make it track log size

1 files changed, 15 insertions, 4 deletions

diff --git a/cowrieder.py b/cowrieder.py
index 2f778ee..4162a40 100644
--- a/cowrieder.py
+++ b/cowrieder.py
@@ -190,7 +190,15 @@ def session_pingcmds(session): # Return array of all "command ping" commands use
           ra.append(event['input'])
   return ra
 
-# Loading da file(s) into da master array
+# File loading stuff
+
+total_filesize = 0
+def sizeof_fmt(num, suffix="B"): # Shamelessly stolen from StackOverflow
+    for unit in ("", "Ki", "Mi", "Gi", "Ti", "Pi", "Ei", "Zi"):
+        if abs(num) < 1024.0:
+            return f"{num:3.1f}{unit}{suffix}"
+        num /= 1024.0
+    return f"{num:.1f}Yi{suffix}"
 
 def load_json_to_array(fp):
   with open(fp, "r") as logfile:
@@ -210,6 +218,7 @@ def load_json_to_array(fp):
 
 if(os.path.isfile(filepath)):
   load_json_to_array(filepath)
+  total_filesize += os.path.getsize(filepath)
 elif(os.path.isdir(filepath)):
   # run load_json_to_array for every json file in the directory
   files = {}
@@ -219,6 +228,7 @@ elif(os.path.isdir(filepath)):
       files.update({fi: os.path.getmtime(fi)})
   for g in dict(sorted(files.items(), key=lambda item: item[1])):
     load_json_to_array(g)
+    total_filesize += os.path.getsize(g)
 else:
   raise FileNotFoundError
 
@@ -462,7 +472,8 @@ if(verbosity == 3): # Maximum verbosity: literally just dump everything we have
       'passwords_by_attempt_count': passwords_by_attempt_count,
       'passwords_by_username': passwords_by_username,
       'login_combos_by_ip': login_combos_by_ip,
-      'log_parsing_time': log_parsing_time
+      'log_parsing_time': log_parsing_time,
+      'total_log_size': total_filesize
     }
     outjson.update({'stats': all_stats})
 
@@ -495,7 +506,7 @@ elif(verbosity == 2): # High verbosity mode: print timestamp, event type and mes
     print("\n--- Credential combos grouped by source IP: ---")
     pprint(login_combos_by_ip, sort_dicts=False)
     print_percentage_stats()
-    print("Log took {}s to parse".format(str(log_parsing_time)[:4]))
+    print("---\nTook {}s to parse {} of logs".format(str(log_parsing_time)[:4], sizeof_fmt(total_filesize)))
 
 
 
@@ -575,4 +586,4 @@ else: # Default (low) verbosity: just the good stuff
       print("\n--- Top IPs with credential combos tried: ---")
       print_dict_of_lists(login_combos_by_ip, truncate=7, limit = 20, key_is_ip=True)
     print_percentage_stats()
-    print("Log took {}s to parse".format(str(log_parsing_time)[:4]))
\ No newline at end of file
+    print("---\nTook {}s to parse {} of logs".format(str(log_parsing_time)[:4], sizeof_fmt(total_filesize)))
\ No newline at end of file