Add the schema upgrade corresponding to r15351 in-db ingest encoding fixes

Ensure MARC::File::XML treats the content as UTF-8; then follow
recommended practice of explicitly decoding the UTF-8 string to a
bytes string before operating on it, then encoding it back to UTF8
when it's returned.

'Québec' now gets naco-normalized to 'quebec' as one would expect. YAY!

git-svn-id: svn://svn.open-ils.org/ILS/trunk@15352 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 f7f29f4..60d97bd 100644 (file)
--- a/Open-ILS/src/sql/Pg/002.schema.config.sql
+++ b/Open-ILS/src/sql/Pg/002.schema.config.sql
@@ -51,7 +51,7 @@ CREATE TABLE config.upgrade_log (
     install_date    TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT NOW()
 );
 
-INSERT INTO config.upgrade_log (version) VALUES ('0137'); -- miker
+INSERT INTO config.upgrade_log (version) VALUES ('0138'); -- dbs
 
 CREATE TABLE config.bib_source (
        id              SERIAL  PRIMARY KEY,

diff --git a/Open-ILS/src/sql/Pg/upgrade/0138.schema.in-db-encoding-fixes.sql b/Open-ILS/src/sql/Pg/upgrade/0138.schema.in-db-encoding-fixes.sql
new file mode 100644 (file)
index 0000000..a62d3b6
--- /dev/null
+++ b/Open-ILS/src/sql/Pg/upgrade/0138.schema.in-db-encoding-fixes.sql
@@ -0,0 +1,95 @@
+
+BEGIN;
+
+INSERT INTO config.upgrade_log (version) VALUES ('0138'); -- dbs
+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\!\"\(\)\-\{\}\<\>\;\:\.\?\xA1\xBF\/\\\@\*\%\=\xB1\+\xAE\xA9\x{2117}\$\xA3\x{FFE1}\xB0\^\_\~\`/LD /;     # 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;
+
+CREATE OR REPLACE FUNCTION biblio.flatten_marc ( TEXT ) RETURNS SETOF metabib.full_rec AS $func$
+
+use MARC::Record;
+use MARC::File::XML (BinaryEncoding => 'UTF-8');
+
+my $xml = shift;
+my $r = MARC::Record->new_from_xml( $xml );
+
+return_next( { tag => 'LDR', value => $r->leader } );
+
+for my $f ( $r->fields ) {
+       if ($f->is_control_field) {
+               return_next({ tag => $f->tag, value => $f->data });
+       } else {
+               for my $s ($f->subfields) {
+                       return_next({
+                               tag      => $f->tag,
+                               ind1     => $f->indicator(1),
+                               ind2     => $f->indicator(2),
+                               subfield => $s->[0],
+                               value    => $s->[1]
+                       });
+
+                       if ( $f->tag eq '245' and $s->[0] eq 'a' ) {
+                               my $trim = $f->indicator(2) || 0;
+                               return_next({
+                                       tag      => 'tnf',
+                                       ind1     => $f->indicator(1),
+                                       ind2     => $f->indicator(2),
+                                       subfield => 'a',
+                                       value    => substr( $s->[1], $trim )
+                               });
+                       }
+               }
+       }
+}
+
+return undef;
+
+$func$ LANGUAGE PLPERLU;
+
+
+COMMIT;
+