#!/usr/bin/perl

# harvest.pl - given a set of URI's, cache MARCXML locally

# Eric Lease Morgan <eric_morgan@infomotions.com>
# March 29, 2016 - first cut


# configure
use constant WGET => 'wget -q -O ##FILE## ##URI##';

# require
use strict;

# sanity check
my $db = $ARGV[ 0 ];
if ( ! $db ) {

	print "Usage: $0 <file>\n";
	exit;
	
}

# open the database file and process each record
open LIST, " < $db" or die "Can't open database file ($db): $!\n";
my $index = 0;
while ( <LIST> ) {

	# parse
	chop;
	my ( $id, $briefName, $date, $fullName, $tries, $total, $authorizedName, $uri, $lccn ) = split /\t/, $_;
	
	# only process found items
	next if ( ! $uri );
	
	# configure
	my $file = $id  . '.xml';
	$uri     = $uri . '.marcxml.xml';
	
	# do the work only if the file does not exist
	next if ( -e $file );
	
	# create the command
	my $cmd =  WGET;
	$cmd    =~ s/##FILE##/$file/e;
	$cmd    =~ s/##URI##/$uri/e;
	
	# increment
	$index++;
	
	# do the work
	print "\t$index\t$cmd\r";
	`$cmd`;
	
}

# done
exit;