Add record summary tests

db/sql/test_4_record_summaries.sql

@@ -0,0 +1,223 @@
+DROP EXTENSION IF EXISTS pgtap CASCADE;
+CREATE EXTENSION IF NOT EXISTS pgtap;
+
+CREATE OR REPLACE FUNCTION __setup_record_summary_data() RETURNS SETOF TEXT AS $$
+BEGIN
+  CREATE TABLE authors (
+    id SERIAL PRIMARY KEY,
+    name TEXT
+  );
+  INSERT INTO authors (name) VALUES
+    ('Alice'),   -- id 1
+    ('Bob'),     -- id 2
+    ('Charlie'); -- id 3
+  CREATE TABLE books (
+    id SERIAL PRIMARY KEY,
+    title TEXT,
+    author_id INT REFERENCES authors(id),
+    year_published INT
+  );
+  INSERT INTO books (title, author_id, year_published) values
+    ('Apple'    , 1, 2001), -- id 1
+    ('Asparagus', 1, 2002), -- id 2
+    ('Banana'   , 2, 2003), -- id 3
+    ('Broccoli' , 2, 2004), -- id 4
+    ('Carrot'   , 3, 2005); -- id 5
+  CREATE TABLE items (
+    id SERIAL PRIMARY KEY,
+    book_id INT REFERENCES books(id),
+    barcode TEXT
+  );
+  INSERT INTO items (book_id, barcode) VALUES
+    (1, 'FH'), -- id 1
+    (1, 'DQ'), -- id 2
+    (2, 'MW'), -- id 3
+    (3, 'BR'), -- id 4
+    (4, 'SN'), -- id 5
+    (5, 'ZA'); -- id 6
+END;
+$$ LANGUAGE plpgsql;
+
+CREATE OR REPLACE FUNCTION test_record_summary_basic() RETURNS SETOF TEXT AS $t$
+BEGIN
+  PERFORM __setup_record_summary_data();
+
+  RETURN NEXT bag_eq(
+    $$SELECT * FROM msar.get_record_summaries('books'::regclass::oid, array[4,4,4,4])$$,
+    $$VALUES (4, 'Broccoli')$$,
+    'Duplicate input ids should be made distinct in the output'
+  );
+
+  RETURN NEXT bag_eq(
+    $$SELECT * FROM msar.get_record_summaries('books'::regclass::oid, array[1,3,5])$$,
+    $$VALUES (1, 'Apple'), (3, 'Banana'), (5, 'Carrot')$$,
+    'Auto-generated record summaries for books'
+  );
+
+  RETURN NEXT bag_eq(
+    $$SELECT * FROM msar.get_record_summaries(
+      'books'::regclass::oid,
+      array[1,3,5],
+      '[[2]," by ", [3,2]]'::jsonb
+    )$$,
+    $$VALUES
+      (1, 'Apple by Alice'),
+      (3, 'Banana by Bob'),
+      (5, 'Carrot by Charlie')
+    $$,
+    'Custom record summaries for books'
+  );
+
+  RETURN NEXT bag_eq(
+    $$SELECT * FROM msar.get_record_summaries(
+      'items'::regclass::oid,
+      array[3,4],
+      '[[3],": ",[2,2]," by ",[2,3,2]]'::jsonb
+    )$$,
+    $$VALUES
+      (3, 'MW: Asparagus by Alice'),
+      (4, 'BR: Banana by Bob')
+    $$,
+    'Custom record summaries for items'
+  );
+END;
+$t$ LANGUAGE plpgsql;
+
+CREATE OR REPLACE FUNCTION test_record_summary_with_empty_template() RETURNS SETOF TEXT AS $t$
+BEGIN
+  PERFORM __setup_record_summary_data();
+  RETURN NEXT bag_eq(
+    $$SELECT * FROM msar.get_record_summaries('books'::regclass::oid, array[1], '[]'::jsonb)$$,
+    $$VALUES (1, '?')$$,
+    'Record summary with empty template'
+  );
+END;
+$t$ LANGUAGE plpgsql;
+
+CREATE OR REPLACE FUNCTION test_record_summary_with_invalid_attnum() RETURNS SETOF TEXT AS $t$
+BEGIN
+  PERFORM __setup_record_summary_data();
+  RETURN NEXT bag_eq(
+    $$SELECT * FROM msar.get_record_summaries(
+      'books'::regclass::oid,
+      array[1],
+      '[[2]," - ",[98],[99,2]," - ",[4]]'::jsonb
+    )$$,
+    $$VALUES (1, 'Apple -  - 2001')$$,
+    'Invalid column attnums in template should be disregarded'
+  );
+END;
+$t$ LANGUAGE plpgsql;
+
+CREATE OR REPLACE FUNCTION test_record_summary_with_invalid_fk_column() RETURNS SETOF TEXT AS $t$
+BEGIN
+  PERFORM __setup_record_summary_data();
+  RETURN NEXT bag_eq(
+    $$SELECT * FROM msar.get_record_summaries(
+      'books'::regclass::oid,
+      array[1],
+      -- [4,2] should be ignored here. 4 is a valid attnum, but it's not an FK column
+      '[[2]," - ",[4,2]]'::jsonb
+    )$$,
+    $$VALUES (1, 'Apple - ')$$,
+    'Invalid FK columns in template should be disregarded'
+  );
+END;
+$t$ LANGUAGE plpgsql;
+
+CREATE OR REPLACE FUNCTION test_record_summary_with_funky_characters() RETURNS SETOF TEXT AS $t$
+BEGIN
+  PERFORM __setup_record_summary_data();
+  RETURN NEXT bag_eq(
+    $q$SELECT * FROM msar.get_record_summaries(
+      'books'::regclass::oid,
+      array[1],
+      '[[2]," $$;''\""]'::jsonb
+    )$q$,
+    $v$VALUES (1, 'Apple $$;''"')$v$,
+    'Record summaries should permit SQL quotes and delimiters as literal values'
+  );
+END;
+$t$ LANGUAGE plpgsql;
+
+
+CREATE OR REPLACE FUNCTION _summarize(id INT) RETURNS TEXT AS $$
+BEGIN
+  UPDATE counter SET count = count + 1;
+  RETURN 'Number: ' || id::text;
+END;
+$$ LANGUAGE plpgsql;
+
+-- This is test is skipped because it's currently failing. We'd like to call the `summarize`
+-- function only 4 times, but instead we're calling it 1000 times. We need to do some work
+-- within get_record_summaries_via_query to figure out how to fix this.
+CREATE OR REPLACE FUNCTION __SKIP__test_record_summary_performance() RETURNS SETOF TEXT AS $t$
+BEGIN
+  CREATE TABLE numbers AS SELECT * FROM generate_series(1, 1000) AS id;
+  ALTER TABLE numbers ADD CONSTRAINT pk__numbers_id PRIMARY KEY (id);
+
+  CREATE TABLE counter AS SELECT 0 AS count;
+
+  RETURN NEXT bag_eq(
+    $r$SELECT * FROM msar.get_record_summaries_via_query(
+      $q$SELECT id, _summarize(id) AS record_summary FROM numbers$q$,
+      array[950,10,400,700]
+    )$r$,
+    $r$VALUES
+      (950, 'Number: 950'),
+      (10,  'Number: 10'),
+      (400, 'Number: 400'),
+      (700, 'Number: 700')
+    $r$,
+    'Performance test values should match'
+  );
+
+  RETURN NEXT results_eq(
+    $$SELECT * FROM COUNTER$$,
+    $$VALUES (4)$$,
+    'Summarization computation should only occur for specified records'
+  );
+END;
+$t$ LANGUAGE plpgsql;
+
+CREATE OR REPLACE FUNCTION test_record_summary_with_text_pk_column() RETURNS SETOF TEXT AS $t$
+BEGIN
+  CREATE TABLE things (
+    name text primary key,
+    description text
+  );
+  INSERT INTO things VALUES
+    ('foo', 'YEAH'),
+    ('bar', 'NO');
+
+  RETURN NEXT bag_eq(
+    $$SELECT * FROM msar.get_record_summaries('things'::regclass::oid, array['foo', 'bar'])$$,
+    $$VALUES ('foo', 'foo'), ('bar', 'bar');$$,
+    'Record summary should auto-select PK column when it''s text'
+  );
+END;
+$t$ LANGUAGE plpgsql;
+
+CREATE OR REPLACE FUNCTION test_record_summary_with_no_text_columns() RETURNS SETOF TEXT AS $t$
+BEGIN
+  CREATE TABLE points (
+    id_value UUID PRIMARY KEY DEFAULT gen_random_uuid(),
+    position point
+  );
+  INSERT INTO points VALUES
+    ('f8d0a39a-70ca-4a11-8ffc-4a44ebc66898'::uuid, point(0,0)),
+    ('cfa72e14-34a1-4ee9-a7bf-9aaf93acc275'::uuid, point(1,1));
+
+  RETURN NEXT bag_eq(
+    $$SELECT * FROM msar.get_record_summaries(
+      'points'::regclass::oid,
+      array[ 'f8d0a39a-70ca-4a11-8ffc-4a44ebc66898'::uuid ]
+    )$$,
+    $$VALUES (
+      'f8d0a39a-70ca-4a11-8ffc-4a44ebc66898'::uuid,
+      'f8d0a39a-70ca-4a11-8ffc-4a44ebc66898'
+    )$$,
+    'Record summary should work even when no columns are text'
+  );
+END;
+$t$ LANGUAGE plpgsql;

db/sql/test_startup.sql

@@ -10,3 +10,5 @@ CREATE DATABASE mathesar_testing;
 \c mathesar_testing
 \ir 0_msar.sql
 \ir test_0_msar.sql
+\ir 4_record_summaries.sql
+\ir test_4_record_summaries.sql