Merge pull request #76 from meilisearch/panic-at-search

Fix panic at search when the index is empty

Author: Kerollmops

src/reader.rs

@@ -203,6 +203,9 @@ impl<'t, D: Distance> Reader<'t, D> {
         search_k: Option<NonZeroUsize>,
         candidates: Option<&RoaringBitmap>,
     ) -> Result<Vec<(ItemId, f32)>> {
+        if self.items.is_empty() {
+            return Ok(Vec::new());
+        }
         // Since the datastructure describes a kind of btree, the capacity is something in the order of:
         // The number of root nodes + log2 of the total number of vectors.
         let mut queue =

src/tests/reader.rs

@@ -203,3 +203,19 @@ fn filtering() {
     id(99): distance(99)
     "###);
 }
+
+#[test]
+fn search_in_empty_database() {
+    // See https://github.com/meilisearch/arroy/issues/74
+    let handle = create_database::<Euclidean>();
+
+    let mut wtxn = handle.env.write_txn().unwrap();
+    let writer = Writer::new(handle.database, 0, 2);
+    writer.build(&mut wtxn, &mut rng(), None).unwrap();
+    wtxn.commit().unwrap();
+
+    let rtxn = handle.env.read_txn().unwrap();
+    let reader = Reader::open(&rtxn, 0, handle.database).unwrap();
+    let ret = reader.nns_by_vector(&rtxn, &[0., 0.], 10, None, None).unwrap();
+    insta::assert_debug_snapshot!(ret, @"[]");
+}

src/tests/writer.rs

@@ -3,7 +3,7 @@ use rand::seq::SliceRandom;
 use rand::Rng;
 
 use super::{create_database, rng};
-use crate::distance::{DotProduct, Euclidean};
+use crate::distance::{Angular, DotProduct, Euclidean};
 use crate::{Database, Reader, Writer};
 
 #[test]
@@ -467,6 +467,43 @@ fn delete_one_leaf_in_a_split() {
     "###);
 }
 
+#[test]
+fn delete_one_item_in_a_single_document_database() {
+    let handle = create_database::<Angular>();
+    let mut rng = rng();
+    let mut wtxn = handle.env.write_txn().unwrap();
+    let writer = Writer::new(handle.database, 0, 2);
+
+    // first, insert a bunch of elements
+    writer.add_item(&mut wtxn, 0, &[0., 0.]).unwrap();
+    writer.build(&mut wtxn, &mut rng, None).unwrap();
+    wtxn.commit().unwrap();
+
+    insta::assert_display_snapshot!(handle, @r###"
+    ==================
+    Dumping index 0
+    Item 0: Leaf(Leaf { header: NodeHeaderAngular { norm: 0.0 }, vector: [0.0000, 0.0000] })
+    Tree 0: Descendants(Descendants { descendants: [0] })
+    Root: Metadata { dimensions: 2, items: RoaringBitmap<[0]>, roots: [0], distance: "angular" }
+    updated_item_ids: RoaringBitmap<[]>
+    "###);
+
+    let mut wtxn = handle.env.write_txn().unwrap();
+    let writer = Writer::new(handle.database, 0, 2);
+
+    writer.del_item(&mut wtxn, 0).unwrap();
+
+    writer.build(&mut wtxn, &mut rng, None).unwrap();
+    wtxn.commit().unwrap();
+
+    insta::assert_display_snapshot!(handle, @r###"
+    ==================
+    Dumping index 0
+    Root: Metadata { dimensions: 2, items: RoaringBitmap<[]>, roots: [], distance: "angular" }
+    updated_item_ids: RoaringBitmap<[]>
+    "###);
+}
+
 #[test]
 fn delete_one_item() {
     let handle = create_database::<Euclidean>();