fix(pgvector): add skipInitializationCheck option to disable schema c… (#8414)

Author: hntrl

libs/langchain-community/src/vectorstores/pgvector.ts

@@ -26,6 +26,7 @@ export interface PGVectorStoreArgs {
   collectionMetadata?: Metadata | null;
   schemaName?: string | null;
   extensionSchemaName?: string | null;
+  skipInitializationCheck?: boolean;
   columns?: {
     idColumnName?: string;
     vectorColumnName?: string;
@@ -213,6 +214,8 @@ export class PGVectorStore extends VectorStore {
 
   extensionSchemaName: string | null;
 
+  skipInitializationCheck: boolean;
+
   metadataColumnName: string;
 
   filter?: Metadata;
@@ -247,6 +250,7 @@ export class PGVectorStore extends VectorStore {
     this.collectionMetadata = config.collectionMetadata ?? null;
     this.schemaName = config.schemaName ?? null;
     this.extensionSchemaName = config.extensionSchemaName ?? null;
+    this.skipInitializationCheck = config.skipInitializationCheck ?? false;
 
     this.filter = config.filter;
 
@@ -741,6 +745,9 @@ export class PGVectorStore extends VectorStore {
    * @returns Promise that resolves when the table has been ensured.
    */
   async ensureTableInDatabase(dimensions?: number): Promise<void> {
+    if (this.skipInitializationCheck) {
+      return;
+    }
     const vectorQuery =
       this.extensionSchemaName == null
         ? "CREATE EXTENSION IF NOT EXISTS vector;"
@@ -772,6 +779,9 @@ export class PGVectorStore extends VectorStore {
    */
   async ensureCollectionTableInDatabase(): Promise<void> {
     try {
+      if (this.skipInitializationCheck) {
+        return;
+      }
       const queryString = `
         CREATE TABLE IF NOT EXISTS ${this.computedCollectionTableName} (
           uuid uuid NOT NULL DEFAULT gen_random_uuid() PRIMARY KEY,

libs/langchain-community/src/vectorstores/tests/pgvector/pgvector.int.test.ts

@@ -821,6 +821,131 @@ describe("PGVectorStore with schema", () => {
   });
 });
 
+describe("PGVectorStore with skipInitializationCheck", () => {
+  let pgvectorVectorStore: PGVectorStore;
+  const tableName = "testlangchain_skip_init";
+
+  afterEach(async () => {
+    const pool = new pg.Pool(postgresConnectionOptions);
+    await pool.query(`DROP TABLE IF EXISTS "${tableName}"`);
+    await pool.end();
+  });
+
+  test("skipInitializationCheck=false (default) should initialize tables", async () => {
+    const config: PGVectorStoreArgs = {
+      postgresConnectionOptions,
+      tableName,
+      columns: {
+        idColumnName: "id",
+        vectorColumnName: "vector",
+        contentColumnName: "content",
+        metadataColumnName: "metadata",
+      },
+    };
+
+    pgvectorVectorStore = await PGVectorStore.initialize(
+      embeddingsEngine,
+      config
+    );
+
+    const result = await pgvectorVectorStore.pool.query(
+      `SELECT EXISTS (
+        SELECT FROM information_schema.tables 
+        WHERE table_schema = 'public' 
+        AND table_name = '${tableName}'
+      )`
+    );
+
+    expect(result.rows[0].exists).toBe(true);
+    await pgvectorVectorStore.end();
+  });
+
+  test("skipInitializationCheck=true should skip table initialization", async () => {
+    const config: PGVectorStoreArgs = {
+      postgresConnectionOptions,
+      tableName,
+      skipInitializationCheck: true,
+      columns: {
+        idColumnName: "id",
+        vectorColumnName: "vector",
+        contentColumnName: "content",
+        metadataColumnName: "metadata",
+      },
+    };
+
+    pgvectorVectorStore = await PGVectorStore.initialize(
+      embeddingsEngine,
+      config
+    );
+
+    const tableExists = await pgvectorVectorStore.pool.query(
+      `SELECT EXISTS (
+        SELECT FROM information_schema.tables 
+        WHERE table_schema = 'public' 
+        AND table_name = '${tableName}'
+      )`
+    );
+
+    expect(tableExists.rows[0].exists).toBe(false);
+
+    await pgvectorVectorStore.end();
+  });
+
+  test("skipInitializationCheck=true should work with addDocuments", async () => {
+    const setupPool = new pg.Pool(postgresConnectionOptions);
+
+    await setupPool.query(`CREATE EXTENSION IF NOT EXISTS vector`);
+    await setupPool.query(`
+      CREATE TABLE "${tableName}" (
+        "id" uuid NOT NULL DEFAULT gen_random_uuid() PRIMARY KEY,
+        "content" text,
+        "metadata" jsonb,
+        "vector" vector(1536)
+      );
+    `);
+
+    await setupPool.end();
+
+    const config: PGVectorStoreArgs = {
+      postgresConnectionOptions,
+      tableName,
+      skipInitializationCheck: true,
+      columns: {
+        idColumnName: "id",
+        vectorColumnName: "vector",
+        contentColumnName: "content",
+        metadataColumnName: "metadata",
+      },
+    };
+
+    pgvectorVectorStore = await PGVectorStore.initialize(
+      embeddingsEngine,
+      config
+    );
+
+    const documents = [
+      { pageContent: "Hello world", metadata: { source: "test" } },
+      {
+        pageContent: "Testing skipInitializationCheck",
+        metadata: { source: "test" },
+      },
+    ];
+
+    await pgvectorVectorStore.addDocuments(documents);
+
+    const query = await embeddingsEngine.embedQuery("Hello");
+    const results = await pgvectorVectorStore.similaritySearchVectorWithScore(
+      query,
+      1
+    );
+
+    expect(results).toHaveLength(1);
+    expect(results[0][0].pageContent).toBe("Hello world");
+
+    await pgvectorVectorStore.end();
+  });
+});
+
 describe("PGVectorStore with HNSW index", () => {
   let pgvectorVectorStore: PGVectorStore;
   const tableName = "testlangchain";