# The user agent is at the end of the line. It may contain space, so we concatenate.
for i in range(12, len(sline)):
sline[11] += ' ' + sline[i]
# Remove leftovers
del sline[12:]
m = PATTERN_IPV4.match(sline[0])
sortkey = None
ip = None
if m is None:
m = PATTERN_IPV6.match(sline[0])
if m is None:
m = PATTERN_ONION.match(sline[0])
if m is None:
return None
else:
net = 'onion'
ipstr = sortkey = m.group(1)
port = int(m.group(2))
else:
net = 'ipv6'
# Not interested in localhost
if m.group(1) in ['::']:
return None
ipstr = m.group(1)
# XXX parse IPv6 into number, could use name_to_ipv6 from generate-seeds
sortkey = ipstr
port = int(m.group(2))
else:
# Do IPv4 sanity check
ip = 0
for i in range(0, 4):
if int(m.group(i + 2)) < 0 or int(m.group(i + 2)) > 255:
return None
ip = ip + (int(m.group(i + 2)) << (8 * (3 - i)))
if ip == 0:
return None
net = 'ipv4'
sortkey = ip
ipstr = m.group(1)
port = int(m.group(6))
# Skip bad results.
if sline[1] == 0:
return None
# Extract uptime %.
uptime30 = float(sline[7][:-1])
# Extract Unix timestamp of last success.
lastsuccess = int(sline[2])
# Extract protocol version.
version = int(sline[10])
# Extract user agent.
agent = sline[11][1:-1]
# Extract service flags.
service = int(sline[9], 16)
# Extract blocks.
blocks = int(sline[8])
# Construct result.
return {
'net': net,
'ip': ipstr,
'port': port,
'ipnum': ip,
'uptime': uptime30,
'lastsuccess': lastsuccess,
'version': version,
'agent': agent,
'service': service,
'blocks': blocks,
'sortkey': sortkey,
}
def filtermultiport(ips):
'''Filter out hosts with more nodes per IP'''
hist = collections.defaultdict(list)
for ip in ips:
hist[ip['sortkey']].append(ip)
return [value[0] for (key, value) in list(hist.items()) if len(value) == 1]
# Based on Greg Maxwell's seed_filter.py
def filterbyasn(ips, max_per_asn, max_total):
# Sift out ips by type
ips_ipv4 = [ip for ip in ips if ip['net'] == 'ipv4']
ips_ipv6 = [ip for ip in ips if ip['net'] == 'ipv6']
ips_onion = [ip for ip in ips if ip['net'] == 'onion']
# Filter IPv4 by ASN
result = []
asn_count = {}
for ip in ips_ipv4:
if len(result) == max_total:
break
try:
asn = int([x.to_text()
for x in dns.resolver.query('.'.join(reversed(ip['ip'].split('.'))) + '.origin.asn.cymru.com', 'TXT').response.answer][0].split('\"')[1].split(' ')[0])
if asn not in asn_count:
asn_count[asn] = 0
if asn_count[asn] == max_per_asn:
continue
asn_count[asn] += 1
result.append(ip)
except:
sys.stderr.write(
'ERR: Could not resolve ASN for "' + ip['ip'] + '"\n')
# TODO: filter IPv6 by ASN
# Add back non-IPv4
result.extend(ips_ipv6)
result.extend(ips_onion)
return result
def main():
lines = sys.stdin.readlines()
ips = [parseline(line) for line in lines]
# Skip entries with valid address.
ips = [ip for ip in ips if ip is not None]
# Skip entries from suspicious hosts.
ips = [ip for ip in ips if ip['ip'] not in SUSPICIOUS_HOSTS]
# Enforce minimal number of blocks.
ips = [ip for ip in ips if ip['blocks'] >= MIN_BLOCKS]
# Require service bit 1.
ips = [ip for ip in ips if (ip['service'] & 1) == 1]
# Require at least 50% 30-day uptime.
ips = [ip for ip in ips if ip['uptime'] > 50]
# Require a known and recent user agent.
ips = [ip for ip in ips if PATTERN_AGENT.match(ip['agent'])]
# Sort by availability (and use last success as tie breaker)