def report_observation_with_db(ndb, service, fp):
"""Insert or update an Observation record in the notary database."""
cur_time = int(time.time())
obs = ndb.get_observations(service)
most_recent_time_by_key = {}
most_recent_key = None
most_recent_time = 0
# calculate the most recently seen key
for (service, key, start, end) in obs:
if key not in most_recent_time_by_key or end > most_recent_time_by_key[key]:
most_recent_time_by_key[key] = end
for k in most_recent_time_by_key:
if most_recent_time_by_key[k] > most_recent_time:
most_recent_key = k
most_recent_time = most_recent_time_by_key[k]
ndb.close_session()
if most_recent_key == fp:
# this key matches the most recently seen key before this observation.
# just update the observation 'end' time.
ndb.update_observation_end_time(service, fp, most_recent_time, cur_time)
ndb.report_metric('ServiceScanKeyUpdated', service)
else:
# the key has changed or no observations exist yet for this service.
# add a new entry for this key with start and end set to the current time
ndb.insert_observation(service, fp, cur_time, cur_time)
if most_recent_key != None:
# if there was a previous key, set its 'end' timespan value to be
# the current time minus one second (ending just before the new key)
ndb.update_observation_end_time(service, most_recent_key, most_recent_time, cur_time -1)
ndb.report_metric('ServiceScanPrevKeyUpdated', service)
def import_records(infile):
"""Read a file of tuples and extract service and observation data."""
lines = infile.readlines()
infile.close()
service_names = {}
observations = []
num_lines = 0
# tuples will be formatted like this:
# (domain.com:443,2, aa:bb:cc:dd:ee:ff, 123, 456)
valid_tuple = re.compile("^\(([\w:,.]+), *([0-9a-fA-F:]+), *(\d)+, *(\d+)\)$")
for line in lines:
# remember: ALL INPUT IS EVIL!
# test each line before passing to the database.
if (valid_tuple.match(line)):
match = valid_tuple.match(line)
service = str(match.group(1))
key = str(match.group(2))
start = int(match.group(3))
end = int(match.group(4))
if (service not in service_names):
service_names[service] = True
observations.append((service, key, start, end))
num_lines += 1
if (num_lines) % 1000 == 0:
print "Finished %s lines..." % num_lines
print "Found %s services. Adding to database." % (len(service_names))
for name in service_names.keys():
ndb.insert_service(name)
if not args.services_only:
print "Found %s observations. Adding to database." % len(observations)
for (service, key, start, end) in observations:
ndb.insert_observation(service, key, start, end)
