Release 0.1.44 (#266)

Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>

.github/workflows/e2e.yml

@@ -20,18 +20,6 @@ jobs:
         os: [macos-latest, windows-latest, ubuntu-22.04]
         frameworks: ["nextjs", "express", "fastapi"]
         datasources: ["--no-files", "--example-file"]
-        templates: ["streaming", "extractor"]
-        exclude:
-          # The extractor template currently only works with FastAPI and files,
-          # and it's not compatible with Windows at the moment.
-          - templates: "extractor"
-            os: windows-latest
-          - templates: "extractor"
-            frameworks: "nextjs"
-          - templates: "extractor"
-            frameworks: "express"
-          - templates: "extractor"
-            datasources: "--no-files"
     defaults:
       run:
         shell: bash
@@ -77,7 +65,6 @@ jobs:
         env:
           OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
           LLAMA_CLOUD_API_KEY: ${{ secrets.LLAMA_CLOUD_API_KEY }}
-          TEMPLATE: ${{ matrix.templates }}
           FRAMEWORK: ${{ matrix.frameworks }}
           DATASOURCE: ${{ matrix.datasources }}
         working-directory: .

CHANGELOG.md

@@ -1,5 +1,27 @@
 # create-llama
 
+## 0.1.44
+
+### Patch Changes
+
+- bedde2b: Change metadata filters to use already existing documents in LlamaCloud Index
+- 5cd12fa: Use one callback manager per request
+- 5cd12fa: Bump llama_index version to 0.11.1
+- fd4abb3: Fix to use filename for uploaded documents in NextJS
+- 2f8feab: Simplify CLI interface
+
+## 0.1.43
+
+### Patch Changes
+
+- 4fa2b76: feat: implement citation for TS
+
+## 0.1.42
+
+### Patch Changes
+
+- 8f670a9: Allow relative URL in documents
+
 ## 0.1.41
 
 ### Patch Changes

README.md

@@ -94,7 +94,7 @@ Need to install the following packages:
   create-llama@latest
 Ok to proceed? (y) y
 ✔ What is your project named? … my-app
-✔ Which template would you like to use? › Agentic RAG (single agent)
+✔ Which template would you like to use? › Agentic RAG (e.g. chat with docs)
 ✔ Which framework would you like to use? › NextJS
 ✔ Would you like to set up observability? › No
 ✔ Please provide your OpenAI API key (leave blank to skip): …

e2e/extractor_template.spec.ts

@@ -3,9 +3,23 @@ import { expect, test } from "@playwright/test";
 import { ChildProcess } from "child_process";
 import fs from "fs";
 import path from "path";
+import { TemplateFramework } from "../helpers";
 import { createTestDir, runCreateLlama } from "./utils";
 
-if (process.env.TEMPLATE === "extractor") {
+const templateFramework: TemplateFramework = process.env.FRAMEWORK
+  ? (process.env.FRAMEWORK as TemplateFramework)
+  : "fastapi";
+const dataSource: string = process.env.DATASOURCE
+  ? process.env.DATASOURCE
+  : "--example-file";
+
+// The extractor template currently only works with FastAPI and files (and not on Windows)
+if (
+  process.platform !== "win32" &&
+  templateFramework !== "nextjs" &&
+  templateFramework !== "express" &&
+  dataSource !== "--no-files"
+) {
   test.describe("Test extractor template", async () => {
     let frontendPort: number;
     let backendPort: number;

e2e/streaming_template.spec.ts

@@ -6,14 +6,10 @@ import path from "path";
 import type {
   TemplateFramework,
   TemplatePostInstallAction,
-  TemplateType,
   TemplateUI,
 } from "../helpers";
 import { createTestDir, runCreateLlama, type AppType } from "./utils";
 
-const templateType: TemplateType | undefined = process.env.TEMPLATE
-  ? (process.env.TEMPLATE as TemplateType)
-  : undefined;
 const templateFramework: TemplateFramework = process.env.FRAMEWORK
   ? (process.env.FRAMEWORK as TemplateFramework)
   : "fastapi";
@@ -30,96 +26,94 @@ const appType: AppType = templateFramework === "nextjs" ? "" : "--frontend";
 const userMessage =
   dataSource !== "--no-files" ? "Physical standard for letters" : "Hello";
 
-if (templateType === "streaming") {
-  test.describe(`Test streaming template ${templateFramework} ${dataSource} ${templateUI} ${appType} ${templatePostInstallAction}`, async () => {
-    let port: number;
-    let externalPort: number;
-    let cwd: string;
-    let name: string;
-    let appProcess: ChildProcess;
-    // Only test without using vector db for now
-    const vectorDb = "none";
+test.describe(`Test streaming template ${templateFramework} ${dataSource} ${templateUI} ${appType} ${templatePostInstallAction}`, async () => {
+  let port: number;
+  let externalPort: number;
+  let cwd: string;
+  let name: string;
+  let appProcess: ChildProcess;
+  // Only test without using vector db for now
+  const vectorDb = "none";
 
-    test.beforeAll(async () => {
-      port = Math.floor(Math.random() * 10000) + 10000;
-      externalPort = port + 1;
-      cwd = await createTestDir();
-      const result = await runCreateLlama(
-        cwd,
-        templateType,
-        templateFramework,
-        dataSource,
-        vectorDb,
-        port,
-        externalPort,
-        templatePostInstallAction,
-        templateUI,
-        appType,
-        llamaCloudProjectName,
-        llamaCloudIndexName,
-      );
-      name = result.projectName;
-      appProcess = result.appProcess;
-    });
-
-    test("App folder should exist", async () => {
-      const dirExists = fs.existsSync(path.join(cwd, name));
-      expect(dirExists).toBeTruthy();
-    });
-    test("Frontend should have a title", async ({ page }) => {
-      test.skip(templatePostInstallAction !== "runApp");
-      await page.goto(`http://localhost:${port}`);
-      await expect(page.getByText("Built by LlamaIndex")).toBeVisible();
-    });
-
-    test("Frontend should be able to submit a message and receive a response", async ({
-      page,
-    }) => {
-      test.skip(templatePostInstallAction !== "runApp");
-      await page.goto(`http://localhost:${port}`);
-      await page.fill("form input", userMessage);
-      const [response] = await Promise.all([
-        page.waitForResponse(
-          (res) => {
-            return res.url().includes("/api/chat") && res.status() === 200;
-          },
-          {
-            timeout: 1000 * 60,
-          },
-        ),
-        page.click("form button[type=submit]"),
-      ]);
-      const text = await response.text();
-      console.log("AI response when submitting message: ", text);
-      expect(response.ok()).toBeTruthy();
-    });
-
-    test("Backend frameworks should response when calling non-streaming chat API", async ({
-      request,
-    }) => {
-      test.skip(templatePostInstallAction !== "runApp");
-      test.skip(templateFramework === "nextjs");
-      const response = await request.post(
-        `http://localhost:${externalPort}/api/chat/request`,
-        {
-          data: {
-            messages: [
-              {
-                role: "user",
-                content: userMessage,
-              },
-            ],
-          },
-        },
-      );
-      const text = await response.text();
-      console.log("AI response when calling API: ", text);
-      expect(response.ok()).toBeTruthy();
-    });
-
-    // clean processes
-    test.afterAll(async () => {
-      appProcess?.kill();
-    });
+  test.beforeAll(async () => {
+    port = Math.floor(Math.random() * 10000) + 10000;
+    externalPort = port + 1;
+    cwd = await createTestDir();
+    const result = await runCreateLlama(
+      cwd,
+      "streaming",
+      templateFramework,
+      dataSource,
+      vectorDb,
+      port,
+      externalPort,
+      templatePostInstallAction,
+      templateUI,
+      appType,
+      llamaCloudProjectName,
+      llamaCloudIndexName,
+    );
+    name = result.projectName;
+    appProcess = result.appProcess;
   });
-}
+
+  test("App folder should exist", async () => {
+    const dirExists = fs.existsSync(path.join(cwd, name));
+    expect(dirExists).toBeTruthy();
+  });
+  test("Frontend should have a title", async ({ page }) => {
+    test.skip(templatePostInstallAction !== "runApp");
+    await page.goto(`http://localhost:${port}`);
+    await expect(page.getByText("Built by LlamaIndex")).toBeVisible();
+  });
+
+  test("Frontend should be able to submit a message and receive a response", async ({
+    page,
+  }) => {
+    test.skip(templatePostInstallAction !== "runApp");
+    await page.goto(`http://localhost:${port}`);
+    await page.fill("form input", userMessage);
+    const [response] = await Promise.all([
+      page.waitForResponse(
+        (res) => {
+          return res.url().includes("/api/chat") && res.status() === 200;
+        },
+        {
+          timeout: 1000 * 60,
+        },
+      ),
+      page.click("form button[type=submit]"),
+    ]);
+    const text = await response.text();
+    console.log("AI response when submitting message: ", text);
+    expect(response.ok()).toBeTruthy();
+  });
+
+  test("Backend frameworks should response when calling non-streaming chat API", async ({
+    request,
+  }) => {
+    test.skip(templatePostInstallAction !== "runApp");
+    test.skip(templateFramework === "nextjs");
+    const response = await request.post(
+      `http://localhost:${externalPort}/api/chat/request`,
+      {
+        data: {
+          messages: [
+            {
+              role: "user",
+              content: userMessage,
+            },
+          ],
+        },
+      },
+    );
+    const text = await response.text();
+    console.log("AI response when calling API: ", text);
+    expect(response.ok()).toBeTruthy();
+  });
+
+  // clean processes
+  test.afterAll(async () => {
+    appProcess?.kill();
+  });
+});

helpers/env-variables.ts

@@ -454,12 +454,7 @@ const getSystemPromptEnv = (
     },
   ];
 
-  // Citation only works with FastAPI along with the chat engine and data source provided for now.
-  if (
-    framework === "fastapi" &&
-    tools?.length == 0 &&
-    (dataSources?.length ?? 0 > 0)
-  ) {
+  if (tools?.length == 0 && (dataSources?.length ?? 0 > 0)) {
     const citationPrompt = `'You have provided information from a knowledge base that has been passed to you in nodes of information.
 Each node has useful metadata such as node ID, file name, page, etc.
 Please add the citation to the data node for each sentence or paragraph that you reference in the provided information.

helpers/python.ts

@@ -12,6 +12,7 @@ import {
   InstallTemplateArgs,
   ModelConfig,
   TemplateDataSource,
+  TemplateType,
   TemplateVectorDB,
 } from "./types";
 
@@ -26,6 +27,7 @@ const getAdditionalDependencies = (
   vectorDb?: TemplateVectorDB,
   dataSources?: TemplateDataSource[],
   tools?: Tool[],
+  templateType?: TemplateType,
 ) => {
   const dependencies: Dependency[] = [];
 
@@ -128,7 +130,7 @@ const getAdditionalDependencies = (
         case "llamacloud":
           dependencies.push({
             name: "llama-index-indices-managed-llama-cloud",
-            version: "^0.2.7",
+            version: "^0.3.0",
           });
           break;
       }
@@ -147,77 +149,99 @@ const getAdditionalDependencies = (
     case "ollama":
       dependencies.push({
         name: "llama-index-llms-ollama",
-        version: "0.1.2",
+        version: "0.3.0",
       });
       dependencies.push({
         name: "llama-index-embeddings-ollama",
-        version: "0.1.2",
+        version: "0.3.0",
       });
       break;
     case "openai":
-      dependencies.push({
-        name: "llama-index-agent-openai",
-        version: "0.2.6",
-      });
+      if (templateType !== "multiagent") {
+        dependencies.push({
+          name: "llama-index-llms-openai",
+          version: "^0.2.0",
+        });
+        dependencies.push({
+          name: "llama-index-embeddings-openai",
+          version: "^0.2.3",
+        });
+        dependencies.push({
+          name: "llama-index-agent-openai",
+          version: "^0.3.0",
+        });
+      }
       break;
     case "groq":
+      // Fastembed==0.2.0 does not support python3.13 at the moment
+      // Fixed the python version less than 3.13
+      dependencies.push({
+        name: "python",
+        version: "^3.11,<3.13",
+      });
       dependencies.push({
         name: "llama-index-llms-groq",
-        version: "0.1.4",
+        version: "0.2.0",
       });
       dependencies.push({
         name: "llama-index-embeddings-fastembed",
-        version: "^0.1.4",
+        version: "^0.2.0",
       });
       break;
     case "anthropic":
+      // Fastembed==0.2.0 does not support python3.13 at the moment
+      // Fixed the python version less than 3.13
+      dependencies.push({
+        name: "python",
+        version: "^3.11,<3.13",
+      });
       dependencies.push({
         name: "llama-index-llms-anthropic",
-        version: "0.1.10",
+        version: "0.3.0",
       });
       dependencies.push({
         name: "llama-index-embeddings-fastembed",
-        version: "^0.1.4",
+        version: "^0.2.0",
       });
       break;
     case "gemini":
       dependencies.push({
         name: "llama-index-llms-gemini",
-        version: "0.1.10",
+        version: "0.3.4",
       });
       dependencies.push({
         name: "llama-index-embeddings-gemini",
-        version: "0.1.6",
+        version: "^0.2.0",
       });
       break;
     case "mistral":
       dependencies.push({
         name: "llama-index-llms-mistralai",
-        version: "0.1.17",
+        version: "0.2.1",
       });
       dependencies.push({
         name: "llama-index-embeddings-mistralai",
-        version: "0.1.4",
+        version: "0.2.0",
       });
       break;
     case "azure-openai":
       dependencies.push({
         name: "llama-index-llms-azure-openai",
-        version: "0.1.10",
+        version: "0.2.0",
       });
       dependencies.push({
         name: "llama-index-embeddings-azure-openai",
-        version: "0.1.11",
+        version: "0.2.4",
       });
       break;
     case "t-systems":
       dependencies.push({
         name: "llama-index-agent-openai",
-        version: "0.2.2",
+        version: "0.3.0",
       });
       dependencies.push({
         name: "llama-index-llms-openai-like",
-        version: "0.1.3",
+        version: "0.2.0",
       });
       break;
   }
@@ -227,7 +251,7 @@ const getAdditionalDependencies = (
 
 const mergePoetryDependencies = (
   dependencies: Dependency[],
-  existingDependencies: Record<string, Omit<Dependency, "name">>,
+  existingDependencies: Record<string, Omit<Dependency, "name"> | string>,
 ) => {
   for (const dependency of dependencies) {
     let value = existingDependencies[dependency.name] ?? {};
@@ -246,7 +270,13 @@ const mergePoetryDependencies = (
       );
     }
 
-    existingDependencies[dependency.name] = value;
+    // Serialize separately only if extras are provided
+    if (value.extras && value.extras.length > 0) {
+      existingDependencies[dependency.name] = value;
+    } else {
+      // Otherwise, serialize just the version string
+      existingDependencies[dependency.name] = value.version;
+    }
   }
 };
 
@@ -388,6 +418,7 @@ export const installPythonTemplate = async ({
     vectorDb,
     dataSources,
     tools,
+    template,
   );
 
   if (observability && observability !== "none") {

index.ts

@@ -173,7 +173,14 @@ const program = new Commander.Command(packageJson.name)
     "--ask-models",
     `
 
-  Select LLM and embedding models.
+  Allow interactive selection of LLM and embedding models of different model providers.
+`,
+  )
+  .option(
+    "--ask-examples",
+    `
+
+  Allow interactive selection of community templates and LlamaPacks.
 `,
   )
   .allowUnknownOption()
@@ -195,6 +202,7 @@ if (
   program.useLlamaParse = false;
 }
 program.askModels = process.argv.includes("--ask-models");
+program.askExamples = process.argv.includes("--ask-examples");
 if (process.argv.includes("--no-files")) {
   program.dataSources = [];
 } else if (process.argv.includes("--example-file")) {

package.json

@@ -1,6 +1,6 @@
 {
   "name": "create-llama",
-  "version": "0.1.41",
+  "version": "0.1.44",
   "description": "Create LlamaIndex-powered apps with one command",
   "keywords": [
     "rag",

questions.ts

@@ -28,6 +28,7 @@ export type QuestionArgs = Omit<
   "appPath" | "packageManager"
 > & {
   askModels?: boolean;
+  askExamples?: boolean;
 };
 const supportedContextFileTypes = [
   ".pdf",
@@ -338,20 +339,24 @@ export const askQuestions = async (
           name: "template",
           message: "Which template would you like to use?",
           choices: [
-            { title: "Agentic RAG (single agent)", value: "streaming" },
+            { title: "Agentic RAG (e.g. chat with docs)", value: "streaming" },
             {
               title: "Multi-agent app (using llama-agents)",
               value: "multiagent",
             },
             { title: "Structured Extractor", value: "extractor" },
-            {
-              title: `Community template from ${styledRepo}`,
-              value: "community",
-            },
-            {
-              title: "Example using a LlamaPack",
-              value: "llamapack",
-            },
+            ...(program.askExamples
+              ? [
+                  {
+                    title: `Community template from ${styledRepo}`,
+                    value: "community",
+                  },
+                  {
+                    title: "Example using a LlamaPack",
+                    value: "llamapack",
+                  },
+                ]
+              : []),
           ],
           initial: 0,
         },

templates/components/engines/python/agent/engine.py

@@ -1,19 +1,22 @@
 import os
 
-from app.engine.index import get_index
+from app.engine.index import IndexConfig, get_index
 from app.engine.tools import ToolFactory
 from llama_index.core.agent import AgentRunner
+from llama_index.core.callbacks import CallbackManager
 from llama_index.core.settings import Settings
 from llama_index.core.tools.query_engine import QueryEngineTool
 
 
-def get_chat_engine(filters=None, params=None):
+def get_chat_engine(filters=None, params=None, event_handlers=None):
     system_prompt = os.getenv("SYSTEM_PROMPT")
     top_k = int(os.getenv("TOP_K", 0))
     tools = []
+    callback_manager = CallbackManager(handlers=event_handlers or [])
 
     # Add query tool if index exists
-    index = get_index()
+    index_config = IndexConfig(callback_manager=callback_manager, **(params or {}))
+    index = get_index(index_config)
     if index is not None:
         query_engine = index.as_query_engine(
             filters=filters, **({"similarity_top_k": top_k} if top_k != 0 else {})
@@ -28,5 +31,6 @@ def get_chat_engine(filters=None, params=None):
         llm=Settings.llm,
         tools=tools,
         system_prompt=system_prompt,
+        callback_manager=callback_manager,
         verbose=True,
     )

templates/components/engines/python/chat/engine.py

@@ -1,22 +1,31 @@
 import os
 
-from app.engine.index import get_index
+from app.engine.index import IndexConfig, get_index
 from app.engine.node_postprocessors import NodeCitationProcessor
 from fastapi import HTTPException
+from llama_index.core.callbacks import CallbackManager
 from llama_index.core.chat_engine import CondensePlusContextChatEngine
+from llama_index.core.memory import ChatMemoryBuffer
+from llama_index.core.settings import Settings
 
 
-def get_chat_engine(filters=None, params=None):
+def get_chat_engine(filters=None, params=None, event_handlers=None):
     system_prompt = os.getenv("SYSTEM_PROMPT")
     citation_prompt = os.getenv("SYSTEM_CITATION_PROMPT", None)
     top_k = int(os.getenv("TOP_K", 0))
+    llm = Settings.llm
+    memory = ChatMemoryBuffer.from_defaults(
+        token_limit=llm.metadata.context_window - 256
+    )
+    callback_manager = CallbackManager(handlers=event_handlers or [])
 
     node_postprocessors = []
     if citation_prompt:
         node_postprocessors = [NodeCitationProcessor()]
         system_prompt = f"{system_prompt}\n{citation_prompt}"
 
-    index = get_index(params)
+    index_config = IndexConfig(callback_manager=callback_manager, **(params or {}))
+    index = get_index(index_config)
     if index is None:
         raise HTTPException(
             status_code=500,
@@ -29,8 +38,11 @@ def get_chat_engine(filters=None, params=None):
         filters=filters, **({"similarity_top_k": top_k} if top_k != 0 else {})
     )
 
-    return CondensePlusContextChatEngine.from_defaults(
+    return CondensePlusContextChatEngine(
+        llm=llm,
+        memory=memory,
         system_prompt=system_prompt,
         retriever=retriever,
         node_postprocessors=node_postprocessors,
+        callback_manager=callback_manager,
     )

templates/components/engines/typescript/chat/chat.ts

@@ -1,5 +1,6 @@
 import { ContextChatEngine, Settings } from "llamaindex";
 import { getDataSource } from "./index";
+import { nodeCitationProcessor } from "./nodePostprocessors";
 import { generateFilters } from "./queryFilter";
 
 export async function createChatEngine(documentIds?: string[], params?: any) {
@@ -14,9 +15,18 @@ export async function createChatEngine(documentIds?: string[], params?: any) {
     filters: generateFilters(documentIds || []),
   });
 
+  const systemPrompt = process.env.SYSTEM_PROMPT;
+  const citationPrompt = process.env.SYSTEM_CITATION_PROMPT;
+  const prompt =
+    [systemPrompt, citationPrompt].filter((p) => p).join("\n") || undefined;
+  const nodePostprocessors = citationPrompt
+    ? [nodeCitationProcessor]
+    : undefined;
+
   return new ContextChatEngine({
     chatModel: Settings.llm,
     retriever,
-    systemPrompt: process.env.SYSTEM_PROMPT,
+    systemPrompt: prompt,
+    nodePostprocessors,
   });
 }

templates/components/engines/typescript/chat/nodePostprocessors.ts

@@ -0,0 +1,26 @@
+import {
+  BaseNodePostprocessor,
+  MessageContent,
+  NodeWithScore,
+} from "llamaindex";
+
+class NodeCitationProcessor implements BaseNodePostprocessor {
+  /**
+   * Append node_id into metadata for citation purpose.
+   * Config SYSTEM_CITATION_PROMPT in your runtime environment variable to enable this feature.
+   */
+  async postprocessNodes(
+    nodes: NodeWithScore[],
+    query?: MessageContent,
+  ): Promise<NodeWithScore[]> {
+    for (const nodeScore of nodes) {
+      if (!nodeScore.node || !nodeScore.node.metadata) {
+        continue; // Skip nodes with missing properties
+      }
+      nodeScore.node.metadata["node_id"] = nodeScore.node.id_;
+    }
+    return nodes;
+  }
+}
+
+export const nodeCitationProcessor = new NodeCitationProcessor();

templates/components/llamaindex/typescript/documents/helper.ts

@@ -1,5 +1,4 @@
 import fs from "fs";
-import crypto from "node:crypto";
 import { getExtractors } from "../../engine/loader";
 
 const MIME_TYPE_TO_EXT: Record<string, string> = {
@@ -11,9 +10,13 @@ const MIME_TYPE_TO_EXT: Record<string, string> = {
 
 const UPLOADED_FOLDER = "output/uploaded";
 
-export async function storeAndParseFile(fileBuffer: Buffer, mimeType: string) {
+export async function storeAndParseFile(
+  filename: string,
+  fileBuffer: Buffer,
+  mimeType: string,
+) {
   const documents = await loadDocuments(fileBuffer, mimeType);
-  const { filename } = await saveDocument(fileBuffer, mimeType);
+  await saveDocument(filename, fileBuffer, mimeType);
   for (const document of documents) {
     document.metadata = {
       ...document.metadata,
@@ -35,11 +38,14 @@ async function loadDocuments(fileBuffer: Buffer, mimeType: string) {
   return await reader.loadDataAsContent(fileBuffer);
 }
 
-async function saveDocument(fileBuffer: Buffer, mimeType: string) {
+async function saveDocument(
+  filename: string,
+  fileBuffer: Buffer,
+  mimeType: string,
+) {
   const fileExt = MIME_TYPE_TO_EXT[mimeType];
   if (!fileExt) throw new Error(`Unsupported document type: ${mimeType}`);
 
-  const filename = `${crypto.randomUUID()}.${fileExt}`;
   const filepath = `${UPLOADED_FOLDER}/${filename}`;
   const fileurl = `${process.env.FILESERVER_URL_PREFIX}/${filepath}`;
 

templates/components/llamaindex/typescript/documents/upload.ts

@@ -27,6 +27,6 @@ export async function uploadDocument(
   }
 
   // run the pipeline for other vector store indexes
-  const documents = await storeAndParseFile(fileBuffer, mimeType);
+  const documents = await storeAndParseFile(filename, fileBuffer, mimeType);
   return runPipeline(index, documents);
 }

templates/components/vectordbs/python/llamacloud/index.py

@@ -1,41 +1,87 @@
 import logging
 import os
-from llama_index.indices.managed.llama_cloud import LlamaCloudIndex
+from typing import Optional
+
+from llama_index.core.callbacks import CallbackManager
 from llama_index.core.ingestion.api_utils import (
     get_client as llama_cloud_get_client,
 )
+from llama_index.indices.managed.llama_cloud import LlamaCloudIndex
+from pydantic import BaseModel, Field, validator
 
 logger = logging.getLogger("uvicorn")
 
 
-def get_client():
-    return llama_cloud_get_client(
-        os.getenv("LLAMA_CLOUD_API_KEY"),
-        os.getenv("LLAMA_CLOUD_BASE_URL"),
+class LlamaCloudConfig(BaseModel):
+    # Private attributes
+    api_key: str = Field(
+        default=os.getenv("LLAMA_CLOUD_API_KEY"),
+        exclude=True,  # Exclude from the model representation
+    )
+    base_url: Optional[str] = Field(
+        default=os.getenv("LLAMA_CLOUD_BASE_URL"),
+        exclude=True,
+    )
+    organization_id: Optional[str] = Field(
+        default=os.getenv("LLAMA_CLOUD_ORGANIZATION_ID"),
+        exclude=True,
+    )
+    # Configuration attributes, can be set by the user
+    pipeline: str = Field(
+        description="The name of the pipeline to use",
+        default=os.getenv("LLAMA_CLOUD_INDEX_NAME"),
+    )
+    project: str = Field(
+        description="The name of the LlamaCloud project",
+        default=os.getenv("LLAMA_CLOUD_PROJECT_NAME"),
     )
 
+    # Validate and throw error if the env variables are not set before starting the app
+    @validator("pipeline", "project", "api_key", pre=True, always=True)
+    @classmethod
+    def validate_env_vars(cls, value):
+        if value is None:
+            raise ValueError(
+                "Please set LLAMA_CLOUD_INDEX_NAME, LLAMA_CLOUD_PROJECT_NAME and LLAMA_CLOUD_API_KEY"
+                " to your environment variables or config them in .env file"
+            )
+        return value
 
-def get_index(params=None):
-    configParams = params or {}
-    pipelineConfig = configParams.get("llamaCloudPipeline", {})
-    name = pipelineConfig.get("pipeline", os.getenv("LLAMA_CLOUD_INDEX_NAME"))
-    project_name = pipelineConfig.get("project", os.getenv("LLAMA_CLOUD_PROJECT_NAME"))
-    api_key = os.getenv("LLAMA_CLOUD_API_KEY")
-    base_url = os.getenv("LLAMA_CLOUD_BASE_URL")
-    organization_id = os.getenv("LLAMA_CLOUD_ORGANIZATION_ID")
+    def to_client_kwargs(self) -> dict:
+        return {
+            "api_key": self.api_key,
+            "base_url": self.base_url,
+        }
 
-    if name is None or project_name is None or api_key is None:
-        raise ValueError(
-            "Please set LLAMA_CLOUD_INDEX_NAME, LLAMA_CLOUD_PROJECT_NAME and LLAMA_CLOUD_API_KEY"
-            " to your environment variables or config them in .env file"
-        )
 
-    index = LlamaCloudIndex(
-        name=name,
-        project_name=project_name,
-        api_key=api_key,
-        base_url=base_url,
-        organization_id=organization_id,
+class IndexConfig(BaseModel):
+    llama_cloud_pipeline_config: LlamaCloudConfig = Field(
+        default=LlamaCloudConfig(),
+        alias="llamaCloudPipeline",
     )
+    callback_manager: Optional[CallbackManager] = Field(
+        default=None,
+    )
+
+    def to_index_kwargs(self) -> dict:
+        return {
+            "name": self.llama_cloud_pipeline_config.pipeline,
+            "project_name": self.llama_cloud_pipeline_config.project,
+            "api_key": self.llama_cloud_pipeline_config.api_key,
+            "base_url": self.llama_cloud_pipeline_config.base_url,
+            "organization_id": self.llama_cloud_pipeline_config.organization_id,
+            "callback_manager": self.callback_manager,
+        }
+
+
+def get_index(config: IndexConfig = None):
+    if config is None:
+        config = IndexConfig()
+    index = LlamaCloudIndex(**config.to_index_kwargs())
 
     return index
+
+
+def get_client():
+    config = LlamaCloudConfig()
+    return llama_cloud_get_client(**config.to_client_kwargs())

templates/components/vectordbs/python/llamacloud/query_filter.py

@@ -5,11 +5,11 @@ def generate_filters(doc_ids):
     """
     Generate public/private document filters based on the doc_ids and the vector store.
     """
-    # Using "nin" filter to include the documents don't have the "private" key because they're uploaded in LlamaCloud UI
+    # Using "is_empty" filter to include the documents don't have the "private" key because they're uploaded in LlamaCloud UI
     public_doc_filter = MetadataFilter(
         key="private",
-        value=["true"],
-        operator="nin",  # type: ignore
+        value=None,
+        operator="is_empty",  # type: ignore
     )
     selected_doc_filter = MetadataFilter(
         key="file_id",  # Note: LLamaCloud uses "file_id" to reference private document ids as "doc_id" is a restricted field in LlamaCloud

templates/components/vectordbs/python/none/index.py

@@ -1,30 +1,43 @@
-import os
 import logging
+import os
 from datetime import timedelta
+from typing import Optional
 
-from cachetools import cached, TTLCache
-from llama_index.core.storage import StorageContext
+from cachetools import TTLCache, cached
+from llama_index.core.callbacks import CallbackManager
 from llama_index.core.indices import load_index_from_storage
+from llama_index.core.storage import StorageContext
+from pydantic import BaseModel, Field
 
 logger = logging.getLogger("uvicorn")
 
 
+class IndexConfig(BaseModel):
+    callback_manager: Optional[CallbackManager] = Field(
+        default=None,
+    )
+
+
+def get_index(config: IndexConfig = None):
+    if config is None:
+        config = IndexConfig()
+    storage_dir = os.getenv("STORAGE_DIR", "storage")
+    # check if storage already exists
+    if not os.path.exists(storage_dir):
+        return None
+    # load the existing index
+    logger.info(f"Loading index from {storage_dir}...")
+    storage_context = get_storage_context(storage_dir)
+    index = load_index_from_storage(
+        storage_context, callback_manager=config.callback_manager
+    )
+    logger.info(f"Finished loading index from {storage_dir}")
+    return index
+
+
 @cached(
     TTLCache(maxsize=10, ttl=timedelta(minutes=5).total_seconds()),
     key=lambda *args, **kwargs: "global_storage_context",
 )
 def get_storage_context(persist_dir: str) -> StorageContext:
     return StorageContext.from_defaults(persist_dir=persist_dir)
-
-
-def get_index(params=None):
-    storage_dir = os.getenv("STORAGE_DIR", "storage")
-    # check if storage already exists
-    if not os.path.exists(storage_dir):
-        return None
-    # load the existing index
-    logger.info(f"Loading index from {storage_dir}...")
-    storage_context = get_storage_context(storage_dir)
-    index = load_index_from_storage(storage_context)
-    logger.info(f"Finished loading index from {storage_dir}")
-    return index

templates/components/vectordbs/typescript/llamacloud/queryFilter.ts

@@ -4,8 +4,8 @@ export function generateFilters(documentIds: string[]): MetadataFilters {
   // public documents don't have the "private" field or it's set to "false"
   const publicDocumentsFilter: MetadataFilter = {
     key: "private",
-    value: ["true"],
-    operator: "nin",
+    value: null,
+    operator: "is_empty",
   };
 
   // if no documentIds are provided, only retrieve information from public documents

templates/types/extractor/fastapi/app/engine/index.py

@@ -1,17 +1,31 @@
 import logging
-from llama_index.core.indices import VectorStoreIndex
-from app.engine.vectordb import get_vector_store
+from typing import Optional
 
+from llama_index.core.callbacks import CallbackManager
+from llama_index.core.indices import VectorStoreIndex
+from pydantic import BaseModel, Field
+
+from app.engine.vectordb import get_vector_store
 
 logger = logging.getLogger("uvicorn")
 
 
-def get_index(params=None):
+class IndexConfig(BaseModel):
+    callback_manager: Optional[CallbackManager] = Field(
+        default=None,
+    )
+
+
+def get_index(config: IndexConfig = None):
+    if config is None:
+        config = IndexConfig()
     logger.info("Connecting vector store...")
     store = get_vector_store()
     # Load the index from the vector store
     # If you are using a vector store that doesn't store text,
     # you must load the index from both the vector store and the document store
-    index = VectorStoreIndex.from_vector_store(store)
+    index = VectorStoreIndex.from_vector_store(
+        store, callback_manager=config.callback_manager
+    )
     logger.info("Finished load index from vector store.")
     return index

templates/types/extractor/fastapi/app/services/extractor.py

@@ -1,4 +1,5 @@
 import logging
+
 from app.engine import get_query_engine
 from app.services.model import IMPORTS
 
@@ -33,4 +34,4 @@ class ExtractorService:
         query_engine = get_query_engine(schema_model)
         response = await query_engine.aquery(query)
         output_data = response.response.dict()
-        return schema_model(**output_data).json(indent=2)
+        return schema_model(**output_data).model_dump_json(indent=2)

templates/types/extractor/fastapi/pyproject.toml

@@ -13,7 +13,7 @@ python = "^3.11,<4.0"
 fastapi = "^0.109.1"
 uvicorn = { extras = ["standard"], version = "^0.23.2" }
 python-dotenv = "^1.0.0"
-llama-index = "^0.10.58"
+llama-index = "^0.11.1"
 cachetools = "^5.3.3"
 reflex = "^0.5.9"
 

templates/types/streaming/express/package.json

@@ -20,7 +20,7 @@
     "dotenv": "^16.3.1",
     "duck-duck-scrape": "^2.2.5",
     "express": "^4.18.2",
-    "llamaindex": "0.5.19",
+    "llamaindex": "0.5.20",
     "pdf2json": "3.0.5",
     "ajv": "^8.12.0",
     "@e2b/code-interpreter": "^0.0.5",

templates/types/streaming/fastapi/app/api/routers/chat.py

@@ -38,11 +38,10 @@ async def chat(
         logger.info(
             f"Creating chat engine with filters: {str(filters)}",
         )
-        chat_engine = get_chat_engine(filters=filters, params=params)
-
         event_handler = EventCallbackHandler()
-        chat_engine.callback_manager.handlers.append(event_handler)  # type: ignore
-
+        chat_engine = get_chat_engine(
+            filters=filters, params=params, event_handlers=[event_handler]
+        )
         response = await chat_engine.astream_chat(last_message_content, messages)
         process_response_nodes(response.source_nodes, background_tasks)
 

templates/types/streaming/fastapi/app/api/services/file.py

@@ -5,8 +5,7 @@ from io import BytesIO
 from pathlib import Path
 from typing import Any, List, Tuple
 
-
-from app.engine.index import get_index
+from app.engine.index import IndexConfig, get_index
 from llama_index.core import VectorStoreIndex
 from llama_index.core.ingestion import IngestionPipeline
 from llama_index.core.readers.file.base import (
@@ -77,7 +76,8 @@ class PrivateFileService:
         file_data, extension = PrivateFileService.preprocess_base64_file(base64_content)
 
         # Add the nodes to the index and persist it
-        current_index = get_index(params)
+        index_config = IndexConfig(**params)
+        current_index = get_index(index_config)
 
         # Insert the documents into the index
         if isinstance(current_index, LlamaCloudIndex):

templates/types/streaming/fastapi/app/engine/index.py

@@ -1,17 +1,31 @@
 import logging
-from llama_index.core.indices import VectorStoreIndex
-from app.engine.vectordb import get_vector_store
+from typing import Optional
 
+from llama_index.core.callbacks import CallbackManager
+from llama_index.core.indices import VectorStoreIndex
+from pydantic import BaseModel, Field
+
+from app.engine.vectordb import get_vector_store
 
 logger = logging.getLogger("uvicorn")
 
 
-def get_index(params=None):
+class IndexConfig(BaseModel):
+    callback_manager: Optional[CallbackManager] = Field(
+        default=None,
+    )
+
+
+def get_index(config: IndexConfig = None):
+    if config is None:
+        config = IndexConfig()
     logger.info("Connecting vector store...")
     store = get_vector_store()
     # Load the index from the vector store
     # If you are using a vector store that doesn't store text,
     # you must load the index from both the vector store and the document store
-    index = VectorStoreIndex.from_vector_store(store)
+    index = VectorStoreIndex.from_vector_store(
+        store, callback_manager=config.callback_manager
+    )
     logger.info("Finished load index from vector store.")
     return index

templates/types/streaming/fastapi/pyproject.toml

@@ -14,7 +14,7 @@ fastapi = "^0.109.1"
 uvicorn = { extras = ["standard"], version = "^0.23.2" }
 python-dotenv = "^1.0.0"
 aiostream = "^0.5.2"
-llama-index = "0.10.58"
+llama-index = "0.11.1"
 cachetools = "^5.3.3"
 
 [build-system]

templates/types/streaming/nextjs/app/components/ui/chat/chat-message/chat-suggestedQuestions.tsx

@@ -1,15 +1,15 @@
-import { useState } from "react";
 import { ChatHandler, SuggestedQuestionsData } from "..";
 
 export function SuggestedQuestions({
   questions,
   append,
+  isLastMessage,
 }: {
   questions: SuggestedQuestionsData;
   append: Pick<ChatHandler, "append">["append"];
+  isLastMessage: boolean;
 }) {
-  const [showQuestions, setShowQuestions] = useState(questions.length > 0);
-
+  const showQuestions = isLastMessage && questions.length > 0;
   return (
     showQuestions &&
     append !== undefined && (
@@ -19,7 +19,6 @@ export function SuggestedQuestions({
             key={index}
             onClick={() => {
               append({ role: "user", content: question });
-              setShowQuestions(false);
             }}
             className="text-sm italic hover:underline cursor-pointer"
           >

templates/types/streaming/nextjs/app/components/ui/chat/chat-message/index.tsx

@@ -34,10 +34,12 @@ function ChatMessageContent({
   message,
   isLoading,
   append,
+  isLastMessage,
 }: {
   message: Message;
   isLoading: boolean;
   append: Pick<ChatHandler, "append">["append"];
+  isLastMessage: boolean;
 }) {
   const annotations = message.annotations as MessageAnnotation[] | undefined;
   if (!annotations?.length) return <Markdown content={message.content} />;
@@ -102,6 +104,7 @@ function ChatMessageContent({
         <SuggestedQuestions
           questions={suggestedQuestionsData[0]}
           append={append}
+          isLastMessage={isLastMessage}
         />
       ) : null,
     },
@@ -122,10 +125,12 @@ export default function ChatMessage({
   chatMessage,
   isLoading,
   append,
+  isLastMessage,
 }: {
   chatMessage: Message;
   isLoading: boolean;
   append: Pick<ChatHandler, "append">["append"];
+  isLastMessage: boolean;
 }) {
   const { isCopied, copyToClipboard } = useCopyToClipboard({ timeout: 2000 });
   return (
@@ -136,6 +141,7 @@ export default function ChatMessage({
           message={chatMessage}
           isLoading={isLoading}
           append={append}
+          isLastMessage={isLastMessage}
         />
         <Button
           onClick={() => copyToClipboard(chatMessage.content)}

templates/types/streaming/nextjs/app/components/ui/chat/chat-messages.tsx

@@ -69,6 +69,7 @@ export default function ChatMessages(
               chatMessage={m}
               isLoading={isLoadingMessage}
               append={props.append!}
+              isLastMessage={i === messageLength - 1}
             />
           );
         })}

templates/types/streaming/nextjs/app/components/ui/chat/index.ts

@@ -1,5 +1,4 @@
 import { JSONValue } from "ai";
-import { isValidUrl } from "../lib/utils";
 import ChatInput from "./chat-input";
 import ChatMessages from "./chat-messages";
 
@@ -113,7 +112,7 @@ function preprocessSourceNodes(nodes: SourceNode[]): SourceNode[] {
   // Filter source nodes has lower score
   nodes = nodes
     .filter((node) => (node.score ?? 1) > NODE_SCORE_THRESHOLD)
-    .filter((node) => isValidUrl(node.url))
+    .filter((node) => node.url && node.url.trim() !== "")
     .sort((a, b) => (b.score ?? 1) - (a.score ?? 1))
     .map((node) => {
       // remove trailing slash for node url if exists

templates/types/streaming/nextjs/app/components/ui/lib/utils.ts

@@ -4,13 +4,3 @@ import { twMerge } from "tailwind-merge";
 export function cn(...inputs: ClassValue[]) {
   return twMerge(clsx(inputs));
 }
-
-export function isValidUrl(url?: string): boolean {
-  if (!url) return false;
-  try {
-    new URL(url);
-    return true;
-  } catch (_) {
-    return false;
-  }
-}

templates/types/streaming/nextjs/next.config.json

@@ -1,8 +1,15 @@
 {
   "experimental": {
     "outputFileTracingIncludes": {
-      "/*": ["./cache/**/*"],
-      "/api/**/*": ["./node_modules/**/*.wasm"]
+      "/*": ["./cache/**/*"]
+    },
+    "outputFileTracingExcludes": {
+      "/api/files/*": [
+        ".next/**/*",
+        "node_modules/**/*",
+        "public/**/*",
+        "app/**/*"
+      ]
     }
   }
 }

templates/types/streaming/nextjs/package.json

@@ -25,7 +25,7 @@
     "duck-duck-scrape": "^2.2.5",
     "formdata-node": "^6.0.3",
     "got": "^14.4.1",
-    "llamaindex": "0.5.19",
+    "llamaindex": "0.5.20",
     "lucide-react": "^0.294.0",
     "next": "^14.2.4",
     "react": "^18.2.0",