nidx-binding 6.9.1.post587__tar.gz → 6.9.1.post596__tar.gz

{nidx_binding-6.9.1.post587 → nidx_binding-6.9.1.post596}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nidx_binding
-Version: 6.9.1.post587
+Version: 6.9.1.post596
 Classifier: Programming Language :: Rust
 Classifier: Programming Language :: Python :: Implementation :: CPython
 Classifier: Programming Language :: Python :: Implementation :: PyPy

{nidx_binding-6.9.1.post587 → nidx_binding-6.9.1.post596}/nidx_protos/pyproject.toml

@@ -10,7 +10,7 @@ build-backend = "pdm.backend"
 
 [project]
 name = "nidx_protos"
-version = "6.9.1.post587"
+version = "6.9.1.post596"
 license = "AGPL-3.0-or-later"
 description = "Protobuf definitions for nucliadb/nidx"
 authors = [{ name = "Nuclia", email = "nucliadb@nuclia.com" }]

{nidx_binding-6.9.1.post587 → nidx_binding-6.9.1.post596}/nidx_vector/src/searcher.rs

@@ -156,11 +156,10 @@ impl<'a> Fssc<'a> {
             buff: HashSet::with_capacity(size),
         }
     }
-    fn add(&mut self, candidate: ScoredParagraph<'a>, vector: &[u8]) {
-        if !self.with_duplicates && self.seen.contains(vector) {
+    fn add(&mut self, candidate: ScoredParagraph<'a>, vector: Vec<u8>) {
+        if !self.with_duplicates && self.seen.contains(&vector) {
             return;
         } else if !self.with_duplicates {
-            let vector = vector.to_vec();
             self.seen.insert(vector);
         }
 
@@ -261,23 +260,37 @@ impl Searcher {
         let min_score = request.min_score();
         let mut ffsv = Fssc::new(request.no_results(), with_duplicates);
 
-        for open_segment in &self.open_segments {
-            // Skip this segment if it doesn't match the segment filter
-            if !segment_filter
-                .as_ref()
-                .is_none_or(|f| segment_matches(f, open_segment.tags()))
-            {
-                continue;
-            }
-            let partial_solution =
-                open_segment.search(query, filter, with_duplicates, no_results, &self.config, min_score);
-
-            for candidate in partial_solution {
-                let addr = candidate.paragraph();
-                let paragraph = open_segment.get_paragraph(addr);
-                let scored_paragraph = ScoredParagraph::new(open_segment, addr, paragraph, candidate.score());
-                ffsv.add(scored_paragraph, candidate.vector());
-            }
+        let pool = rayon::ThreadPoolBuilder::new().build().unwrap();
+        let results: Vec<_> = pool.install(|| {
+            self.open_segments
+                .par_iter()
+                .flat_map(|open_segment| {
+                    // Skip this segment if it doesn't match the segment filter
+                    if !segment_filter
+                        .as_ref()
+                        .is_none_or(|f| segment_matches(f, open_segment.tags()))
+                    {
+                        return vec![];
+                    }
+                    let partial_solution =
+                        open_segment.search(query, filter, with_duplicates, no_results, &self.config, min_score);
+
+                    partial_solution
+                        .map(|candidate| {
+                            let addr = candidate.paragraph();
+                            let paragraph = open_segment.get_paragraph(addr);
+                            (
+                                ScoredParagraph::new(open_segment, addr, paragraph, candidate.score()),
+                                candidate.vector().to_vec(),
+                            )
+                        })
+                        .collect()
+                })
+                .collect()
+        });
+
+        for (paragraph, vector) in results {
+            ffsv.add(paragraph, vector);
         }
 
         Ok(ffsv.into())

{nidx_binding-6.9.1.post587 → nidx_binding-6.9.1.post596}/pyproject.toml

@@ -13,7 +13,7 @@ classifiers = [
     "Programming Language :: Python :: Implementation :: CPython",
     "Programming Language :: Python :: Implementation :: PyPy",
 ]
-version = "6.9.1.post587"
+version = "6.9.1.post596"
 
 [project.urls]
 Homepage = "https://nuclia.com"

{nidx_binding-6.9.1.post587 → nidx_binding-6.9.1.post596}/src/searcher/shard_selector.rs

@@ -215,23 +215,21 @@ impl ShardSelector {
             .collect()
     }
 
-    /// List the nodes that contain a shard. If the local node is in the list, it will always be returned first
+    /// List the nodes that contain a shard in order of preference.
+    /// This way we preferentially send all requests for a shard to the same node to improve caching.
     pub fn nodes_for_shard(&self, shard: &Uuid) -> Vec<SearcherNode> {
         let nodes = self._nodes_for_shard(self.nodes.list_nodes(), shard);
-        // See if some of this nodes are me, and put them first
-        let mut searcher_nodes = Vec::new();
-        let mut has_this_node = false;
-        for node in nodes {
-            if node == self.nodes.this_node() {
-                has_this_node = true;
-            } else {
-                searcher_nodes.push(SearcherNode::Remote(node));
-            }
-        }
-        if has_this_node {
-            searcher_nodes.insert(0, SearcherNode::This);
-        }
-        searcher_nodes
+
+        nodes
+            .into_iter()
+            .map(|n| {
+                if n == self.nodes.this_node() {
+                    SearcherNode::This
+                } else {
+                    SearcherNode::Remote(n)
+                }
+            })
+            .collect()
     }
 
     fn _nodes_for_shard(&self, mut nodes: Vec<String>, shard: &Uuid) -> Vec<String> {