s3vectors: report the Euclidean distance, not its square (#12318)

A euclidean index answered QueryVectors with Lance's L2 distance, which is
the squared Euclidean distance: [3, 4] was 25 from the origin, not 5. AWS
defines the metric as the straight-line distance between the vectors. The
gateway now returns its square root, so the single engine and the enterprise
coordinator, which both hand back the raw Lance value, report the same
distance. Ordering is unchanged.


Claude-Session: https://claude.ai/code/session_011UpvafvK8JpagF94MHWqwK

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

test/s3vectors/s3vectors_test.py

@@ -58,7 +58,7 @@ def test_vector_lifecycle(vectors):
     assert [h["key"] for h in hits] == ["a", "c"]
     distances = {h["key"]: h["distance"] for h in hits}
     assert distances["a"] == pytest.approx(0.0, abs=1e-6)
-    assert distances["c"] == pytest.approx(0.02, abs=1e-4)
+    assert distances["c"] == pytest.approx(0.02 ** 0.5, abs=1e-4)
     assert all(h["metadata"]["color"] == "red" for h in hits)
 
     vectors.delete_vectors(vectorBucketName=BUCKET, indexName=INDEX, keys=["b"])
@@ -223,6 +223,13 @@ def test_filter_arrays(vectors, index):
     assert matches({"$nin": [1, "2.5"]}) == ["flag", "scalar", "space"]
 
 
+def test_euclidean_distance(vectors, index):
+    vectors.put_vectors(**index, vectors=[{"key": "p", "data": {"float32": [3, 4, 0, 0]}}])
+    result = vectors.query_vectors(**index, queryVector={"float32": [0, 0, 0, 0]}, topK=1, returnDistance=True)
+    assert result["distanceMetric"] == "euclidean"
+    assert result["vectors"][0]["distance"] == pytest.approx(5.0)
+
+
 def test_cosine_metric(vectors):
     bucket = "cosine-" + uuid.uuid4().hex[:16]
     vectors.create_vector_bucket(vectorBucketName=bucket)