review fixes

fix: use cli framework in CLI
Merge branch 'main' into ms/simplify
2026-07-13 21:01:55 -04:00 · 2024-10-14 12:34:06 +07:00 · 2024-10-14 12:27:53 +07:00 · 2024-10-14 12:08:45 +07:00 · 2024-10-14 12:05:49 +07:00 · 2024-10-14 12:03:03 +07:00
82 changed files with 1664 additions and 1357 deletions
@@ -0,0 +1,5 @@
+---
+"create-llama": patch
+---
+
+docs: chroma env variables
@@ -19,7 +19,7 @@ jobs:
        python-version: ["3.11"]
        os: [macos-latest, windows-latest, ubuntu-22.04]
        frameworks: ["fastapi"]
-        datasources: ["--no-files", "--example-file"]
+        datasources: ["--no-files", "--example-file", "--llamacloud"]
    defaults:
      run:
        shell: bash
@@ -17,6 +17,9 @@ jobs:

      - uses: pnpm/action-setup@v3

+      - name: Install uv
+        uses: astral-sh/setup-uv@v3
+
      - name: Setup Node.js
        uses: actions/setup-node@v4
        with:
@@ -1,2 +1,3 @@
 pnpm format
 pnpm lint
+uvx ruff format --check templates/
@@ -1,5 +1,36 @@
 # create-llama

+## 0.2.19
+
+### Patch Changes
+
+- 3d41488: feat: use selected llamacloud for multiagent
+
+## 0.2.18
+
+### Patch Changes
+
+- 75e1f61: Fix cannot query public document from llamacloud
+- 88220f1: fix workflow doesn't stop when user presses stop generation button
+- 75e1f61: Fix typescript templates cannot upload file to llamacloud
+- 88220f1: Bump llama_index@0.11.17
+
+## 0.2.17
+
+### Patch Changes
+
+- cd3fcd0: bump: use LlamaIndexTS 0.6.18
+- 6335de1: Fix using LlamaCloud selector does not use the configured values in the environment (Python)
+
+## 0.2.16
+
+### Patch Changes
+
+- 0e78ba4: Fix: programmatically ensure index for LlamaCloud
+- 0e78ba4: Fix .env not loaded on poetry run generate
+- 7f4ac22: Don't need to run generate script for LlamaCloud
+- 5263bde: Use selected LlamaCloud index in multi-agent template
+
 ## 0.2.15

 ### Patch Changes
@@ -15,6 +15,8 @@ const dataSource: string = process.env.DATASOURCE
  ? process.env.DATASOURCE
  : "--example-file";

+// TODO: add support for other templates
+
 if (
  dataSource === "--example-file" // XXX: this test provides its own data source - only trigger it on one data source (usually the CI matrix will trigger multiple data sources)
 ) {
@@ -45,14 +47,86 @@ if (

  const observabilityOptions = ["llamatrace", "traceloop"];

-  // Run separate tests for each observability option to reduce CI runtime
-  test.describe("Test resolve python dependencies with observability", () => {
-    // Testing with streaming template, vectorDb: none, tools: none, and dataSource: --example-file
-    for (const observability of observabilityOptions) {
-      test(`observability: ${observability}`, async () => {
-        const cwd = await createTestDir();
+  test.describe("Mypy check", () => {
+    test.describe.configure({ retries: 0 });

-        await createAndCheckLlamaProject({
+    // Test vector databases
+    for (const vectorDb of vectorDbs) {
+      test(`Mypy check for vectorDB: ${vectorDb}`, async () => {
+        const cwd = await createTestDir();
+        const { pyprojectPath } = await createAndCheckLlamaProject({
+          options: {
+            cwd,
+            templateType: "streaming",
+            templateFramework,
+            dataSource: "--example-file",
+            vectorDb,
+            tools: "none",
+            port: 3000,
+            externalPort: 8000,
+            postInstallAction: "none",
+            templateUI: undefined,
+            appType: "--no-frontend",
+            llamaCloudProjectName: undefined,
+            llamaCloudIndexName: undefined,
+            observability: undefined,
+          },
+        });
+
+        const pyprojectContent = fs.readFileSync(pyprojectPath, "utf-8");
+        if (vectorDb !== "none") {
+          if (vectorDb === "pg") {
+            expect(pyprojectContent).toContain(
+              "llama-index-vector-stores-postgres",
+            );
+          } else {
+            expect(pyprojectContent).toContain(
+              `llama-index-vector-stores-${vectorDb}`,
+            );
+          }
+        }
+      });
+    }
+
+    // Test tools
+    for (const tool of toolOptions) {
+      test(`Mypy check for tool: ${tool}`, async () => {
+        const cwd = await createTestDir();
+        const { pyprojectPath } = await createAndCheckLlamaProject({
+          options: {
+            cwd,
+            templateType: "streaming",
+            templateFramework,
+            dataSource: "--example-file",
+            vectorDb: "none",
+            tools: tool,
+            port: 3000,
+            externalPort: 8000,
+            postInstallAction: "none",
+            templateUI: undefined,
+            appType: "--no-frontend",
+            llamaCloudProjectName: undefined,
+            llamaCloudIndexName: undefined,
+            observability: undefined,
+          },
+        });
+
+        const pyprojectContent = fs.readFileSync(pyprojectPath, "utf-8");
+        if (tool === "wikipedia.WikipediaToolSpec") {
+          expect(pyprojectContent).toContain("wikipedia");
+        }
+        if (tool === "google.GoogleSearchToolSpec") {
+          expect(pyprojectContent).toContain("google");
+        }
+      });
+    }
+
+    // Test data sources
+    for (const dataSource of dataSources) {
+      const dataSourceType = dataSource.split(" ")[0];
+      test(`Mypy check for data source: ${dataSourceType}`, async () => {
+        const cwd = await createTestDir();
+        const { pyprojectPath } = await createAndCheckLlamaProject({
          options: {
            cwd,
            templateType: "streaming",
@@ -60,89 +134,53 @@ if (
            dataSource,
            vectorDb: "none",
            tools: "none",
-            port: 3000, // port, not used
-            externalPort: 8000, // externalPort, not used
-            postInstallAction: "none", // postInstallAction
-            templateUI: undefined, // ui
-            appType: "--no-frontend", // appType
-            llamaCloudProjectName: undefined, // llamaCloudProjectName
-            llamaCloudIndexName: undefined, // llamaCloudIndexName
+            port: 3000,
+            externalPort: 8000,
+            postInstallAction: "none",
+            templateUI: undefined,
+            appType: "--no-frontend",
+            llamaCloudProjectName: undefined,
+            llamaCloudIndexName: undefined,
+            observability: undefined,
+          },
+        });
+
+        const pyprojectContent = fs.readFileSync(pyprojectPath, "utf-8");
+        if (dataSource.includes("--web-source")) {
+          expect(pyprojectContent).toContain("llama-index-readers-web");
+        }
+        if (dataSource.includes("--db-source")) {
+          expect(pyprojectContent).toContain("llama-index-readers-database");
+        }
+      });
+    }
+
+    // Test observability options
+    for (const observability of observabilityOptions) {
+      test(`Mypy check for observability: ${observability}`, async () => {
+        const cwd = await createTestDir();
+
+        const { pyprojectPath } = await createAndCheckLlamaProject({
+          options: {
+            cwd,
+            templateType: "streaming",
+            templateFramework,
+            dataSource: "--example-file",
+            vectorDb: "none",
+            tools: "none",
+            port: 3000,
+            externalPort: 8000,
+            postInstallAction: "none",
+            templateUI: undefined,
+            appType: "--no-frontend",
+            llamaCloudProjectName: undefined,
+            llamaCloudIndexName: undefined,
            observability,
          },
        });
      });
    }
  });
-
-  test.describe("Test resolve python dependencies", () => {
-    for (const vectorDb of vectorDbs) {
-      for (const tool of toolOptions) {
-        for (const dataSource of dataSources) {
-          const dataSourceType = dataSource.split(" ")[0];
-          const toolDescription = tool === "none" ? "no tools" : tool;
-          const optionDescription = `vectorDb: ${vectorDb}, ${toolDescription}, dataSource: ${dataSourceType}`;
-
-          test(`options: ${optionDescription}`, async () => {
-            const cwd = await createTestDir();
-
-            const { pyprojectPath, projectPath } =
-              await createAndCheckLlamaProject({
-                options: {
-                  cwd,
-                  templateType: "streaming",
-                  templateFramework,
-                  dataSource,
-                  vectorDb,
-                  tools: tool,
-                  port: 3000, // port, not used
-                  externalPort: 8000, // externalPort, not used
-                  postInstallAction: "none", // postInstallAction
-                  templateUI: undefined, // ui
-                  appType: "--no-frontend", // appType
-                  llamaCloudProjectName: undefined, // llamaCloudProjectName
-                  llamaCloudIndexName: undefined, // llamaCloudIndexName
-                  observability: undefined, // observability
-                },
-              });
-
-            // Additional checks for specific dependencies
-
-            // Verify that specific dependencies are in pyproject.toml
-            const pyprojectContent = fs.readFileSync(pyprojectPath, "utf-8");
-            if (vectorDb !== "none") {
-              if (vectorDb === "pg") {
-                expect(pyprojectContent).toContain(
-                  "llama-index-vector-stores-postgres",
-                );
-              } else {
-                expect(pyprojectContent).toContain(
-                  `llama-index-vector-stores-${vectorDb}`,
-                );
-              }
-            }
-            if (tool !== "none") {
-              if (tool === "wikipedia.WikipediaToolSpec") {
-                expect(pyprojectContent).toContain("wikipedia");
-              }
-              if (tool === "google.GoogleSearchToolSpec") {
-                expect(pyprojectContent).toContain("google");
-              }
-            }
-
-            // Check for data source specific dependencies
-            if (dataSource.includes("--web-source")) {
-              expect(pyprojectContent).toContain("llama-index-readers-web");
-            }
-            if (dataSource.includes("--db-source")) {
-              expect(pyprojectContent).toContain(
-                "llama-index-readers-database ",
-              );
-            }
-          });
-        }
-      }
-    }
-  });
 }

 async function createAndCheckLlamaProject({
@@ -161,21 +199,39 @@ async function createAndCheckLlamaProject({
  const pyprojectPath = path.join(projectPath, "pyproject.toml");
  expect(fs.existsSync(pyprojectPath)).toBeTruthy();

-  // Run poetry lock
+  const env = {
+    ...process.env,
+    POETRY_VIRTUALENVS_IN_PROJECT: "true",
+  };
+
+  // Run poetry install
  try {
-    const { stdout, stderr } = await execAsync(
-      "poetry config virtualenvs.in-project true && poetry lock --no-update",
-      { cwd: projectPath },
+    const { stdout: installStdout, stderr: installStderr } = await execAsync(
+      "poetry install",
+      { cwd: projectPath, env },
    );
-    console.log("poetry lock stdout:", stdout);
-    console.error("poetry lock stderr:", stderr);
+    console.log("poetry install stdout:", installStdout);
+    console.error("poetry install stderr:", installStderr);
  } catch (error) {
-    console.error("Error running poetry lock:", error);
+    console.error("Error running poetry install:", error);
    throw error;
  }

-  // Check if poetry.lock file was created
-  expect(fs.existsSync(path.join(projectPath, "poetry.lock"))).toBeTruthy();
+  // Run poetry run mypy
+  try {
+    const { stdout: mypyStdout, stderr: mypyStderr } = await execAsync(
+      "poetry run mypy .",
+      { cwd: projectPath, env },
+    );
+    console.log("poetry run mypy stdout:", mypyStdout);
+    console.error("poetry run mypy stderr:", mypyStderr);
+  } catch (error) {
+    console.error("Error running mypy:", error);
+    throw error;
+  }
+
+  // If we reach this point without throwing an error, the test passes
+  expect(true).toBeTruthy();

  return { pyprojectPath, projectPath };
 }
@@ -27,6 +27,13 @@ const userMessage =
  dataSource !== "--no-files" ? "Physical standard for letters" : "Hello";

 test.describe(`Test streaming template ${templateFramework} ${dataSource} ${templateUI} ${appType} ${templatePostInstallAction}`, async () => {
+  const isNode18 = process.version.startsWith("v18");
+  const isLlamaCloud = dataSource === "--llamacloud";
+  // llamacloud is using File API which is not supported on node 18
+  if (isNode18 && isLlamaCloud) {
+    test.skip(true, "Skipping tests for Node 18 and LlamaCloud data source");
+  }
+
  let port: number;
  let externalPort: number;
  let cwd: string;
@@ -65,7 +65,7 @@ const getVectorDBEnvs = (
        {
          name: "PG_CONNECTION_STRING",
          description:
-            "For generating a connection URI, see https://docs.timescale.com/use-timescale/latest/services/create-a-service\nThe PostgreSQL connection string.",
+            "For generating a connection URI, see https://supabase.com/vector\nThe PostgreSQL connection string.",
        },
      ];

@@ -182,11 +182,11 @@ const getVectorDBEnvs = (
        },
        {
          name: "CHROMA_HOST",
-          description: "The API endpoint for your Chroma database",
+          description: "The hostname for your Chroma database. Eg: localhost",
        },
        {
          name: "CHROMA_PORT",
-          description: "The port for your Chroma database",
+          description: "The port for your Chroma database. Eg: 8000",
        },
      ];
      // TS Version doesn't support config local storage path
@@ -1,7 +1,7 @@
 import ciInfo from "ci-info";
 import prompts from "prompts";
 import { ModelConfigParams } from ".";
-import { questionHandlers, toChoice } from "../../questions";
+import { questionHandlers, toChoice } from "../../questions/utils";

 const MODELS = [
  "claude-3-opus",
@@ -1,7 +1,7 @@
 import ciInfo from "ci-info";
 import prompts from "prompts";
 import { ModelConfigParams, ModelConfigQuestionsParams } from ".";
-import { questionHandlers } from "../../questions";
+import { questionHandlers } from "../../questions/utils";

 const ALL_AZURE_OPENAI_CHAT_MODELS: Record<string, { openAIModel: string }> = {
  "gpt-35-turbo": { openAIModel: "gpt-3.5-turbo" },
@@ -1,7 +1,7 @@
 import ciInfo from "ci-info";
 import prompts from "prompts";
 import { ModelConfigParams } from ".";
-import { questionHandlers, toChoice } from "../../questions";
+import { questionHandlers, toChoice } from "../../questions/utils";

 const MODELS = ["gemini-1.5-pro-latest", "gemini-pro", "gemini-pro-vision"];
 type ModelData = {
@@ -1,7 +1,7 @@
 import ciInfo from "ci-info";
 import prompts from "prompts";
 import { ModelConfigParams } from ".";
-import { questionHandlers, toChoice } from "../../questions";
+import { questionHandlers, toChoice } from "../../questions/utils";

 import got from "got";
 import ora from "ora";
@@ -1,6 +1,5 @@
-import ciInfo from "ci-info";
 import prompts from "prompts";
-import { questionHandlers } from "../../questions";
+import { questionHandlers } from "../../questions/utils";
 import { ModelConfig, ModelProvider, TemplateFramework } from "../types";
 import { askAnthropicQuestions } from "./anthropic";
 import { askAzureQuestions } from "./azure";
@@ -27,7 +26,7 @@ export async function askModelConfig({
  framework,
 }: ModelConfigQuestionsParams): Promise<ModelConfig> {
  let modelProvider: ModelProvider = DEFAULT_MODEL_PROVIDER;
-  if (askModels && !ciInfo.isCI) {
+  if (askModels) {
    let choices = [
      { title: "OpenAI", value: "openai" },
      { title: "Groq", value: "groq" },
@@ -4,7 +4,7 @@ import ora from "ora";
 import { red } from "picocolors";
 import prompts from "prompts";
 import { ModelConfigParams } from ".";
-import { questionHandlers } from "../../questions";
+import { questionHandlers } from "../../questions/utils";

 export const TSYSTEMS_LLMHUB_API_URL =
  "https://llm-server.llmhub.t-systems.net/v2";
@@ -1,7 +1,7 @@
 import ciInfo from "ci-info";
 import prompts from "prompts";
 import { ModelConfigParams } from ".";
-import { questionHandlers, toChoice } from "../../questions";
+import { questionHandlers, toChoice } from "../../questions/utils";

 const MODELS = ["mistral-tiny", "mistral-small", "mistral-medium"];
 type ModelData = {
@@ -3,7 +3,7 @@ import ollama, { type ModelResponse } from "ollama";
 import { red } from "picocolors";
 import prompts from "prompts";
 import { ModelConfigParams } from ".";
-import { questionHandlers, toChoice } from "../../questions";
+import { questionHandlers, toChoice } from "../../questions/utils";

 type ModelData = {
  dimensions: number;
@@ -4,7 +4,7 @@ import ora from "ora";
 import { red } from "picocolors";
 import prompts from "prompts";
 import { ModelConfigParams, ModelConfigQuestionsParams } from ".";
-import { questionHandlers } from "../../questions";
+import { questionHandlers } from "../../questions/utils";

 const OPENAI_API_URL = "https://api.openai.com/v1";

@@ -93,6 +93,12 @@ const getAdditionalDependencies = (
      });
      break;
    }
+    case "llamacloud":
+      dependencies.push({
+        name: "llama-index-indices-managed-llama-cloud",
+        version: "^0.3.1",
+      });
+      break;
  }

  // Add data source dependencies
@@ -123,16 +129,10 @@ const getAdditionalDependencies = (
            extras: ["rsa"],
          });
          dependencies.push({
-            name: "psycopg2",
+            name: "psycopg2-binary",
            version: "^2.9.9",
          });
          break;
-        case "llamacloud":
-          dependencies.push({
-            name: "llama-index-indices-managed-llama-cloud",
-            version: "^0.3.1",
-          });
-          break;
      }
    }
  }
@@ -46,7 +46,7 @@ export type TemplateDataSource = {
  type: TemplateDataSourceType;
  config: TemplateDataSourceConfig;
 };
-export type TemplateDataSourceType = "file" | "web" | "db" | "llamacloud";
+export type TemplateDataSourceType = "file" | "web" | "db";
 export type TemplateObservability = "none" | "traceloop" | "llamatrace";
 // Config for both file and folder
 export type FileSourceConfig = {
@@ -1,7 +1,6 @@
 /* eslint-disable import/no-extraneous-dependencies */
 import { execSync } from "child_process";
-import Commander from "commander";
-import Conf from "conf";
+import { Command } from "commander";
 import fs from "fs";
 import path from "path";
 import { bold, cyan, green, red, yellow } from "picocolors";
@@ -17,8 +16,9 @@ import { runApp } from "./helpers/run-app";
 import { getTools } from "./helpers/tools";
 import { validateNpmName } from "./helpers/validate-pkg";
 import packageJson from "./package.json";
-import { QuestionArgs, askQuestions, onPromptState } from "./questions";
-
+import { askQuestions } from "./questions/index";
+import { QuestionArgs } from "./questions/types";
+import { onPromptState } from "./questions/utils";
 // Run the initialization function
 initializeGlobalAgent();

@@ -29,12 +29,14 @@ const handleSigTerm = () => process.exit(0);
 process.on("SIGINT", handleSigTerm);
 process.on("SIGTERM", handleSigTerm);

-const program = new Commander.Command(packageJson.name)
+const program = new Command(packageJson.name)
  .version(packageJson.version)
-  .arguments("<project-directory>")
-  .usage(`${green("<project-directory>")} [options]`)
+  .arguments("[project-directory]")
+  .usage(`${green("[project-directory]")} [options]`)
  .action((name) => {
-    projectPath = name;
+    if (name) {
+      projectPath = name;
+    }
  })
  .option(
    "--use-npm",
@@ -55,13 +57,6 @@ const program = new Commander.Command(packageJson.name)
    `

  Explicitly tell the CLI to bootstrap the application using Yarn
-`,
-  )
-  .option(
-    "--reset-preferences",
-    `
-
-  Explicitly tell the CLI to reset any stored preferences
 `,
  )
  .option(
@@ -124,7 +119,14 @@ const program = new Commander.Command(packageJson.name)
    "--frontend",
    `

-  Whether to generate a frontend for your backend.
+  Generate a frontend for your backend.
+`,
+  )
+  .option(
+    "--no-frontend",
+    `
+
+  Do not generate a frontend for your backend.
 `,
  )
  .option(
@@ -161,6 +163,13 @@ const program = new Commander.Command(packageJson.name)

  Specify the tools you want to use by providing a comma-separated list. For example, 'wikipedia.WikipediaToolSpec,google.GoogleSearchToolSpec'. Use 'none' to not using any tools.
 `,
+    (tools, _) => {
+      if (tools === "none") {
+        return [];
+      } else {
+        return getTools(tools.split(","));
+      }
+    },
  )
  .option(
    "--use-llama-parse",
@@ -189,86 +198,66 @@ const program = new Commander.Command(packageJson.name)

  Allow interactive selection of LLM and embedding models of different model providers.
 `,
+    false,
  )
  .option(
-    "--ask-examples",
+    "--pro",
    `

-  Allow interactive selection of community templates and LlamaPacks.
+  Allow interactive selection of all features.
 `,
+    false,
  )
  .allowUnknownOption()
  .parse(process.argv);
-if (process.argv.includes("--no-frontend")) {
-  program.frontend = false;
-}
-if (process.argv.includes("--tools")) {
-  if (program.tools === "none") {
-    program.tools = [];
-  } else {
-    program.tools = getTools(program.tools.split(","));
-  }
-}
+
+const options = program.opts();
+
 if (
  process.argv.includes("--no-llama-parse") ||
-  program.template === "extractor"
+  options.template === "extractor"
 ) {
-  program.useLlamaParse = false;
+  options.useLlamaParse = false;
 }
-program.askModels = process.argv.includes("--ask-models");
-program.askExamples = process.argv.includes("--ask-examples");
 if (process.argv.includes("--no-files")) {
-  program.dataSources = [];
+  options.dataSources = [];
 } else if (process.argv.includes("--example-file")) {
-  program.dataSources = getDataSources(program.files, program.exampleFile);
+  options.dataSources = getDataSources(options.files, options.exampleFile);
 } else if (process.argv.includes("--llamacloud")) {
-  program.dataSources = [
-    {
-      type: "llamacloud",
-      config: {},
-    },
-    EXAMPLE_FILE,
-  ];
+  options.dataSources = [EXAMPLE_FILE];
+  options.vectorDb = "llamacloud";
 } else if (process.argv.includes("--web-source")) {
-  program.dataSources = [
+  options.dataSources = [
    {
      type: "web",
      config: {
-        baseUrl: program.webSource,
-        prefix: program.webSource,
+        baseUrl: options.webSource,
+        prefix: options.webSource,
        depth: 1,
      },
    },
  ];
 } else if (process.argv.includes("--db-source")) {
-  program.dataSources = [
+  options.dataSources = [
    {
      type: "db",
      config: {
-        uri: program.dbSource,
-        queries: program.dbQuery || "SELECT * FROM mytable",
+        uri: options.dbSource,
+        queries: options.dbQuery || "SELECT * FROM mytable",
      },
    },
  ];
 }

-const packageManager = !!program.useNpm
+const packageManager = !!options.useNpm
  ? "npm"
-  : !!program.usePnpm
+  : !!options.usePnpm
    ? "pnpm"
-    : !!program.useYarn
+    : !!options.useYarn
      ? "yarn"
      : getPkgManager();

 async function run(): Promise<void> {
-  const conf = new Conf({ projectName: "create-llama" });
-
-  if (program.resetPreferences) {
-    conf.clear();
-    console.log(`Preferences reset successfully`);
-    return;
-  }
-
  if (typeof projectPath === "string") {
    projectPath = projectPath.trim();
  }
@@ -331,35 +320,16 @@ async function run(): Promise<void> {
    process.exit(1);
  }

-  const preferences = (conf.get("preferences") || {}) as QuestionArgs;
-  await askQuestions(
-    program as unknown as QuestionArgs,
-    preferences,
-    program.openAiKey,
-  );
+  const answers = await askQuestions(options as unknown as QuestionArgs);

  await createApp({
-    template: program.template,
-    framework: program.framework,
-    ui: program.ui,
+    ...answers,
    appPath: resolvedProjectPath,
    packageManager,
-    frontend: program.frontend,
-    modelConfig: program.modelConfig,
-    llamaCloudKey: program.llamaCloudKey,
-    communityProjectConfig: program.communityProjectConfig,
-    llamapack: program.llamapack,
-    vectorDb: program.vectorDb,
-    externalPort: program.externalPort,
-    postInstallAction: program.postInstallAction,
-    dataSources: program.dataSources,
-    tools: program.tools,
-    useLlamaParse: program.useLlamaParse,
-    observability: program.observability,
+    externalPort: options.externalPort,
  });
-  conf.set("preferences", preferences);

-  if (program.postInstallAction === "VSCode") {
+  if (answers.postInstallAction === "VSCode") {
    console.log(`Starting VSCode in ${root}...`);
    try {
      execSync(`code . --new-window --goto README.md`, {
@@ -383,15 +353,15 @@ Please check ${cyan(
        )} for more information.`,
      );
    }
-  } else if (program.postInstallAction === "runApp") {
+  } else if (answers.postInstallAction === "runApp") {
    console.log(`Running app in ${root}...`);
    await runApp(
      root,
-      program.template,
-      program.frontend,
-      program.framework,
-      program.port,
-      program.externalPort,
+      answers.template,
+      answers.frontend,
+      answers.framework,
+      options.port,
+      options.externalPort,
    );
  }
 }
@@ -1,6 +1,6 @@
 {
  "name": "create-llama",
-  "version": "0.2.15",
+  "version": "0.2.19",
  "description": "Create LlamaIndex-powered apps with one command",
  "keywords": [
    "rag",
@@ -49,8 +49,7 @@
    "async-retry": "1.3.1",
    "async-sema": "3.0.1",
    "ci-info": "github:watson/ci-info#f43f6a1cefff47fb361c88cf4b943fdbcaafe540",
-    "commander": "2.20.0",
-    "conf": "10.2.0",
+    "commander": "12.1.0",
    "cross-spawn": "7.0.3",
    "fast-glob": "3.3.1",
    "fs-extra": "11.2.0",
@@ -59,7 +58,7 @@
    "ollama": "^0.5.0",
    "ora": "^8.0.1",
    "picocolors": "1.0.0",
-    "prompts": "2.1.0",
+    "prompts": "2.4.2",
    "smol-toml": "^1.1.4",
    "tar": "6.1.15",
    "terminal-link": "^3.0.0",
@@ -42,11 +42,8 @@ importers:
        specifier: github:watson/ci-info#f43f6a1cefff47fb361c88cf4b943fdbcaafe540
        version: https://codeload.github.com/watson/ci-info/tar.gz/f43f6a1cefff47fb361c88cf4b943fdbcaafe540
      commander:
-        specifier: 2.20.0
-        version: 2.20.0
-      conf:
-        specifier: 10.2.0
-        version: 10.2.0
+        specifier: 12.1.0
+        version: 12.1.0
      cross-spawn:
        specifier: 7.0.3
        version: 7.0.3
@@ -72,8 +69,8 @@ importers:
        specifier: 1.0.0
        version: 1.0.0
      prompts:
-        specifier: 2.1.0
-        version: 2.1.0
+        specifier: 2.4.2
+        version: 2.4.2
      smol-toml:
        specifier: ^1.1.4
        version: 1.1.4
@@ -336,20 +333,9 @@ packages:
    engines: {node: '>=0.4.0'}
    hasBin: true

-  ajv-formats@2.1.1:
-    resolution: {integrity: sha512-Wx0Kx52hxE7C18hkMEggYlEifqWZtYaRgouJor+WMdPnQyEK13vgEWyVNup7SoeeoLMsr4kf5h6dOW11I15MUA==}
-    peerDependencies:
-      ajv: ^8.0.0
-    peerDependenciesMeta:
-      ajv:
-        optional: true
-
  ajv@6.12.6:
    resolution: {integrity: sha512-j3fVLgvTo527anyYyJOGTYJbG+vnnQYvE0m5mmkc1TK+nxAppkCLMIL0aZ4dblVCNoGShhm+kzE4ZUykBoMg4g==}

-  ajv@8.13.0:
-    resolution: {integrity: sha512-PRA911Blj99jR5RMeTunVbNXMF6Lp4vZXnk5GQjcnUWUTsrXtekg/pnmFFI2u/I36Y/2bITGS30GZCXei6uNkA==}
-
  ansi-colors@4.1.3:
    resolution: {integrity: sha512-/6w/C21Pm1A7aZitlI5Ni/2J6FFQN8i1Cvz3kHABAAbw93v/NlvKdVOqz7CCWz/3iv/JplRSEEZ83XION15ovw==}
    engines: {node: '>=6'}
@@ -410,10 +396,6 @@ packages:
  async-sema@3.0.1:
    resolution: {integrity: sha512-fKT2riE8EHAvJEfLJXZiATQWqZttjx1+tfgnVshCDrH8vlw4YC8aECe0B8MU184g+aVRFVgmfxFlKZKaozSrNw==}

-  atomically@1.7.0:
-    resolution: {integrity: sha512-Xcz9l0z7y9yQ9rdDaxlmaI4uJHf/T8g9hOEzJcsEqX2SjCj4J20uK7+ldkDHMbpJDK76wF7xEIgxc/vSlsfw5w==}
-    engines: {node: '>=10.12.0'}
-
  available-typed-arrays@1.0.7:
    resolution: {integrity: sha512-wvUjBtSGN7+7SjNpq/9M2Tg350UZD3q62IFZLbRAR1bSMlCo1ZaeW+BJ+D090e4hIIZLBcTDWe4Mh4jvUDajzQ==}
    engines: {node: '>= 0.4'}
@@ -530,8 +512,9 @@ packages:
  color-name@1.1.4:
    resolution: {integrity: sha512-dOy+3AuW3a2wNbZHIuMZpTcgjGuLU/uBL/ubcZF9OXbDo8ff4O8yVp5Bf0efS8uEoYo5q4Fx7dY9OgQGXgAsQA==}

-  commander@2.20.0:
-    resolution: {integrity: sha512-7j2y+40w61zy6YC2iRNpUe/NwhNyoXrYpHMrSunaMG64nRnaf96zO/KMQR4OyN/UnE5KLyEBnKHd4aG3rskjpQ==}
+  commander@12.1.0:
+    resolution: {integrity: sha512-Vw8qHK3bZM9y/P10u3Vib8o/DdkvA2OtPtZvD871QKjy74Wj1WSKFILMPRPSdUSx5RFK1arlJzEtA4PkFgnbuA==}
+    engines: {node: '>=18'}

  commander@9.5.0:
    resolution: {integrity: sha512-KRs7WVDKg86PWiuAqhDrAQnTXZKraVcCc6vFdL14qrZ/DcWwuRo7VoiYXalXO7S5GKpqYiVEwCbgFDfxNHKJBQ==}
@@ -540,10 +523,6 @@ packages:
  concat-map@0.0.1:
    resolution: {integrity: sha512-/Srv4dswyQNBfohGpz9o6Yb3Gz3SrUDqBH5rTuhGR7ahtlbYKnVxw2bCFMRljaA7EXHaXZ8wsHdodFvbkhKmqg==}

-  conf@10.2.0:
-    resolution: {integrity: sha512-8fLl9F04EJqjSqH+QjITQfJF8BrOVaYr1jewVgSRAEWePfxT0sku4w2hrGQ60BC/TNLGQ2pgxNlTbWQmMPFvXg==}
-    engines: {node: '>=12'}
-
  cross-spawn@5.1.0:
    resolution: {integrity: sha512-pTgQJ5KC0d2hcY8eyL1IzlBPYjTkyH72XRZPnLyKus2mBfNjQs3klqbJU2VILqZryAZUt9JOb3h/mWMy23/f5A==}

@@ -576,10 +555,6 @@ packages:
    resolution: {integrity: sha512-t/Ygsytq+R995EJ5PZlD4Cu56sWa8InXySaViRzw9apusqsOO2bQP+SbYzAhR0pFKoB+43lYy8rWban9JSuXnA==}
    engines: {node: '>= 0.4'}

-  debounce-fn@4.0.0:
-    resolution: {integrity: sha512-8pYCQiL9Xdcg0UPSD3d+0KMlOjp+KGU5EPwYddgzQ7DATsg4fuUDjQtsYLmWjnk2obnNHgV3vE2Y4jejSOJVBQ==}
-    engines: {node: '>=10'}
-
  debug@4.3.4:
    resolution: {integrity: sha512-PRWFHuSU3eDtQJPvnNY7Jcket1j0t5OuOsFzPPzsekD52Zl8qUfFIPEiswXqIvHWGVHOgX+7G/vCNNhehwxfkQ==}
    engines: {node: '>=6.0'}
@@ -638,10 +613,6 @@ packages:
    resolution: {integrity: sha512-yS+Q5i3hBf7GBkd4KG8a7eBNNWNGLTaEwwYWUijIYM7zrlYDM0BFXHjjPWlWZ1Rg7UaddZeIDmi9jF3HmqiQ2w==}
    engines: {node: '>=6.0.0'}

-  dot-prop@6.0.1:
-    resolution: {integrity: sha512-tE7ztYzXHIeyvc7N+hR3oi7FIbf/NIjVP9hmAt3yMXzrQ072/fpjGLx2GxNxGxUl5V73MEqYzioOMoVhGMJ5cA==}
-    engines: {node: '>=10'}
-
  duplexer3@0.1.5:
    resolution: {integrity: sha512-1A8za6ws41LQgv9HrE/66jyC5yuSjQ3L/KOpFtoBilsAK2iA2wuS5rTt1OCzIvtS2V7nVmedsUU+DGRcjBmOYA==}

@@ -664,10 +635,6 @@ packages:
    resolution: {integrity: sha512-rRqJg/6gd538VHvR3PSrdRBb/1Vy2YfzHqzvbhGIQpDRKIa4FgV/54b5Q1xYSxOOwKvjXweS26E0Q+nAMwp2pQ==}
    engines: {node: '>=8.6'}

-  env-paths@2.2.1:
-    resolution: {integrity: sha512-+h1lkLKhZMTYjog1VEpJNG7NZJWcuc2DDk/qsqSTRRCOXiLjeQ1d1/udrUGhqMxUgAlwKNZ0cf2uqan5GLuS2A==}
-    engines: {node: '>=6'}
-
  error-ex@1.3.2:
    resolution: {integrity: sha512-7dFHNmqeFSEt2ZBsCriorKnn3Z2pj+fd9kmI6QoWw4//DL+icEBfc0U7qJCisqrTsKTjw4fNFy2pW9OqStD84g==}

@@ -788,10 +755,6 @@ packages:
    resolution: {integrity: sha512-qOo9F+dMUmC2Lcb4BbVvnKJxTPjCm+RRpe4gDuGrzkL7mEVl/djYSu2OdQ2Pa302N4oqkSg9ir6jaLWJ2USVpQ==}
    engines: {node: '>=8'}

-  find-up@3.0.0:
-    resolution: {integrity: sha512-1yD6RmLI1XBfxugvORwlck6f75tYL+iR0jqwsOrOxMZyGYqUuDhJ0l4AXdO1iX/FTs9cBAMEk1gWSEx1kSbylg==}
-    engines: {node: '>=6'}
-
  find-up@4.1.0:
    resolution: {integrity: sha512-PpOwAdQ/YlXQ2vj8a3h8IipDuYRi3wceVQQGYWxNINccq40Anw7BlsEXCMbt1Zt+OLA6Fq9suIpIWD0OsnISlw==}
    engines: {node: '>=8'}
@@ -1057,10 +1020,6 @@ packages:
    resolution: {integrity: sha512-41Cifkg6e8TylSpdtTpeLVMqvSBEVzTttHvERD741+pnZ8ANv0004MRL43QKPDlK9cGvNp6NZWZUBlbGXYxxng==}
    engines: {node: '>=0.12.0'}

-  is-obj@2.0.0:
-    resolution: {integrity: sha512-drqDG3cbczxxEJRoOXcOjtdp1J/lyp1mNn0xaznRs8+muBhgQcrnbspox5X5fOw0HnMnbfDzvnEMEtqDEJEo8w==}
-    engines: {node: '>=8'}
-
  is-path-inside@3.0.3:
    resolution: {integrity: sha512-Fd4gABb+ycGAmKou8eMftCupSir5lRxqf4aD/vd0cD2qc4HL07OjCeuHMr8Ro4CoMaeCKDB0/ECBOVWjTwUvPQ==}
    engines: {node: '>=8'}
@@ -1138,12 +1097,6 @@ packages:
  json-schema-traverse@0.4.1:
    resolution: {integrity: sha512-xbbCH5dCYU5T8LcEhhuh7HJ88HXuW3qsI3Y0zOZFKfZEHcpWiHU/Jxzk629Brsab/mMiHQti9wMP+845RPe3Vg==}

-  json-schema-traverse@1.0.0:
-    resolution: {integrity: sha512-NM8/P9n3XjXhIZn1lLhkFaACTOURQXjWhV4BA/RnOv8xvgqtqpAX9IO4mRQxSx1Rlo4tqzeqb0sOlruaOy3dug==}
-
-  json-schema-typed@7.0.3:
-    resolution: {integrity: sha512-7DE8mpG+/fVw+dTpjbxnx47TaMnDfOI1jwft9g1VybltZCduyRQPJPvc+zzKY9WPHxhPWczyFuYa6I8Mw4iU5A==}
-
  json-stable-stringify-without-jsonify@1.0.1:
    resolution: {integrity: sha512-Bdboy+l7tA3OGW6FjyFHWkP5LuByj1Tk33Ljyq0axyzdk9//JSi2u3fP1QSmd1KNwq6VOKYGlAu87CisVir6Pw==}

@@ -1182,10 +1135,6 @@ packages:
    resolution: {integrity: sha512-OfCBkGEw4nN6JLtgRidPX6QxjBQGQf72q3si2uvqyFEMbycSFFHwAZeXx6cJgFM9wmLrf9zBwCP3Ivqa+LLZPw==}
    engines: {node: '>=6'}

-  locate-path@3.0.0:
-    resolution: {integrity: sha512-7AO748wWnIhNqAuaty2ZWHkQHRSNfPVIsPIfwEOWO22AmaoVrWavlOcMR5nzTLNYvp36X220/maaRsrec1G65A==}
-    engines: {node: '>=6'}
-
  locate-path@5.0.0:
    resolution: {integrity: sha512-t7hw9pI+WvuwNJXwk5zVHpyhIqzg2qTlklJOf0mVxGSbe3Fp2VieZcduNYjaLDoy6p9uGpQEGWG87WpMKlNq8g==}
    engines: {node: '>=8'}
@@ -1243,10 +1192,6 @@ packages:
    resolution: {integrity: sha512-OqbOk5oEQeAZ8WXWydlu9HJjz9WVdEIvamMCcXmuqUYjTknH/sqsWvhQ3vgwKFRR1HpjvNBKQ37nbJgYzGqGcg==}
    engines: {node: '>=6'}

-  mimic-fn@3.1.0:
-    resolution: {integrity: sha512-Ysbi9uYW9hFyfrThdDEQuykN4Ey6BuwPD2kpI5ES/nFTDn/98yxYNLZJcgUAKPT/mcrLLKaGzJR9YVxJrIdASQ==}
-    engines: {node: '>=8'}
-
  mimic-response@1.0.1:
    resolution: {integrity: sha512-j5EctnkH7amfV/q5Hgmoal1g2QHFJRraOtmx0JpIqkxhBhI/lJSl1nMpQ45hVarwNETOoWEimndZ4QK0RHxuxQ==}
    engines: {node: '>=4'}
@@ -1375,10 +1320,6 @@ packages:
    resolution: {integrity: sha512-TYOanM3wGwNGsZN2cVTYPArw454xnXj5qmWF1bEoAc4+cU/ol7GVh7odevjp1FNHduHc3KZMcFduxU5Xc6uJRQ==}
    engines: {node: '>=10'}

-  p-locate@3.0.0:
-    resolution: {integrity: sha512-x+12w/To+4GFfgJhBEpiDcLozRJGegY+Ei7/z0tSLkMmxGZNybVMSfWj9aJn8Z5Fc7dBUNJOOVgPv2H7IwulSQ==}
-    engines: {node: '>=6'}
-
  p-locate@4.1.0:
    resolution: {integrity: sha512-R79ZZ/0wAxKGu3oYMlz8jy/kbhsNrS7SKZ7PxEHBgJ5+F2mtFW2fK2cOtBh1cHYkQsbzFV7I+EoRKe6Yt0oK7A==}
    engines: {node: '>=8'}
@@ -1407,10 +1348,6 @@ packages:
    resolution: {integrity: sha512-ayCKvm/phCGxOkYRSCM82iDwct8/EonSEgCSxWxD7ve6jHggsFl4fZVQBPRNgQoKiuV/odhFrGzQXZwbifC8Rg==}
    engines: {node: '>=8'}

-  path-exists@3.0.0:
-    resolution: {integrity: sha512-bpC7GYwiDYQ4wYLe+FA8lhRjhQCMcQGuSgGGqDkg/QerRWw9CmGRT0iSOVRSZJ29NMLZgIzqaljJ63oaL4NIJQ==}
-    engines: {node: '>=4'}
-
  path-exists@4.0.0:
    resolution: {integrity: sha512-ak9Qy5Q7jYb2Wwcey5Fpvg2KoAc/ZIhLSLOSBmRmygPsGwkVVt0fZa0qrtMz+m6tJTAHfZQ8FnmB4MG4LWy7/w==}
    engines: {node: '>=8'}
@@ -1449,10 +1386,6 @@ packages:
    resolution: {integrity: sha512-HRDzbaKjC+AOWVXxAU/x54COGeIv9eb+6CkDSQoNTt4XyWoIJvuPsXizxu/Fr23EiekbtZwmh1IcIG/l/a10GQ==}
    engines: {node: '>=8'}

-  pkg-up@3.1.0:
-    resolution: {integrity: sha512-nDywThFk1i4BQK4twPQ6TA4RT8bDY96yeuCVBWL3ePARCiEKDRSrNGbFIgUJpLp+XeIR65v8ra7WuJOFUBtkMA==}
-    engines: {node: '>=8'}
-
  playwright-core@1.44.0:
    resolution: {integrity: sha512-ZTbkNpFfYcGWohvTTl+xewITm7EOuqIqex0c7dNZ+aXsbrLj0qI8XlGKfPpipjm0Wny/4Lt4CJsWJk1stVS5qQ==}
    engines: {node: '>=16'}
@@ -1498,8 +1431,8 @@ packages:
    engines: {node: '>=14'}
    hasBin: true

-  prompts@2.1.0:
-    resolution: {integrity: sha512-+x5TozgqYdOwWsQFZizE/Tra3fKvAoy037kOyU6cgz84n8f6zxngLOV4O32kTwt9FcLCxAqw0P/c8rOr9y+Gfg==}
+  prompts@2.4.2:
+    resolution: {integrity: sha512-NxNv/kLguCA7p3jE8oL2aEBsrJWgAakBpgmgK6lpPWV+WuOmY6r2/zbAVnP+T8bQlA0nzHXSJSJW0Hq7ylaD2Q==}
    engines: {node: '>= 6'}

  pseudomap@1.0.2:
@@ -1557,10 +1490,6 @@ packages:
    resolution: {integrity: sha512-fGxEI7+wsG9xrvdjsrlmL22OMTTiHRwAMroiEeMgq8gzoLC/PQr7RsRDSTLUg/bZAZtF+TVIkHc6/4RIKrui+Q==}
    engines: {node: '>=0.10.0'}

-  require-from-string@2.0.2:
-    resolution: {integrity: sha512-Xf0nWe6RseziFMu+Ap9biiUbmplq6S9/p+7w7YXP/JBHhrUDDUhwa+vANyubuqfZWTveU//DYVGsDG7RKL/vEw==}
-    engines: {node: '>=0.10.0'}
-
  require-main-filename@2.0.0:
    resolution: {integrity: sha512-NKN5kMDylKuldxYLSUfrbo5Tuzh4hd+2E8NPPX02mZtn1VuREQToYe/ZdlJy+J3uCpfaiGF05e7B8W0iXbQHmg==}

@@ -2306,10 +2235,6 @@ snapshots:

  acorn@8.11.3: {}

-  ajv-formats@2.1.1(ajv@8.13.0):
-    optionalDependencies:
-      ajv: 8.13.0
-
  ajv@6.12.6:
    dependencies:
      fast-deep-equal: 3.1.3
@@ -2317,13 +2242,6 @@ snapshots:
      json-schema-traverse: 0.4.1
      uri-js: 4.4.1

-  ajv@8.13.0:
-    dependencies:
-      fast-deep-equal: 3.1.3
-      json-schema-traverse: 1.0.0
-      require-from-string: 2.0.2
-      uri-js: 4.4.1
-
  ansi-colors@4.1.3: {}

  ansi-escapes@5.0.0:
@@ -2383,8 +2301,6 @@ snapshots:

  async-sema@3.0.1: {}

-  atomically@1.7.0: {}
-
  available-typed-arrays@1.0.7:
    dependencies:
      possible-typed-array-names: 1.0.0
@@ -2506,25 +2422,12 @@ snapshots:

  color-name@1.1.4: {}

-  commander@2.20.0: {}
+  commander@12.1.0: {}

  commander@9.5.0: {}

  concat-map@0.0.1: {}

-  conf@10.2.0:
-    dependencies:
-      ajv: 8.13.0
-      ajv-formats: 2.1.1(ajv@8.13.0)
-      atomically: 1.7.0
-      debounce-fn: 4.0.0
-      dot-prop: 6.0.1
-      env-paths: 2.2.1
-      json-schema-typed: 7.0.3
-      onetime: 5.1.2
-      pkg-up: 3.1.0
-      semver: 7.6.1
-
  cross-spawn@5.1.0:
    dependencies:
      lru-cache: 4.1.5
@@ -2568,10 +2471,6 @@ snapshots:
      es-errors: 1.3.0
      is-data-view: 1.0.1

-  debounce-fn@4.0.0:
-    dependencies:
-      mimic-fn: 3.1.0
-
  debug@4.3.4:
    dependencies:
      ms: 2.1.2
@@ -2621,10 +2520,6 @@ snapshots:
    dependencies:
      esutils: 2.0.3

-  dot-prop@6.0.1:
-    dependencies:
-      is-obj: 2.0.0
-
  duplexer3@0.1.5: {}

  eastasianwidth@0.2.0: {}
@@ -2644,8 +2539,6 @@ snapshots:
      ansi-colors: 4.1.3
      strip-ansi: 6.0.1

-  env-paths@2.2.1: {}
-
  error-ex@1.3.2:
    dependencies:
      is-arrayish: 0.2.1
@@ -2841,10 +2734,6 @@ snapshots:
    dependencies:
      to-regex-range: 5.0.1

-  find-up@3.0.0:
-    dependencies:
-      locate-path: 3.0.0
-
  find-up@4.1.0:
    dependencies:
      locate-path: 5.0.0
@@ -3129,8 +3018,6 @@ snapshots:

  is-number@7.0.0: {}

-  is-obj@2.0.0: {}
-
  is-path-inside@3.0.3: {}

  is-plain-obj@1.1.0: {}
@@ -3197,10 +3084,6 @@ snapshots:

  json-schema-traverse@0.4.1: {}

-  json-schema-traverse@1.0.0: {}
-
-  json-schema-typed@7.0.3: {}
-
  json-stable-stringify-without-jsonify@1.0.1: {}

  json-stringify-safe@5.0.1: {}
@@ -3239,11 +3122,6 @@ snapshots:
      pify: 4.0.1
      strip-bom: 3.0.0

-  locate-path@3.0.0:
-    dependencies:
-      p-locate: 3.0.0
-      path-exists: 3.0.0
-
  locate-path@5.0.0:
    dependencies:
      p-locate: 4.1.0
@@ -3301,8 +3179,6 @@ snapshots:

  mimic-fn@2.1.0: {}

-  mimic-fn@3.1.0: {}
-
  mimic-response@1.0.1: {}

  mimic-response@2.1.0: {}
@@ -3425,10 +3301,6 @@ snapshots:
    dependencies:
      yocto-queue: 0.1.0

-  p-locate@3.0.0:
-    dependencies:
-      p-limit: 2.3.0
-
  p-locate@4.1.0:
    dependencies:
      p-limit: 2.3.0
@@ -3456,8 +3328,6 @@ snapshots:
      json-parse-even-better-errors: 2.3.1
      lines-and-columns: 1.2.4

-  path-exists@3.0.0: {}
-
  path-exists@4.0.0: {}

  path-is-absolute@1.0.1: {}
@@ -3483,10 +3353,6 @@ snapshots:
    dependencies:
      find-up: 4.1.0

-  pkg-up@3.1.0:
-    dependencies:
-      find-up: 3.0.0
-
  playwright-core@1.44.0: {}

  playwright@1.44.0:
@@ -3515,7 +3381,7 @@ snapshots:

  prettier@3.2.5: {}

-  prompts@2.1.0:
+  prompts@2.4.2:
    dependencies:
      kleur: 3.0.3
      sisteransi: 1.0.5
@@ -3585,8 +3451,6 @@ snapshots:

  require-directory@2.1.1: {}

-  require-from-string@2.0.2: {}
-
  require-main-filename@2.0.0: {}

  resolve-from@4.0.0: {}
@@ -1,769 +0,0 @@
-import { execSync } from "child_process";
-import ciInfo from "ci-info";
-import fs from "fs";
-import path from "path";
-import { blue, green, red } from "picocolors";
-import prompts from "prompts";
-import { InstallAppArgs } from "./create-app";
-import {
-  TemplateDataSource,
-  TemplateDataSourceType,
-  TemplateFramework,
-  TemplateType,
-} from "./helpers";
-import { COMMUNITY_OWNER, COMMUNITY_REPO } from "./helpers/constant";
-import { EXAMPLE_FILE } from "./helpers/datasources";
-import { templatesDir } from "./helpers/dir";
-import { getAvailableLlamapackOptions } from "./helpers/llama-pack";
-import { askModelConfig } from "./helpers/providers";
-import { getProjectOptions } from "./helpers/repo";
-import {
-  supportedTools,
-  toolRequiresConfig,
-  toolsRequireConfig,
-} from "./helpers/tools";
-
-export type QuestionArgs = Omit<
-  InstallAppArgs,
-  "appPath" | "packageManager"
-> & {
-  askModels?: boolean;
-  askExamples?: boolean;
-};
-const supportedContextFileTypes = [
-  ".pdf",
-  ".doc",
-  ".docx",
-  ".xls",
-  ".xlsx",
-  ".csv",
-];
-const MACOS_FILE_SELECTION_SCRIPT = `
-osascript -l JavaScript -e '
-  a = Application.currentApplication();
-  a.includeStandardAdditions = true;
-  a.chooseFile({ withPrompt: "Please select files to process:", multipleSelectionsAllowed: true }).map(file => file.toString())
-'`;
-const MACOS_FOLDER_SELECTION_SCRIPT = `
-osascript -l JavaScript -e '
-  a = Application.currentApplication();
-  a.includeStandardAdditions = true;
-  a.chooseFolder({ withPrompt: "Please select folders to process:", multipleSelectionsAllowed: true }).map(folder => folder.toString())
-'`;
-const WINDOWS_FILE_SELECTION_SCRIPT = `
-Add-Type -AssemblyName System.Windows.Forms
-$openFileDialog = New-Object System.Windows.Forms.OpenFileDialog
-$openFileDialog.InitialDirectory = [Environment]::GetFolderPath('Desktop')
-$openFileDialog.Multiselect = $true
-$result = $openFileDialog.ShowDialog()
-if ($result -eq 'OK') {
-  $openFileDialog.FileNames
-}
-`;
-const WINDOWS_FOLDER_SELECTION_SCRIPT = `
-Add-Type -AssemblyName System.windows.forms
-$folderBrowser = New-Object System.Windows.Forms.FolderBrowserDialog
-$dialogResult = $folderBrowser.ShowDialog()
-if ($dialogResult -eq [System.Windows.Forms.DialogResult]::OK)
-{
-    $folderBrowser.SelectedPath
-}
-`;
-
-const defaults: Omit<QuestionArgs, "modelConfig"> = {
-  template: "streaming",
-  framework: "nextjs",
-  ui: "shadcn",
-  frontend: false,
-  llamaCloudKey: "",
-  useLlamaParse: false,
-  communityProjectConfig: undefined,
-  llamapack: "",
-  postInstallAction: "dependencies",
-  dataSources: [],
-  tools: [],
-};
-
-export const questionHandlers = {
-  onCancel: () => {
-    console.error("Exiting.");
-    process.exit(1);
-  },
-};
-
-const getVectorDbChoices = (framework: TemplateFramework) => {
-  const choices = [
-    {
-      title: "No, just store the data in the file system",
-      value: "none",
-    },
-    { title: "MongoDB", value: "mongo" },
-    { title: "PostgreSQL", value: "pg" },
-    { title: "Pinecone", value: "pinecone" },
-    { title: "Milvus", value: "milvus" },
-    { title: "Astra", value: "astra" },
-    { title: "Qdrant", value: "qdrant" },
-    { title: "ChromaDB", value: "chroma" },
-    { title: "Weaviate", value: "weaviate" },
-  ];
-
-  const vectordbLang = framework === "fastapi" ? "python" : "typescript";
-  const compPath = path.join(templatesDir, "components");
-  const vectordbPath = path.join(compPath, "vectordbs", vectordbLang);
-
-  const availableChoices = fs
-    .readdirSync(vectordbPath)
-    .filter((file) => fs.statSync(path.join(vectordbPath, file)).isDirectory());
-
-  const displayedChoices = choices.filter((choice) =>
-    availableChoices.includes(choice.value),
-  );
-
-  return displayedChoices;
-};
-
-export const getDataSourceChoices = (
-  framework: TemplateFramework,
-  selectedDataSource: TemplateDataSource[],
-  template?: TemplateType,
-) => {
-  // If LlamaCloud is already selected, don't show any other options
-  if (selectedDataSource.find((s) => s.type === "llamacloud")) {
-    return [];
-  }
-
-  const choices = [];
-
-  if (selectedDataSource.length > 0) {
-    choices.push({
-      title: "No",
-      value: "no",
-    });
-  }
-  if (selectedDataSource === undefined || selectedDataSource.length === 0) {
-    choices.push({
-      title: "No datasource",
-      value: "none",
-    });
-    choices.push({
-      title:
-        process.platform !== "linux"
-          ? "Use an example PDF"
-          : "Use an example PDF (you can add your own data files later)",
-      value: "exampleFile",
-    });
-  }
-
-  // Linux has many distros so we won't support file/folder picker for now
-  if (process.platform !== "linux") {
-    choices.push(
-      {
-        title: `Use local files (${supportedContextFileTypes.join(", ")})`,
-        value: "file",
-      },
-      {
-        title:
-          process.platform === "win32"
-            ? "Use a local folder"
-            : "Use local folders",
-        value: "folder",
-      },
-    );
-  }
-
-  if (framework === "fastapi" && template !== "extractor") {
-    choices.push({
-      title: "Use website content (requires Chrome)",
-      value: "web",
-    });
-    choices.push({
-      title: "Use data from a database (Mysql, PostgreSQL)",
-      value: "db",
-    });
-  }
-
-  if (!selectedDataSource.length && template !== "extractor") {
-    choices.push({
-      title: "Use managed index from LlamaCloud",
-      value: "llamacloud",
-    });
-  }
-  return choices;
-};
-
-const selectLocalContextData = async (type: TemplateDataSourceType) => {
-  try {
-    let selectedPath: string = "";
-    let execScript: string;
-    let execOpts: any = {};
-    switch (process.platform) {
-      case "win32": // Windows
-        execScript =
-          type === "file"
-            ? WINDOWS_FILE_SELECTION_SCRIPT
-            : WINDOWS_FOLDER_SELECTION_SCRIPT;
-        execOpts = { shell: "powershell.exe" };
-        break;
-      case "darwin": // MacOS
-        execScript =
-          type === "file"
-            ? MACOS_FILE_SELECTION_SCRIPT
-            : MACOS_FOLDER_SELECTION_SCRIPT;
-        break;
-      default: // Unsupported OS
-        console.log(red("Unsupported OS error!"));
-        process.exit(1);
-    }
-    selectedPath = execSync(execScript, execOpts).toString().trim();
-    const paths =
-      process.platform === "win32"
-        ? selectedPath.split("\r\n")
-        : selectedPath.split(", ");
-
-    for (const p of paths) {
-      if (
-        fs.statSync(p).isFile() &&
-        !supportedContextFileTypes.includes(path.extname(p))
-      ) {
-        console.log(
-          red(
-            `Please select a supported file type: ${supportedContextFileTypes}`,
-          ),
-        );
-        process.exit(1);
-      }
-    }
-    return paths;
-  } catch (error) {
-    console.log(
-      red(
-        "Got an error when trying to select local context data! Please try again or select another data source option.",
-      ),
-    );
-    process.exit(1);
-  }
-};
-
-export const onPromptState = (state: any) => {
-  if (state.aborted) {
-    // If we don't re-enable the terminal cursor before exiting
-    // the program, the cursor will remain hidden
-    process.stdout.write("\x1B[?25h");
-    process.stdout.write("\n");
-    process.exit(1);
-  }
-};
-
-export const askQuestions = async (
-  program: QuestionArgs,
-  preferences: QuestionArgs,
-  openAiKey?: string,
-) => {
-  const getPrefOrDefault = <K extends keyof Omit<QuestionArgs, "modelConfig">>(
-    field: K,
-  ): Omit<QuestionArgs, "modelConfig">[K] =>
-    preferences[field] ?? defaults[field];
-
-  // Ask for next action after installation
-  async function askPostInstallAction() {
-    if (program.postInstallAction === undefined) {
-      if (ciInfo.isCI) {
-        program.postInstallAction = getPrefOrDefault("postInstallAction");
-      } else {
-        const actionChoices = [
-          {
-            title: "Just generate code (~1 sec)",
-            value: "none",
-          },
-          {
-            title: "Start in VSCode (~1 sec)",
-            value: "VSCode",
-          },
-          {
-            title: "Generate code and install dependencies (~2 min)",
-            value: "dependencies",
-          },
-        ];
-
-        const modelConfigured =
-          !program.llamapack && program.modelConfig.isConfigured();
-        // If using LlamaParse, require LlamaCloud API key
-        const llamaCloudKeyConfigured = program.useLlamaParse
-          ? program.llamaCloudKey || process.env["LLAMA_CLOUD_API_KEY"]
-          : true;
-        const hasVectorDb = program.vectorDb && program.vectorDb !== "none";
-        // Can run the app if all tools do not require configuration
-        if (
-          !hasVectorDb &&
-          modelConfigured &&
-          llamaCloudKeyConfigured &&
-          !toolsRequireConfig(program.tools)
-        ) {
-          actionChoices.push({
-            title:
-              "Generate code, install dependencies, and run the app (~2 min)",
-            value: "runApp",
-          });
-        }
-
-        const { action } = await prompts(
-          {
-            type: "select",
-            name: "action",
-            message: "How would you like to proceed?",
-            choices: actionChoices,
-            initial: 1,
-          },
-          questionHandlers,
-        );
-
-        program.postInstallAction = action;
-      }
-    }
-  }
-
-  if (!program.template) {
-    if (ciInfo.isCI) {
-      program.template = getPrefOrDefault("template");
-    } else {
-      const styledRepo = blue(
-        `https://github.com/${COMMUNITY_OWNER}/${COMMUNITY_REPO}`,
-      );
-      const { template } = await prompts(
-        {
-          type: "select",
-          name: "template",
-          message: "Which template would you like to use?",
-          choices: [
-            { title: "Agentic RAG (e.g. chat with docs)", value: "streaming" },
-            {
-              title: "Multi-agent app (using workflows)",
-              value: "multiagent",
-            },
-            { title: "Structured Extractor", value: "extractor" },
-            ...(program.askExamples
-              ? [
-                  {
-                    title: `Community template from ${styledRepo}`,
-                    value: "community",
-                  },
-                  {
-                    title: "Example using a LlamaPack",
-                    value: "llamapack",
-                  },
-                ]
-              : []),
-          ],
-          initial: 0,
-        },
-        questionHandlers,
-      );
-      program.template = template;
-      preferences.template = template;
-    }
-  }
-
-  if (program.template === "community") {
-    const projectOptions = await getProjectOptions(
-      COMMUNITY_OWNER,
-      COMMUNITY_REPO,
-    );
-    const { communityProjectConfig } = await prompts(
-      {
-        type: "select",
-        name: "communityProjectConfig",
-        message: "Select community template",
-        choices: projectOptions.map(({ title, value }) => ({
-          title,
-          value: JSON.stringify(value), // serialize value to string in terminal
-        })),
-        initial: 0,
-      },
-      questionHandlers,
-    );
-    const projectConfig = JSON.parse(communityProjectConfig);
-    program.communityProjectConfig = projectConfig;
-    preferences.communityProjectConfig = projectConfig;
-    return; // early return - no further questions needed for community projects
-  }
-
-  if (program.template === "llamapack") {
-    const availableLlamaPacks = await getAvailableLlamapackOptions();
-    const { llamapack } = await prompts(
-      {
-        type: "select",
-        name: "llamapack",
-        message: "Select LlamaPack",
-        choices: availableLlamaPacks.map((pack) => ({
-          title: pack.name,
-          value: pack.folderPath,
-        })),
-        initial: 0,
-      },
-      questionHandlers,
-    );
-    program.llamapack = llamapack;
-    preferences.llamapack = llamapack;
-    await askPostInstallAction();
-    return; // early return - no further questions needed for llamapack projects
-  }
-
-  if (program.template === "extractor") {
-    // Extractor template only supports FastAPI, empty data sources, and llamacloud
-    // So we just use example file for extractor template, this allows user to choose vector database later
-    program.dataSources = [EXAMPLE_FILE];
-    program.framework = preferences.framework = "fastapi";
-  }
-  if (!program.framework) {
-    if (ciInfo.isCI) {
-      program.framework = getPrefOrDefault("framework");
-    } else {
-      const choices = [
-        { title: "NextJS", value: "nextjs" },
-        { title: "Express", value: "express" },
-        { title: "FastAPI (Python)", value: "fastapi" },
-      ];
-
-      const { framework } = await prompts(
-        {
-          type: "select",
-          name: "framework",
-          message: "Which framework would you like to use?",
-          choices,
-          initial: 0,
-        },
-        questionHandlers,
-      );
-      program.framework = framework;
-      preferences.framework = framework;
-    }
-  }
-
-  if (
-    (program.framework === "express" || program.framework === "fastapi") &&
-    (program.template === "streaming" || program.template === "multiagent")
-  ) {
-    // if a backend-only framework is selected, ask whether we should create a frontend
-    if (program.frontend === undefined) {
-      if (ciInfo.isCI) {
-        program.frontend = getPrefOrDefault("frontend");
-      } else {
-        const styledNextJS = blue("NextJS");
-        const styledBackend = green(
-          program.framework === "express"
-            ? "Express "
-            : program.framework === "fastapi"
-              ? "FastAPI (Python) "
-              : "",
-        );
-        const { frontend } = await prompts({
-          onState: onPromptState,
-          type: "toggle",
-          name: "frontend",
-          message: `Would you like to generate a ${styledNextJS} frontend for your ${styledBackend}backend?`,
-          initial: getPrefOrDefault("frontend"),
-          active: "Yes",
-          inactive: "No",
-        });
-        program.frontend = Boolean(frontend);
-        preferences.frontend = Boolean(frontend);
-      }
-    }
-  } else {
-    program.frontend = false;
-  }
-
-  if (program.framework === "nextjs" || program.frontend) {
-    if (!program.ui) {
-      program.ui = defaults.ui;
-    }
-  }
-
-  if (!program.observability && program.template === "streaming") {
-    if (ciInfo.isCI) {
-      program.observability = getPrefOrDefault("observability");
-    } else {
-      const { observability } = await prompts(
-        {
-          type: "select",
-          name: "observability",
-          message: "Would you like to set up observability?",
-          choices: [
-            { title: "No", value: "none" },
-            ...(program.framework === "fastapi"
-              ? [{ title: "LlamaTrace", value: "llamatrace" }]
-              : []),
-            { title: "Traceloop", value: "traceloop" },
-          ],
-          initial: 0,
-        },
-        questionHandlers,
-      );
-
-      program.observability = observability;
-      preferences.observability = observability;
-    }
-  }
-
-  if (!program.modelConfig) {
-    const modelConfig = await askModelConfig({
-      openAiKey,
-      askModels: program.askModels ?? false,
-      framework: program.framework,
-    });
-    program.modelConfig = modelConfig;
-    preferences.modelConfig = modelConfig;
-  }
-
-  if (!program.dataSources) {
-    if (ciInfo.isCI) {
-      program.dataSources = getPrefOrDefault("dataSources");
-    } else {
-      program.dataSources = [];
-      // continue asking user for data sources if none are initially provided
-      while (true) {
-        const firstQuestion = program.dataSources.length === 0;
-        const choices = getDataSourceChoices(
-          program.framework,
-          program.dataSources,
-          program.template,
-        );
-        if (choices.length === 0) break;
-        const { selectedSource } = await prompts(
-          {
-            type: "select",
-            name: "selectedSource",
-            message: firstQuestion
-              ? "Which data source would you like to use?"
-              : "Would you like to add another data source?",
-            choices,
-            initial: firstQuestion ? 1 : 0,
-          },
-          questionHandlers,
-        );
-
-        if (selectedSource === "no" || selectedSource === "none") {
-          // user doesn't want another data source or any data source
-          break;
-        }
-        switch (selectedSource) {
-          case "exampleFile": {
-            program.dataSources.push(EXAMPLE_FILE);
-            break;
-          }
-          case "file":
-          case "folder": {
-            const selectedPaths = await selectLocalContextData(selectedSource);
-            for (const p of selectedPaths) {
-              program.dataSources.push({
-                type: "file",
-                config: {
-                  path: p,
-                },
-              });
-            }
-            break;
-          }
-          case "web": {
-            const { baseUrl } = await prompts(
-              {
-                type: "text",
-                name: "baseUrl",
-                message: "Please provide base URL of the website: ",
-                initial: "https://www.llamaindex.ai",
-                validate: (value: string) => {
-                  if (!value.includes("://")) {
-                    value = `https://${value}`;
-                  }
-                  const urlObj = new URL(value);
-                  if (
-                    urlObj.protocol !== "https:" &&
-                    urlObj.protocol !== "http:"
-                  ) {
-                    return `URL=${value} has invalid protocol, only allow http or https`;
-                  }
-                  return true;
-                },
-              },
-              questionHandlers,
-            );
-
-            program.dataSources.push({
-              type: "web",
-              config: {
-                baseUrl,
-                prefix: baseUrl,
-                depth: 1,
-              },
-            });
-            break;
-          }
-          case "db": {
-            const dbPrompts: prompts.PromptObject<string>[] = [
-              {
-                type: "text",
-                name: "uri",
-                message:
-                  "Please enter the connection string (URI) for the database.",
-                initial: "mysql+pymysql://user:pass@localhost:3306/mydb",
-                validate: (value: string) => {
-                  if (!value) {
-                    return "Please provide a valid connection string";
-                  } else if (
-                    !(
-                      value.startsWith("mysql+pymysql://") ||
-                      value.startsWith("postgresql+psycopg://")
-                    )
-                  ) {
-                    return "The connection string must start with 'mysql+pymysql://' for MySQL or 'postgresql+psycopg://' for PostgreSQL";
-                  }
-                  return true;
-                },
-              },
-              // Only ask for a query, user can provide more complex queries in the config file later
-              {
-                type: (prev) => (prev ? "text" : null),
-                name: "queries",
-                message: "Please enter the SQL query to fetch data:",
-                initial: "SELECT * FROM mytable",
-              },
-            ];
-            program.dataSources.push({
-              type: "db",
-              config: await prompts(dbPrompts, questionHandlers),
-            });
-            break;
-          }
-          case "llamacloud": {
-            program.dataSources.push({
-              type: "llamacloud",
-              config: {},
-            });
-            program.dataSources.push(EXAMPLE_FILE);
-            break;
-          }
-        }
-      }
-    }
-  }
-
-  const isUsingLlamaCloud = program.dataSources.some(
-    (ds) => ds.type === "llamacloud",
-  );
-
-  // Asking for LlamaParse if user selected file data source
-  if (isUsingLlamaCloud) {
-    // default to use LlamaParse if using LlamaCloud
-    program.useLlamaParse = preferences.useLlamaParse = true;
-  } else {
-    // Extractor template doesn't support LlamaParse and LlamaCloud right now (cannot use asyncio loop in Reflex)
-    if (
-      program.useLlamaParse === undefined &&
-      program.template !== "extractor"
-    ) {
-      // if already set useLlamaParse, don't ask again
-      if (program.dataSources.some((ds) => ds.type === "file")) {
-        if (ciInfo.isCI) {
-          program.useLlamaParse = getPrefOrDefault("useLlamaParse");
-        } else {
-          const { useLlamaParse } = await prompts(
-            {
-              type: "toggle",
-              name: "useLlamaParse",
-              message:
-                "Would you like to use LlamaParse (improved parser for RAG - requires API key)?",
-              initial: false,
-              active: "yes",
-              inactive: "no",
-            },
-            questionHandlers,
-          );
-          program.useLlamaParse = useLlamaParse;
-          preferences.useLlamaParse = useLlamaParse;
-        }
-      }
-    }
-  }
-
-  // Ask for LlamaCloud API key when using a LlamaCloud index or LlamaParse
-  if (isUsingLlamaCloud || program.useLlamaParse) {
-    if (!program.llamaCloudKey) {
-      // if already set, don't ask again
-      if (ciInfo.isCI) {
-        program.llamaCloudKey = getPrefOrDefault("llamaCloudKey");
-      } else {
-        // Ask for LlamaCloud API key
-        const { llamaCloudKey } = await prompts(
-          {
-            type: "text",
-            name: "llamaCloudKey",
-            message:
-              "Please provide your LlamaCloud API key (leave blank to skip):",
-          },
-          questionHandlers,
-        );
-        program.llamaCloudKey = preferences.llamaCloudKey =
-          llamaCloudKey || process.env.LLAMA_CLOUD_API_KEY;
-      }
-    }
-  }
-
-  if (isUsingLlamaCloud) {
-    // When using a LlamaCloud index, don't ask for vector database and use code in `llamacloud` folder for vector database
-    const vectorDb = "llamacloud";
-    program.vectorDb = vectorDb;
-    preferences.vectorDb = vectorDb;
-  } else if (program.dataSources.length > 0 && !program.vectorDb) {
-    if (ciInfo.isCI) {
-      program.vectorDb = getPrefOrDefault("vectorDb");
-    } else {
-      const { vectorDb } = await prompts(
-        {
-          type: "select",
-          name: "vectorDb",
-          message: "Would you like to use a vector database?",
-          choices: getVectorDbChoices(program.framework),
-          initial: 0,
-        },
-        questionHandlers,
-      );
-      program.vectorDb = vectorDb;
-      preferences.vectorDb = vectorDb;
-    }
-  }
-
-  if (
-    !program.tools &&
-    (program.template === "streaming" || program.template === "multiagent")
-  ) {
-    if (ciInfo.isCI) {
-      program.tools = getPrefOrDefault("tools");
-    } else {
-      const options = supportedTools.filter((t) =>
-        t.supportedFrameworks?.includes(program.framework),
-      );
-      const toolChoices = options.map((tool) => ({
-        title: `${tool.display}${toolRequiresConfig(tool) ? " (needs configuration)" : ""}`,
-        value: tool.name,
-      }));
-      const { toolsName } = await prompts({
-        type: "multiselect",
-        name: "toolsName",
-        message:
-          "Would you like to build an agent using tools? If so, select the tools here, otherwise just press enter",
-        choices: toolChoices,
-      });
-      const tools = toolsName?.map((tool: string) =>
-        supportedTools.find((t) => t.name === tool),
-      );
-      program.tools = tools;
-      preferences.tools = tools;
-    }
-  }
-
-  await askPostInstallAction();
-};
-
-export const toChoice = (value: string) => {
-  return { title: value, value };
-};
@@ -0,0 +1,30 @@
+import { askModelConfig } from "../helpers/providers";
+import { QuestionArgs, QuestionResults } from "./types";
+
+const defaults: Omit<QuestionArgs, "modelConfig"> = {
+  template: "streaming",
+  framework: "nextjs",
+  ui: "shadcn",
+  frontend: false,
+  llamaCloudKey: "",
+  useLlamaParse: false,
+  communityProjectConfig: undefined,
+  llamapack: "",
+  postInstallAction: "dependencies",
+  dataSources: [],
+  tools: [],
+};
+
+export async function getCIQuestionResults(
+  program: QuestionArgs,
+): Promise<QuestionResults> {
+  return {
+    ...defaults,
+    ...program,
+    modelConfig: await askModelConfig({
+      openAiKey: program.openAiKey,
+      askModels: false,
+      framework: program.framework,
+    }),
+  };
+}
@@ -0,0 +1,64 @@
+import {
+  TemplateDataSource,
+  TemplateFramework,
+  TemplateType,
+} from "../helpers";
+import { supportedContextFileTypes } from "./utils";
+
+export const getDataSourceChoices = (
+  framework: TemplateFramework,
+  selectedDataSource: TemplateDataSource[],
+  template?: TemplateType,
+) => {
+  const choices = [];
+
+  if (selectedDataSource.length > 0) {
+    choices.push({
+      title: "No",
+      value: "no",
+    });
+  }
+  if (selectedDataSource === undefined || selectedDataSource.length === 0) {
+    choices.push({
+      title: "No datasource",
+      value: "none",
+    });
+    choices.push({
+      title:
+        process.platform !== "linux"
+          ? "Use an example PDF"
+          : "Use an example PDF (you can add your own data files later)",
+      value: "exampleFile",
+    });
+  }
+
+  // Linux has many distros so we won't support file/folder picker for now
+  if (process.platform !== "linux") {
+    choices.push(
+      {
+        title: `Use local files (${supportedContextFileTypes.join(", ")})`,
+        value: "file",
+      },
+      {
+        title:
+          process.platform === "win32"
+            ? "Use a local folder"
+            : "Use local folders",
+        value: "folder",
+      },
+    );
+  }
+
+  if (framework === "fastapi" && template !== "extractor") {
+    choices.push({
+      title: "Use website content (requires Chrome)",
+      value: "web",
+    });
+    choices.push({
+      title: "Use data from a database (Mysql, PostgreSQL)",
+      value: "db",
+    });
+  }
+
+  return choices;
+};
@@ -0,0 +1,18 @@
+import ciInfo from "ci-info";
+import { getCIQuestionResults } from "./ci";
+import { askProQuestions } from "./questions";
+import { askSimpleQuestions } from "./simple";
+import { QuestionArgs, QuestionResults } from "./types";
+
+export const askQuestions = async (
+  args: QuestionArgs,
+): Promise<QuestionResults> => {
+  if (ciInfo.isCI) {
+    return await getCIQuestionResults(args);
+  } else if (args.pro) {
+    // TODO: refactor pro questions to return a result object
+    await askProQuestions(args);
+    return args as unknown as QuestionResults;
+  }
+  return await askSimpleQuestions(args);
+};
@@ -0,0 +1,400 @@
+import { blue, green } from "picocolors";
+import prompts from "prompts";
+import { COMMUNITY_OWNER, COMMUNITY_REPO } from "../helpers/constant";
+import { EXAMPLE_FILE } from "../helpers/datasources";
+import { getAvailableLlamapackOptions } from "../helpers/llama-pack";
+import { askModelConfig } from "../helpers/providers";
+import { getProjectOptions } from "../helpers/repo";
+import { supportedTools, toolRequiresConfig } from "../helpers/tools";
+import { getDataSourceChoices } from "./datasources";
+import { getVectorDbChoices } from "./stores";
+import { QuestionArgs } from "./types";
+import {
+  askPostInstallAction,
+  onPromptState,
+  questionHandlers,
+  selectLocalContextData,
+} from "./utils";
+
+export const askProQuestions = async (program: QuestionArgs) => {
+  if (!program.template) {
+    const styledRepo = blue(
+      `https://github.com/${COMMUNITY_OWNER}/${COMMUNITY_REPO}`,
+    );
+    const { template } = await prompts(
+      {
+        type: "select",
+        name: "template",
+        message: "Which template would you like to use?",
+        choices: [
+          { title: "Agentic RAG (e.g. chat with docs)", value: "streaming" },
+          {
+            title: "Multi-agent app (using workflows)",
+            value: "multiagent",
+          },
+          { title: "Structured Extractor", value: "extractor" },
+          {
+            title: `Community template from ${styledRepo}`,
+            value: "community",
+          },
+          {
+            title: "Example using a LlamaPack",
+            value: "llamapack",
+          },
+        ],
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    program.template = template;
+  }
+
+  if (program.template === "community") {
+    const projectOptions = await getProjectOptions(
+      COMMUNITY_OWNER,
+      COMMUNITY_REPO,
+    );
+    const { communityProjectConfig } = await prompts(
+      {
+        type: "select",
+        name: "communityProjectConfig",
+        message: "Select community template",
+        choices: projectOptions.map(({ title, value }) => ({
+          title,
+          value: JSON.stringify(value), // serialize value to string in terminal
+        })),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    const projectConfig = JSON.parse(communityProjectConfig);
+    program.communityProjectConfig = projectConfig;
+    return; // early return - no further questions needed for community projects
+  }
+
+  if (program.template === "llamapack") {
+    const availableLlamaPacks = await getAvailableLlamapackOptions();
+    const { llamapack } = await prompts(
+      {
+        type: "select",
+        name: "llamapack",
+        message: "Select LlamaPack",
+        choices: availableLlamaPacks.map((pack) => ({
+          title: pack.name,
+          value: pack.folderPath,
+        })),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    program.llamapack = llamapack;
+    program.postInstallAction = await askPostInstallAction(program);
+    return; // early return - no further questions needed for llamapack projects
+  }
+
+  if (program.template === "extractor") {
+    // Extractor template only supports FastAPI, empty data sources, and llamacloud
+    // So we just use example file for extractor template, this allows user to choose vector database later
+    program.dataSources = [EXAMPLE_FILE];
+    program.framework = "fastapi";
+  }
+
+  if (!program.framework) {
+    const choices = [
+      { title: "NextJS", value: "nextjs" },
+      { title: "Express", value: "express" },
+      { title: "FastAPI (Python)", value: "fastapi" },
+    ];
+
+    const { framework } = await prompts(
+      {
+        type: "select",
+        name: "framework",
+        message: "Which framework would you like to use?",
+        choices,
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    program.framework = framework;
+  }
+
+  if (
+    (program.framework === "express" || program.framework === "fastapi") &&
+    (program.template === "streaming" || program.template === "multiagent")
+  ) {
+    // if a backend-only framework is selected, ask whether we should create a frontend
+    if (program.frontend === undefined) {
+      const styledNextJS = blue("NextJS");
+      const styledBackend = green(
+        program.framework === "express"
+          ? "Express "
+          : program.framework === "fastapi"
+            ? "FastAPI (Python) "
+            : "",
+      );
+      const { frontend } = await prompts({
+        onState: onPromptState,
+        type: "toggle",
+        name: "frontend",
+        message: `Would you like to generate a ${styledNextJS} frontend for your ${styledBackend}backend?`,
+        initial: false,
+        active: "Yes",
+        inactive: "No",
+      });
+      program.frontend = Boolean(frontend);
+    }
+  } else {
+    program.frontend = false;
+  }
+
+  if (program.framework === "nextjs" || program.frontend) {
+    if (!program.ui) {
+      program.ui = "shadcn";
+    }
+  }
+
+  if (!program.observability && program.template === "streaming") {
+    const { observability } = await prompts(
+      {
+        type: "select",
+        name: "observability",
+        message: "Would you like to set up observability?",
+        choices: [
+          { title: "No", value: "none" },
+          ...(program.framework === "fastapi"
+            ? [{ title: "LlamaTrace", value: "llamatrace" }]
+            : []),
+          { title: "Traceloop", value: "traceloop" },
+        ],
+        initial: 0,
+      },
+      questionHandlers,
+    );
+
+    program.observability = observability;
+  }
+
+  if (!program.modelConfig) {
+    const modelConfig = await askModelConfig({
+      openAiKey: program.openAiKey,
+      askModels: program.askModels ?? false,
+      framework: program.framework,
+    });
+    program.modelConfig = modelConfig;
+  }
+
+  if (!program.vectorDb) {
+    const { vectorDb } = await prompts(
+      {
+        type: "select",
+        name: "vectorDb",
+        message: "Would you like to use a vector database?",
+        choices: getVectorDbChoices(program.framework),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    program.vectorDb = vectorDb;
+  }
+
+  if (program.vectorDb === "llamacloud") {
+    // When using a LlamaCloud index, don't ask for data sources just copy an example file
+    program.dataSources = [EXAMPLE_FILE];
+  }
+
+  if (!program.dataSources) {
+    program.dataSources = [];
+    // continue asking user for data sources if none are initially provided
+    while (true) {
+      const firstQuestion = program.dataSources.length === 0;
+      const choices = getDataSourceChoices(
+        program.framework,
+        program.dataSources,
+        program.template,
+      );
+      if (choices.length === 0) break;
+      const { selectedSource } = await prompts(
+        {
+          type: "select",
+          name: "selectedSource",
+          message: firstQuestion
+            ? "Which data source would you like to use?"
+            : "Would you like to add another data source?",
+          choices,
+          initial: firstQuestion ? 1 : 0,
+        },
+        questionHandlers,
+      );
+
+      if (selectedSource === "no" || selectedSource === "none") {
+        // user doesn't want another data source or any data source
+        break;
+      }
+      switch (selectedSource) {
+        case "exampleFile": {
+          program.dataSources.push(EXAMPLE_FILE);
+          break;
+        }
+        case "file":
+        case "folder": {
+          const selectedPaths = await selectLocalContextData(selectedSource);
+          for (const p of selectedPaths) {
+            program.dataSources.push({
+              type: "file",
+              config: {
+                path: p,
+              },
+            });
+          }
+          break;
+        }
+        case "web": {
+          const { baseUrl } = await prompts(
+            {
+              type: "text",
+              name: "baseUrl",
+              message: "Please provide base URL of the website: ",
+              initial: "https://www.llamaindex.ai",
+              validate: (value: string) => {
+                if (!value.includes("://")) {
+                  value = `https://${value}`;
+                }
+                const urlObj = new URL(value);
+                if (
+                  urlObj.protocol !== "https:" &&
+                  urlObj.protocol !== "http:"
+                ) {
+                  return `URL=${value} has invalid protocol, only allow http or https`;
+                }
+                return true;
+              },
+            },
+            questionHandlers,
+          );
+
+          program.dataSources.push({
+            type: "web",
+            config: {
+              baseUrl,
+              prefix: baseUrl,
+              depth: 1,
+            },
+          });
+          break;
+        }
+        case "db": {
+          const dbPrompts: prompts.PromptObject<string>[] = [
+            {
+              type: "text",
+              name: "uri",
+              message:
+                "Please enter the connection string (URI) for the database.",
+              initial: "mysql+pymysql://user:pass@localhost:3306/mydb",
+              validate: (value: string) => {
+                if (!value) {
+                  return "Please provide a valid connection string";
+                } else if (
+                  !(
+                    value.startsWith("mysql+pymysql://") ||
+                    value.startsWith("postgresql+psycopg://")
+                  )
+                ) {
+                  return "The connection string must start with 'mysql+pymysql://' for MySQL or 'postgresql+psycopg://' for PostgreSQL";
+                }
+                return true;
+              },
+            },
+            // Only ask for a query, user can provide more complex queries in the config file later
+            {
+              type: (prev) => (prev ? "text" : null),
+              name: "queries",
+              message: "Please enter the SQL query to fetch data:",
+              initial: "SELECT * FROM mytable",
+            },
+          ];
+          program.dataSources.push({
+            type: "db",
+            config: await prompts(dbPrompts, questionHandlers),
+          });
+          break;
+        }
+      }
+    }
+  }
+
+  const isUsingLlamaCloud = program.vectorDb === "llamacloud";
+
+  // Asking for LlamaParse if user selected file data source
+  if (isUsingLlamaCloud) {
+    // default to use LlamaParse if using LlamaCloud
+    program.useLlamaParse = true;
+  } else {
+    // Extractor template doesn't support LlamaParse and LlamaCloud right now (cannot use asyncio loop in Reflex)
+    if (
+      program.useLlamaParse === undefined &&
+      program.template !== "extractor"
+    ) {
+      // if already set useLlamaParse, don't ask again
+      if (program.dataSources.some((ds) => ds.type === "file")) {
+        const { useLlamaParse } = await prompts(
+          {
+            type: "toggle",
+            name: "useLlamaParse",
+            message:
+              "Would you like to use LlamaParse (improved parser for RAG - requires API key)?",
+            initial: false,
+            active: "Yes",
+            inactive: "No",
+          },
+          questionHandlers,
+        );
+        program.useLlamaParse = useLlamaParse;
+      }
+    }
+  }
+
+  // Ask for LlamaCloud API key when using a LlamaCloud index or LlamaParse
+  if (isUsingLlamaCloud || program.useLlamaParse) {
+    if (!program.llamaCloudKey) {
+      // if already set, don't ask again
+      // Ask for LlamaCloud API key
+      const { llamaCloudKey } = await prompts(
+        {
+          type: "text",
+          name: "llamaCloudKey",
+          message:
+            "Please provide your LlamaCloud API key (leave blank to skip):",
+        },
+        questionHandlers,
+      );
+      program.llamaCloudKey = llamaCloudKey || process.env.LLAMA_CLOUD_API_KEY;
+    }
+  }
+
+  if (
+    !program.tools &&
+    (program.template === "streaming" || program.template === "multiagent")
+  ) {
+    const options = supportedTools.filter((t) =>
+      t.supportedFrameworks?.includes(program.framework),
+    );
+    const toolChoices = options.map((tool) => ({
+      title: `${tool.display}${toolRequiresConfig(tool) ? " (needs configuration)" : ""}`,
+      value: tool.name,
+    }));
+    const { toolsName } = await prompts({
+      type: "multiselect",
+      name: "toolsName",
+      message:
+        "Would you like to build an agent using tools? If so, select the tools here, otherwise just press enter",
+      choices: toolChoices,
+    });
+    const tools = toolsName?.map((tool: string) =>
+      supportedTools.find((t) => t.name === tool),
+    );
+    program.tools = tools;
+  }
+
+  program.postInstallAction = await askPostInstallAction(program);
+};
@@ -0,0 +1,148 @@
+import prompts from "prompts";
+import { EXAMPLE_FILE } from "../helpers/datasources";
+import { askModelConfig } from "../helpers/providers";
+import { getTools } from "../helpers/tools";
+import { ModelConfig, TemplateFramework } from "../helpers/types";
+import { PureQuestionArgs, QuestionResults } from "./types";
+import { askPostInstallAction, questionHandlers } from "./utils";
+type AppType = "rag" | "code_artifact" | "multiagent" | "extractor";
+
+type SimpleAnswers = {
+  appType: AppType;
+  language: TemplateFramework;
+  useLlamaCloud: boolean;
+  llamaCloudKey?: string;
+  modelConfig: ModelConfig;
+};
+
+export const askSimpleQuestions = async (
+  args: PureQuestionArgs,
+): Promise<QuestionResults> => {
+  const { appType } = await prompts(
+    {
+      type: "select",
+      name: "appType",
+      message: "What app do you want to build?",
+      choices: [
+        { title: "Agentic RAG", value: "rag" },
+        { title: "Code Artifact Agent", value: "code_artifact" },
+        { title: "Multi-Agent Report Gen", value: "multiagent" },
+        { title: "Structured extraction", value: "extractor" },
+      ],
+    },
+    questionHandlers,
+  );
+
+  let language: TemplateFramework = "fastapi";
+  if (appType !== "extractor") {
+    const res = await prompts(
+      {
+        type: "select",
+        name: "language",
+        message: "What language do you want to use?",
+        choices: [
+          { title: "Python (FastAPI)", value: "fastapi" },
+          { title: "Typescript (NextJS)", value: "nextjs" },
+        ],
+      },
+      questionHandlers,
+    );
+    language = res.language;
+  }
+
+  const { useLlamaCloud } = await prompts(
+    {
+      type: "toggle",
+      name: "useLlamaCloud",
+      message: "Do you want to use LlamaCloud services?",
+      initial: false,
+      active: "Yes",
+      inactive: "No",
+      hint: "see https://www.llamaindex.ai/enterprise for more info",
+    },
+    questionHandlers,
+  );
+
+  let llamaCloudKey = args.llamaCloudKey;
+  if (useLlamaCloud && !llamaCloudKey) {
+    // Ask for LlamaCloud API key, if not set
+    const { llamaCloudKey: newLlamaCloudKey } = await prompts(
+      {
+        type: "text",
+        name: "llamaCloudKey",
+        message:
+          "Please provide your LlamaCloud API key (leave blank to skip):",
+      },
+      questionHandlers,
+    );
+    llamaCloudKey = newLlamaCloudKey || process.env.LLAMA_CLOUD_API_KEY;
+  }
+
+  const modelConfig = await askModelConfig({
+    openAiKey: args.openAiKey,
+    askModels: args.askModels ?? false,
+    framework: language,
+  });
+
+  const results = convertAnswers({
+    appType,
+    language,
+    useLlamaCloud,
+    llamaCloudKey,
+    modelConfig,
+  });
+
+  results.postInstallAction = await askPostInstallAction(results);
+  return results;
+};
+
+const convertAnswers = (answers: SimpleAnswers): QuestionResults => {
+  const lookup: Record<
+    AppType,
+    Pick<QuestionResults, "template" | "tools" | "frontend" | "dataSources">
+  > = {
+    rag: {
+      template: "streaming",
+      tools: getTools(["duckduckgo"]),
+      frontend: true,
+      dataSources: [EXAMPLE_FILE],
+    },
+    code_artifact: {
+      template: "streaming",
+      tools: getTools(["artifact"]),
+      frontend: true,
+      dataSources: [],
+    },
+    multiagent: {
+      template: "multiagent",
+      tools: getTools([
+        "document_generator",
+        "wikipedia.WikipediaToolSpec",
+        "duckduckgo",
+        "img_gen",
+      ]),
+      frontend: true,
+      dataSources: [EXAMPLE_FILE],
+    },
+    extractor: {
+      template: "extractor",
+      tools: [],
+      frontend: false,
+      dataSources: [EXAMPLE_FILE],
+    },
+  };
+  const results = lookup[answers.appType];
+  return {
+    framework: answers.language,
+    ui: "shadcn",
+    llamaCloudKey: answers.llamaCloudKey,
+    useLlamaParse: answers.useLlamaCloud,
+    llamapack: "",
+    postInstallAction: "none",
+    vectorDb: answers.useLlamaCloud ? "llamacloud" : "none",
+    modelConfig: answers.modelConfig,
+    observability: "none",
+    ...results,
+    frontend: answers.language === "nextjs" ? false : results.frontend,
+  };
+};
@@ -0,0 +1,36 @@
+import fs from "fs";
+import path from "path";
+import { TemplateFramework } from "../helpers";
+import { templatesDir } from "../helpers/dir";
+
+export const getVectorDbChoices = (framework: TemplateFramework) => {
+  const choices = [
+    {
+      title: "No, just store the data in the file system",
+      value: "none",
+    },
+    { title: "MongoDB", value: "mongo" },
+    { title: "PostgreSQL", value: "pg" },
+    { title: "Pinecone", value: "pinecone" },
+    { title: "Milvus", value: "milvus" },
+    { title: "Astra", value: "astra" },
+    { title: "Qdrant", value: "qdrant" },
+    { title: "ChromaDB", value: "chroma" },
+    { title: "Weaviate", value: "weaviate" },
+    { title: "LlamaCloud (use Managed Index)", value: "llamacloud" },
+  ];
+
+  const vectordbLang = framework === "fastapi" ? "python" : "typescript";
+  const compPath = path.join(templatesDir, "components");
+  const vectordbPath = path.join(compPath, "vectordbs", vectordbLang);
+
+  const availableChoices = fs
+    .readdirSync(vectordbPath)
+    .filter((file) => fs.statSync(path.join(vectordbPath, file)).isDirectory());
+
+  const displayedChoices = choices.filter((choice) =>
+    availableChoices.includes(choice.value),
+  );
+
+  return displayedChoices;
+};
@@ -0,0 +1,15 @@
+import { InstallAppArgs } from "../create-app";
+
+export type QuestionResults = Omit<
+  InstallAppArgs,
+  "appPath" | "packageManager" | "externalPort"
+>;
+
+export type PureQuestionArgs = {
+  askModels?: boolean;
+  pro?: boolean;
+  openAiKey?: string;
+  llamaCloudKey?: string;
+};
+
+export type QuestionArgs = QuestionResults & PureQuestionArgs;
@@ -0,0 +1,178 @@
+import { execSync } from "child_process";
+import fs from "fs";
+import path from "path";
+import { red } from "picocolors";
+import prompts from "prompts";
+import { TemplateDataSourceType, TemplatePostInstallAction } from "../helpers";
+import { toolsRequireConfig } from "../helpers/tools";
+import { QuestionResults } from "./types";
+
+export const supportedContextFileTypes = [
+  ".pdf",
+  ".doc",
+  ".docx",
+  ".xls",
+  ".xlsx",
+  ".csv",
+];
+
+const MACOS_FILE_SELECTION_SCRIPT = `
+osascript -l JavaScript -e '
+  a = Application.currentApplication();
+  a.includeStandardAdditions = true;
+  a.chooseFile({ withPrompt: "Please select files to process:", multipleSelectionsAllowed: true }).map(file => file.toString())
+'`;
+
+const MACOS_FOLDER_SELECTION_SCRIPT = `
+osascript -l JavaScript -e '
+  a = Application.currentApplication();
+  a.includeStandardAdditions = true;
+  a.chooseFolder({ withPrompt: "Please select folders to process:", multipleSelectionsAllowed: true }).map(folder => folder.toString())
+'`;
+
+const WINDOWS_FILE_SELECTION_SCRIPT = `
+Add-Type -AssemblyName System.Windows.Forms
+$openFileDialog = New-Object System.Windows.Forms.OpenFileDialog
+$openFileDialog.InitialDirectory = [Environment]::GetFolderPath('Desktop')
+$openFileDialog.Multiselect = $true
+$result = $openFileDialog.ShowDialog()
+if ($result -eq 'OK') {
+  $openFileDialog.FileNames
+}
+`;
+
+const WINDOWS_FOLDER_SELECTION_SCRIPT = `
+Add-Type -AssemblyName System.windows.forms
+$folderBrowser = New-Object System.Windows.Forms.FolderBrowserDialog
+$dialogResult = $folderBrowser.ShowDialog()
+if ($dialogResult -eq [System.Windows.Forms.DialogResult]::OK)
+{
+    $folderBrowser.SelectedPath
+}
+`;
+
+export const selectLocalContextData = async (type: TemplateDataSourceType) => {
+  try {
+    let selectedPath: string = "";
+    let execScript: string;
+    let execOpts: any = {};
+    switch (process.platform) {
+      case "win32": // Windows
+        execScript =
+          type === "file"
+            ? WINDOWS_FILE_SELECTION_SCRIPT
+            : WINDOWS_FOLDER_SELECTION_SCRIPT;
+        execOpts = { shell: "powershell.exe" };
+        break;
+      case "darwin": // MacOS
+        execScript =
+          type === "file"
+            ? MACOS_FILE_SELECTION_SCRIPT
+            : MACOS_FOLDER_SELECTION_SCRIPT;
+        break;
+      default: // Unsupported OS
+        console.log(red("Unsupported OS error!"));
+        process.exit(1);
+    }
+    selectedPath = execSync(execScript, execOpts).toString().trim();
+    const paths =
+      process.platform === "win32"
+        ? selectedPath.split("\r\n")
+        : selectedPath.split(", ");
+
+    for (const p of paths) {
+      if (
+        fs.statSync(p).isFile() &&
+        !supportedContextFileTypes.includes(path.extname(p))
+      ) {
+        console.log(
+          red(
+            `Please select a supported file type: ${supportedContextFileTypes}`,
+          ),
+        );
+        process.exit(1);
+      }
+    }
+    return paths;
+  } catch (error) {
+    console.log(
+      red(
+        "Got an error when trying to select local context data! Please try again or select another data source option.",
+      ),
+    );
+    process.exit(1);
+  }
+};
+
+export const onPromptState = (state: any) => {
+  if (state.aborted) {
+    // If we don't re-enable the terminal cursor before exiting
+    // the program, the cursor will remain hidden
+    process.stdout.write("\x1B[?25h");
+    process.stdout.write("\n");
+    process.exit(1);
+  }
+};
+
+export const toChoice = (value: string) => {
+  return { title: value, value };
+};
+
+export const questionHandlers = {
+  onCancel: () => {
+    console.error("Exiting.");
+    process.exit(1);
+  },
+};
+
+// Ask for next action after installation
+export async function askPostInstallAction(
+  args: QuestionResults,
+): Promise<TemplatePostInstallAction> {
+  const actionChoices = [
+    {
+      title: "Just generate code (~1 sec)",
+      value: "none",
+    },
+    {
+      title: "Start in VSCode (~1 sec)",
+      value: "VSCode",
+    },
+    {
+      title: "Generate code and install dependencies (~2 min)",
+      value: "dependencies",
+    },
+  ];
+
+  const modelConfigured = !args.llamapack && args.modelConfig.isConfigured();
+  // If using LlamaParse, require LlamaCloud API key
+  const llamaCloudKeyConfigured = args.useLlamaParse
+    ? args.llamaCloudKey || process.env["LLAMA_CLOUD_API_KEY"]
+    : true;
+  const hasVectorDb = args.vectorDb && args.vectorDb !== "none";
+  // Can run the app if all tools do not require configuration
+  if (
+    !hasVectorDb &&
+    modelConfigured &&
+    llamaCloudKeyConfigured &&
+    !toolsRequireConfig(args.tools)
+  ) {
+    actionChoices.push({
+      title: "Generate code, install dependencies, and run the app (~2 min)",
+      value: "runApp",
+    });
+  }
+
+  const { action } = await prompts(
+    {
+      type: "select",
+      name: "action",
+      message: "How would you like to proceed?",
+      choices: actionChoices,
+      initial: 1,
+    },
+    questionHandlers,
+  );
+
+  return action;
+}
@@ -1,17 +1,19 @@
 import os
+from typing import List

 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 import BaseTool
 from llama_index.core.tools.query_engine import QueryEngineTool


 def get_chat_engine(filters=None, params=None, event_handlers=None, **kwargs):
    system_prompt = os.getenv("SYSTEM_PROMPT")
    top_k = int(os.getenv("TOP_K", 0))
-    tools = []
+    tools: List[BaseTool] = []
    callback_manager = CallbackManager(handlers=event_handlers or [])

    # Add query tool if index exists
@@ -25,7 +27,8 @@ def get_chat_engine(filters=None, params=None, event_handlers=None, **kwargs):
        tools.append(query_engine_tool)

    # Add additional tools
-    tools += ToolFactory.from_env()
+    configured_tools: List[BaseTool] = ToolFactory.from_env()
+    tools.extend(configured_tools)

    return AgentRunner.from_llm(
        llm=Settings.llm,
@@ -1,7 +1,8 @@
 import importlib
 import os
+from typing import Dict, List, Union

-import yaml
+import yaml  # type: ignore
 from llama_index.core.tools.function_tool import FunctionTool
 from llama_index.core.tools.tool_spec.base import BaseToolSpec

@@ -17,7 +18,8 @@ class ToolFactory:
        ToolType.LOCAL: "app.engine.tools",
    }

-    def load_tools(tool_type: str, tool_name: str, config: dict) -> list[FunctionTool]:
+    @staticmethod
+    def load_tools(tool_type: str, tool_name: str, config: dict) -> List[FunctionTool]:
        source_package = ToolFactory.TOOL_SOURCE_PACKAGE_MAP[tool_type]
        try:
            if "ToolSpec" in tool_name:
@@ -43,24 +45,32 @@ class ToolFactory:
    @staticmethod
    def from_env(
        map_result: bool = False,
-    ) -> list[FunctionTool] | dict[str, FunctionTool]:
+    ) -> Union[Dict[str, List[FunctionTool]], List[FunctionTool]]:
        """
        Load tools from the configured file.
-        Params:
-            - use_map: if True, return map of tool name and the tool itself
+
+        Args:
+            map_result: If True, return a map of tool names to their corresponding tools.
+
+        Returns:
+            A dictionary of tool names to lists of FunctionTools if map_result is True,
+            otherwise a list of FunctionTools.
        """
-        if map_result:
-            tools = {}
-        else:
-            tools = []
+        tools: Union[Dict[str, List[FunctionTool]], List[FunctionTool]] = (
+            {} if map_result else []
+        )
+
        if os.path.exists("config/tools.yaml"):
            with open("config/tools.yaml", "r") as f:
                tool_configs = yaml.safe_load(f)
                for tool_type, config_entries in tool_configs.items():
                    for tool_name, config in config_entries.items():
-                        tool = ToolFactory.load_tools(tool_type, tool_name, config)
+                        loaded_tools = ToolFactory.load_tools(
+                            tool_type, tool_name, config
+                        )
                        if map_result:
-                            tools[tool_name] = tool
+                            tools[tool_name] = loaded_tools  # type: ignore
                        else:
-                            tools.extend(tool)
+                            tools.extend(loaded_tools)  # type: ignore
+
        return tools
@@ -87,7 +87,7 @@ class CodeGeneratorTool:
            ChatMessage(role="user", content=user_message),
        ]
        try:
-            sllm = Settings.llm.as_structured_llm(output_cls=CodeArtifact)
+            sllm = Settings.llm.as_structured_llm(output_cls=CodeArtifact)  # type: ignore
            response = sllm.chat(messages)
            data: CodeArtifact = response.raw
            return data.model_dump()
@@ -21,14 +21,15 @@ def duckduckgo_search(
            "Please install it by running: `poetry add duckduckgo_search` or `pip install duckduckgo_search`"
        )

-    params = {
-        "keywords": query,
-        "region": region,
-        "max_results": max_results,
-    }
    results = []
    with DDGS() as ddg:
-        results = list(ddg.text(**params))
+        results = list(
+            ddg.text(
+                keywords=query,
+                region=region,
+                max_results=max_results,
+            )
+        )
    return results


@@ -51,13 +52,14 @@ def duckduckgo_image_search(
            "duckduckgo_search package is required to use this function."
            "Please install it by running: `poetry add duckduckgo_search` or `pip install duckduckgo_search`"
        )
-    params = {
-        "keywords": query,
-        "region": region,
-        "max_results": max_results,
-    }
    with DDGS() as ddg:
-        results = list(ddg.images(**params))
+        results = list(
+            ddg.images(
+                keywords=query,
+                region=region,
+                max_results=max_results,
+            )
+        )
    return results


@@ -43,6 +43,6 @@ def get_chat_engine(filters=None, params=None, event_handlers=None, **kwargs):
        memory=memory,
        system_prompt=system_prompt,
        retriever=retriever,
-        node_postprocessors=node_postprocessors,
+        node_postprocessors=node_postprocessors,  # type: ignore
        callback_manager=callback_manager,
    )
@@ -1,6 +1,6 @@
 import {
+  BaseChatEngine,
  BaseToolWithCall,
-  ChatEngine,
  OpenAIAgent,
  QueryEngineTool,
 } from "llamaindex";
@@ -45,7 +45,7 @@ export async function createChatEngine(documentIds?: string[], params?: any) {
  const agent = new OpenAIAgent({
    tools,
    systemPrompt: process.env.SYSTEM_PROMPT,
-  }) as unknown as ChatEngine;
+  }) as unknown as BaseChatEngine;

  return agent;
 }
@@ -16,14 +16,26 @@ export async function uploadDocument(
    // trigger LlamaCloudIndex API to upload the file and run the pipeline
    const projectId = await index.getProjectId();
    const pipelineId = await index.getPipelineId();
-    return [
-      await LLamaCloudFileService.addFileToPipeline(
-        projectId,
-        pipelineId,
-        new File([fileBuffer], filename, { type: mimeType }),
-        { private: "true" },
-      ),
-    ];
+    try {
+      return [
+        await LLamaCloudFileService.addFileToPipeline(
+          projectId,
+          pipelineId,
+          new File([fileBuffer], filename, { type: mimeType }),
+          { private: "true" },
+        ),
+      ];
+    } catch (error) {
+      if (
+        error instanceof ReferenceError &&
+        error.message.includes("File is not defined")
+      ) {
+        throw new Error(
+          "File class is not supported in the current Node.js version. Please use Node.js 20 or higher.",
+        );
+      }
+      throw error;
+    }
  }

  // run the pipeline for other vector store indexes
@@ -172,3 +172,26 @@ function getValidAnnotation(annotation: JSONValue): Annotation {
  }
  return { type: annotation.type, data: annotation.data };
 }
+
+// validate and get all annotations of a specific type or role from the frontend messages
+export function getAnnotations<
+  T extends Annotation["data"] = Annotation["data"],
+>(
+  messages: Message[],
+  options?: {
+    role?: Message["role"]; // message role
+    type?: Annotation["type"]; // annotation type
+  },
+): {
+  type: string;
+  data: T;
+}[] {
+  const messagesByRole = options?.role
+    ? messages.filter((msg) => msg.role === options?.role)
+    : messages;
+  const annotations = getAllAnnotations(messagesByRole);
+  const annotationsByType = options?.type
+    ? annotations.filter((a) => a.type === options.type)
+    : annotations;
+  return annotationsByType as { type: string; data: T }[];
+}
@@ -75,7 +75,7 @@ export function createCallbackManager(stream: StreamData) {
  callbackManager.on("retrieve-end", (data) => {
    const { nodes, query } = data.detail;
    appendSourceData(stream, nodes);
-    appendEventData(stream, `Retrieving context for query: '${query}'`);
+    appendEventData(stream, `Retrieving context for query: '${query.query}'`);
    appendEventData(
      stream,
      `Retrieved ${nodes.length} sources to use as context for the query`,
@@ -1,20 +1,22 @@
 import logging
+from typing import Any, Dict, List

-import yaml
+import yaml  # type: ignore
 from app.engine.loaders.db import DBLoaderConfig, get_db_documents
 from app.engine.loaders.file import FileLoaderConfig, get_file_documents
 from app.engine.loaders.web import WebLoaderConfig, get_web_documents
+from llama_index.core import Document

 logger = logging.getLogger(__name__)


-def load_configs():
+def load_configs() -> Dict[str, Any]:
    with open("config/loaders.yaml") as f:
        configs = yaml.safe_load(f)
    return configs


-def get_documents():
+def get_documents() -> List[Document]:
    documents = []
    config = load_configs()
    for loader_type, loader_config in config.items():
@@ -1,5 +1,6 @@
 import logging
 from typing import List
+
 from pydantic import BaseModel

 logger = logging.getLogger(__name__)
@@ -11,7 +12,13 @@ class DBLoaderConfig(BaseModel):


 def get_db_documents(configs: list[DBLoaderConfig]):
-    from llama_index.readers.database import DatabaseReader
+    try:
+        from llama_index.readers.database import DatabaseReader
+    except ImportError:
+        logger.error(
+            "Failed to import DatabaseReader. Make sure llama_index is installed."
+        )
+        raise

    docs = []
    for entry in configs:
@@ -1,3 +1,5 @@
+from typing import List, Optional
+
 from pydantic import BaseModel, Field


@@ -8,8 +10,8 @@ class CrawlUrl(BaseModel):


 class WebLoaderConfig(BaseModel):
-    driver_arguments: list[str] = Field(default=None)
-    urls: list[CrawlUrl]
+    driver_arguments: Optional[List[str]] = Field(default_factory=list)
+    urls: List[CrawlUrl]


 def get_web_documents(config: WebLoaderConfig):
@@ -1,11 +1,10 @@
 import logging

-from app.api.routers.events import EventCallbackHandler
 from app.api.routers.models import (
    ChatData,
 )
 from app.api.routers.vercel_response import VercelStreamResponse
-from app.engine import get_chat_engine
+from app.engine.engine import get_chat_engine
 from fastapi import APIRouter, BackgroundTasks, HTTPException, Request, status

 chat_router = r = APIRouter()
@@ -23,13 +22,12 @@ async def chat(
        last_message_content = data.get_last_message_content()
        messages = data.get_history_messages(include_agent_messages=True)

-        event_handler = EventCallbackHandler()
        # The chat API supports passing private document filters and chat params
        # but agent workflow does not support them yet
        # ignore chat params and use all documents for now
        # TODO: generate filters based on doc_ids
-        # TODO: use chat params
-        engine = get_chat_engine(chat_history=messages)
+        params = data.data or {}
+        engine = get_chat_engine(chat_history=messages, params=params)

        event_handler = engine.run(input=last_message_content, streaming=True)
        return VercelStreamResponse(
@@ -1,6 +1,6 @@
+import asyncio
 import json
 import logging
-from abc import ABC
 from typing import AsyncGenerator, List

 from aiostream import stream
@@ -13,7 +13,7 @@ from fastapi.responses import StreamingResponse
 logger = logging.getLogger("uvicorn")


-class VercelStreamResponse(StreamingResponse, ABC):
+class VercelStreamResponse(StreamingResponse):
    """
    Base class to convert the response from the chat engine to the streaming format expected by Vercel
    """
@@ -23,26 +23,34 @@ class VercelStreamResponse(StreamingResponse, ABC):

    def __init__(self, request: Request, chat_data: ChatData, *args, **kwargs):
        self.request = request
-
-        stream = self._create_stream(request, chat_data, *args, **kwargs)
-        content = self.content_generator(stream)
-
+        self.chat_data = chat_data
+        content = self.content_generator(*args, **kwargs)
        super().__init__(content=content)

-    async def content_generator(self, stream):
+    async def content_generator(self, event_handler, events):
+        logger.info("Starting content_generator")
+        stream = self._create_stream(
+            self.request, self.chat_data, event_handler, events
+        )
        is_stream_started = False
+        try:
+            async with stream.stream() as streamer:
+                async for output in streamer:
+                    if not is_stream_started:
+                        is_stream_started = True
+                        # Stream a blank message to start the stream
+                        yield self.convert_text("")

-        async with stream.stream() as streamer:
-            async for output in streamer:
-                if not is_stream_started:
-                    is_stream_started = True
-                    # Stream a blank message to start the stream
-                    yield self.convert_text("")
-
-                yield output
-
-                if await self.request.is_disconnected():
-                    break
+                    yield output
+        except asyncio.CancelledError:
+            logger.info("Stopping workflow")
+            await event_handler.cancel_run()
+        except Exception as e:
+            logger.error(
+                f"Unexpected error in content_generator: {str(e)}", exc_info=True
+            )
+        finally:
+            logger.info("The stream has been stopped!")

    def _create_stream(
        self,
@@ -18,11 +18,11 @@ def get_chat_engine(
    agent_type = os.getenv("EXAMPLE_TYPE", "").lower()
    match agent_type:
        case "choreography":
-            agent = create_choreography(chat_history)
+            agent = create_choreography(chat_history, **kwargs)
        case "orchestrator":
-            agent = create_orchestrator(chat_history)
+            agent = create_orchestrator(chat_history, **kwargs)
        case _:
-            agent = create_workflow(chat_history)
+            agent = create_workflow(chat_history, **kwargs)

    logger.info(f"Using agent pattern: {agent_type}")

@@ -8,8 +8,8 @@ from app.examples.researcher import create_researcher
 from llama_index.core.chat_engine.types import ChatMessage


-def create_choreography(chat_history: Optional[List[ChatMessage]] = None):
-    researcher = create_researcher(chat_history)
+def create_choreography(chat_history: Optional[List[ChatMessage]] = None, **kwargs):
+    researcher = create_researcher(chat_history, **kwargs)
    publisher = create_publisher(chat_history)
    reviewer = FunctionCallingAgent(
        name="reviewer",
@@ -21,12 +21,14 @@ def create_choreography(chat_history: Optional[List[ChatMessage]] = None):
        name="writer",
        agents=[researcher, reviewer, publisher],
        description="expert in writing blog posts, needs researched information and images to write a blog post",
-        system_prompt=dedent("""
+        system_prompt=dedent(
+            """
            You are an expert in writing blog posts. You are given a task to write a blog post. Before starting to write the post, consult the researcher agent to get the information you need. Don't make up any information yourself.
            After creating a draft for the post, send it to the reviewer agent to receive feedback and make sure to incorporate the feedback from the reviewer.
            You can consult the reviewer and researcher a maximum of two times. Your output should contain only the blog post.
            Finally, always request the publisher to create a document (PDF, HTML) and publish the blog post.
-        """),
+        """
+        ),
        # TODO: add chat_history support to AgentCallingAgent
        # chat_history=chat_history,
    )
@@ -8,28 +8,32 @@ from app.examples.researcher import create_researcher
 from llama_index.core.chat_engine.types import ChatMessage


-def create_orchestrator(chat_history: Optional[List[ChatMessage]] = None):
-    researcher = create_researcher(chat_history)
+def create_orchestrator(chat_history: Optional[List[ChatMessage]] = None, **kwargs):
+    researcher = create_researcher(chat_history, **kwargs)
    writer = FunctionCallingAgent(
        name="writer",
        description="expert in writing blog posts, need information and images to write a post",
-        system_prompt=dedent("""
+        system_prompt=dedent(
+            """
            You are an expert in writing blog posts.
            You are given a task to write a blog post. Do not make up any information yourself.
            If you don't have the necessary information to write a blog post, reply "I need information about the topic to write the blog post".
            If you need to use images, reply "I need images about the topic to write the blog post". Do not use any dummy images made up by you.
            If you have all the information needed, write the blog post.
-        """),
+        """
+        ),
        chat_history=chat_history,
    )
    reviewer = FunctionCallingAgent(
        name="reviewer",
        description="expert in reviewing blog posts, needs a written blog post to review",
-        system_prompt=dedent("""
+        system_prompt=dedent(
+            """
            You are an expert in reviewing blog posts. You are given a task to review a blog post. Review the post and fix any issues found yourself. You must output a final blog post.
            A post must include at least one valid image. If not, reply "I need images about the topic to write the blog post". An image URL starting with "example" or "your website" is not valid.
            Especially check for logical inconsistencies and proofread the post for grammar and spelling errors.
-        """),
+        """
+        ),
        chat_history=chat_history,
    )
    publisher = create_publisher(chat_history)
@@ -3,17 +3,19 @@ from textwrap import dedent
 from typing import List

 from app.agents.single import FunctionCallingAgent
-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.chat_engine.types import ChatMessage
 from llama_index.core.tools import QueryEngineTool, ToolMetadata


-def _create_query_engine_tool() -> QueryEngineTool:
+def _create_query_engine_tool(params=None) -> QueryEngineTool:
    """
    Provide an agent worker that can be used to query the index.
    """
-    index = get_index()
+    # Add query tool if index exists
+    index_config = IndexConfig(**(params or {}))
+    index = get_index(index_config)
    if index is None:
        return None
    top_k = int(os.getenv("TOP_K", 0))
@@ -31,13 +33,13 @@ def _create_query_engine_tool() -> QueryEngineTool:
    )


-def _get_research_tools() -> QueryEngineTool:
+def _get_research_tools(**kwargs) -> QueryEngineTool:
    """
    Researcher take responsibility for retrieving information.
    Try init wikipedia or duckduckgo tool if available.
    """
    tools = []
-    query_engine_tool = _create_query_engine_tool()
+    query_engine_tool = _create_query_engine_tool(**kwargs)
    if query_engine_tool is not None:
        tools.append(query_engine_tool)
    researcher_tool_names = ["duckduckgo", "wikipedia.WikipediaToolSpec"]
@@ -48,16 +50,17 @@ def _get_research_tools() -> QueryEngineTool:
    return tools


-def create_researcher(chat_history: List[ChatMessage]):
+def create_researcher(chat_history: List[ChatMessage], **kwargs):
    """
    Researcher is an agent that take responsibility for using tools to complete a given task.
    """
-    tools = _get_research_tools()
+    tools = _get_research_tools(**kwargs)
    return FunctionCallingAgent(
        name="researcher",
        tools=tools,
        description="expert in retrieving any unknown content or searching for images from the internet",
-        system_prompt=dedent("""
+        system_prompt=dedent(
+            """
            You are a researcher agent. You are given a research task.
            
            If the conversation already includes the information and there is no new request for additional information from the user, you should return the appropriate content to the writer.
@@ -77,6 +80,7 @@ def create_researcher(chat_history: List[ChatMessage]):

            If you use the tools but don't find any related information, please return "I didn't find any new information for {the topic}." along with the content you found. Don't try to make up information yourself.
            If the request doesn't need any new information because it was in the conversation history, please return "The task doesn't need any new information. Please reuse the existing content in the conversation history."
-        """),
+        """
+        ),
        chat_history=chat_history,
    )
@@ -17,9 +17,10 @@ from llama_index.core.workflow import (
 )


-def create_workflow(chat_history: Optional[List[ChatMessage]] = None):
+def create_workflow(chat_history: Optional[List[ChatMessage]] = None, **kwargs):
    researcher = create_researcher(
        chat_history=chat_history,
+        **kwargs,
    )
    publisher = create_publisher(
        chat_history=chat_history,
@@ -127,7 +128,8 @@ class BlogPostWorkflow(Workflow):
        self, input: str, chat_history: List[ChatMessage]
    ) -> str:
        prompt_template = PromptTemplate(
-            dedent("""
+            dedent(
+                """
                You are an expert in decision-making, helping people write and publish blog posts.
                If the user is asking for a file or to publish content, respond with 'publish'.
                If the user requests to write or update a blog post, respond with 'not_publish'.
@@ -140,7 +142,8 @@ class BlogPostWorkflow(Workflow):

                Given the chat history and the new user request, decide whether to publish based on existing information.
                Decision (respond with either 'not_publish' or 'publish'):
-            """)
+            """
+            )
        )

        chat_history_str = "\n".join(
@@ -1,13 +1,13 @@
 import { StopEvent } from "@llamaindex/core/workflow";
 import { Message, streamToResponse } from "ai";
 import { Request, Response } from "express";
-import { ChatMessage, ChatResponseChunk } from "llamaindex";
+import { ChatResponseChunk } from "llamaindex";
 import { createWorkflow } from "./workflow/factory";
 import { toDataStream, workflowEventsToStreamData } from "./workflow/stream";

 export const chat = async (req: Request, res: Response) => {
  try {
-    const { messages }: { messages: Message[] } = req.body;
+    const { messages, data }: { messages: Message[]; data?: any } = req.body;
    const userMessage = messages.pop();
    if (!messages || !userMessage || userMessage.role !== "user") {
      return res.status(400).json({
@@ -16,8 +16,7 @@ export const chat = async (req: Request, res: Response) => {
      });
    }

-    const chatHistory = messages as ChatMessage[];
-    const agent = createWorkflow(chatHistory);
+    const agent = createWorkflow(messages, data);
    const result = agent.run<AsyncGenerator<ChatResponseChunk>>(
      userMessage.content,
    ) as unknown as Promise<StopEvent<AsyncGenerator<ChatResponseChunk>>>;
@@ -1,7 +1,7 @@
 import { initObservability } from "@/app/observability";
 import { StopEvent } from "@llamaindex/core/workflow";
 import { Message, StreamingTextResponse } from "ai";
-import { ChatMessage, ChatResponseChunk } from "llamaindex";
+import { ChatResponseChunk } from "llamaindex";
 import { NextRequest, NextResponse } from "next/server";
 import { initSettings } from "./engine/settings";
 import { createWorkflow } from "./workflow/factory";
@@ -16,7 +16,7 @@ export const dynamic = "force-dynamic";
 export async function POST(request: NextRequest) {
  try {
    const body = await request.json();
-    const { messages }: { messages: Message[] } = body;
+    const { messages, data }: { messages: Message[]; data?: any } = body;
    const userMessage = messages.pop();
    if (!messages || !userMessage || userMessage.role !== "user") {
      return NextResponse.json(
@@ -28,8 +28,7 @@ export async function POST(request: NextRequest) {
      );
    }

-    const chatHistory = messages as ChatMessage[];
-    const agent = createWorkflow(chatHistory);
+    const agent = createWorkflow(messages, data);
    // TODO: fix type in agent.run in LITS
    const result = agent.run<AsyncGenerator<ChatResponseChunk>>(
      userMessage.content,
@@ -1,14 +1,19 @@
 import { ChatMessage } from "llamaindex";
 import { FunctionCallingAgent } from "./single-agent";
-import { lookupTools } from "./tools";
+import { getQueryEngineTool, lookupTools } from "./tools";

-export const createResearcher = async (chatHistory: ChatMessage[]) => {
-  const tools = await lookupTools([
-    "query_index",
-    "wikipedia_tool",
-    "duckduckgo_search",
-    "image_generator",
-  ]);
+export const createResearcher = async (
+  chatHistory: ChatMessage[],
+  params?: any,
+) => {
+  const queryEngineTool = await getQueryEngineTool(params);
+  const tools = (
+    await lookupTools([
+      "wikipedia_tool",
+      "duckduckgo_search",
+      "image_generator",
+    ])
+  ).concat(queryEngineTool ? [queryEngineTool] : []);

  return new FunctionCallingAgent({
    name: "researcher",
@@ -5,7 +5,9 @@ import {
  Workflow,
  WorkflowEvent,
 } from "@llamaindex/core/workflow";
+import { Message } from "ai";
 import { ChatMessage, ChatResponseChunk, Settings } from "llamaindex";
+import { getAnnotations } from "../llamaindex/streaming/annotations";
 import {
  createPublisher,
  createResearcher,
@@ -25,19 +27,15 @@ class WriteEvent extends WorkflowEvent<{
 class ReviewEvent extends WorkflowEvent<{ input: string }> {}
 class PublishEvent extends WorkflowEvent<{ input: string }> {}

-const prepareChatHistory = (chatHistory: ChatMessage[]) => {
+const prepareChatHistory = (chatHistory: Message[]): ChatMessage[] => {
  // By default, the chat history only contains the assistant and user messages
  // all the agents messages are stored in annotation data which is not visible to the LLM

  const MAX_AGENT_MESSAGES = 10;
-
-  // Construct a new agent message from agent messages
-  // Get annotations from assistant messages
-  const agentAnnotations = chatHistory
-    .filter((msg) => msg.role === "assistant")
-    .flatMap((msg) => msg.annotations || [])
-    .filter((annotation) => annotation.type === "agent")
-    .slice(-MAX_AGENT_MESSAGES);
+  const agentAnnotations = getAnnotations<{ agent: string; text: string }>(
+    chatHistory,
+    { role: "assistant", type: "agent" },
+  ).slice(-MAX_AGENT_MESSAGES);

  const agentMessages = agentAnnotations
    .map(
@@ -59,13 +57,13 @@ const prepareChatHistory = (chatHistory: ChatMessage[]) => {
      ...chatHistory.slice(0, -1),
      agentMessage,
      chatHistory.slice(-1)[0],
-    ];
+    ] as ChatMessage[];
  }
-  return chatHistory;
+  return chatHistory as ChatMessage[];
 };

-export const createWorkflow = (chatHistory: ChatMessage[]) => {
-  const chatHistoryWithAgentMessages = prepareChatHistory(chatHistory);
+export const createWorkflow = (messages: Message[], params?: any) => {
+  const chatHistoryWithAgentMessages = prepareChatHistory(messages);
  const runAgent = async (
    context: Context,
    agent: Workflow,
@@ -123,7 +121,10 @@ Decision (respond with either 'not_publish' or 'publish'):`;
  };

  const research = async (context: Context, ev: ResearchEvent) => {
-    const researcher = await createResearcher(chatHistoryWithAgentMessages);
+    const researcher = await createResearcher(
+      chatHistoryWithAgentMessages,
+      params,
+    );
    const researchRes = await runAgent(context, researcher, {
      message: ev.data.input,
    });
@@ -143,7 +143,7 @@ export class FunctionCallingAgent extends Workflow {
        fullResponse = chunk;
      }

-      if (fullResponse) {
+      if (fullResponse?.options && Object.keys(fullResponse.options).length) {
        memory.put({
          role: "assistant",
          content: "",
@@ -4,8 +4,10 @@ import path from "path";
 import { getDataSource } from "../engine";
 import { createTools } from "../engine/tools/index";

-const getQueryEngineTool = async (): Promise<QueryEngineTool | null> => {
-  const index = await getDataSource();
+export const getQueryEngineTool = async (
+  params?: any,
+): Promise<QueryEngineTool | null> => {
+  const index = await getDataSource(params);
  if (!index) {
    return null;
  }
@@ -1,7 +1,11 @@
-from llama_index.embeddings.openai import OpenAIEmbedding
-from llama_index.core.settings import Settings
-from typing import Dict
+import logging
 import os
+from typing import Dict
+
+from llama_index.core.settings import Settings
+from llama_index.embeddings.openai import OpenAIEmbedding
+
+logger = logging.getLogger(__name__)

 DEFAULT_MODEL = "gpt-3.5-turbo"
 DEFAULT_EMBEDDING_MODEL = "text-embedding-3-large"
@@ -50,7 +54,11 @@ def embedding_config_from_env() -> Dict:


 def init_llmhub():
-    from llama_index.llms.openai_like import OpenAILike
+    try:
+        from llama_index.llms.openai_like import OpenAILike
+    except ImportError:
+        logger.error("Failed to import OpenAILike. Make sure llama_index is installed.")
+        raise

    llm_configs = llm_config_from_env()
    embedding_configs = embedding_config_from_env()
@@ -33,8 +33,13 @@ def init_settings():


 def init_ollama():
-    from llama_index.embeddings.ollama import OllamaEmbedding
-    from llama_index.llms.ollama.base import DEFAULT_REQUEST_TIMEOUT, Ollama
+    try:
+        from llama_index.embeddings.ollama import OllamaEmbedding
+        from llama_index.llms.ollama.base import DEFAULT_REQUEST_TIMEOUT, Ollama
+    except ImportError:
+        raise ImportError(
+            "Ollama support is not installed. Please install it with `poetry add llama-index-llms-ollama` and `poetry add llama-index-embeddings-ollama`"
+        )

    base_url = os.getenv("OLLAMA_BASE_URL") or "http://127.0.0.1:11434"
    request_timeout = float(
@@ -55,25 +60,29 @@ def init_openai():
    from llama_index.llms.openai import OpenAI

    max_tokens = os.getenv("LLM_MAX_TOKENS")
-    config = {
-        "model": os.getenv("MODEL"),
-        "temperature": float(os.getenv("LLM_TEMPERATURE", DEFAULT_TEMPERATURE)),
-        "max_tokens": int(max_tokens) if max_tokens is not None else None,
-    }
-    Settings.llm = OpenAI(**config)
+    Settings.llm = OpenAI(
+        model=os.getenv("MODEL", "gpt-4o-mini"),
+        temperature=float(os.getenv("LLM_TEMPERATURE", DEFAULT_TEMPERATURE)),
+        max_tokens=int(max_tokens) if max_tokens is not None else None,
+    )

    dimensions = os.getenv("EMBEDDING_DIM")
-    config = {
-        "model": os.getenv("EMBEDDING_MODEL"),
-        "dimensions": int(dimensions) if dimensions is not None else None,
-    }
-    Settings.embed_model = OpenAIEmbedding(**config)
+    Settings.embed_model = OpenAIEmbedding(
+        model=os.getenv("EMBEDDING_MODEL", "text-embedding-3-small"),
+        dimensions=int(dimensions) if dimensions is not None else None,
+    )


 def init_azure_openai():
    from llama_index.core.constants import DEFAULT_TEMPERATURE
-    from llama_index.embeddings.azure_openai import AzureOpenAIEmbedding
-    from llama_index.llms.azure_openai import AzureOpenAI
+
+    try:
+        from llama_index.embeddings.azure_openai import AzureOpenAIEmbedding
+        from llama_index.llms.azure_openai import AzureOpenAI
+    except ImportError:
+        raise ImportError(
+            "Azure OpenAI support is not installed. Please install it with `poetry add llama-index-llms-azure-openai` and `poetry add llama-index-embeddings-azure-openai`"
+        )

    llm_deployment = os.environ["AZURE_OPENAI_LLM_DEPLOYMENT"]
    embedding_deployment = os.environ["AZURE_OPENAI_EMBEDDING_DEPLOYMENT"]
@@ -105,26 +114,37 @@ def init_azure_openai():


 def init_fastembed():
-    """
-    Use Qdrant Fastembed as the local embedding provider.
-    """
-    from llama_index.embeddings.fastembed import FastEmbedEmbedding
+    try:
+        from llama_index.embeddings.fastembed import FastEmbedEmbedding
+    except ImportError:
+        raise ImportError(
+            "FastEmbed support is not installed. Please install it with `poetry add llama-index-embeddings-fastembed`"
+        )

    embed_model_map: Dict[str, str] = {
        # Small and multilingual
        "all-MiniLM-L6-v2": "sentence-transformers/all-MiniLM-L6-v2",
        # Large and multilingual
-        "paraphrase-multilingual-mpnet-base-v2": "sentence-transformers/paraphrase-multilingual-mpnet-base-v2",  # noqa: E501
+        "paraphrase-multilingual-mpnet-base-v2": "sentence-transformers/paraphrase-multilingual-mpnet-base-v2",
    }

+    embedding_model = os.getenv("EMBEDDING_MODEL")
+    if embedding_model is None:
+        raise ValueError("EMBEDDING_MODEL environment variable is not set")
+
    # This will download the model automatically if it is not already downloaded
    Settings.embed_model = FastEmbedEmbedding(
-        model_name=embed_model_map[os.getenv("EMBEDDING_MODEL")]
+        model_name=embed_model_map[embedding_model]
    )


 def init_groq():
-    from llama_index.llms.groq import Groq
+    try:
+        from llama_index.llms.groq import Groq
+    except ImportError:
+        raise ImportError(
+            "Groq support is not installed. Please install it with `poetry add llama-index-llms-groq`"
+        )

    Settings.llm = Groq(model=os.getenv("MODEL"))
    # Groq does not provide embeddings, so we use FastEmbed instead
@@ -132,7 +152,12 @@ def init_groq():


 def init_anthropic():
-    from llama_index.llms.anthropic import Anthropic
+    try:
+        from llama_index.llms.anthropic import Anthropic
+    except ImportError:
+        raise ImportError(
+            "Anthropic support is not installed. Please install it with `poetry add llama-index-llms-anthropic`"
+        )

    model_map: Dict[str, str] = {
        "claude-3-opus": "claude-3-opus-20240229",
@@ -148,8 +173,13 @@ def init_anthropic():


 def init_gemini():
-    from llama_index.embeddings.gemini import GeminiEmbedding
-    from llama_index.llms.gemini import Gemini
+    try:
+        from llama_index.embeddings.gemini import GeminiEmbedding
+        from llama_index.llms.gemini import Gemini
+    except ImportError:
+        raise ImportError(
+            "Gemini support is not installed. Please install it with `poetry add llama-index-llms-gemini` and `poetry add llama-index-embeddings-gemini`"
+        )

    model_name = f"models/{os.getenv('MODEL')}"
    embed_model_name = f"models/{os.getenv('EMBEDDING_MODEL')}"
@@ -15,6 +15,6 @@ def get_vector_store():
        token=token,
        api_endpoint=endpoint,
        collection_name=collection,
-        embedding_dimension=int(os.getenv("EMBEDDING_DIM")),
+        embedding_dimension=int(os.getenv("EMBEDDING_DIM", 768)),
    )
    return store
@@ -1,4 +1,5 @@
 import os
+
 from llama_index.vector_stores.chroma import ChromaVectorStore


@@ -18,7 +19,7 @@ def get_vector_store():
            )
        store = ChromaVectorStore.from_params(
            host=os.getenv("CHROMA_HOST"),
-            port=int(os.getenv("CHROMA_PORT")),
+            port=os.getenv("CHROMA_PORT", "8001"),
            collection_name=collection_name,
        )
    return store
@@ -1,22 +1,66 @@
 # flake8: noqa: E402
-from dotenv import load_dotenv
+import os

-from app.engine.index import get_index
+from dotenv import load_dotenv

 load_dotenv()

 import logging
-from llama_index.core.readers import SimpleDirectoryReader
+
+from app.engine.index import get_client, get_index
 from app.engine.service import LLamaCloudFileService
+from app.settings import init_settings
+from llama_cloud import PipelineType
+from llama_index.core.readers import SimpleDirectoryReader
+from llama_index.core.settings import Settings

 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger()


+def ensure_index(index):
+    project_id = index._get_project_id()
+    client = get_client()
+    pipelines = client.pipelines.search_pipelines(
+        project_id=project_id,
+        pipeline_name=index.name,
+        pipeline_type=PipelineType.MANAGED.value,
+    )
+    if len(pipelines) == 0:
+        from llama_index.embeddings.openai import OpenAIEmbedding
+
+        if not isinstance(Settings.embed_model, OpenAIEmbedding):
+            raise ValueError(
+                "Creating a new pipeline with a non-OpenAI embedding model is not supported."
+            )
+        client.pipelines.upsert_pipeline(
+            project_id=project_id,
+            request={
+                "name": index.name,
+                "embedding_config": {
+                    "type": "OPENAI_EMBEDDING",
+                    "component": {
+                        "api_key": os.getenv("OPENAI_API_KEY"),  # editable
+                        "model_name": os.getenv("EMBEDDING_MODEL"),
+                    },
+                },
+                "transform_config": {
+                    "mode": "auto",
+                    "config": {
+                        "chunk_size": Settings.chunk_size,  # editable
+                        "chunk_overlap": Settings.chunk_overlap,  # editable
+                    },
+                },
+            },
+        )
+
+
 def generate_datasource():
+    init_settings()
    logger.info("Generate index for the provided data")

    index = get_index()
+    ensure_index(index)
    project_id = index._get_project_id()
    pipeline_id = index._get_pipeline_id()

@@ -34,13 +78,7 @@ def generate_datasource():
                f"Adding file {input_file} to pipeline {index.name} in project {index.project_name}"
            )
            LLamaCloudFileService.add_file_to_pipeline(
-                project_id,
-                pipeline_id,
-                f,
-                custom_metadata={
-                    # Set private=false to mark the document as public (required for filtering)
-                    "private": "false",
-                },
+                project_id, pipeline_id, f, custom_metadata={}
            )

    logger.info("Finished generating the index")
@@ -7,7 +7,7 @@ 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
+from pydantic import BaseModel, Field, field_validator

 logger = logging.getLogger("uvicorn")

@@ -15,31 +15,39 @@ logger = logging.getLogger("uvicorn")
 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"),
    )

+    def __init__(self, **kwargs):
+        if "api_key" not in kwargs:
+            kwargs["api_key"] = os.getenv("LLAMA_CLOUD_API_KEY")
+        if "base_url" not in kwargs:
+            kwargs["base_url"] = os.getenv("LLAMA_CLOUD_BASE_URL")
+        if "organization_id" not in kwargs:
+            kwargs["organization_id"] = os.getenv("LLAMA_CLOUD_ORGANIZATION_ID")
+        if "pipeline" not in kwargs:
+            kwargs["pipeline"] = os.getenv("LLAMA_CLOUD_INDEX_NAME")
+        if "project" not in kwargs:
+            kwargs["project"] = os.getenv("LLAMA_CLOUD_PROJECT_NAME")
+        super().__init__(**kwargs)
+
    # Validate and throw error if the env variables are not set before starting the app
-    @validator("pipeline", "project", "api_key", pre=True, always=True)
+    @field_validator("pipeline", "project", "api_key", mode="before")
    @classmethod
-    def validate_env_vars(cls, value):
+    def validate_fields(cls, value):
        if value is None:
            raise ValueError(
                "Please set LLAMA_CLOUD_INDEX_NAME, LLAMA_CLOUD_PROJECT_NAME and LLAMA_CLOUD_API_KEY"
@@ -56,7 +64,7 @@ class LlamaCloudConfig(BaseModel):

 class IndexConfig(BaseModel):
    llama_cloud_pipeline_config: LlamaCloudConfig = Field(
-        default=LlamaCloudConfig(),
+        default_factory=lambda: LlamaCloudConfig(),
        alias="llamaCloudPipeline",
    )
    callback_manager: Optional[CallbackManager] = Field(
@@ -5,7 +5,7 @@ def generate_filters(doc_ids):
    """
    Generate public/private document filters based on the doc_ids and the vector store.
    """
-    # Using "is_empty" filter to include the documents don't have the "private" key because they're uploaded in LlamaCloud UI
+    # public documents (ingested by "poetry run generate" or in the LlamaCloud UI) don't have the "private" field
    public_doc_filter = MetadataFilter(
        key="private",
        value=None,
@@ -1,4 +1,5 @@
 import os
+
 from llama_index.vector_stores.milvus import MilvusVectorStore


@@ -15,6 +16,6 @@ def get_vector_store():
        user=os.getenv("MILVUS_USERNAME"),
        password=os.getenv("MILVUS_PASSWORD"),
        collection_name=collection,
-        dim=int(os.getenv("EMBEDDING_DIM")),
+        dim=int(os.getenv("EMBEDDING_DIM", 768)),
    )
    return store
@@ -3,7 +3,7 @@ import os
 from datetime import timedelta
 from typing import Optional

-from cachetools import TTLCache, cached
+from cachetools import TTLCache, cached  # type: ignore
 from llama_index.core.callbacks import CallbackManager
 from llama_index.core.indices import load_index_from_storage
 from llama_index.core.storage import StorageContext
@@ -25,6 +25,8 @@ async function* walk(dir: string): AsyncGenerator<string> {

 async function loadAndIndex() {
  const index = await getDataSource();
+  // ensure the index is available or create a new one
+  await index.ensureIndex();
  const projectId = await index.getProjectId();
  const pipelineId = await index.getPipelineId();

@@ -32,10 +34,23 @@ async function loadAndIndex() {
  for await (const filePath of walk(DATA_DIR)) {
    const buffer = await fs.readFile(filePath);
    const filename = path.basename(filePath);
-    const file = new File([buffer], filename);
-    await LLamaCloudFileService.addFileToPipeline(projectId, pipelineId, file, {
-      private: "false",
-    });
+    try {
+      await LLamaCloudFileService.addFileToPipeline(
+        projectId,
+        pipelineId,
+        new File([buffer], filename),
+      );
+    } catch (error) {
+      if (
+        error instanceof ReferenceError &&
+        error.message.includes("File is not defined")
+      ) {
+        throw new Error(
+          "File class is not supported in the current Node.js version. Please use Node.js 20 or higher.",
+        );
+      }
+      throw error;
+    }
  }

  console.log(`Successfully uploaded documents to LlamaCloud!`);
@@ -1,14 +1,17 @@
-import { MetadataFilter, MetadataFilters } from "llamaindex";
+import { CloudRetrieveParams, MetadataFilter } from "llamaindex";

-export function generateFilters(documentIds: string[]): MetadataFilters {
-  // public documents don't have the "private" field or it's set to "false"
+export function generateFilters(documentIds: string[]) {
+  // public documents (ingested by "npm run generate" or in the LlamaCloud UI) don't have the "private" field
  const publicDocumentsFilter: MetadataFilter = {
    key: "private",
    operator: "is_empty",
  };

  // if no documentIds are provided, only retrieve information from public documents
-  if (!documentIds.length) return { filters: [publicDocumentsFilter] };
+  if (!documentIds.length)
+    return {
+      filters: [publicDocumentsFilter],
+    } as CloudRetrieveParams["filters"];

  const privateDocumentsFilter: MetadataFilter = {
    key: "file_id", // Note: LLamaCloud uses "file_id" to reference private document ids as "doc_id" is a restricted field in LlamaCloud
@@ -20,5 +23,5 @@ export function generateFilters(documentIds: string[]): MetadataFilters {
  return {
    filters: [publicDocumentsFilter, privateDocumentsFilter],
    condition: "or",
-  };
+  } as CloudRetrieveParams["filters"];
 }
@@ -18,7 +18,9 @@ async function loadAndIndex() {

  // create postgres vector store
  const vectorStore = new PGVectorStore({
-    connectionString: process.env.PG_CONNECTION_STRING,
+    clientConfig: {
+      connectionString: process.env.PG_CONNECTION_STRING,
+    },
    schemaName: PGVECTOR_SCHEMA,
    tableName: PGVECTOR_TABLE,
  });
@@ -9,7 +9,9 @@ import {
 export async function getDataSource(params?: any) {
  checkRequiredEnvVars();
  const pgvs = new PGVectorStore({
-    connectionString: process.env.PG_CONNECTION_STRING,
+    clientConfig: {
+      connectionString: process.env.PG_CONNECTION_STRING,
+    },
    schemaName: PGVECTOR_SCHEMA,
    tableName: PGVECTOR_TABLE,
  });
@@ -6,7 +6,7 @@ load_dotenv()
 import logging
 import os

-from llama_index.core.ingestion import IngestionPipeline
+from llama_index.core.ingestion import DocstoreStrategy, IngestionPipeline
 from llama_index.core.node_parser import SentenceSplitter
 from llama_index.core.settings import Settings
 from llama_index.core.storage import StorageContext
@@ -41,7 +41,7 @@ def run_pipeline(docstore, vector_store, documents):
            Settings.embed_model,
        ],
        docstore=docstore,
-        docstore_strategy="upserts_and_delete",
+        docstore_strategy=DocstoreStrategy.UPSERTS_AND_DELETE,  # type: ignore
        vector_store=vector_store,
    )

@@ -16,7 +16,7 @@ class IndexConfig(BaseModel):
    )


-def get_index(config: IndexConfig = None):
+def get_index(config: Optional[IndexConfig] = None) -> VectorStoreIndex:
    if config is None:
        config = IndexConfig()
    logger.info("Connecting vector store...")
@@ -21,7 +21,7 @@
    "dotenv": "^16.3.1",
    "duck-duck-scrape": "^2.2.5",
    "express": "^4.18.2",
-    "llamaindex": "0.6.2",
+    "llamaindex": "0.6.19",
    "pdf2json": "3.0.5",
    "ajv": "^8.12.0",
    "@e2b/code-interpreter": "0.0.9-beta.3",
@@ -13,7 +13,7 @@ from app.api.routers.models import (
    SourceNodes,
 )
 from app.api.routers.vercel_response import VercelStreamResponse
-from app.engine import get_chat_engine
+from app.engine.engine import get_chat_engine
 from app.engine.query_filter import generate_filters

 chat_router = r = APIRouter()
@@ -1,11 +1,10 @@
 import logging
 import os

-from fastapi import APIRouter
+from fastapi import APIRouter, HTTPException

 from app.api.routers.models import ChatConfig

-
 config_router = r = APIRouter()

 logger = logging.getLogger("uvicorn")
@@ -23,10 +22,14 @@ async def chat_config() -> ChatConfig:
 try:
    from app.engine.service import LLamaCloudFileService

-    logger.info("LlamaCloud is configured. Adding /config/llamacloud route.")
+    print("LlamaCloud is configured. Adding /config/llamacloud route.")

    @r.get("/llamacloud")
    async def chat_llama_cloud_config():
+        if not os.getenv("LLAMA_CLOUD_API_KEY"):
+            raise HTTPException(
+                status_code=500, detail="LlamaCloud API KEY is not configured"
+            )
        projects = LLamaCloudFileService.get_all_projects_with_pipelines()
        pipeline = os.getenv("LLAMA_CLOUD_INDEX_NAME")
        project = os.getenv("LLAMA_CLOUD_PROJECT_NAME")
@@ -42,7 +45,5 @@ try:
        }

 except ImportError:
-    logger.debug(
-        "LlamaCloud is not configured. Skipping adding /config/llamacloud route."
-    )
+    print("LlamaCloud is not configured. Skipping adding /config/llamacloud route.")
    pass
@@ -1,13 +1,13 @@
-import json
 import asyncio
+import json
 import logging
-from typing import AsyncGenerator, Dict, Any, List, Optional
+from typing import Any, AsyncGenerator, Dict, List, Optional
+
 from llama_index.core.callbacks.base import BaseCallbackHandler
 from llama_index.core.callbacks.schema import CBEventType
 from llama_index.core.tools.types import ToolOutput
 from pydantic import BaseModel

-
 logger = logging.getLogger(__name__)


@@ -31,15 +31,20 @@ class CallbackEvent(BaseModel):
            return None

    def get_tool_message(self) -> dict | None:
+        if self.payload is None:
+            return None
        func_call_args = self.payload.get("function_call")
        if func_call_args is not None and "tool" in self.payload:
            tool = self.payload.get("tool")
+            if tool is None:
+                return None
            return {
                "type": "events",
                "data": {
                    "title": f"Calling tool: {tool.name} with inputs: {func_call_args}",
                },
            }
+        return None

    def _is_output_serializable(self, output: Any) -> bool:
        try:
@@ -49,6 +54,8 @@ class CallbackEvent(BaseModel):
            return False

    def get_agent_tool_response(self) -> dict | None:
+        if self.payload is None:
+            return None
        response = self.payload.get("response")
        if response is not None:
            sources = response.sources
@@ -74,6 +81,7 @@ class CallbackEvent(BaseModel):
                            },
                        },
                    }
+        return None

    def to_response(self):
        try:
@@ -114,11 +122,13 @@ class EventCallbackHandler(BaseCallbackHandler):
        event_type: CBEventType,
        payload: Optional[Dict[str, Any]] = None,
        event_id: str = "",
+        parent_id: str = "",
        **kwargs: Any,
    ) -> str:
        event = CallbackEvent(event_id=event_id, event_type=event_type, payload=payload)
        if event.to_response() is not None:
            self._aqueue.put_nowait(event)
+        return event_id

    def on_event_end(
        self,
@@ -1,6 +1,6 @@
 import logging
 import os
-from typing import Any, Dict, List, Literal, Optional
+from typing import Any, Dict, List, Literal, Optional, Union

 from llama_index.core.llms import ChatMessage, MessageRole
 from llama_index.core.schema import NodeWithScore
@@ -62,15 +62,23 @@ class ArtifactAnnotation(BaseModel):

 class Annotation(BaseModel):
    type: str
-    data: AnnotationFileData | List[str] | AgentAnnotation | ArtifactAnnotation
+    data: Union[AnnotationFileData, List[str], AgentAnnotation, ArtifactAnnotation]

-    def to_content(self) -> str | None:
+    def to_content(self) -> Optional[str]:
        if self.type == "document_file":
-            # We only support generating context content for CSV files for now
-            csv_files = [file for file in self.data.files if file.filetype == "csv"]
-            if len(csv_files) > 0:
-                return "Use data from following CSV raw content\n" + "\n".join(
-                    [f"```csv\n{csv_file.content.value}\n```" for csv_file in csv_files]
+            if isinstance(self.data, AnnotationFileData):
+                # We only support generating context content for CSV files for now
+                csv_files = [file for file in self.data.files if file.filetype == "csv"]
+                if len(csv_files) > 0:
+                    return "Use data from following CSV raw content\n" + "\n".join(
+                        [
+                            f"```csv\n{csv_file.content.value}\n```"
+                            for csv_file in csv_files
+                        ]
+                    )
+            else:
+                logger.warning(
+                    f"Unexpected data type for document_file annotation: {type(self.data)}"
                )
        else:
            logger.warning(
@@ -213,6 +221,7 @@ class ChatData(BaseModel):
                for annotation in message.annotations:
                    if (
                        annotation.type == "document_file"
+                        and isinstance(annotation.data, AnnotationFileData)
                        and annotation.data.files is not None
                    ):
                        for fi in annotation.data.files:
@@ -242,7 +251,7 @@ class SourceNodes(BaseModel):
        )

    @classmethod
-    def get_url_from_metadata(cls, metadata: Dict[str, Any]) -> str:
+    def get_url_from_metadata(cls, metadata: Dict[str, Any]) -> Optional[str]:
        url_prefix = os.getenv("FILESERVER_URL_PREFIX")
        if not url_prefix:
            logger.warning(
@@ -1 +0,0 @@
-from .engine import get_chat_engine as get_chat_engine
@@ -6,15 +6,16 @@ load_dotenv()
 import logging
 import os

-from app.engine.loaders import get_documents
-from app.engine.vectordb import get_vector_store
-from app.settings import init_settings
-from llama_index.core.ingestion import IngestionPipeline
+from llama_index.core.ingestion import DocstoreStrategy, IngestionPipeline
 from llama_index.core.node_parser import SentenceSplitter
 from llama_index.core.settings import Settings
 from llama_index.core.storage import StorageContext
 from llama_index.core.storage.docstore import SimpleDocumentStore

+from app.engine.loaders import get_documents
+from app.engine.vectordb import get_vector_store
+from app.settings import init_settings
+
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger()

@@ -40,7 +41,7 @@ def run_pipeline(docstore, vector_store, documents):
            Settings.embed_model,
        ],
        docstore=docstore,
-        docstore_strategy="upserts_and_delete",
+        docstore_strategy=DocstoreStrategy.UPSERTS_AND_DELETE,  # type: ignore
        vector_store=vector_store,
    )

@@ -15,8 +15,25 @@ uvicorn = { extras = ["standard"], version = "^0.23.2" }
 python-dotenv = "^1.0.0"
 aiostream = "^0.5.2"
 cachetools = "^5.3.3"
-llama-index = "0.11.6"
+llama-index = "^0.11.17"
+
+[tool.poetry.group.dev.dependencies]
+mypy = "^1.8.0"

 [build-system]
 requires = ["poetry-core"]
 build-backend = "poetry.core.masonry.api"
+
+[tool.mypy]
+python_version = "3.11"
+plugins = "pydantic.mypy"
+exclude = [ "tests", "venv", ".venv", "output", "config" ]
+check_untyped_defs = true
+warn_unused_ignores = false
+show_error_codes = true
+namespace_packages = true
+ignore_missing_imports = true
+follow_imports = "silent"
+implicit_optional = true
+strict_optional = false
+disable_error_code = ["return-value", "import-untyped", "assignment"]
@@ -45,7 +45,7 @@ export function LlamaCloudSelector({
  setRequestData,
  onSelect,
  defaultPipeline,
-  shouldCheckValid = true,
+  shouldCheckValid = false,
 }: LlamaCloudSelectorProps) {
  const { backend } = useClientConfig();
  const [config, setConfig] = useState<LlamaCloudConfig>();
@@ -66,7 +66,16 @@ export function LlamaCloudSelector({
  useEffect(() => {
    if (process.env.NEXT_PUBLIC_USE_LLAMACLOUD === "true" && !config) {
      fetch(`${backend}/api/chat/config/llamacloud`)
-        .then((response) => response.json())
+        .then((response) => {
+          if (!response.ok) {
+            return response.json().then((errorData) => {
+              window.alert(
+                `Error: ${JSON.stringify(errorData) || "Unknown error occurred"}`,
+              );
+            });
+          }
+          return response.json();
+        })
        .then((data) => {
          const pipeline = defaultPipeline ?? data.pipeline; // defaultPipeline will override pipeline in .env
          setConfig({ ...data, pipeline });
@@ -95,7 +104,8 @@ export function LlamaCloudSelector({
      </div>
    );
  }
-  if (!isValid(config) && shouldCheckValid) {
+
+  if (shouldCheckValid && !isValid(config.projects, config.pipeline)) {
    return (
      <p className="text-red-500">
        Invalid LlamaCloud configuration. Check console logs.
@@ -107,7 +117,11 @@ export function LlamaCloudSelector({
  return (
    <Select
      onValueChange={handlePipelineSelect}
-      defaultValue={JSON.stringify(pipeline)}
+      defaultValue={
+        isValid(projects, pipeline, false)
+          ? JSON.stringify(pipeline)
+          : undefined
+      }
    >
      <SelectTrigger className="w-[200px]">
        <SelectValue placeholder="Select a pipeline" />
@@ -137,26 +151,33 @@ export function LlamaCloudSelector({
  );
 }

-function isValid(config: LlamaCloudConfig): boolean {
-  const { projects, pipeline } = config;
+function isValid(
+  projects: LLamaCloudProject[] | undefined,
+  pipeline: PipelineConfig | undefined,
+  logErrors: boolean = true,
+): boolean {
  if (!projects?.length) return false;
  if (!pipeline) return false;
  const matchedProject = projects.find(
    (project: LLamaCloudProject) => project.name === pipeline.project,
  );
  if (!matchedProject) {
-    console.error(
-      `LlamaCloud project ${pipeline.project} not found. Check LLAMA_CLOUD_PROJECT_NAME variable`,
-    );
+    if (logErrors) {
+      console.error(
+        `LlamaCloud project ${pipeline.project} not found. Check LLAMA_CLOUD_PROJECT_NAME variable`,
+      );
+    }
    return false;
  }
  const pipelineExists = matchedProject.pipelines.some(
    (p) => p.name === pipeline.pipeline,
  );
  if (!pipelineExists) {
-    console.error(
-      `LlamaCloud pipeline ${pipeline.pipeline} not found. Check LLAMA_CLOUD_INDEX_NAME variable`,
-    );
+    if (logErrors) {
+      console.error(
+        `LlamaCloud pipeline ${pipeline.pipeline} not found. Check LLAMA_CLOUD_INDEX_NAME variable`,
+      );
+    }
    return false;
  }
  return true;
@@ -27,7 +27,7 @@
    "duck-duck-scrape": "^2.2.5",
    "formdata-node": "^6.0.3",
    "got": "^14.4.1",
-    "llamaindex": "0.6.2",
+    "llamaindex": "0.6.19",
    "lucide-react": "^0.294.0",
    "next": "^14.2.4",
    "react": "^18.2.0",
@@ -18,7 +18,7 @@
    "create-app.ts",
    "index.ts",
    "./helpers",
-    "questions.ts",
+    "./questions",
    "package.json",
    "types/**/*"
  ],
Author	SHA1	Message	Date
Marcus Schiesser	f886e3132b	review fixes	2024-10-14 12:34:06 +07:00
Marcus Schiesser	4015c157b6	fix: use cli framework in CLI	2024-10-14 12:27:53 +07:00
Marcus Schiesser	8539c7ddec	Merge branch 'main' into ms/simplify	2024-10-14 12:08:45 +07:00
Marcus Schiesser	24fcde52f2	add data sources for simple templates	2024-10-14 12:05:49 +07:00
Marcus Schiesser	40dc3a48b4	fix: llamacloud is not a datasource	2024-10-14 12:03:03 +07:00
Marcus Schiesser	ff2d1a1e87	ask for llamacloud key in simple mode	2024-10-14 11:22:45 +07:00
Marcus Schiesser	40b9c01b0d	fix: don't override cli defaults in pro mode	2024-10-14 10:51:54 +07:00
Marcus Schiesser	e3b54e60fd	factor out CI defaults	2024-10-11 16:58:29 +07:00
Marcus Schiesser	4e600778e8	remove askExamples	2024-10-11 16:39:36 +07:00
Marcus Schiesser	56088bd3d7	remove preferences	2024-10-11 16:37:50 +07:00
Marcus Schiesser	f00cb5aa03	use pro for CI	2024-10-11 16:33:16 +07:00
Marcus Schiesser	3bb5af8d5f	add postinstall question to simple questions	2024-10-11 16:19:49 +07:00
Marcus Schiesser	b7af20762c	chore: order types	2024-10-11 15:43:05 +07:00
Marcus Schiesser	6e51bcd81a	feat: add simple questions	2024-10-11 15:30:05 +07:00
Marcus Schiesser	5ba4b3bcf1	chore: update prompts	2024-10-11 11:56:19 +07:00
Marcus Schiesser	571a8c3a12	use new commander features	2024-10-11 11:44:14 +07:00
Marcus Schiesser	c89e85bf4e	chore: use latest commander	2024-10-11 11:25:45 +07:00
Thuc Pham	98a82b0b25	docs: chroma env variables (#367 )	2024-10-11 11:10:29 +07:00
github-actions[bot]	7db72b6f2e	Release 0.2.19 (#365 ) Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>	2024-10-10 18:41:25 +07:00
Thuc Pham	3d41488301	feat: use selected llamacloud for multiagent (#359 )	2024-10-10 18:37:55 +07:00
github-actions[bot]	1ee05eaf4b	Release 0.2.18 (#364 ) Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>	2024-10-10 18:03:43 +07:00
Huu Le	75e1f6104c	fix: TypeScript templates do not create a new LlamaCloud index or upload a file to an existing index. (#356 )	2024-10-10 17:58:12 +07:00
Huu Le	88220f1dd2	feat: add canceling workflow for multiagent (#361 )	2024-10-10 15:24:43 +07:00
github-actions[bot]	6304114ef5	Release 0.2.17 (#357 ) Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>	2024-10-09 16:31:50 +07:00
Marcus Schiesser	6335de1174	docs: changeset	2024-10-09 16:18:11 +07:00
Huu Le	b9184ff59a	fix: (FastAPI) Using LlamaCloud parameters does not use the configured value in the environment. (#358 )	2024-10-09 16:13:35 +07:00
Thuc Pham	cd3fcd0512	bump: use latest LITS (#343 )	2024-10-09 13:40:04 +07:00
github-actions[bot]	a47d778602	Release 0.2.16 (#349 ) Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>	2024-10-08 17:28:40 +07:00
Marcus Schiesser	7f4ac228ee	Don't need to run generate script for LlamaCloud (#352 )	2024-10-08 16:56:12 +07:00
Marcus Schiesser	5263bde8e7	feat: Use selected LlamaCloud index in multi-agent template (#350 )	2024-10-08 16:54:14 +07:00
Huu Le	4dee65b93d	add astral's uv tool to github action (#351 )	2024-10-08 16:19:20 +07:00
Huu Le	c60182a925	Add mypy checker (#346 )	2024-10-08 15:17:38 +07:00
Marcus Schiesser	0e78ba4603	fix: .env not loaded on poetry run generate (#348 ) -------- Co-authored-by: leehuwuj <leehuwuj@gmail.com>	2024-10-08 13:41:37 +07:00
				`@@ -1 +0,0 @@`
				`from .engine import get_chat_engine as get_chat_engine`