From: gmc <gmc@dcc99617-32d9-48b4-a31d-7c20da2025e4>
Date: Mon, 18 Oct 2010 22:56:48 +0000 (+0000)
Subject: fix mapping of the crossed D character in naco_normalize
X-Git-Url: https://old-git.evergreen-ils.org/?a=commitdiff_plain;h=632fee9376fc4a2ea889ab3dcab5576908d50f78;p=evergreen%2Fbjwebb.git

fix mapping of the crossed D character in naco_normalize

Signed-off-by: Galen Charlton <gmc@esilibrary.com>


git-svn-id: svn://svn.open-ils.org/ILS/trunk@18388 dcc99617-32d9-48b4-a31d-7c20da2025e4
---

diff --git a/Open-ILS/src/sql/Pg/002.schema.config.sql b/Open-ILS/src/sql/Pg/002.schema.config.sql
index ad2b8ca83..bcc7a48f4 100644
--- a/Open-ILS/src/sql/Pg/002.schema.config.sql
+++ b/Open-ILS/src/sql/Pg/002.schema.config.sql
@@ -70,7 +70,7 @@ CREATE TABLE config.upgrade_log (
     install_date    TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT NOW()
 );
 
-INSERT INTO config.upgrade_log (version) VALUES ('0440'); -- dbs
+INSERT INTO config.upgrade_log (version) VALUES ('0441'); -- gmc
 
 CREATE TABLE config.bib_source (
 	id		SERIAL	PRIMARY KEY,
diff --git a/Open-ILS/src/sql/Pg/020.schema.functions.sql b/Open-ILS/src/sql/Pg/020.schema.functions.sql
index aab35caf0..1f6b39cb3 100644
--- a/Open-ILS/src/sql/Pg/020.schema.functions.sql
+++ b/Open-ILS/src/sql/Pg/020.schema.functions.sql
@@ -53,7 +53,7 @@ CREATE OR REPLACE FUNCTION public.naco_normalize( TEXT, TEXT ) RETURNS TEXT AS $
 	$txt =~ tr/\x{2080}\x{2081}\x{2082}\x{2083}\x{2084}\x{2085}\x{2086}\x{2087}\x{2088}\x{2089}\x{208A}\x{208B}/0123456889+-/;# Convert subscript numbers
 
 	$txt =~ tr/\x{0251}\x{03B1}\x{03B2}\x{0262}\x{03B3}/AABGG/;	 	# Convert Latin and Greek
-	$txt =~ tr/\x{2113}\xF0\!\"\(\)\-\{\}\<\>\;\:\.\?\xA1\xBF\/\\\@\*\%\=\xB1\+\xAE\xA9\x{2117}\$\xA3\x{FFE1}\xB0\^\_\~\`/LD /;	# Convert Misc
+	$txt =~ tr/\x{2113}\xF0\x{111}\!\"\(\)\-\{\}\<\>\;\:\.\?\xA1\xBF\/\\\@\*\%\=\xB1\+\xAE\xA9\x{2117}\$\xA3\x{FFE1}\xB0\^\_\~\`/LDD /;	# Convert Misc
 	$txt =~ tr/\'\[\]\|//d;							# Remove Misc
 
 	if ($sf && $sf =~ /^a/o) {
diff --git a/Open-ILS/src/sql/Pg/upgrade/0441.schema.naco-normalized-crossed-d.sql b/Open-ILS/src/sql/Pg/upgrade/0441.schema.naco-normalized-crossed-d.sql
new file mode 100644
index 000000000..f7c9ef5b9
--- /dev/null
+++ b/Open-ILS/src/sql/Pg/upgrade/0441.schema.naco-normalized-crossed-d.sql
@@ -0,0 +1,52 @@
+BEGIN;
+
+INSERT INTO config.upgrade_log (version) VALUES ('0441'); -- gmc
+
+CREATE OR REPLACE FUNCTION public.naco_normalize( TEXT, TEXT ) RETURNS TEXT AS $func$
+	use Unicode::Normalize;
+	use Encode;
+
+	# When working with Unicode data, the first step is to decode it to
+	# a byte string; after that, lowercasing is safe
+	my $txt = lc(decode_utf8(shift));
+	my $sf = shift;
+
+	$txt = NFD($txt);
+	$txt =~ s/\pM+//go;	# Remove diacritics
+
+	$txt =~ s/\xE6/AE/go;	# Convert ae digraph
+	$txt =~ s/\x{153}/OE/go;# Convert oe digraph
+	$txt =~ s/\xFE/TH/go;	# Convert Icelandic thorn
+
+	$txt =~ tr/\x{2070}\x{2071}\x{2072}\x{2073}\x{2074}\x{2075}\x{2076}\x{2077}\x{2078}\x{2079}\x{207A}\x{207B}/0123456789+-/;# Convert superscript numbers
+	$txt =~ tr/\x{2080}\x{2081}\x{2082}\x{2083}\x{2084}\x{2085}\x{2086}\x{2087}\x{2088}\x{2089}\x{208A}\x{208B}/0123456889+-/;# Convert subscript numbers
+
+	$txt =~ tr/\x{0251}\x{03B1}\x{03B2}\x{0262}\x{03B3}/AABGG/;	 	# Convert Latin and Greek
+	$txt =~ tr/\x{2113}\xF0\x{111}\!\"\(\)\-\{\}\<\>\;\:\.\?\xA1\xBF\/\\\@\*\%\=\xB1\+\xAE\xA9\x{2117}\$\xA3\x{FFE1}\xB0\^\_\~\`/LDD /;	# Convert Misc
+	$txt =~ tr/\'\[\]\|//d;							# Remove Misc
+
+	if ($sf && $sf =~ /^a/o) {
+		my $commapos = index($txt,',');
+		if ($commapos > -1) {
+			if ($commapos != length($txt) - 1) {
+				my @list = split /,/, $txt;
+				my $first = shift @list;
+				$txt = $first . ',' . join(' ', @list);
+			} else {
+				$txt =~ s/,/ /go;
+			}
+		}
+	} else {
+		$txt =~ s/,/ /go;
+	}
+
+	$txt =~ s/\s+/ /go;	# Compress multiple spaces
+	$txt =~ s/^\s+//o;	# Remove leading space
+	$txt =~ s/\s+$//o;	# Remove trailing space
+
+	# Encoding the outgoing string is good practice, but not strictly
+	# necessary in this case because we've stripped everything from it
+	return encode_utf8($txt);
+$func$ LANGUAGE 'plperlu' STRICT IMMUTABLE;
+
+COMMIT;