Force facet values to Unicode NFC to reduce spurious non-matches

git-svn-id: svn://svn.open-ils.org/ILS/trunk@19899 dcc99617-32d9-48b4-a31d-7c20da2025e4

diff --git a/Open-ILS/src/sql/Pg/002.functions.config.sql b/Open-ILS/src/sql/Pg/002.functions.config.sql
index 47461e2..afd2b2b 100644 (file)
--- a/Open-ILS/src/sql/Pg/002.functions.config.sql
+++ b/Open-ILS/src/sql/Pg/002.functions.config.sql
@@ -482,6 +482,11 @@ BEGIN
 END;
 $func$ LANGUAGE PLPGSQL;
 
+CREATE OR REPLACE FUNCTION force_unicode_normal_form(string TEXT, form TEXT) RETURNS TEXT AS $func$
+use Unicode::Normalize 'normalize';
+return normalize($_[1],$_[0]); # reverse the params
+$func$ LANGUAGE PLPERLU;
+
 CREATE OR REPLACE FUNCTION maintain_control_numbers() RETURNS TRIGGER AS $func$
 use strict;
 use MARC::Record;

diff --git a/Open-ILS/src/sql/Pg/002.schema.config.sql b/Open-ILS/src/sql/Pg/002.schema.config.sql
index 72ce6c5..9fb53d7 100644 (file)
--- 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 ('0504'); -- miker for tsbere
+INSERT INTO config.upgrade_log (version) VALUES ('0505'); -- miker
 
 CREATE TABLE config.bib_source (
        id              SERIAL  PRIMARY KEY,

diff --git a/Open-ILS/src/sql/Pg/030.schema.metabib.sql b/Open-ILS/src/sql/Pg/030.schema.metabib.sql
index eaa48e9..864ea65 100644 (file)
--- a/Open-ILS/src/sql/Pg/030.schema.metabib.sql
+++ b/Open-ILS/src/sql/Pg/030.schema.metabib.sql
@@ -135,6 +135,17 @@ CREATE INDEX metabib_facet_entry_field_idx ON metabib.facet_entry (field);
 CREATE INDEX metabib_facet_entry_value_idx ON metabib.facet_entry (SUBSTRING(value,1,1024));
 CREATE INDEX metabib_facet_entry_source_idx ON metabib.facet_entry (source);
 
+CREATE OR REPLACE FUNCTION facet_force_nfc() RETURNS TRIGGER AS $$
+BEGIN
+    NEW.value := force_unicode_normal_form(NEW.value,'NFC');
+    RETURN NEW;
+END;
+$$ LANUAGE PLPGSQL;
+
+CREATE TRIGGER facet_force_nfc_tgr
+       BEFORE UPDATE OR INSERT ON metabib.facet_entry
+       FOR EACH ROW EXECUTE PROCEDURE facet_force_nfc();
+
 CREATE TABLE metabib.record_attr (
        id              BIGINT  PRIMARY KEY REFERENCES biblio.record_entry (id) ON DELETE CASCADE,
        attrs   HSTORE  NOT NULL DEFAULT ''::HSTORE

diff --git a/Open-ILS/src/sql/Pg/upgrade/0505.schema.force_facets_to_NFC.sql b/Open-ILS/src/sql/Pg/upgrade/0505.schema.force_facets_to_NFC.sql
new file mode 100644 (file)
index 0000000..57e7933
--- /dev/null
+++ b/Open-ILS/src/sql/Pg/upgrade/0505.schema.force_facets_to_NFC.sql
@@ -0,0 +1,24 @@
+BEGIN;
+
+INSERT INTO config.upgrade_log (version) VALUES ('0505'); --miker
+
+CREATE OR REPLACE FUNCTION force_unicode_normal_form(string TEXT, form TEXT) RETURNS TEXT AS $func$
+use Unicode::Normalize 'normalize';
+return normalize($_[1],$_[0]); # reverse the params
+$func$ LANGUAGE PLPERLU;
+
+UPDATE metabib.facet_entry SET value = force_unicode_normal_form(value,'NFC');
+
+CREATE OR REPLACE FUNCTION facet_force_nfc() RETURNS TRIGGER AS $$
+BEGIN
+    NEW.value := force_unicode_normal_form(NEW.value,'NFC');
+    RETURN NEW;
+END;
+$$ LANUAGE PLPGSQL;
+
+CREATE TRIGGER facet_force_nfc_tgr
+       BEFORE UPDATE OR INSERT ON metabib.facet_entry
+       FOR EACH ROW EXECUTE PROCEDURE facet_force_nfc();
+
+COMMIT;
+