Added burst statistics and improved error checking.

src/feature-extractor/sample.py

@@ -28,7 +28,7 @@ class Sample:
     def set_activity_thresholds(lower_bound: float, upper_bound: float,
                                 lookback: float):
         assert(lower_bound < upper_bound)
-        assert(0 < lower_bound)
+        assert(lookback <= lower_bound)
         assert(0 < lookback)
         Sample.high_act_threshold = upper_bound
         Sample.low_act_threshold = lower_bound
@@ -90,32 +90,41 @@ class Sample:
         }
 
     def __count_activity_stats(arr):
-        (time_spent, iat) = Sample.__get_time_spent(arr)
+        (time_spent, iat, burst_count) = Sample.__get_time_spent(arr)
         return {
             "total_packets": len(arr),
             "time_spent": time_spent,
-            "average_iat": iat
+            "average_iat": iat,
+            "burst_count": burst_count,
+            "avg_burst_size": burst_count/len(arr)
         }
 
     def __is_continuous(p1, p2):
         return int(p1.index) + 1 == int(p2.index)
     
+    def __time_since_last_packet(p):
+        return Sample.__packet_time(p) - min(p.delta, Sample.lookback)
+        
     def __get_time_spent(arr):
         if not arr:
             return (0.0, 0.0)
         else:
             time_spent = 0.0
             p_in_seg = 0
-            start = Sample.__packet_time(arr[0]) - min(arr[0].delta, Sample.lookback)
+            burst_count = 1
+            start = Sample.__time_since_last_packet(arr[0])
             for prev, cur in window(arr):
                 if not Sample.__is_continuous(prev, cur):
                     time_spent += Sample.__packet_time(prev) - start
-                    start = Sample.__packet_time(cur) - min(cur.delta, Sample.lookback)
+                    start = Sample.__time_since_last_packet(cur)
+                    burst_count += 1
                 else:
                     # Helps deal with "Lone Wolf" packets
                     p_in_seg += 1
             time_spent += Sample.__packet_time(arr[-1]) - start
-            return (time_spent, p_in_seg / time_spent if time_spent != 0 else 0)
+            return (time_spent,
+                    p_in_seg / time_spent if time_spent != 0 else 0,
+                    burst_count)
     
     def __extract_time_stats(self, pcap):
         start = Sample.__packet_time(pcap[0])