gpg-crawler.py

#!/usr/bin/python

import urllib
import base64
import subprocess
import os
import fcntl
import sys

KEYSERVER_URL = "http://pgp.mit.edu/pks/lookup?op=get&search=0x%s"

def get_key(keyid):
    filename = os.sep.join(["keys", keyid])
    try:
        f = open(filename)
        ret = f.read()
        f.close()
    except:
        # Download the key from the key server and extract it from the
        # response.
        raw_html = urllib.urlopen(KEYSERVER_URL % keyid).read()
        key_start_pos = raw_html.find('\n\n') + 2
        key_end_pos = raw_html.find('\n-----END PGP PUBLIC KEY BLOCK-----')
        key = raw_html[key_start_pos:key_end_pos]
        ret = base64.b64decode(key)
        f = open(filename, "w")
        f.write(ret)
        f.close()
    return ret

def parse_key(decoded):
    # The --batch mode will prevent GnuPG from asking us for a password in case
    # it recognizes the key as encrypted data or something.
    s = subprocess.Popen(["/usr/bin/gpg", "--list-packets", "--batch"],
                         stdin=subprocess.PIPE,
                         stdout=subprocess.PIPE)

    # The key servers are full of keys corrupt in many strange ways - it might
    # happen that GnuPG will report us some problem and won't read from stdin
    # any more. To solve the problem, I decided to switch it to non-blocking
    # mode.
    fd = s.stdin.fileno()
    fl = fcntl.fcntl(fd, fcntl.F_GETFL)
    fcntl.fcntl(fd, fcntl.F_SETFL, fl | os.O_NONBLOCK)

    try:
        s.stdin.write(decoded)
    except IOError:
        s.terminate()
        return {'uid': '', 'sigs': []}
    s.stdin.flush()
    s.stdin.close()
    out = s.stdout.read()
    s.stdout.close()
    s.poll()

    sigs = []
    mode = ""
    uid = "?"
    for line in out.split('\n'):
        if line.startswith(':user ID packet: '):
            uid_raw = line.split(':user ID packet: ')[1]
            uid_without_quotes = uid_raw[1:-1]
            uid = uid_without_quotes.decode('string-escape')
        elif line.startswith(':signature packet: '):
            sig = line.split('keyid ')[1]
            sigs += [sig]

    # Return the UID and unique signatures.
    sigs = list(set(sigs))
    return {'uid': uid, 'sigs': sigs}

def traverse_DFS(start):
    todo = [[start, [start]]]
    visited = {}
    while 0 < len(todo):
        (node, path) = todo.pop(0)
        if node in visited:
            continue
        else:
            visited[node] = True
        nodes = parse_key(get_key(node))
        print("n=%d, uid=%s [%s]" % (len(todo), nodes['uid'], node))
        for next_node in nodes['sigs']:
            if next_node in path:
                continue
            else:
                todo.append([next_node, path + [next_node]])

if __name__ == "__main__":
    traverse_DFS('7909761da0a42745')