How to clean up masscan output (-oG)

I have a problem with the output produced by the masscan utility with the -oG options (“grep-able” output); for instance, it outputs this:

# Masscan 1.0.3 scan initiated Wed Jun  4 01:35:02 2014
# Ports scanned: TCP(3;21-23,) UDP(0;) SCTP(0;) PROTOCOLS(0;)
Host: 192.168.100.19 () Ports: 2222/open/tcp////
Host: 192.168.100.13 () Ports: 2222/open/tcp////
Host: 192.168.100.16 () Ports: 443/open/tcp////
Host: 192.168.100.8 ()  Ports: 21/open/tcp////
Host: 192.168.100.5 ()  Ports: 22/open/tcp////
Host: 192.168.100.5 ()  Ports: 443/open/tcp////
Host: 192.168.100.16 () Ports: 80/open/tcp////
Host: 192.168.100.19 () Ports: 22/open/tcp////
Host: 192.168.100.7 ()  Ports: 80/open/tcp////
Host: 192.168.100.8 ()  Ports: 80/open/tcp////
Host: 192.168.100.12 () Ports: 2222/open/tcp////
Host: 192.168.100.13 () Ports: 22/open/tcp////
# Masscan done at Wed Jun  4 01:35:16 2014

The above is neither very readable nor easy to understand.

How can I use Linux command-line utilities, e.g. sed, awk, or grep, to output something as follows, using the file above?

Host: 192.168.100.5
Ports: 22, 443

Host: 192.168.100.7
Ports: 80

Host: 192.168.100.8
Ports: 21, 80

Host: 192.168.100.12
Ports: 2222

Host: 192.168.100.13
Ports: 2222, 22

......

As you can see, the output is much more readable in this layout: sorted by IP address, with all associated ports listed below, consolidated across multiple input lines with the same IP address.

Answer

Try this:

awk -F' +|/' '
  !/s*#/ {    # ignore comment lines
      # Add the port on the current line to the associative array 
      # element for the IP address on the current line.
    ips[$2] = ips[$2] (ips[$2] == "" ? $5 : ", " $5)
  }
  END {
      # Enumerate all IPs and the ports for each.
      # Since the IPs will be listed in no specific order, the output
      # is piped as a _single_ line to "sort" in order to sort by IP address,
      # and then expanded into 2 lines via "tr".
    for (ip in ips) {
      printf "Host: %s@Ports: %s@n", ip, ips[ip] | 
        "sort -t. -n -k 1.6,1 -k 2,2 -k 3,3 -k 4,4 | tr @ "n""
    } 
  }
  ' file

• This solution properly sorts the output by IP address and separates the ports with commas.
• By contrast, for a given IP address, the port numbers are listed in the order they were encountered in the input (as in the sample output data in the question).