stored procedures providing NACO normalization for controlled headings, esp authority...

git-svn-id: svn://svn.open-ils.org/ILS/trunk@7761 dcc99617-32d9-48b4-a31d-7c20da2025e4

diff --git a/Open-ILS/src/sql/Pg/020.schema.functions.sql b/Open-ILS/src/sql/Pg/020.schema.functions.sql
index 9a17029..c2a85bd 100644 (file)
--- a/Open-ILS/src/sql/Pg/020.schema.functions.sql
+++ b/Open-ILS/src/sql/Pg/020.schema.functions.sql
@@ -16,6 +16,49 @@ CREATE OR REPLACE FUNCTION public.non_filing_normalize ( TEXT, "char" ) RETURNS
                );
 $$ LANGUAGE SQL STRICT IMMUTABLE;
 
+CREATE OR REPLACE FUNCTION public.naco_normalize( TEXT, TEXT ) RETURNS TEXT AS $func$
+       my $txt = lc(shift);
+       my $sf = shift;
+
+       $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\!\"\(\)\-\{\}\<\>\;\:\.\?\xA1\xBF\/\\\@\*\%\=\xB1\+\xAE\xA9\x{2117}\$\xA3\x{FFE1}\xB0\^\_\~\`/LD /;     # Convert Misc
+       $txt =~ tr/\'\[\]\|//d;                                                 # Remove Misc
+
+       if ($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
+
+       return $txt;
+$func$ LANGUAGE 'plperlu' STRICT IMMUTABLE;
+
+CREATE OR REPLACE FUNCTION public.naco_normalize( TEXT ) RETURNS TEXT AS $func$
+       SELECT public.naco_normalize($1,'');
+$func$ LANGUAGE 'sql' STRICT IMMUTABLE;
+
 CREATE OR REPLACE FUNCTION public.call_number_dewey( TEXT ) RETURNS TEXT AS $$
        my $txt = shift;
        $txt =~ s/^\s+//o;