How to extract the text between two patterns using REGEX perl - regex

Question

In the following lines how can I store the lines between "Description:" and "Tag:" in a variable using REGEX PERL and what would be a good datatype to use, string or list or something else?
(I am trying to write a program in Perl to extract the information of a text file with Debian package information and convert it into a RDF(OWL) file(ontology).)
Description: library for decoding ATSC A/52 streams (development)
liba52 is a free library for decoding ATSC A/52 streams. The A/52 standard is
used in a variety of applications, including digital television and DVD. It is
also known as AC-3.
This package contains the development files.
Homepage: http://liba52.sourceforge.net/
Tag: devel::library, role::devel-lib
The code I have written so far is:
#!/usr/bin/perl
open(DEB,"Packages");
open(ONT,">>debianmodelling.txt");
$i=0;
while(my $line = <DEB>)
{
if($line =~ /Package/)
{
$line =~ s/Package: //;
print ONT ' <package rdf:ID="instance'.$i.'">';
print ONT ' <name rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</name>'."\n";
}
elsif($line =~ /Priority/)
{
$line =~ s/Priority: //;
print ONT ' <priority rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</priority>'."\n";
}
elsif($line =~ /Section/)
{
$line =~ s/Section: //;
print ONT ' <Section rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</Section>'."\n";
}
elsif($line =~ /Maintainer/)
{
$line =~ s/Maintainer: //;
print ONT ' <maintainer rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</maintainer>'."\n";
}
elsif($line =~ /Architecture/)
{
$line =~ s/Architecture: //;
print ONT ' <architecture rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</architecture>'."\n";
}
elsif($line =~ /Version/)
{
$line =~ s/Version: //;
print ONT ' <version rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</version>'."\n";
}
elsif($line =~ /Provides/)
{
$line =~ s/Provides: //;
print ONT ' <provides rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</provides>'."\n";
}
elsif($line =~ /Depends/)
{
$line =~ s/Depends: //;
print ONT ' <depends rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</depends>'."\n";
}
elsif($line =~ /Suggests/)
{
$line =~ s/Suggests: //;
print ONT ' <suggests rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</suggests>'."\n";
}
elsif($line =~ /Description/)
{
$line =~ s/Description: //;
print ONT ' <Description rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</Description>'."\n";
}
elsif($line =~ /Tag/)
{
$line =~ s/Tag: //;
print ONT ' <Tag rdf:datatype="http://www.w3.org/2001/XMLSchema#string">'.$line.'</Tag>'."\n";
print ONT ' </Package>'."\n\n";
}
$i=$i+1;
}

Answer 1

my $desc = "Description:";
my $tag = "Tag:";
$line =~ /$desc(.*?)$tag/;
my $matched = $1;
print $matched;
or
my $desc = "Description:";
my $tag = "Tag:";
my #matched = $line =~ /$desc(.*?)$tag/;
print $matched[0];
or
my $desc = "Description:";
my $tag = "Tag:";
(my $matched = $line) =~ s/$desc(.*?)$tag/$1/;
print $matched;
Additional
If your Description and Tag may be on separate lines, you may need to use the /s modifier, to treat it as a single line, so the \n won't wreck it. Example:
$_=qq{Description:foo
more description on
new line Tag: some
tag};
s/Description:(.*?)Tag:/$1/s; #notice the trailing slash
print;

Answer 2

Assuming:
my $example; # holds the example text above
You could:
(my $result=$example)=~s/^.*?\n(Description:)/$1/s; # strip up to first marker
$result=~s/(\nTag:[^\n]*\n).+$/$1/s; # strip everything after second marker line
Or
(my $result=$example)=~s/^.*?\n(Description:.+?Tag:[^\n]*\n).*$/$1/s;
Both assume the Tag: value is contained on a single line.
If this is not the case, you might try:
(my $result=$example)=~s/
( # start capture
Description: # literal 'Description:'
.+? # any chars (non-greedy) up to
Tag: # literal 'Tag:'
.+? # any chars up to
)
(?: # either
\n[A-Z][a-z]+\: # another tagged value name
| # or
$ # end of string
)
/$1/sx;

Answer 3

I believe that the problem is caused by using a line reading loop for data structured by paragraphs. If you can slurp the file into memory and and apply split with a captured delimiter, the processing will be much smoother:
#!/usr/bin/perl -w
use strict;
use diagnostics;
use warnings;
use English;
# simple sample sub
my $printhead = sub {
printf "%5s got the tag '%s ...'\n", '', substr( shift, 0, 30 );
};
# map keys/tags? to functions
my %tagsoups = (
'PackageName' => sub {printf "%5s got the name '%s'\n", '', shift;}
, 'Description' => sub {printf "%5s got the description:\n---------\n%s\n----------\n", '', shift;}
, 'Tag' => $printhead
);
# slurp Packages (fallback: parse using $INPUT_RECORD_SEPARATOR = "Package:")
open my $fh, "<", './Packages-00.txt' or die $!;
local $/; # enable localized slurp mode
my $all = <$fh>;
my #pks = split /^(Package):\s+/ms, $all;
close $fh;
# outer loop: Packages
for (my $p = 1, my $n = 0; $p < scalar #pks; $p +=2) {
my $blk = "PackageName: " . $pks[$p + 1];
my #inf = split /\s*^([\w-]+):\s+/ms, $blk;
printf "%3d %s named %s\n", ++$n, $pks[$p], $inf[ 2 ];
# outer loop: key-value-pairs (or whatever they are called)
for (my $x = 1; $x < scalar #inf; $x += 2) {
if (exists($tagsoups{$inf[ $x ]})) {
$tagsoups{$inf[ $x ]}($inf[$x + 1]);
}
}
}
output for a shortened Packages file from my Ubuntu Linux:
3 Package named abrowser-3.5-branding
got the PackageName:
---------
abrowser-3.5-branding
----------
got the Description:
---------
dummy upgrade package for firefox-3.5 -> firefox
This is a transitional package so firefox-3.5 users get firefox on
upgrades. It can be safely removed.
----------
4 Package named casper
got the PackageName:
---------
casper
----------
got the Description:
---------
Run a "live" preinstalled system from read-only media
----------
got the Tag:
---------
admin::boot, admin::filesystem, implemented-in::shell, protocol::smb, role::plugin, scope::utility, special::c
ompletely-tagged, works-with-format::iso9660
----------
Using a hash for the functions to apply to the extracted parts will keep the details of generating xml out of the parser loops.