LP#1276222: apply correct TS maps when building field_entry index vectors

This patch ensures that only the approriate class- and field-level
TS maps get used when populating the index_vector column in
metabib.*_field_entry.  Prior to this patch, the active flag in each
config.metabib_*_ts_maps tables was ignored, meaning that one
could not disable a mapping without deleting it outright.  Furthermore,
defects in the query to fetch the TS maps to apply meant that *all*
ts_config in the class-level TS map table were used, not just
the ones that applied to the particular class.

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

diff --git a/Open-ILS/src/sql/Pg/030.schema.metabib.sql b/Open-ILS/src/sql/Pg/030.schema.metabib.sql
index 83200e2..79536cd 100644 (file)
--- a/Open-ILS/src/sql/Pg/030.schema.metabib.sql
+++ b/Open-ILS/src/sql/Pg/030.schema.metabib.sql
@@ -1768,13 +1768,15 @@ BEGIN
             SELECT ts_config, index_weight
             FROM config.metabib_class_ts_map
             WHERE field_class = TG_ARGV[0]
-                AND index_lang IS NULL OR EXISTS (SELECT 1 FROM metabib.record_attr WHERE id = NEW.source AND index_lang IN(attrs->'item_lang',attrs->'language'))
-                AND always OR NOT EXISTS (SELECT 1 FROM config.metabib_field_ts_map WHERE metabib_field = NEW.field)
+                AND active
+                AND (index_lang IS NULL OR EXISTS (SELECT 1 FROM metabib.record_attr WHERE id = NEW.source AND index_lang IN(attrs->'item_lang',attrs->'language')))
+                AND (always OR NOT EXISTS (SELECT 1 FROM config.metabib_field_ts_map WHERE metabib_field = NEW.field))
             UNION
             SELECT ts_config, index_weight
             FROM config.metabib_field_ts_map
             WHERE metabib_field = NEW.field
-               AND index_lang IS NULL OR EXISTS (SELECT 1 FROM metabib.record_attr WHERE id = NEW.source AND index_lang IN(attrs->'item_lang',attrs->'language'))
+               AND active
+               AND (index_lang IS NULL OR EXISTS (SELECT 1 FROM metabib.record_attr WHERE id = NEW.source AND index_lang IN(attrs->'item_lang',attrs->'language')))
             ORDER BY index_weight ASC
         LOOP
             IF cur_weight IS NOT NULL AND cur_weight != ts_rec.index_weight THEN

diff --git a/Open-ILS/src/sql/Pg/upgrade/XXXX.function.grab_correct_ts_maps.sql b/Open-ILS/src/sql/Pg/upgrade/XXXX.function.grab_correct_ts_maps.sql
new file mode 100644 (file)
index 0000000..f2dcbf0
--- /dev/null
+++ b/Open-ILS/src/sql/Pg/upgrade/XXXX.function.grab_correct_ts_maps.sql
@@ -0,0 +1,103 @@
+-- Evergreen DB patch XXXX.function.grab_correct_ts_maps.sql
+--
+-- update public.oils_tsearch2() to ensure that it
+-- uses only the applicable, active TS maps
+--
+BEGIN;
+
+-- check whether patch can be applied
+SELECT evergreen.upgrade_deps_block_check('XXXX', :eg_version);
+
+CREATE OR REPLACE FUNCTION public.oils_tsearch2 () RETURNS TRIGGER AS $$
+DECLARE
+    normalizer      RECORD;
+    value           TEXT := '';
+    temp_vector     TEXT := '';
+    ts_rec          RECORD;
+    cur_weight      "char";
+BEGIN
+
+    value := NEW.value;
+    NEW.index_vector = ''::tsvector;
+
+    IF TG_TABLE_NAME::TEXT ~ 'field_entry$' THEN
+        FOR normalizer IN
+            SELECT  n.func AS func,
+                    n.param_count AS param_count,
+                    m.params AS params
+              FROM  config.index_normalizer n
+                    JOIN config.metabib_field_index_norm_map m ON (m.norm = n.id)
+              WHERE field = NEW.field AND m.pos < 0
+              ORDER BY m.pos LOOP
+                EXECUTE 'SELECT ' || normalizer.func || '(' ||
+                    quote_literal( value ) ||
+                    CASE
+                        WHEN normalizer.param_count > 0
+                            THEN ',' || REPLACE(REPLACE(BTRIM(normalizer.params,'[]'),E'\'',E'\\\''),E'"',E'\'')
+                            ELSE ''
+                        END ||
+                    ')' INTO value;
+
+        END LOOP;
+
+        NEW.value = value;
+
+        FOR normalizer IN
+            SELECT  n.func AS func,
+                    n.param_count AS param_count,
+                    m.params AS params
+              FROM  config.index_normalizer n
+                    JOIN config.metabib_field_index_norm_map m ON (m.norm = n.id)
+              WHERE field = NEW.field AND m.pos >= 0
+              ORDER BY m.pos LOOP
+                EXECUTE 'SELECT ' || normalizer.func || '(' ||
+                    quote_literal( value ) ||
+                    CASE
+                        WHEN normalizer.param_count > 0
+                            THEN ',' || REPLACE(REPLACE(BTRIM(normalizer.params,'[]'),E'\'',E'\\\''),E'"',E'\'')
+                            ELSE ''
+                        END ||
+                    ')' INTO value;
+
+        END LOOP;
+   END IF;
+
+    IF TG_TABLE_NAME::TEXT ~ 'browse_entry$' THEN
+        value :=  ARRAY_TO_STRING(
+            evergreen.regexp_split_to_array(value, E'\\W+'), ' '
+        );
+        value := public.search_normalize(value);
+        NEW.index_vector = to_tsvector(TG_ARGV[0]::regconfig, value);
+    ELSIF TG_TABLE_NAME::TEXT ~ 'field_entry$' THEN
+        FOR ts_rec IN
+            SELECT ts_config, index_weight
+            FROM config.metabib_class_ts_map
+            WHERE field_class = TG_ARGV[0]
+                AND active
+                AND (index_lang IS NULL OR EXISTS (SELECT 1 FROM metabib.record_attr WHERE id = NEW.source AND index_lang IN(attrs->'item_lang',attrs->'language')))
+                AND (always OR NOT EXISTS (SELECT 1 FROM config.metabib_field_ts_map WHERE metabib_field = NEW.field))
+            UNION
+            SELECT ts_config, index_weight
+            FROM config.metabib_field_ts_map
+            WHERE metabib_field = NEW.field
+               AND active 
+               AND (index_lang IS NULL OR EXISTS (SELECT 1 FROM metabib.record_attr WHERE id = NEW.source AND index_lang IN(attrs->'item_lang',attrs->'language')))
+            ORDER BY index_weight ASC
+        LOOP
+            IF cur_weight IS NOT NULL AND cur_weight != ts_rec.index_weight THEN
+                NEW.index_vector = NEW.index_vector || setweight(temp_vector::tsvector,cur_weight);
+                temp_vector = '';
+            END IF;
+            cur_weight = ts_rec.index_weight;
+            SELECT INTO temp_vector temp_vector || ' ' || to_tsvector(ts_rec.ts_config::regconfig, value)::TEXT;
+        END LOOP;
+        NEW.index_vector = NEW.index_vector || setweight(temp_vector::tsvector,cur_weight);
+    ELSE
+        NEW.index_vector = to_tsvector(TG_ARGV[0]::regconfig, value);
+    END IF;
+
+    RETURN NEW;
+END;
+$$ LANGUAGE PLPGSQL;
+
+COMMIT;