community[patch]: add ?| (arrayContains) filter on metadata to PGVect…

…or search (#5381) * add ?| (arrayContains) filter on metadata to PGVector search * fix: align to style of IN filter * Format * Update build artifacts * Format --------- Co-authored-by: jacoblee93 <[email protected]>
langchain-ai · May 26, 2024 · 2494ed7 · 2494ed7
1 parent 28785cf
commit 2494ed7
Show file tree

Hide file tree

Showing 3 changed files with 65 additions and 2 deletions.
diff --git a/examples/src/indexes/vector_stores/pgvector_vectorstore/pgvector.ts b/examples/src/indexes/vector_stores/pgvector_vectorstore/pgvector.ts
@@ -34,8 +34,8 @@ const pgvectorStore = await PGVectorStore.initialize(
 );
 
 await pgvectorStore.addDocuments([
-  { pageContent: "what's this", metadata: { a: 2 } },
-  { pageContent: "Cat drinks milk", metadata: { a: 1 } },
+  { pageContent: "what's this", metadata: { a: 2, b: ["tag1", "tag2"] } },
+  { pageContent: "Cat drinks milk", metadata: { a: 1, b: ["tag2"] } },
 ]);
 
 const results = await pgvectorStore.similaritySearch("water", 1);
@@ -84,4 +84,17 @@ console.log(results4);
   [ Document { pageContent: 'what's this', metadata: { a: 2 } } ]
 */
 
+// Filtering using arrayContains (?|) is supported
+const results5 = await pgvectorStore.similaritySearch("water", 1, {
+  b: {
+    arrayContains: ["tag1"],
+  },
+});
+
+console.log(results5);
+
+/*
+  [ Document { pageContent: "what's this", metadata: { a: 2, b: ['tag1', 'tag2'] } } } ]
+*/
+
 await pgvectorStore.end();
diff --git a/libs/langchain-community/src/vectorstores/pgvector.ts b/libs/langchain-community/src/vectorstores/pgvector.ts
@@ -501,6 +501,18 @@ export class PGVectorStore extends VectorStore {
           parameters.push(..._value.in);
           paramCount += _value.in.length;
         }
+        if (Array.isArray(_value.arrayContains)) {
+          const placeholders = _value.arrayContains
+            .map(
+              (_: unknown, index: number) => `$${currentParamCount + index + 1}`
+            )
+            .join(",");
+          whereClauses.push(
+            `${this.metadataColumnName}->'${key}' ?| array[${placeholders}]`
+          );
+          parameters.push(..._value.arrayContains);
+          paramCount += _value.arrayContains.length;
+        }
       } else {
         paramCount += 1;
         whereClauses.push(

diff --git a/libs/langchain-community/src/vectorstores/tests/pgvector/pgvector.int.test.ts b/libs/langchain-community/src/vectorstores/tests/pgvector/pgvector.int.test.ts
@@ -141,6 +141,44 @@ describe("PGVectorStore", () => {
     expect(result3.length).toEqual(3);
   });
 
+  test("PGvector supports arrayContains (?|) in metadata filter ", async () => {
+    const documents = [
+      { pageContent: "Lorem Ipsum", metadata: { a: ["tag1", "tag2"] } },
+      { pageContent: "Lorem Ipsum", metadata: { a: ["tag2"] } },
+      { pageContent: "Lorem Ipsum", metadata: { a: ["tag1"] } },
+    ];
+
+    await pgvectorVectorStore.addDocuments(documents);
+
+    const result = await pgvectorVectorStore.similaritySearch("hello", 2, {
+      a: {
+        arrayContains: ["tag1"],
+      },
+    });
+
+    expect(result.length).toEqual(2);
+    expect(result).toEqual([
+      { pageContent: "Lorem Ipsum", metadata: { a: ["tag1", "tag2"] } },
+      { pageContent: "Lorem Ipsum", metadata: { a: ["tag1"] } },
+    ]);
+
+    const result2 = await pgvectorVectorStore.similaritySearch("hello", 2, {
+      a: {
+        arrayContains: ["tag2"],
+      },
+    });
+    expect(result2.length).toEqual(2);
+    expect(result2).toEqual([
+      { pageContent: "Lorem Ipsum", metadata: { a: ["tag1", "tag2"] } },
+      { pageContent: "Lorem Ipsum", metadata: { a: ["tag2"] } },
+    ]);
+
+    const result3 = await pgvectorVectorStore.similaritySearch("hello", 3);
+
+    expect(result3.length).toEqual(3);
+    expect(result3).toEqual(documents);
+  });
+
   test("PGvector can delete document by id", async () => {
     try {
       const documents = [