daniel99 has asked for the wisdom of the Perl Monks concerning the following question:

I need to extract the data elements in a data file containing curly brace neted data to produce a list of NAT pairs. For example, here is a dump from our Palo Alto FW: 

------------------
Sample input

         {
            {
                {
                Prod1_Core_10.2.6.1 {
                  to WAN;
                  from WAN;
                  source any;
                  destination 8.39.63.3;
                  service service-https;
                  to-interface any;
                  destination-translation {
                    translated-address 10.2.6.1;
                  }
                  description "Static NAT to Core 10.2.6.1";
                }
                Prod1_Core_10.2.6.2 {
                  to WAN;
                  from WAN;
                  source any;
                  destination 8.39.63.4;
                  service service-https;
                  to-interface any;
                  destination-translation {
                    translated-address 10.2.6.2;
                  }
                  description "Static NAT to Core 10.2.6.2";
                }
            }
         }

end sample input
---------------

[download]

Essentially, we want to extract these values:

    >>>>     Prod1_Core_10.2.6.2 {
                  to WAN;
                  from WAN;
                  source any;
    >>>>     destination 8.39.63.4; 
                  service service-https;
                  to-interface any;
                  destination-translation {
    >>>>     translated-address 10.2.6.2;  <<<<

[download]

From this data we need to produce a list of the nat pairs like:

queue = [
  {:corename => "Prod1_Core_10.2.6.1", :pubip => "8.39.63.3",  :privip
+ => "10.2.6.1"},
  {:corename => "Prod1_Core_10.2.6.2", :pubip => "8.39.63.4", :privip 
+=> "10.2.6.2"},
  ]

[download]

Is there a parser that will be able to map the data blocks in the nested data into single entities like a string?

any guidance is greatly appreciated :)

Daniel

2018-05-02 Athanasius changed pre to paragraph and code tags

Replies are listed 'Best First'.
Re: Parsing { } nested data in a file
by NetWallah (Canon) on May 02, 2018 at 00:56 UTC
    Here you go: 
    use strict;
use warnings;

local $/="}\n";
my ($currentitem, %item);

while (<>){
   #print "---$_;;;\n";
   for my $line (split "\n"){
     if ($currentitem){
        if ( $line=~ m/^\s*(\S+)\s*\{\s*$/){
           #ignore "destination-translation"# ;
         }elsif (  $line=~/(\S+)\s+(\S+)\s*;/){
            $item{$currentitem}{$1}=$2;
         }
     }elsif( $line=~ m/^\s*(\S+)\s*\{\s*$/){
       $currentitem = $1
     }
   }
   $currentitem=undef;
}

for my $name (sort keys %item){
  print qq|  \{:corename => "$name", :pubip => "$item{$name}{destinati
+on}",  :privip => "$item{$name}{'translated-address'}"},\n|;

}

    [download]
    Run: 
    >perl 1213915.pl YourFileName.txt
  {:corename => "Prod1_Core_10.2.6.1", :pubip => "8.39.63.3",  :privip
+ => "10.2.6.1"},
  {:corename => "Prod1_Core_10.2.6.2", :pubip => "8.39.63.4",  :privip
+ => "10.2.6.2"},

    [download]
    Note: This depends on the format of the file having consistent braces. Not the smartest parser - just a hack.

                    Memory fault   --   brain fried

[reply]
[d/l]
[select]
      Beautiful hack! thanks so much NetWallah :)
[reply]
Re: Parsing { } nested data in a file
by Anonymous Monk on May 02, 2018 at 15:38 UTC
    It looks to me like a very simple awk-like approach would serve you just as well, if the data appears with one item per-line. Ignoring the braces completely, you can simply look for an identifier, followed by something, followed by a semicolon, grabbing the identifier and the something. (Or, an identifier followed by a left-brace.) Since you know the sequence of the items in the file you can simply process the file line-by-line in this way.
[reply]

      It If it would be very simple, please take 5 minutes to write the code to demonstrate the idea. Perl does awk quite nicely after all and the more approaches demonstrated, the better.

[reply]
        Instead, it is quite reasonable to assume that the OP was not born yesterday. Every significant \n-delimited line of any interest within the file as presented contains an instantly-identifiable, regex-friendly pattern. Long before a man named Larry Wall decided that he could do awk one-better, awk was solving pragmatic problems more-or-less exactly like this one. Translation of this concept into actual Perl source-code is left as an exercise to the reader, and s/he has about thirty years' worth of source-code to get ideas from.
[reply]

Back to Seekers of Perl Wisdom