nidx-binding 6.6.1.post421__tar.gz → 6.6.1.post424__tar.gz

{nidx_binding-6.6.1.post421 → nidx_binding-6.6.1.post424}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nidx_binding
-Version: 6.6.1.post421
+Version: 6.6.1.post424
 Classifier: Programming Language :: Rust
 Classifier: Programming Language :: Python :: Implementation :: CPython
 Classifier: Programming Language :: Python :: Implementation :: PyPy

{nidx_binding-6.6.1.post421 → nidx_binding-6.6.1.post424}/nidx_paragraph/src/reader.rs

@@ -212,6 +212,11 @@ impl ParagraphReaderService {
         Ok(response)
     }
     fn adapt_text(&self, parser: &QueryParser, text: &str) -> String {
+        // FIXME: after migrating from tantivy 0.22 -> 0.24, the query grammar
+        // now includes single quotes as special character. Queries having a
+        // single single quote now fail to parse. As a quick fix, we remove them
+        // all and replace them by a space.
+        let text = text.replace('\'', " ");
         match text.trim() {
             "" => text.to_string(),
             text => parser

{nidx_binding-6.6.1.post421 → nidx_binding-6.6.1.post424}/nidx_paragraph/tests/reader.rs

@@ -465,3 +465,46 @@ fn test_new_paragraph() -> anyhow::Result<()> {
     assert_eq!(count, 4);
     Ok(())
 }
+
+#[test]
+fn test_query_parsing() -> anyhow::Result<()> {
+    // This test validates we are properly removing single quotes from queries
+    // to avoid issues with tantivy + our own parsing.
+    //
+    // For further context, see:
+    // https://github.com/nuclia/nucliadb/pull/3216
+    //
+    const UUID: &str = "f56c58ac-b4f9-4d61-a077-ffccaadd0001";
+
+    let seconds = SystemTime::now()
+        .duration_since(SystemTime::UNIX_EPOCH)
+        .map(|t| t.as_secs() as i64)
+        .unwrap();
+    let timestamp = Timestamp { seconds, nanos: 0 };
+
+    let resource1 = create_resource("shard1".to_string(), timestamp);
+    let paragraph_reader_service = test_reader(&resource1);
+
+    // Only one paragraph  matches
+    let mut search = ParagraphSearchRequest {
+        id: "shard1".to_string(),
+        uuid: UUID.to_string(),
+        body: "".to_string(),
+        faceted: None,
+        order: None,
+        page_number: 0,
+        result_per_page: 20,
+        only_faceted: false,
+        ..Default::default()
+    };
+
+    search.body = "some document".to_string();
+    let result = paragraph_reader_service.search(&search, &PrefilterResult::All).unwrap();
+    assert_eq!(result.total, 1);
+
+    search.body = "some ' document".to_string();
+    let result = paragraph_reader_service.search(&search, &PrefilterResult::All).unwrap();
+    assert_eq!(result.total, 1);
+
+    Ok(())
+}

{nidx_binding-6.6.1.post421 → nidx_binding-6.6.1.post424}/nidx_protos/pyproject.toml

@@ -10,7 +10,7 @@ build-backend = "pdm.backend"
 
 [project]
 name = "nidx_protos"
-version = "6.6.1.post421"
+version = "6.6.1.post424"
 license = "AGPL-3.0-or-later"
 description = "Protobuf definitions for nucliadb/nidx"
 authors = [{ name = "Nuclia", email = "nucliadb@nuclia.com" }]

{nidx_binding-6.6.1.post421 → nidx_binding-6.6.1.post424}/pyproject.toml

@@ -13,7 +13,7 @@ classifiers = [
     "Programming Language :: Python :: Implementation :: CPython",
     "Programming Language :: Python :: Implementation :: PyPy",
 ]
-version = "6.6.1.post421"
+version = "6.6.1.post424"
 
 [project.urls]
 Homepage = "https://nuclia.com"