子流程的输出操作

Question

I'd like to manipulate the output from a subprocess call and only print specific details. Now I was originally going down the PIPE route and piping the output to another subprocess of grep but this seemed too limiting, meaning I could only print a few lines (some of which I didn't want), by using the -A or -B grep arguments. So now I'm taking the stdout.readline route. My question is, is there a better way to do achieve what I am trying to do? Is pattern matching with regular expressions and re the best option?

Below is the code for both trails of thought, just to give some context.

PIPE to another subprocess:

def nmap_ips(ip_list):

    for ip in ip_list:
        cmd1 = subprocess.Popen(["nmap", "-sC", "--open", "-Pn", "-p80,443", ip], stderr=subprocess.STDOUT, stdout=subprocess.PIPE)
        cmd2 = subprocess.Popen(["grep", "title", "-A", "1"], stdin=cmd1.stdout)
        cmd1.stdout.close()
        output = cmd2.communicate()[0]

stdout read lines: #Im not actually getting output from this function at the moment, I'm guessing it's my regex

def nmap_ips(ip_list):
   regs = ["title", "commonName"]
   combined = "|".join(regs)

   for ip in ip_list:
      p = subprocess.Popen(["nmap", "-sC", "--open", "-Pn", "-p80,443", ip], stderr=subprocess.STDOUT, stdout=subprocess.PIPE)
      for line in iter(p.stdout.readline, b''):
         ml = re.match(combined, line)
         if ml:
            print ml.group()

Example of data I am trying to manipulate

Starting Nmap 6.40 ( http://nmap.org ) at 2014-10-10 23:47 BST
Nmap scan report for BTHomeHub.home (192.168.1.254)
Host is up (0.0054s latency).
PORT    STATE SERVICE
80/tcp  open  http
| http-title: Home Hub Manager - Please Reset Your Password
|_Requested resource was http://BTHomeHub.home/html/home/a_firstlogin_configuration.html
443/tcp open  https
| ssl-cert: Subject: commonName=bthomehub.home/countryName=UK
| Not valid before: 2011-01-15T16:52:23+00:00
|_Not valid after:  2024-01-17T16:52:23+00:00
|_ssl-date: 2014-10-10T22:47:18+00:00; +14s from local time.

Answer 1

Your approach is fine. Both grep 'title\\|commonName' and re.search(b'title|commonName', line) should work.

Note: re.match() starts matching at the begining of its input string. re.search() finds the match anywhere in the string.

If you want to save these lines in a variable then the grep process is unnecessary. It is enough to start only nmap .

from subprocess import Popen, PIPE, STDOUT

def nmap_something(ip_list):
    match = re.compile(b"title|commonName").search
    p = Popen(["nmap", "-sC", "--open", "-Pn", "-p80,443"] + ip_list,
              stdout=PIPE, stderr=STDOUT, bufsize=1)
    result = list(filter(match, p.stdout))
    p.stdout.close()
    p.wait()
    return result

I'm not sure about the exact syntax to pass multiple ips to a single nmap child process but it supports it ie, you don't need to start multiple processes to scan multiple ips.

If you want to print ssl certificates for web servers then you don't need nmap , you could use a pure Python solution eg, see How to get server's ssl certificate in a human readable form?

Answer 2

After some advice from JFSebastian, I changed to re.search and found printing the line directly achieved what I was after.

def nmap_ips(ip_list):
    regs = ["title", "commonName"]
    combined = "|".join(regs)

    for ip in ip_list:
        p = subprocess.Popen(["nmap", "-sC", "--open", "-Pn", "-p80,443", ip], stderr=subprocess.STDOUT, stdout=subprocess.PIPE)
        for line in iter(p.stdout.readline, b''):
            ml = re.search(combined, line)
            if ml:
                print line