TShark to Python DPI and save in Mongo

dpi_parser.py

__author__ = 'cloudjunky@gmail.com'

import sys
import hashlib
import pymongo

"""
:usage: tshark -T fields -e ip.src -e tcp.srcport -e ip.dst -e tcp.dstport -e frame.protocols -r capture.pcap | python dpi_parser.py

"""
print "Processing stdin....."
print sys.argv

flows = {}

for line in sys.stdin:
    try:
        src, sport, dst, dport, proto = line.strip().split('\t')
        flow = (src, int(sport)), (dst, int(dport))
        flow_hash = hashlib.sha1(str(flow)).hexdigest()
        #print "{} => {}".format(flow, flow_hash)
        if flow_hash in flows:
            pass
        else:
            if proto:
                layers = proto.split(':')
                if 'tcp' in layers:
                    flows[flow_hash] = dict(hash=flow_hash, src=src, sport=sport,
                                            dst=dst, dport=dport, flow=flow, proto=layers[3])
                    if layers[4:]:
                        flows[flow_hash].update(app=layers[4:])
    except Exception, e:
        #print "{}\nerror in line {}".format(e, line)
        continue

print "Found application data in {} flows".format(len(flows.keys()))

conn = pymongo.Connection('localhost', 27017)
db = conn.packetloop
collection = db['captures']
dpis = db.dpis

print "Writing to Mongo"
for id, flow in flows.iteritems():
    db.dpis.insert(flow)
print "Done!"