def main(P, src, model):
samples = []
# read data
for line in src:
line = line.strip()
if not line[0].isdigit(): continue
d = line.split()
proto = 1 if d[1] == "TCP" else 2
port = int(d[2])
gt = d[-1]
b = int((len(d)-4) / 2)
pl_up = [int(x) for x in d[3:3+b]][0:P.i]
pl_down = [int(x) for x in d[3+b:-1]][0:P.i]
if pl_up[0] == -1 or pl_down[0] == -1: continue
v = [proto,port] + pl_up + pl_down
samples.append((v, gt))
# load model
cls = DT(verb=True)
cls.load(model)
# test
(acc, ratio, err) = cls.score([x[0] for x in samples], [x[1] for x in samples])
print("ok %.3f%%\tin %.3f%%\tof %d K total (%d errors)" %
(acc * 100.0, ratio * 100.0, len(samples)/1000.0, err))
def main(P, src, model):
samples = []
# read data
for line in src:
line = line.strip()
if not line[0].isdigit(): continue
d = line.split()
proto = 1 if d[1] == "TCP" else 2
port = int(d[2])
gt = d[-1]
stats = [int(x) for x in d[3:-1]]
if stats[4] == 0 or stats[12] == 0: continue
v = [proto,port] + stats
samples.append((v, gt))
# load model
cls = DT()
cls.load(model)
cls.algo.set_params(n_jobs=-1)
# test
(acc, ratio, err) = cls.score([x[0] for x in samples], [x[1] for x in samples])
print("ok %.3f%%\tin %.3f%%\tof %d K total (%d errors)" %
(acc * 100.0, ratio * 100.0, len(samples)/1000.0, err))
def main(P, src, model):
samples = []
# read data
for line in src:
line = line.strip()
if not line[0].isdigit(): continue
d = line.split()
proto = 1 if d[1] == "TCP" else 2
port = int(d[2])
gt = d[-1]
b = int((len(d)-4) / 2)
szup = [int(x) for x in d[3:3+b]][0:P.i]
szdown = [int(x) for x in d[3+b:-1]][0:P.i]
if szup[0] == 0 or szdown[0] == 0: continue
v = [proto,port] + szup + szdown
samples.append((v, gt))
# load model
#cls = kNN(k=P.k, verb=True)
#cls = kNN(k=P.k)
cls = DT()
cls.load(model)
cls.algo.set_params(n_jobs=-1)
# test
(acc, ratio, err) = cls.score([x[0] for x in samples], [x[1] for x in samples])
print("ok %.3f%%\tin %.3f%%\tof %d K total (%d errors)" %
(acc * 100.0, ratio * 100.0, len(samples)/1000.0, err))
def main(P, src, dst):
samples = []
total = 0
# read data
for line in src:
line = line.strip()
if not line[0].isdigit(): continue
total += 1
d = line.split()
proto = 1 if d[1] == "TCP" else 2
port = int(d[2])
gt = d[-1]
b = int((len(d)-4) / 2)
pl_up = [int(x) for x in d[3:3+b]][0:P.i]
pl_down = [int(x) for x in d[3+b:-1]][0:P.i]
if pl_up[0] == -1 or pl_down[0] == -1: continue
v = [proto,port] + pl_up + pl_down
samples.append((v, gt))
print("read %d samples out of %d total (%.2f%%)" % (len(samples), total, 100.0*len(samples)/total))
# take random samples
if P.t > 0:
samples = random.sample(samples, P.t+P.T)
train = samples[:P.t]
test = samples[P.t:]
else:
train = samples
test = []
# train
knc = DT()
knc.fit([x[0] for x in train], [x[1] for x in train])
# test
if len(test) > 0:
(acc, ratio, err) = knc.score([x[0] for x in test], [x[1] for x in test])
print("ok %.3f%%\tin %.3f%%\tof %d K total (%d errors)" %
(acc * 100.0, ratio * 100.0, len(test)/1000.0, err))
# store model
if dst: knc.store(dst)