additional target, only one has been detected"
+        flat_array_user_action_prob = action_probs.values
+        column = 1
+        row = 0
+    return flat_array_user_action_prob, column, row
+
+
+
+def interpret_action_output(action_id, col, row, targets):
+    '''
+       Given the id of the action selected from the probabilistic inference model and the target variables
+        return the action (user act + react time) or (robot ass and robot feedback)
+       Args
+           action_id 1d array of probs
+           col: #col of array action id
+           row: #row of array action id
+           targets the targets we aim to evaluate
+       Return:
+           user_action
+           user_react_time
+       '''
+    #N.B it assumes that the query is performed passing as first argument robot_assistance
+    # and as a second robot_feedback
+    robot_assistance = 0
+    robot_feedback = 0
     user_action = 0
     user_react_time = 0
 
-    if action_id == 0:
-        user_action = 0; user_react_time = 0
-    elif action_id == 1:
-        user_action = 1; user_react_time = 0
-    elif action_id == 2:
-        user_action = 2; user_react_time = 0
-    elif action_id == 3:
-        user_action = 0; user_react_time = 1
-    elif action_id == 4:
-        user_action = 1; user_react_time = 1
-    elif action_id == 5:
-        user_action = 2; user_react_time = 1
-    elif action_id == 6:
-        user_action = 0; user_react_time = 2
-    elif action_id == 7:
-        user_action = 1; user_react_time = 2
-    elif action_id == 8:
-        user_action = 2; user_react_time = 2
-
-    return user_action, user_react_time
+    if targets[0] == 'user_action':
+        user_action = int(action_id / row)
+        user_react_time = int(action_id % row)
+        print("user_action ", user_action, ' user_react ', user_react_time)
+        return user_action, user_react_time
+    elif targets[1] == 'user_action':
+        user_action = int(action_id % row)
+        user_react_time = int(action_id / row)
+        print("user_action ", user_action, ' user_react ', user_react_time)
+        return user_action, user_react_time
+    elif targets[0] == "robot_assistance":
+        robot_assistance = int(action_id / row)
+        robot_feedback = int(action_id % row)
+        print("robot_ass ", robot_assistance, 'robot_feed ', robot_feedback)
+        return robot_assistance, robot_feedback
+    elif targets[1] == "robot_assistance":
+        robot_assistance = int(action_id % row)
+        robot_feedback = int(action_id / row)
+        print("robot_feed ", robot_assistance, 'robot_ass ', robot_feedback)
+        return robot_assistance, robot_feedback
\ No newline at end of file
+