simulate the dass

simulate.py

+"""..."""
+
+from logic.states import init, progress, absorbing, verify_states
+from logic.ads import verify_ads, eligible_in_state
+from util.validate import verify, seq, prob
+from util.stats import describe
+
+
+def sample_impressibilites(impress_function_per_add, n):
+    """
+    :param impress_function_per_add: dictionary of functions for each ad that will
+           sample impressibility for given user numbers
+    :param n: number of users to sample for
+    """
+    imp = {ad: f(n) for ad, f in impress_function_per_add.items()}
+    verify(seq(seq(prob)), list(imp.values()))
+    return imp
+
+
+def streams(states, ads, impress, n=10):
+    """Generate n streams based on states
+    :param states: dictionary as defined verify_states
+    :param ads: dictionary as defined in verify_ads
+    :param impress: input function for sample_impressibilites
+    :param n: number of streams"""
+
+    verify_states(states)
+    verify_ads(ads)
+
+    # impressibility per ad per user
+    impress_ad_user = sample_impressibilites(impress, n)
+
+    absorbing_states = absorbing(states)
+    streams_all_users = []
+    views_all_users = []
+
+    for user in range(n):
+        state = init(states)  # current state
+        stream = [state]      # names of states user traverses
+        views = []            # ad views as (ad,index)
+        i = 0
+        while state not in absorbing_states:
+
+            # pick first ad eligible in current state
+            ad_options = eligible_in_state(ads, states['names'][state])
+            if ad_options:
+                ad = ad_options[0]
+                ad_name = ads['names'][ad]
+
+                # If user targeted by ad type?"
+                # print(ad, user, impress_ad_user)
+                if impress_ad_user[ad_name][user] >= ads['min_impress'][ad]:
+                    views.append((ad_name, i))
+
+            state = progress(state, states)
+            stream.append(state)
+
+            i += 1
+
+        streams_all_users.append(stream)
+        views_all_users.append(views)
+
+    views_stats = describe(list(map(len, views_all_users)))
+
+    return {
+        'stats': {'users': n, 'views': views_stats},
+        'streams': streams_all_users,
+        'views': views_all_users,
+        'impressibilities': impress_ad_user
+    }
+
+