Add clean_marc function to OpenILS::Utils::Normalize.

Add a library function to clean up MARC records for how we like to
store them in the biblio.record_entry table. Having this in a library
will reduce code duplication.

Also, replace nearly identical code in OpenILS::Application::Vandelay
and OpenILS::Application::Acq::Order with calls to this new function.

Signed-off-by: Jason Stephenson <jstephenson@mvlc.org>

diff --git a/Open-ILS/src/perlmods/lib/OpenILS/Application/Acq/Order.pm b/Open-ILS/src/perlmods/lib/OpenILS/Application/Acq/Order.pm
index 8e478ca..aa4ae5d 100644 (file)
--- a/Open-ILS/src/perlmods/lib/OpenILS/Application/Acq/Order.pm
+++ b/Open-ILS/src/perlmods/lib/OpenILS/Application/Acq/Order.pm
@@ -181,6 +181,7 @@ use OpenSRF::Utils::JSON;
 use OpenSRF::AppSession;
 use OpenILS::Utils::Fieldmapper;
 use OpenILS::Utils::CStoreEditor q/:funcs/;
+use OpenILS::Utils::Normalize qw/clean_marc/;
 use OpenILS::Const qw/:const/;
 use OpenSRF::EX q/:try/;
 use OpenILS::Application::AppUtils;
@@ -1254,13 +1255,7 @@ sub upload_records {
         last unless $r;
 
                try {
-            ($xml = $r->as_xml_record()) =~ s/\n//sog;
-            $xml =~ s/^<\?xml.+\?\s*>//go;
-            $xml =~ s/>\s+</></go;
-            $xml =~ s/\p{Cc}//go;
-            $xml = $U->entityize($xml);
-            $xml =~ s/[\x00-\x1f]//go;
-
+            $xml = clean_marc($r);
                } catch Error with {
                        $err = shift;
                        $logger->warn("Proccessing XML of record $count in set $key failed with error $err.  Skipping this record");

diff --git a/Open-ILS/src/perlmods/lib/OpenILS/Application/Vandelay.pm b/Open-ILS/src/perlmods/lib/OpenILS/Application/Vandelay.pm
index 62dca2e..59a2fc0 100644 (file)
--- a/Open-ILS/src/perlmods/lib/OpenILS/Application/Vandelay.pm
+++ b/Open-ILS/src/perlmods/lib/OpenILS/Application/Vandelay.pm
@@ -9,6 +9,7 @@ use OpenSRF::Utils::SettingsClient;
 use OpenSRF::Utils::Cache;
 use OpenILS::Utils::Fieldmapper;
 use OpenILS::Utils::CStoreEditor qw/:funcs/;
+use OpenILS::Utils::Normalize qw/clean_marc/;
 use MARC::Batch;
 use MARC::Record;
 use MARC::File::XML ( BinaryEncoding => 'UTF-8' );
@@ -286,12 +287,7 @@ sub process_spool {
         $logger->info("processing record $count");
 
         try {
-            (my $xml = $r->as_xml_record()) =~ s/\n//sog;
-            $xml =~ s/^<\?xml.+\?\s*>//go;
-            $xml =~ s/>\s+</></go;
-            $xml =~ s/\p{Cc}//go;
-            $xml = $U->entityize($xml);
-            $xml =~ s/[\x00-\x1f]//go;
+            my $xml = clean_marc($r);
 
             my $qrec;
             # Check the leader to ensure we've got something resembling the expected

diff --git a/Open-ILS/src/perlmods/lib/OpenILS/Utils/Normalize.pm b/Open-ILS/src/perlmods/lib/OpenILS/Utils/Normalize.pm
index d71503c..fae6403 100644 (file)
--- a/Open-ILS/src/perlmods/lib/OpenILS/Utils/Normalize.pm
+++ b/Open-ILS/src/perlmods/lib/OpenILS/Utils/Normalize.pm
@@ -3,9 +3,13 @@ use strict;
 use warnings;
 use Unicode::Normalize;
 use Encode;
+use UNIVERSAL qw/isa/;
+use MARC::Record;
+use MARC::File::XML ( BinaryEncoding => 'UTF-8' );
+use OpenILS::Application::AppUtils;
 
 use Exporter 'import';
-our @EXPORT_OK = qw( naco_normalize );
+our @EXPORT_OK = qw( naco_normalize clean_marc );
 
 sub naco_normalize {
 
@@ -67,4 +71,24 @@ sub naco_normalize {
     return lc $str;
 }
 
+# Cleans up a MARC::Record or MARCXML string for storage in the
+# Open-ILS database.
+#
+# Takes either a MARC::Record or a string of MARCXML.
+#
+# Returns a string of MARCXML as Open-ILS likes to store it.
+#
+# Assumes input is already in UTF-8.
+sub clean_marc {
+    my $input = shift;
+    my $xml = (isa $input, 'MARC::Record') ? $input->as_xml_record() : $input;
+    $xml =~ s/\n//sog;
+    $xml =~ s/^<\?xml.+\?\s*>//go;
+    $xml =~ s/>\s+</></go;
+    $xml =~ s/\p{Cc}//go;
+    $xml = OpenILS::Application::AppUtils->entityize($xml);
+    $xml =~ s/[\x00-\x1f]//go;
+    return $xml;
+}
+
 1;