adding code for analyzing overlap

1 files changed, 63 insertions, 7 deletions

diff --git a/monitor/josef_mover.py b/monitor/josef_mover.py
index 08ec05d..ac79448 100755
--- a/monitor/josef_mover.py
+++ b/monitor/josef_mover.py
@@ -50,10 +50,10 @@ def print_reply(rep, entry):
 def is_new_timestamp(ts):
     MAX_TIMEDIFF = 300 # 5 min, allows for some clock skew
     ts_time = datetime.datetime.fromtimestamp(ts / 1000, UTC()).strftime('%Y-%m-%d %H:%M:%S')
-    start_time = datetime.datetime.utcnow().strftime('2013-10-19 00:00:00')
-    delta_time = datetime.datetime.strptime(start_time, '%Y-%m-%d %H:%M:%S') - datetime.datetime.strptime(ts_time, '%Y-%m-%d %H:%M:%S')
+    start_time = datetime.datetime.utcnow().strftime('2015-10-19 00:00:00')
+    # delta_time = datetime.datetime.strptime(start_time, '%Y-%m-%d %H:%M:%S') - datetime.datetime.strptime(ts_time, '%Y-%m-%d %H:%M:%S')
     # print delta_time.seconds
-    if delta_time.seconds > MAX_TIMEDIFF:
+    if ts_time < start_time:
         return False
     else:
         return True
@@ -123,17 +123,73 @@ def move_entry(first, last, source, dest):
                     print "FAILED!\n"
             print s_log["name"] + "[" + str(first + i) + "] found in " + str(len(inclusions)) + " logs: ", inclusions
 
+def check_inclusion_by_submission(first, last, source, dest):
+    # print entries
+    for s_log in source:
+        try:
+            entries = get_entries(s_log["url"], first, last)["entries"]
+        except:
+            log("Failed to get entries from " + s_log["name"])
+
+        # print "\n\nSource: " + s_log["name"] + "\n"
+        for i in range(len(entries)):
+        # for item in entries:
+            item = entries[i]
+            inclusions = []
+            for d_log in dests:
+                # print "Log: " + d_log["name"]
+                try:
+                    entry = extract_original_entry(item)
+                    if entry[2]:
+                        precert = True
+                    else:
+                        precert = False
+                    submission = []
+
+                    for e in entry[0]:
+                        submission.append(base64.b64encode(e))
+
+                    if entry[2]:
+                        res = add_prechain(d_log["url"], {"chain" : submission})
+                    else:
+                        res = add_chain(d_log["url"], {"chain" : submission})
+                    # print_reply(res, entry)
+
+                    if not is_new_timestamp(res["timestamp"]):
+                        inclusions.append(d_log["name"])
+
+                    # time.sleep(5)
+                except KeyboardInterrupt:
+                    sys.exit()
+                except:
+                    # print "FAILED!\n"
+                    pass
+            s = s_log["name"] + "[" + str(first + i) + "] found in " + str(len(inclusions)) + " logs: " + str(inclusions)
+            print s
+            logfile = OUTPUT_DIR + s_log["name"] + "_overlap.log"
+            log(logfile, s)
+
+def log(fn, string):
+    logfile = fn
+    s = time_str() + " " + string
+    with open(logfile, 'a') as f:
+        f.write(s + "\n")
+        f.close()
 
 if __name__ == "__main__":
-    source = [CTLOGS[0]]
+    source = [CTLOGS[9]]
     dests = CTLOGS
     # source = ctlogs
     # dests = ctlogs
+    for tmp_log in source:
+        sth = get_sth(tmp_log["url"])
+
 
-    first = 100
-    last = 102
+        first = 0
+        last = int(sth["tree_size"]) - 1
+        # print last
 
-    move_entry(first, last, source,dests)
+        check_inclusion_by_submission(first, last, [tmp_log], dests)
     # check_inclusion_all(first,last,source, dests)