From 33369e07446d3d8161c6226c48f06eaa570ac47e Mon Sep 17 00:00:00 2001
From: Andrew Kane <andrew@ankane.org>
Date: Wed, 10 Apr 2024 13:23:20 -0700
Subject: [PATCH] Added tests for HNSW duplicates for bit, halfvec, and
 sparsevec

---
 test/t/026_hnsw_bit_duplicates.pl       | 58 +++++++++++++++++++++++++
 test/t/027_hnsw_halfvec_duplicates.pl   | 58 +++++++++++++++++++++++++
 test/t/028_hnsw_sparsevec_duplicates.pl | 58 +++++++++++++++++++++++++
 3 files changed, 174 insertions(+)
 create mode 100644 test/t/026_hnsw_bit_duplicates.pl
 create mode 100644 test/t/027_hnsw_halfvec_duplicates.pl
 create mode 100644 test/t/028_hnsw_sparsevec_duplicates.pl

diff --git a/test/t/026_hnsw_bit_duplicates.pl b/test/t/026_hnsw_bit_duplicates.pl
new file mode 100644
index 0000000..cdeea4f
--- /dev/null
+++ b/test/t/026_hnsw_bit_duplicates.pl
@@ -0,0 +1,58 @@
+use strict;
+use warnings;
+use PostgresNode;
+use TestLib;
+use Test::More;
+
+# Initialize node
+my $node = get_new_node('node');
+$node->init;
+$node->start;
+
+# Create table
+$node->safe_psql("postgres", "CREATE EXTENSION vector;");
+$node->safe_psql("postgres", "CREATE TABLE tst (v bit(3));");
+
+sub insert_vectors
+{
+	for my $i (1 .. 20)
+	{
+		$node->safe_psql("postgres", "INSERT INTO tst VALUES ('111');");
+	}
+}
+
+sub test_duplicates
+{
+	my $res = $node->safe_psql("postgres", qq(
+		SET enable_seqscan = off;
+		SET hnsw.ef_search = 1;
+		SELECT COUNT(*) FROM (SELECT * FROM tst ORDER BY v <~> '111') t;
+	));
+	is($res, 10);
+}
+
+# Test duplicates with build
+insert_vectors();
+$node->safe_psql("postgres", "CREATE INDEX idx ON tst USING hnsw (v bit_hamming_ops);");
+test_duplicates();
+
+# Reset
+$node->safe_psql("postgres", "TRUNCATE tst;");
+
+# Test duplicates with inserts
+insert_vectors();
+test_duplicates();
+
+# Test fallback path for inserts
+$node->pgbench(
+	"--no-vacuum --client=5 --transactions=100",
+	0,
+	[qr{actually processed}],
+	[qr{^$}],
+	"concurrent INSERTs",
+	{
+		"026_hnsw_bit_duplicates" => "INSERT INTO tst VALUES ('111');"
+	}
+);
+
+done_testing();
diff --git a/test/t/027_hnsw_halfvec_duplicates.pl b/test/t/027_hnsw_halfvec_duplicates.pl
new file mode 100644
index 0000000..47a952a
--- /dev/null
+++ b/test/t/027_hnsw_halfvec_duplicates.pl
@@ -0,0 +1,58 @@
+use strict;
+use warnings;
+use PostgresNode;
+use TestLib;
+use Test::More;
+
+# Initialize node
+my $node = get_new_node('node');
+$node->init;
+$node->start;
+
+# Create table
+$node->safe_psql("postgres", "CREATE EXTENSION vector;");
+$node->safe_psql("postgres", "CREATE TABLE tst (v halfvec(3));");
+
+sub insert_vectors
+{
+	for my $i (1 .. 20)
+	{
+		$node->safe_psql("postgres", "INSERT INTO tst VALUES ('[1,1,1]');");
+	}
+}
+
+sub test_duplicates
+{
+	my $res = $node->safe_psql("postgres", qq(
+		SET enable_seqscan = off;
+		SET hnsw.ef_search = 1;
+		SELECT COUNT(*) FROM (SELECT * FROM tst ORDER BY v <-> '[1,1,1]') t;
+	));
+	is($res, 10);
+}
+
+# Test duplicates with build
+insert_vectors();
+$node->safe_psql("postgres", "CREATE INDEX idx ON tst USING hnsw (v halfvec_l2_ops);");
+test_duplicates();
+
+# Reset
+$node->safe_psql("postgres", "TRUNCATE tst;");
+
+# Test duplicates with inserts
+insert_vectors();
+test_duplicates();
+
+# Test fallback path for inserts
+$node->pgbench(
+	"--no-vacuum --client=5 --transactions=100",
+	0,
+	[qr{actually processed}],
+	[qr{^$}],
+	"concurrent INSERTs",
+	{
+		"027_hnsw_halfvec_duplicates" => "INSERT INTO tst VALUES ('[1,1,1]');"
+	}
+);
+
+done_testing();
diff --git a/test/t/028_hnsw_sparsevec_duplicates.pl b/test/t/028_hnsw_sparsevec_duplicates.pl
new file mode 100644
index 0000000..499900f
--- /dev/null
+++ b/test/t/028_hnsw_sparsevec_duplicates.pl
@@ -0,0 +1,58 @@
+use strict;
+use warnings;
+use PostgresNode;
+use TestLib;
+use Test::More;
+
+# Initialize node
+my $node = get_new_node('node');
+$node->init;
+$node->start;
+
+# Create table
+$node->safe_psql("postgres", "CREATE EXTENSION vector;");
+$node->safe_psql("postgres", "CREATE TABLE tst (v sparsevec(3));");
+
+sub insert_vectors
+{
+	for my $i (1 .. 20)
+	{
+		$node->safe_psql("postgres", "INSERT INTO tst VALUES ('{1:1,2:1,3:1}/3');");
+	}
+}
+
+sub test_duplicates
+{
+	my $res = $node->safe_psql("postgres", qq(
+		SET enable_seqscan = off;
+		SET hnsw.ef_search = 1;
+		SELECT COUNT(*) FROM (SELECT * FROM tst ORDER BY v <-> '{1:1,2:1,3:1}/3') t;
+	));
+	is($res, 10);
+}
+
+# Test duplicates with build
+insert_vectors();
+$node->safe_psql("postgres", "CREATE INDEX idx ON tst USING hnsw (v sparsevec_l2_ops);");
+test_duplicates();
+
+# Reset
+$node->safe_psql("postgres", "TRUNCATE tst;");
+
+# Test duplicates with inserts
+insert_vectors();
+test_duplicates();
+
+# Test fallback path for inserts
+$node->pgbench(
+	"--no-vacuum --client=5 --transactions=100",
+	0,
+	[qr{actually processed}],
+	[qr{^$}],
+	"concurrent INSERTs",
+	{
+		"028_hnsw_sparsevec_duplicates" => "INSERT INTO tst VALUES ('{1:1,2:1,3:1}/3');"
+	}
+);
+
+done_testing();