bump @llamaindex/server

fix llamacloud api and markdown issue
2026-07-19 15:03:35 -04:00 · 2025-04-02 17:20:32 +07:00 · 2025-04-02 16:51:21 +07:00
60 changed files with 297 additions and 3771 deletions
@@ -0,0 +1,5 @@
+---
+"create-llama": patch
+---
+
+fix: add trycatch for generating error
@@ -0,0 +1,5 @@
+---
+"create-llama": patch
+---
+
+bump: chat-ui and tailwind v4
@@ -67,8 +67,6 @@ jobs:
          LLAMA_CLOUD_API_KEY: ${{ secrets.LLAMA_CLOUD_API_KEY }}
          FRAMEWORK: ${{ matrix.frameworks }}
          DATASOURCE: ${{ matrix.datasources }}
-          PYTHONIOENCODING: utf-8
-          PYTHONLEGACYWINDOWSSTDIO: utf-8
        working-directory: .

      - uses: actions/upload-artifact@v4
@@ -88,7 +86,7 @@ jobs:
        node-version: [18, 20]
        python-version: ["3.11"]
        os: [macos-latest, windows-latest, ubuntu-22.04]
-        frameworks: ["nextjs"]
+        frameworks: ["nextjs", "express"]
        datasources: ["--no-files", "--example-file", "--llamacloud"]
    defaults:
      run:
@@ -1,28 +1,5 @@
 # create-llama

-## 0.5.2
-
-### Patch Changes
-
- c9f8f8d: Use custom component for deep research use case
-
-## 0.5.1
-
-### Patch Changes
-
- 08b3e07: Simplify the local index code.
-
-## 0.5.0
-
-### Minor Changes
-
- 54c9e2f: Simplified generated code using LlamaIndexServer
-
-### Patch Changes
-
- 0e4ecfa: fix: add trycatch for generating error
- ee69ce7: bump: chat-ui and tailwind v4
-
 ## 0.4.0

 ### Minor Changes
@@ -90,7 +90,7 @@ export async function createApp({
  // Install backend
  await installTemplate({ ...args, backend: true });

-  if (frontend && framework === "fastapi" && template !== "llamaindexserver") {
+  if (frontend && framework === "fastapi") {
    // install frontend
    const frontendRoot = path.join(root, ".frontend");
    await makeDir(frontendRoot);
@@ -110,7 +110,7 @@ export async function createApp({
    console.log();
  }

-  if (toolsRequireConfig(tools) && template !== "llamaindexserver") {
+  if (toolsRequireConfig(tools)) {
    const configFile =
      framework === "fastapi" ? "config/tools.yaml" : "config/tools.json";
    console.log(
@@ -16,17 +16,15 @@ const templateFramework: TemplateFramework = process.env.FRAMEWORK
 const dataSource: string = "--example-file";
 const templateUI: TemplateUI = "shadcn";
 const templatePostInstallAction: TemplatePostInstallAction = "runApp";
-const appType: AppType = "--frontend";
+const appType: AppType = templateFramework === "fastapi" ? "--frontend" : "";
 const userMessage = "Write a blog post about physical standards for letters";
-const templateUseCases = ["financial_report", "agentic_rag", "deep_research"];
+const templateUseCases = ["financial_report", "blog", "form_filling"];

 for (const useCase of templateUseCases) {
-  test.describe(`Test use case ${useCase} ${templateFramework} ${dataSource} ${templateUI} ${appType} ${templatePostInstallAction}`, async () => {
+  test.describe(`Test multiagent template ${useCase} ${templateFramework} ${dataSource} ${templateUI} ${appType} ${templatePostInstallAction}`, async () => {
    test.skip(
-      process.platform !== "linux" ||
-        process.env.DATASOURCE === "--no-files" ||
-        templateFramework === "express",
-      "The llamaindexserver template currently only works with nextjs, fastapi. We also only run on Linux to speed up tests.",
+      process.platform !== "linux" || process.env.DATASOURCE === "--no-files",
+      "The multiagent template currently only works with files. We also only run on Linux to speed up tests.",
    );
    let port: number;
    let cwd: string;
@@ -40,7 +38,7 @@ for (const useCase of templateUseCases) {
      cwd = await createTestDir();
      const result = await runCreateLlama({
        cwd,
-        templateType: "llamaindexserver",
+        templateType: "multiagent",
        templateFramework,
        dataSource,
        vectorDb,
@@ -74,9 +72,9 @@ for (const useCase of templateUseCases) {
      test.skip(
        templatePostInstallAction !== "runApp" ||
          useCase === "financial_report" ||
-          useCase === "deep_research" ||
+          useCase === "form_filling" ||
          templateFramework === "express",
-        "Skip chat tests for financial report and deep research.",
+        "Skip chat tests for financial report and form filling.",
      );
      await page.goto(`http://localhost:${port}`);
      await page.fill("form textarea", userMessage);
@@ -88,12 +86,6 @@ for (const useCase of templateUseCases) {
      await page.click("form button[type=submit]");

      const response = await responsePromise;
-      console.log(`Response status: ${response.status()}`);
-      const responseBody = await response
-        .text()
-        .catch((e) => `Error reading body: ${e}`);
-      console.log(`Response body: ${responseBody}`);
-
      expect(response.ok()).toBeTruthy();
    });

@@ -113,12 +113,7 @@ export async function runCreateLlama({
  if (observability) {
    commandArgs.push("--observability", observability);
  }
-  if (
-    (templateType === "multiagent" ||
-      templateType === "reflex" ||
-      templateType === "llamaindexserver") &&
-    useCase
-  ) {
+  if ((templateType === "multiagent" || templateType === "reflex") && useCase) {
    commandArgs.push("--use-case", useCase);
  }

@@ -44,7 +44,6 @@ const renderEnvVar = (envVars: EnvVar[]): string => {
 const getVectorDBEnvs = (
  vectorDb?: TemplateVectorDB,
  framework?: TemplateFramework,
-  template?: TemplateType,
 ): EnvVar[] => {
  if (!vectorDb || !framework) {
    return [];
@@ -169,7 +168,7 @@ const getVectorDBEnvs = (
          description:
            "The organization ID for the LlamaCloud project (uses default organization if not specified)",
        },
-        ...(framework === "nextjs" && template !== "llamaindexserver"
+        ...(framework === "nextjs"
          ? // activate index selector per default (not needed for non-NextJS backends as it's handled by createFrontendEnvFile)
            [
              {
@@ -224,15 +223,13 @@ Otherwise, use CHROMA_HOST and CHROMA_PORT config above`,
        },
      ];
    default:
-      return template !== "llamaindexserver"
-        ? [
-            {
-              name: "STORAGE_CACHE_DIR",
-              description: "The directory to store the local storage cache.",
-              value: ".cache",
-            },
-          ]
-        : [];
+      return [
+        {
+          name: "STORAGE_CACHE_DIR",
+          description: "The directory to store the local storage cache.",
+          value: ".cache",
+        },
+      ];
  }
 };

@@ -385,42 +382,38 @@ const getModelEnvs = (modelConfig: ModelConfig): EnvVar[] => {

 const getFrameworkEnvs = (
  framework: TemplateFramework,
-  template: TemplateType,
  port?: number,
 ): EnvVar[] => {
  const sPort = port?.toString() || "8000";
-  const result: EnvVar[] =
-    template !== "llamaindexserver"
-      ? [
-          {
-            name: "FILESERVER_URL_PREFIX",
-            description:
-              "FILESERVER_URL_PREFIX is the URL prefix of the server storing the images generated by the interpreter.",
-            value:
-              framework === "nextjs"
-                ? // FIXME: if we are using nextjs, port should be 3000
-                  "http://localhost:3000/api/files"
-                : `http://localhost:${sPort}/api/files`,
-          },
-        ]
-      : [];
+  const result: EnvVar[] = [
+    {
+      name: "FILESERVER_URL_PREFIX",
+      description:
+        "FILESERVER_URL_PREFIX is the URL prefix of the server storing the images generated by the interpreter.",
+      value:
+        framework === "nextjs"
+          ? // FIXME: if we are using nextjs, port should be 3000
+            "http://localhost:3000/api/files"
+          : `http://localhost:${sPort}/api/files`,
+    },
+  ];
  if (framework === "fastapi") {
    result.push(
      ...[
        {
          name: "APP_HOST",
-          description: "The address to start the FastAPI app.",
+          description: "The address to start the backend app.",
          value: "0.0.0.0",
        },
        {
          name: "APP_PORT",
-          description: "The port to start the FastAPI app.",
+          description: "The port to start the backend app.",
          value: sPort,
        },
      ],
    );
  }
-  if (framework === "nextjs" && template !== "llamaindexserver") {
+  if (framework === "nextjs") {
    result.push({
      name: "NEXT_PUBLIC_CHAT_API",
      description:
@@ -576,41 +569,25 @@ export const createBackendEnvFile = async (
    | "port"
    | "tools"
    | "observability"
-    | "useLlamaParse"
  >,
 ) => {
  // Init env values
  const envFileName = ".env";
  const envVars: EnvVar[] = [
-    ...(opts.useLlamaParse
-      ? [
-          {
-            name: "LLAMA_CLOUD_API_KEY",
-            description: `The Llama Cloud API key.`,
-            value: opts.llamaCloudKey,
-          },
-        ]
-      : []),
-    ...getVectorDBEnvs(opts.vectorDb, opts.framework, opts.template),
-    ...getToolEnvs(opts.tools),
-    ...getFrameworkEnvs(opts.framework, opts.template, opts.port),
+    {
+      name: "LLAMA_CLOUD_API_KEY",
+      description: `The Llama Cloud API key.`,
+      value: opts.llamaCloudKey,
+    },
    // Add environment variables of each component
-    ...(opts.template === "llamaindexserver"
-      ? [
-          {
-            name: "OPENAI_API_KEY",
-            description: "The OpenAI API key to use.",
-            value: opts.modelConfig.apiKey,
-          },
-        ]
-      : [
-          // don't use this stuff for llama-indexserver
-          ...getModelEnvs(opts.modelConfig),
-          ...getEngineEnvs(),
-          ...getTemplateEnvs(opts.template),
-          ...getObservabilityEnvs(opts.observability),
-          ...getSystemPromptEnv(opts.tools, opts.dataSources, opts.template),
-        ]),
+    ...getModelEnvs(opts.modelConfig),
+    ...getEngineEnvs(),
+    ...getVectorDBEnvs(opts.vectorDb, opts.framework),
+    ...getFrameworkEnvs(opts.framework, opts.port),
+    ...getToolEnvs(opts.tools),
+    ...getTemplateEnvs(opts.template),
+    ...getObservabilityEnvs(opts.observability),
+    ...getSystemPromptEnv(opts.tools, opts.dataSources, opts.template),
  ];
  // Render and write env file
  const content = renderEnvVar(envVars);
@@ -1,7 +1,7 @@
 import { callPackageManager } from "./install";

 import path from "path";
-import picocolors, { cyan } from "picocolors";
+import { cyan } from "picocolors";

 import fsExtra from "fs-extra";
 import { writeLoadersConfig } from "./datasources";
@@ -41,11 +41,7 @@ const checkForGenerateScript = (
    missingSettings.push("your LLAMA_CLOUD_API_KEY");
  }

-  if (
-    vectorDb !== undefined &&
-    vectorDb !== "none" &&
-    vectorDb !== "llamacloud"
-  ) {
+  if (vectorDb !== "none" && vectorDb !== "llamacloud") {
    missingSettings.push("your Vector DB environment variables");
  }

@@ -96,7 +92,7 @@ async function generateContextData(
    }

    const settingsMessage = `After setting ${missingSettings.join(" and ")}, run ${runGenerate} to generate the context data.`;
-    console.log(picocolors.yellow(`\n${settingsMessage}\n\n`));
+    console.log(`\n${settingsMessage}\n\n`);
  }
 }

@@ -170,17 +166,6 @@ export const installTemplate = async (

  if (props.framework === "fastapi") {
    await installPythonTemplate(props);
-  } else {
-    await installTSTemplate(props);
-  }
-
-  // write configurations
-  if (props.template !== "llamaindexserver") {
-    await writeToolsConfig(
-      props.root,
-      props.tools,
-      props.framework === "fastapi" ? ConfigFileType.YAML : ConfigFileType.JSON,
-    );
    if (props.vectorDb !== "llamacloud") {
      // write loaders configuration (currently Python only)
      // not needed for LlamaCloud as it has its own loaders
@@ -190,13 +175,26 @@ export const installTemplate = async (
        props.useLlamaParse,
      );
    }
+  } else {
+    await installTSTemplate(props);
  }

+  // write tools configuration
+  await writeToolsConfig(
+    props.root,
+    props.tools,
+    props.framework === "fastapi" ? ConfigFileType.YAML : ConfigFileType.JSON,
+  );
+
  if (props.backend) {
    // This is a backend, so we need to copy the test data and create the env file.

    // Copy the environment file to the target directory.
-    if (props.template !== "community" && props.template !== "llamapack") {
+    if (
+      props.template === "streaming" ||
+      props.template === "multiagent" ||
+      props.template === "reflex"
+    ) {
      await createBackendEnvFile(props.root, props);
    }

@@ -12,7 +12,6 @@ import {
  InstallTemplateArgs,
  ModelConfig,
  TemplateDataSource,
-  TemplateObservability,
  TemplateType,
  TemplateVectorDB,
 } from "./types";
@@ -30,7 +29,6 @@ const getAdditionalDependencies = (
  dataSources?: TemplateDataSource[],
  tools?: Tool[],
  templateType?: TemplateType,
-  observability?: TemplateObservability,
 ) => {
  const dependencies: Dependency[] = [];

@@ -105,7 +103,7 @@ const getAdditionalDependencies = (
    case "llamacloud":
      dependencies.push({
        name: "llama-index-indices-managed-llama-cloud",
-        version: "0.6.3",
+        version: "^0.6.3",
      });
      break;
  }
@@ -270,21 +268,6 @@ const getAdditionalDependencies = (
      break;
  }

-  if (observability && observability !== "none") {
-    if (observability === "traceloop") {
-      dependencies.push({
-        name: "traceloop-sdk",
-        version: "^0.15.11",
-      });
-    }
-    if (observability === "llamatrace") {
-      dependencies.push({
-        name: "llama-index-callbacks-arize-phoenix",
-        version: "^0.3.0",
-      });
-    }
-  }
-
  return dependencies;
 };

@@ -396,24 +379,47 @@ export const installPythonDependencies = (
  }
 };

-const installLegacyPythonTemplate = async ({
+export const installPythonTemplate = async ({
+  appName,
  root,
  template,
+  framework,
  vectorDb,
+  postInstallAction,
+  modelConfig,
  dataSources,
  tools,
+  useLlamaParse,
  useCase,
  observability,
 }: Pick<
  InstallTemplateArgs,
+  | "appName"
  | "root"
  | "template"
+  | "framework"
  | "vectorDb"
+  | "postInstallAction"
+  | "modelConfig"
  | "dataSources"
  | "tools"
+  | "useLlamaParse"
  | "useCase"
  | "observability"
 >) => {
+  console.log("\nInitializing Python project with template:", template, "\n");
+  let templatePath;
+  if (template === "reflex") {
+    templatePath = path.join(templatesDir, "types", "reflex");
+  } else {
+    templatePath = path.join(templatesDir, "types", "streaming", framework);
+  }
+  await copy("**", root, {
+    parents: true,
+    cwd: templatePath,
+    rename: assetRelocator,
+  });
+
  const compPath = path.join(templatesDir, "components");
  const enginePath = path.join(root, "app", "engine");

@@ -503,7 +509,34 @@ const installLegacyPythonTemplate = async ({
    }
  }

+  console.log("Adding additional dependencies");
+
+  const addOnDependencies = getAdditionalDependencies(
+    modelConfig,
+    vectorDb,
+    dataSources,
+    tools,
+    template,
+  );
+
  if (observability && observability !== "none") {
+    if (observability === "traceloop") {
+      addOnDependencies.push({
+        name: "traceloop-sdk",
+        version: "^0.15.11",
+      });
+    }
+
+    if (observability === "llamatrace") {
+      addOnDependencies.push({
+        name: "llama-index-callbacks-arize-phoenix",
+        version: "^0.3.0",
+        constraints: {
+          python: ">=3.11,<3.13",
+        },
+      });
+    }
+
    const templateObservabilityPath = path.join(
      templatesDir,
      "components",
@@ -515,133 +548,6 @@ const installLegacyPythonTemplate = async ({
      cwd: templateObservabilityPath,
    });
  }
-};
-
-const installLlamaIndexServerTemplate = async ({
-  root,
-  useCase,
-  useLlamaParse,
-}: Pick<InstallTemplateArgs, "root" | "useCase" | "useLlamaParse">) => {
-  if (!useCase) {
-    console.log(
-      red(
-        `There is no use case selected. Please pick a use case to use via --use-case flag.`,
-      ),
-    );
-    process.exit(1);
-  }
-
-  await copy("workflow.py", path.join(root, "app"), {
-    parents: true,
-    cwd: path.join(templatesDir, "components", "workflows", "python", useCase),
-  });
-
-  // Copy custom UI component code
-  await copy(`*`, path.join(root, "components"), {
-    parents: true,
-    cwd: path.join(templatesDir, "components", "ui", "workflows", useCase),
-  });
-
-  if (useLlamaParse) {
-    await copy("index.py", path.join(root, "app"), {
-      parents: true,
-      cwd: path.join(
-        templatesDir,
-        "components",
-        "vectordbs",
-        "llamaindexserver",
-        "llamacloud",
-        "python",
-      ),
-    });
-    // TODO: Consider moving generate.py to app folder.
-    await copy("generate.py", path.join(root), {
-      parents: true,
-      cwd: path.join(
-        templatesDir,
-        "components",
-        "vectordbs",
-        "llamaindexserver",
-        "llamacloud",
-        "python",
-      ),
-    });
-  }
-  // Copy README.md
-  await copy("README-template.md", path.join(root), {
-    parents: true,
-    cwd: path.join(templatesDir, "components", "workflows", "python", useCase),
-    rename: assetRelocator,
-  });
-};
-
-export const installPythonTemplate = async ({
-  appName,
-  root,
-  template,
-  framework,
-  vectorDb,
-  postInstallAction,
-  modelConfig,
-  dataSources,
-  tools,
-  useLlamaParse,
-  useCase,
-  observability,
-}: Pick<
-  InstallTemplateArgs,
-  | "appName"
-  | "root"
-  | "template"
-  | "framework"
-  | "vectorDb"
-  | "postInstallAction"
-  | "modelConfig"
-  | "dataSources"
-  | "tools"
-  | "useLlamaParse"
-  | "useCase"
-  | "observability"
->) => {
-  console.log("\nInitializing Python project with template:", template, "\n");
-  let templatePath;
-  if (template === "reflex") {
-    templatePath = path.join(templatesDir, "types", "reflex");
-  } else {
-    templatePath = path.join(templatesDir, "types", template, framework);
-  }
-  await copy("**", root, {
-    parents: true,
-    cwd: templatePath,
-    rename: assetRelocator,
-  });
-
-  if (template === "llamaindexserver") {
-    await installLlamaIndexServerTemplate({
-      root,
-      useCase,
-      useLlamaParse,
-    });
-  } else {
-    await installLegacyPythonTemplate({
-      root,
-      template,
-      vectorDb,
-      dataSources,
-      tools,
-      useCase,
-      observability,
-    });
-  }
-
-  console.log("Adding additional dependencies");
-  const addOnDependencies = getAdditionalDependencies(
-    modelConfig,
-    vectorDb,
-    dataSources,
-    tools,
-    template,
-  );

  await addDependencies(root, addOnDependencies);

@@ -124,7 +124,7 @@ For better results, you can specify the region parameter to get results from a s
    dependencies: [
      {
        name: "e2b_code_interpreter",
-        version: "1.1.1",
+        version: "1.0.3",
      },
    ],
    supportedFrameworks: ["fastapi", "express", "nextjs"],
@@ -155,7 +155,7 @@ For better results, you can specify the region parameter to get results from a s
    dependencies: [
      {
        name: "e2b_code_interpreter",
-        version: "1.1.1",
+        version: "1.0.3",
      },
    ],
    supportedFrameworks: ["fastapi", "express", "nextjs"],
@@ -24,8 +24,7 @@ export type TemplateType =
  | "community"
  | "llamapack"
  | "multiagent"
-  | "reflex"
-  | "llamaindexserver";
+  | "reflex";
 export type TemplateFramework = "nextjs" | "express" | "fastapi";
 export type TemplateUI = "html" | "shadcn";
 export type TemplateVectorDB =
@@ -56,8 +55,7 @@ export type TemplateUseCase =
  | "deep_research"
  | "form_filling"
  | "extractor"
-  | "contract_review"
-  | "agentic_rag";
+  | "contract_review";
 // Config for both file and folder
 export type FileSourceConfig =
  | {
@@ -8,104 +8,42 @@ import { templatesDir } from "./dir";
 import { PackageManager } from "./get-pkg-manager";
 import { InstallTemplateArgs, ModelProvider, TemplateVectorDB } from "./types";

-const installLlamaIndexServerTemplate = async ({
-  root,
-  useCase,
-  vectorDb,
-}: Pick<InstallTemplateArgs, "root" | "useCase" | "vectorDb">) => {
-  if (!useCase) {
-    console.log(
-      red(
-        `There is no use case selected. Please pick a use case to use via --use-case flag.`,
-      ),
-    );
-    process.exit(1);
-  }
-
-  if (!vectorDb) {
-    console.log(
-      red(
-        `There is no vector db selected. Please pick a vector db to use via --vector-db flag.`,
-      ),
-    );
-    process.exit(1);
-  }
-
-  await copy("workflow.ts", path.join(root, "src", "app"), {
-    parents: true,
-    cwd: path.join(
-      templatesDir,
-      "components",
-      "workflows",
-      "typescript",
-      useCase,
-    ),
-  });
-
-  // copy workflow UI components to output/components folder
-  await copy("*", path.join(root, "components"), {
-    parents: true,
-    cwd: path.join(templatesDir, "components", "ui", "workflows", useCase),
-  });
-
-  if (vectorDb === "llamacloud") {
-    await copy("generate.ts", path.join(root, "src"), {
-      parents: true,
-      cwd: path.join(
-        templatesDir,
-        "components",
-        "vectordbs",
-        "llamaindexserver",
-        "llamacloud",
-        "typescript",
-      ),
-    });
-
-    await copy("index.ts", path.join(root, "src", "app"), {
-      parents: true,
-      cwd: path.join(
-        templatesDir,
-        "components",
-        "vectordbs",
-        "llamaindexserver",
-        "llamacloud",
-        "typescript",
-      ),
-      rename: () => "data.ts",
-    });
-  }
-  // Copy README.md
-  await copy("README-template.md", path.join(root), {
-    parents: true,
-    cwd: path.join(
-      templatesDir,
-      "components",
-      "workflows",
-      "typescript",
-      useCase,
-    ),
-    rename: assetRelocator,
-  });
-};
-
-const installLegacyTSTemplate = async ({
+/**
+ * Install a LlamaIndex internal template to a given `root` directory.
+ */
+export const installTSTemplate = async ({
+  appName,
  root,
+  packageManager,
+  isOnline,
  template,
-  backend,
  framework,
  ui,
  vectorDb,
+  postInstallAction,
+  backend,
  observability,
  tools,
  dataSources,
  useLlamaParse,
  useCase,
  modelConfig,
-  relativeEngineDestPath,
-}: InstallTemplateArgs & {
-  backend: boolean;
-  relativeEngineDestPath: string;
-}) => {
+}: InstallTemplateArgs & { backend: boolean }) => {
+  console.log(bold(`Using ${packageManager}.`));
+
+  /**
+   * Copy the template files to the target directory.
+   */
+  console.log("\nInitializing project with template:", template, "\n");
+  const templatePath = path.join(templatesDir, "types", "streaming", framework);
+  const copySource = ["**"];
+
+  await copy(copySource, root, {
+    parents: true,
+    cwd: templatePath,
+    rename: assetRelocator,
+  });
+
  /**
   * If next.js is used, update its configuration if necessary
   */
@@ -160,6 +98,10 @@ const installLegacyTSTemplate = async ({
  }

  const compPath = path.join(templatesDir, "components");
+  const relativeEngineDestPath =
+    framework === "nextjs"
+      ? path.join("app", "api", "chat")
+      : path.join("src", "controllers");
  const enginePath = path.join(root, relativeEngineDestPath, "engine");

  // copy llamaindex code for TS templates
@@ -294,75 +236,6 @@ const installLegacyTSTemplate = async ({
    await fs.rm(path.join(root, "app", "api"), { recursive: true });
    await fs.rm(path.join(root, "config"), { recursive: true, force: true });
  }
-};
-
-/**
- * Install a LlamaIndex internal template to a given `root` directory.
- */
-export const installTSTemplate = async ({
-  appName,
-  root,
-  packageManager,
-  isOnline,
-  template,
-  framework,
-  ui,
-  vectorDb,
-  postInstallAction,
-  backend,
-  observability,
-  tools,
-  dataSources,
-  useLlamaParse,
-  useCase,
-  modelConfig,
-}: InstallTemplateArgs & { backend: boolean }) => {
-  console.log(bold(`Using ${packageManager}.`));
-
-  /**
-   * Copy the template files to the target directory.
-   */
-  console.log("\nInitializing project with template:", template, "\n");
-  const templatePath = path.join(templatesDir, "types", template, framework);
-  const copySource = ["**"];
-
-  await copy(copySource, root, {
-    parents: true,
-    cwd: templatePath,
-    rename: assetRelocator,
-  });
-
-  const relativeEngineDestPath =
-    framework === "nextjs"
-      ? path.join("app", "api", "chat")
-      : path.join("src", "controllers");
-
-  if (template === "llamaindexserver") {
-    await installLlamaIndexServerTemplate({
-      root,
-      useCase,
-      vectorDb,
-    });
-  } else {
-    await installLegacyTSTemplate({
-      appName,
-      root,
-      packageManager,
-      isOnline,
-      template,
-      backend,
-      framework,
-      ui,
-      vectorDb,
-      observability,
-      tools,
-      dataSources,
-      useLlamaParse,
-      useCase,
-      modelConfig,
-      relativeEngineDestPath,
-    });
-  }

  const packageJson = await updatePackageJson({
    root,
@@ -375,7 +248,6 @@ export const installTSTemplate = async ({
    vectorDb,
    backend,
    modelConfig,
-    template,
  });

  if (
@@ -390,27 +262,27 @@ const providerDependencies: {
  [key in ModelProvider]?: Record<string, string>;
 } = {
  openai: {
-    "@llamaindex/openai": "^0.2.0",
+    "@llamaindex/openai": "^0.1.52",
  },
  gemini: {
-    "@llamaindex/google": "^0.2.0",
+    "@llamaindex/google": "^0.0.7",
  },
  ollama: {
-    "@llamaindex/ollama": "^0.1.0",
+    "@llamaindex/ollama": "^0.0.40",
  },
  mistral: {
-    "@llamaindex/mistral": "^0.2.0",
+    "@llamaindex/mistral": "^0.0.5",
  },
  "azure-openai": {
-    "@llamaindex/openai": "^0.2.0",
+    "@llamaindex/openai": "^0.1.52",
  },
  groq: {
-    "@llamaindex/groq": "^0.0.61",
-    "@llamaindex/huggingface": "^0.1.0", // groq uses huggingface as default embedding model
+    "@llamaindex/groq": "^0.0.51",
+    "@llamaindex/huggingface": "^0.0.36", // groq uses huggingface as default embedding model
  },
  anthropic: {
-    "@llamaindex/anthropic": "^0.3.0",
-    "@llamaindex/huggingface": "^0.1.0", // anthropic uses huggingface as default embedding model
+    "@llamaindex/anthropic": "^0.1.0",
+    "@llamaindex/huggingface": "^0.0.36", // anthropic uses huggingface as default embedding model
  },
 };

@@ -459,7 +331,6 @@ async function updatePackageJson({
  vectorDb,
  backend,
  modelConfig,
-  template,
 }: Pick<
  InstallTemplateArgs,
  | "root"
@@ -470,7 +341,6 @@ async function updatePackageJson({
  | "observability"
  | "vectorDb"
  | "modelConfig"
-  | "template"
 > & {
  relativeEngineDestPath: string;
  backend: boolean;
@@ -482,7 +352,7 @@ async function updatePackageJson({
  packageJson.name = appName;
  packageJson.version = "0.1.0";

-  if (relativeEngineDestPath && template !== "llamaindexserver") {
+  if (relativeEngineDestPath) {
    // TODO: move script to {root}/scripts for all frameworks
    // add generate script if using context engine
    packageJson.scripts = {
@@ -6,7 +6,7 @@ LlamaIndexServer is a FastAPI-based application that allows you to quickly launc

 - Serving a workflow as a chatbot
 - Built on FastAPI for high performance and easy API development
- Optional built-in chat UI with extendable UI components
+- Optional built-in chat UI
 - Prebuilt development code

 ## Installation
@@ -75,7 +75,6 @@ The LlamaIndexServer accepts the following configuration parameters:
 - `use_default_routers`: Whether to include default routers (chat, static file serving)
 - `env`: Environment setting ('dev' enables CORS and UI by default)
 - `include_ui`: Whether to include the chat UI
- `component_dir`: The directory for custom UI components rendering events emitted by the workflow. The default is None, which does not render custom UI components.
 - `starter_questions`: List of starter questions for the chat UI
 - `verbose`: Enable verbose logging
 - `api_prefix`: API route prefix (default: "/api")
@@ -102,11 +101,6 @@ When enabled, the server provides a chat interface at the root path (`/`) with:
 - Real-time chat interface
 - API endpoint integration

-### Custom UI Components
-
-You can add custom UI components for your workflow by providing `component_dir` config and adding custom .jsx or .tsx files to the directory.
-See [Custom UI Components](docs/custom_ui_component.md) for more details.
-
 ## Development Mode

 In development mode (`env="dev"`), the server:
@@ -1,67 +0,0 @@
-# Custom UI Components
-
-The LlamaIndex server provides support for rendering workflow events using custom UI components, allowing you to extend and customize the chat interface.
-
-## Overview
-
-Custom UI components are a powerful feature that enables you to:
-
- Add custom interface elements to the chat UI using React JSX or TSX files
- Extend the default chat interface functionality
- Create specialized visualizations or interactions
-
-## Configuration
-
-### Workflow events
-
-Your workflow must emit events that fit this structure, allowing the LlamaIndex server to display the right UI components based on the event type.
-
-```json
-{
-    "type": "<event_name>",
-    "data": <data model>
-}
-```
-
-In Pydantic, this is equivalent to:
-
-```python
-from pydantic import BaseModel
-from typing import Literal, Any
-
-class MyCustomEvent(BaseModel):
-    type: Literal["<my_custom_event_name>"]
-    data: dict | Any
-
-    def to_response(self):
-        return self.model_dump()
-```
-
-### Server Setup
-
-1. Initialize the LlamaIndex server with a component directory:
-
-```python
-server = LlamaIndexServer(
-    workflow_factory=your_workflow,
-    component_dir="path/to/components",
-    include_ui=True
-)
-```
-
-2. Add the custom component code to the directory following the naming pattern:
-
-   - File Extension: `.jsx` and `.tsx` for React components
-   - File Name: Should match the event type from your workflow (e.g., `deep_research_event.jsx` for handling `deep_research_event` type that you defined in your workflow). If there are TSX and JSX files with the same name, the TSX file will be used.
-   - Component Name: Export a default React component named `Component` that receives props from the event data
-
-   Example component structure:
-
-   ```jsx
-   function Component({ events }) {
-       // Your component logic here
-       return (
-           // Your UI code here
-       );
-   }
-   ```
@@ -134,9 +134,3 @@ class SourceNodes(BaseModel):
        cls, source_nodes: List[NodeWithScore]
    ) -> List["SourceNodes"]:
        return [cls.from_source_node(node) for node in source_nodes]
-
-
-class ComponentDefinition(BaseModel):
-    type: str
-    code: str
-    filename: str
@@ -1,4 +0,0 @@
-from llama_index.server.api.routers.chat import chat_router
-from llama_index.server.api.routers.ui import custom_components_router
-
-__all__ = ["chat_router", "custom_components_router"]
@@ -1,20 +0,0 @@
-import logging
-from typing import List
-
-from fastapi import APIRouter
-from llama_index.server.api.models import ComponentDefinition
-from llama_index.server.services.custom_ui import CustomUI
-
-
-def custom_components_router(
-    component_dir: str,
-    logger: logging.Logger,
-) -> APIRouter:
-    router = APIRouter(prefix="/components")
-
-    @router.get("")
-    async def components() -> List[ComponentDefinition]:
-        custom_ui = CustomUI(component_dir=component_dir, logger=logger)
-        return custom_ui.get_components()
-
-    return router
@@ -5,7 +5,7 @@ from typing import Optional

 import requests

-CHAT_UI_VERSION = "0.0.9"
+CHAT_UI_VERSION = "0.0.6"


 def download_chat_ui(
@@ -7,7 +7,7 @@ from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.staticfiles import StaticFiles
 from llama_index.core.workflow import Workflow
-from llama_index.server.api.routers import chat_router, custom_components_router
+from llama_index.server.api.routers.chat import chat_router
 from llama_index.server.chat_ui import download_chat_ui
 from llama_index.server.settings import server_settings

@@ -18,7 +18,6 @@ class LlamaIndexServer(FastAPI):
    starter_questions: Optional[list[str]]
    verbose: bool = False
    ui_path: str = ".ui"
-    component_dir: Optional[str] = None

    def __init__(
        self,
@@ -27,7 +26,6 @@ class LlamaIndexServer(FastAPI):
        use_default_routers: Optional[bool] = True,
        env: Optional[str] = None,
        include_ui: Optional[bool] = None,
-        component_dir: Optional[str] = None,
        starter_questions: Optional[list[str]] = None,
        server_url: Optional[str] = None,
        api_prefix: Optional[str] = None,
@@ -44,7 +42,6 @@ class LlamaIndexServer(FastAPI):
            use_default_routers: Whether to use the default routers (chat, mount `data` and `output` directories).
            env: The environment to run the server in.
            include_ui: Whether to show an chat UI in the root path.
-            component_dir: The directory to custom UI components code.
            starter_questions: A list of starter questions to display in the chat UI.
            server_url: The URL of the server.
            api_prefix: The prefix for the API endpoints.
@@ -58,8 +55,6 @@ class LlamaIndexServer(FastAPI):
        self.include_ui = include_ui  # Store the explicitly passed value first
        self.starter_questions = starter_questions
        self.use_default_routers = use_default_routers or True
-        if component_dir:
-            self.component_dir = component_dir

        # Update the settings
        if server_url:
@@ -74,7 +69,6 @@ class LlamaIndexServer(FastAPI):
            self.allow_cors("*")
            if self.include_ui is None:
                self.include_ui = True
-
        if self.include_ui is None:
            self.include_ui = False

@@ -92,8 +86,6 @@ class LlamaIndexServer(FastAPI):
            config["LLAMA_CLOUD_API"] = (
                f"{server_settings.api_url}/chat/config/llamacloud"
            )
-        if self.component_dir:
-            config["COMPONENTS_API"] = f"{server_settings.api_url}/components"
        return config

    # Default routers
@@ -114,28 +106,12 @@ class LlamaIndexServer(FastAPI):
            prefix=server_settings.api_prefix,
        )

-    def add_components_router(self) -> None:
-        """
-        Add the UI router.
-        """
-        if self.component_dir is None:
-            raise ValueError("component_dir must be specified to add components router")
-
-        self.include_router(
-            custom_components_router(self.component_dir, self.logger),
-            prefix=server_settings.api_prefix,
-        )
-
    def mount_ui(self) -> None:
        """
        Mount the UI.
        """
+        # Check if the static folder exists
        if self.include_ui:
-            if self.component_dir:
-                if not os.path.exists(self.component_dir):
-                    os.makedirs(self.component_dir)
-                self.add_components_router()
-            # Check if the static folder exists
            if not os.path.exists(self.ui_path):
                self.logger.warning(
                    f"UI files not found, downloading UI to {self.ui_path}"
@@ -1,81 +0,0 @@
-import logging
-import os
-from typing import List, Optional
-
-from llama_index.server.api.models import ComponentDefinition
-
-
-class CustomUI:
-    def __init__(
-        self, component_dir: str, logger: Optional[logging.Logger] = None
-    ) -> None:
-        self.component_dir = component_dir
-        self.logger = logger or logging.getLogger(__name__)
-
-    def get_components(self) -> List[ComponentDefinition]:
-        """
-        List all js files in the component directory and return a list of ComponentDefinition objects.
-        Ignores files that fail to load and logs the error.
-        TSX files take precedence over JSX files when duplicate component names are found.
-        """
-        components_dict: dict[str, ComponentDefinition] = {}
-        if not os.path.exists(self.component_dir):
-            self.logger.warning(
-                f"Component directory {self.component_dir} does not exist"
-            )
-            return []
-        try:
-            for file in os.listdir(self.component_dir):
-                if not file.endswith((".jsx", ".tsx")):
-                    continue
-
-                component_name = file.split(".")[0]
-                file_path = os.path.join(self.component_dir, file)
-                file_ext = os.path.splitext(file)[1]
-
-                try:
-                    with open(file_path, "r") as f:
-                        code = f.read()
-                        new_component = ComponentDefinition(
-                            type=component_name,
-                            code=code,
-                            filename=file,
-                        )
-
-                        if component_name in components_dict:
-                            existing_ext = os.path.splitext(
-                                components_dict[component_name].filename
-                            )[1]
-
-                            # If existing is TSX and new is JSX, skip and warn
-                            if existing_ext == ".tsx" and file_ext == ".jsx":
-                                self.logger.warning(
-                                    f"Skipping duplicate JSX component {file} as TSX version already exists"
-                                )
-                                continue
-
-                            # If both are same extension, warn and skip
-                            if existing_ext == file_ext:
-                                self.logger.warning(
-                                    f"Skipping duplicate component {file} with same extension"
-                                )
-                                continue
-
-                            # If existing is JSX and new is TSX, replace and warn
-                            if existing_ext == ".jsx" and file_ext == ".tsx":
-                                self.logger.warning(
-                                    f"Replacing JSX component {components_dict[component_name].filename} with TSX version {file}"
-                                )
-                                components_dict[component_name] = new_component
-                                continue
-
-                        components_dict[component_name] = new_component
-
-                except Exception as e:
-                    self.logger.error(f"Failed to load component {file}: {str(e)}")
-                    continue
-
-        except Exception as e:
-            self.logger.error(f"Error reading component directory: {str(e)}")
-
-        return list(components_dict.values())
@@ -149,8 +149,7 @@ def _create_index(
                    "type": "OPENAI_EMBEDDING",
                    "component": {
                        "api_key": os.getenv("OPENAI_API_KEY"),  # editable
-                        "model_name": Settings.embed_model.model_name
-                        or "text-embedding-3-small",
+                        "model_name": os.getenv("EMBEDDING_MODEL"),
                    },
                },
                "transform_config": {
@@ -376,28 +376,6 @@ files = [
    {file = "backcall-0.2.0.tar.gz", hash = "sha256:5cbdbf27be5e7cfadb448baf0aa95508f91f2bbc6c6437cd9cd06e2a4c215e1e"},
 ]

-[[package]]
-name = "banks"
-version = "2.1.1"
-description = "A prompt programming language"
-optional = false
-python-versions = ">=3.9"
-files = [
-    {file = "banks-2.1.1-py3-none-any.whl", hash = "sha256:06e4ee46a0ff2fcdf5f64a5f028a7b7ceb719d5c7b9339f5aa90b24936fbb7f5"},
-    {file = "banks-2.1.1.tar.gz", hash = "sha256:95ec9c8f3c173c9f1c21eb2451ba0e21dda87f1ceb738854fabadb54bc387b86"},
-]
-
-[package.dependencies]
-deprecated = "*"
-eval-type-backport = {version = "*", markers = "python_version < \"3.10\""}
-griffe = "*"
-jinja2 = "*"
-platformdirs = "*"
-pydantic = "*"
-
-[package.extras]
-all = ["litellm", "redis"]
-
 [[package]]
 name = "beautifulsoup4"
 version = "4.13.3"
@@ -1480,20 +1458,6 @@ files = [
 docs = ["Sphinx", "furo"]
 test = ["objgraph", "psutil"]

-[[package]]
-name = "griffe"
-version = "1.7.2"
-description = "Signatures for entire Python programs. Extract the structure, the frame, the skeleton of your project, to generate API documentation or find breaking changes in your API."
-optional = false
-python-versions = ">=3.9"
-files = [
-    {file = "griffe-1.7.2-py3-none-any.whl", hash = "sha256:1ed9c2e338a75741fc82083fe5a1bc89cb6142efe126194cc313e34ee6af5423"},
-    {file = "griffe-1.7.2.tar.gz", hash = "sha256:98d396d803fab3b680c2608f300872fd57019ed82f0672f5b5323a9ad18c540c"},
-]
-
-[package.dependencies]
-colorama = ">=0.4"
-
 [[package]]
 name = "h11"
 version = "0.14.0"
@@ -2252,18 +2216,17 @@ pydantic = ">=1.10"

 [[package]]
 name = "llama-index-core"
-version = "0.12.28"
+version = "0.12.25"
 description = "Interface between LLMs and your data"
 optional = false
 python-versions = "<4.0,>=3.9"
 files = [
-    {file = "llama_index_core-0.12.28-py3-none-any.whl", hash = "sha256:9bd24224bd57dd5e97bb0a2550f31f6c08b7dc396305f35bbe9714d17b1409a6"},
-    {file = "llama_index_core-0.12.28.tar.gz", hash = "sha256:bf4a9697525e1294855d2df5c3a56b9720c185cde0eae2da713e7629c456c643"},
+    {file = "llama_index_core-0.12.25-py3-none-any.whl", hash = "sha256:affd0b240daf3c1200ad8ca86696ce339f411c3028981b0b5b71b59d78f281ae"},
+    {file = "llama_index_core-0.12.25.tar.gz", hash = "sha256:e0145617f0ab5358c30a7a716fa46dd3c40183bad1b87dd0cbfd3d196fbee8c4"},
 ]

 [package.dependencies]
 aiohttp = ">=3.8.6,<4.0.0"
-banks = ">=2.0.0,<3.0.0"
 dataclasses-json = "*"
 deprecated = ">=1.2.9.3"
 dirtyjson = ">=1.0.8,<2.0.0"
@@ -6134,4 +6097,4 @@ type = ["pytest-mypy"]
 [metadata]
 lock-version = "2.0"
 python-versions = ">=3.9,<4.0"
-content-hash = "61f7e3d0861287b293d3c5839cde36bfc032262f38e6314875ffaf456b77d300"
+content-hash = "27ab493aba5c297223b335c422184f902abd56eaf35e8be2271d22a76d6e8b3a"
@@ -26,7 +26,7 @@ license = "MIT"
 name = "llama-index-server"
 packages = [{include = "llama_index/"}]
 readme = "README.md"
-version = "0.1.9"
+version = "0.1.6"

 [tool.poetry.dependencies]
 python = ">=3.9,<4.0"
@@ -34,7 +34,7 @@ fastapi = {extras = ["standard"], version = "^0.115.11"}
 cachetools = "^5.5.2"
 requests = "^2.32.3"
 pydantic-settings = "^2.8.1"
-llama-index-core = "0.12.28"
+llama-index-core = "0.12.25"
 llama-index-readers-file = "^0.4.6"
 llama-index-indices-managed-llama-cloud = "0.6.3"

@@ -1,5 +1,6 @@
 import pytest
 from httpx import ASGITransport, AsyncClient
+
 from llama_index.core.agent.workflow import AgentWorkflow
 from llama_index.core.llms import MockLLM
 from llama_index.server import LlamaIndexServer
@@ -103,99 +104,3 @@ async def test_ui_is_accessible(server: LlamaIndexServer) -> None:
        response = await ac.get("/")
        assert response.status_code == 200
        assert "text/html" in response.headers["content-type"]
-
-
-@pytest.mark.asyncio()
-async def test_component_dir_creation(server: LlamaIndexServer) -> None:
-    """
-    Test if the component directory is created when specified and doesn't exist.
-    """
-    import os
-    import shutil
-
-    test_component_dir = "./test_components"
-
-    # Clean up any existing directory
-    if os.path.exists(test_component_dir):
-        shutil.rmtree(test_component_dir)
-
-    # Create server with component directory
-    _ = LlamaIndexServer(
-        workflow_factory=_agent_workflow,
-        verbose=True,
-        component_dir=test_component_dir,
-        include_ui=True,
-    )
-
-    # Verify directory was created
-    assert os.path.exists(test_component_dir), "Component directory was not created"
-    assert os.path.isdir(test_component_dir), "Component path is not a directory"
-
-    # Clean up after test
-    shutil.rmtree(test_component_dir)
-
-
-@pytest.mark.asyncio()
-async def test_component_router_addition(server: LlamaIndexServer, tmp_path) -> None:
-    """
-    Test if the component router is added when component directory is specified.
-    """
-    test_component_dir = tmp_path / "test_components"
-
-    # Create server with component directory
-    component_server = LlamaIndexServer(
-        workflow_factory=_agent_workflow,
-        verbose=True,
-        component_dir=str(test_component_dir),
-        include_ui=True,
-    )
-
-    # Verify component route exists
-    component_route_exists = any(
-        route.path == "/api/components" for route in component_server.routes
-    )
-    assert component_route_exists, "Component API route not found in server routes"
-
-
-@pytest.mark.asyncio()
-async def test_ui_config_includes_components_api(
-    server: LlamaIndexServer, tmp_path
-) -> None:
-    """
-    Test if the UI config includes components API when component directory is set.
-    """
-    test_component_dir = tmp_path / "test_components"
-
-    # Create server with component directory
-    component_server = LlamaIndexServer(
-        workflow_factory=_agent_workflow,
-        verbose=True,
-        component_dir=str(test_component_dir),
-        include_ui=True,
-    )
-
-    # Check if components API is in UI config
-    ui_config = component_server._ui_config
-    assert "COMPONENTS_API" in ui_config, "Components API not found in UI config"
-    assert ui_config["COMPONENTS_API"].endswith("/components"), (
-        "Incorrect components API path"
-    )
-
-
-@pytest.mark.asyncio()
-async def test_component_router_requires_component_dir(
-    server: LlamaIndexServer,
-) -> None:
-    """
-    Test that adding components router without component_dir raises an error.
-    """
-    server_without_component_dir = LlamaIndexServer(
-        workflow_factory=_agent_workflow,
-        verbose=True,
-        include_ui=True,
-    )
-
-    with pytest.raises(
-        ValueError, match="component_dir must be specified to add components router"
-    ):
-        server_without_component_dir.add_components_router()
@@ -1,6 +1,6 @@
 {
  "name": "create-llama",
-  "version": "0.5.2",
+  "version": "0.4.0",
  "description": "Create LlamaIndex-powered apps with one command",
  "keywords": [
    "rag",
@@ -16,6 +16,5 @@ export const askQuestions = async (
    await askProQuestions(args);
    return args as unknown as QuestionResults;
  }
-  const results = await askSimpleQuestions(args);
-  return results;
+  return await askSimpleQuestions(args);
 };
@@ -1,12 +1,25 @@
 import prompts from "prompts";
-import { EXAMPLE_10K_SEC_FILES, EXAMPLE_FILE } from "../helpers/datasources";
+import {
+  AI_REPORTS,
+  EXAMPLE_10K_SEC_FILES,
+  EXAMPLE_FILE,
+  EXAMPLE_GDPR,
+} 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 = "agentic_rag" | "financial_report" | "deep_research";
+type AppType =
+  | "rag"
+  | "code_artifact"
+  | "financial_report_agent"
+  | "form_filling"
+  | "extractor"
+  | "contract_review"
+  | "data_scientist"
+  | "deep_research";

 type SimpleAnswers = {
  appType: AppType;
@@ -22,22 +35,53 @@ export const askSimpleQuestions = async (
    {
      type: "select",
      name: "appType",
-      message: "What use case do you want to build?",
+      message: "What app do you want to build?",
+      hint: "🤖: Agent, 🔀: Workflow",
      choices: [
        {
-          title: "Agentic RAG",
-          value: "agentic_rag",
+          title: "🤖 Agentic RAG",
+          value: "rag",
          description:
            "Chatbot that answers questions based on provided documents.",
        },
        {
-          title: "Financial Report",
-          value: "financial_report",
+          title: "🤖 Data Scientist",
+          value: "data_scientist",
          description:
            "Agent that analyzes data and generates visualizations by using a code interpreter.",
        },
        {
-          title: "Deep Research",
+          title: "🤖 Code Artifact Agent",
+          value: "code_artifact",
+          description:
+            "Agent that writes code, runs it in a sandbox, and shows the output in the chat UI.",
+        },
+        {
+          title: "🤖 Information Extractor",
+          value: "extractor",
+          description:
+            "Extracts information from documents and returns it as a structured JSON object.",
+        },
+        {
+          title: "🔀 Financial Report Generator",
+          value: "financial_report_agent",
+          description:
+            "Generates a financial report by analyzing the provided 10-K SEC data. Uses a code interpreter to create charts or to conduct further analysis.",
+        },
+        {
+          title: "🔀 Financial 10k SEC Form Filler",
+          value: "form_filling",
+          description:
+            "Extracts information from 10k SEC data and uses it to fill out a CSV form.",
+        },
+        {
+          title: "🔀 Contract Reviewer",
+          value: "contract_review",
+          description:
+            "Extracts and reviews contracts to ensure compliance with GDPR regulations",
+        },
+        {
+          title: "🔀 Deep Researcher",
          value: "deep_research",
          description:
            "Researches and analyzes provided documents from multiple perspectives, generating a comprehensive report with citations to support key findings and insights.",
@@ -49,10 +93,13 @@ export const askSimpleQuestions = async (

  let language: TemplateFramework = "fastapi";
  let llamaCloudKey = args.llamaCloudKey;
-
  let useLlamaCloud = false;

-  if (appType !== "extractor" && appType !== "contract_review") {
+  if (
+    appType !== "extractor" &&
+    appType !== "contract_review" &&
+    appType !== "deep_research"
+  ) {
    const { language: newLanguage } = await prompts(
      {
        type: "select",
@@ -123,36 +170,80 @@ const convertAnswers = async (
  };
  const lookup: Record<
    AppType,
-    Pick<QuestionResults, "template" | "tools" | "dataSources" | "useCase"> & {
+    Pick<
+      QuestionResults,
+      "template" | "tools" | "frontend" | "dataSources" | "useCase"
+    > & {
      modelConfig?: ModelConfig;
    }
  > = {
-    agentic_rag: {
-      template: "llamaindexserver",
+    rag: {
+      template: "streaming",
+      tools: getTools(["weather"]),
+      frontend: true,
      dataSources: [EXAMPLE_FILE],
    },
-    financial_report: {
-      template: "llamaindexserver",
-      dataSources: EXAMPLE_10K_SEC_FILES,
+    data_scientist: {
+      template: "streaming",
      tools: getTools(["interpreter", "document_generator"]),
+      frontend: true,
+      dataSources: [],
      modelConfig: MODEL_GPT4o,
    },
+    code_artifact: {
+      template: "streaming",
+      tools: getTools(["artifact"]),
+      frontend: true,
+      dataSources: [],
+      modelConfig: MODEL_GPT4o,
+    },
+    financial_report_agent: {
+      template: "multiagent",
+      useCase: "financial_report",
+      tools: getTools(["document_generator", "interpreter"]),
+      dataSources: EXAMPLE_10K_SEC_FILES,
+      frontend: true,
+      modelConfig: MODEL_GPT4o,
+    },
+    form_filling: {
+      template: "multiagent",
+      useCase: "form_filling",
+      tools: getTools(["form_filling"]),
+      dataSources: EXAMPLE_10K_SEC_FILES,
+      frontend: true,
+      modelConfig: MODEL_GPT4o,
+    },
+    extractor: {
+      template: "reflex",
+      useCase: "extractor",
+      tools: [],
+      frontend: false,
+      dataSources: [EXAMPLE_FILE],
+    },
+    contract_review: {
+      template: "reflex",
+      useCase: "contract_review",
+      tools: [],
+      frontend: false,
+      dataSources: [EXAMPLE_GDPR],
+    },
    deep_research: {
-      template: "llamaindexserver",
-      dataSources: EXAMPLE_10K_SEC_FILES,
+      template: "multiagent",
+      useCase: "deep_research",
      tools: [],
-      modelConfig: MODEL_GPT4o,
+      frontend: true,
+      dataSources: [AI_REPORTS],
    },
  };
-
  const results = lookup[answers.appType];
  return {
    framework: answers.language,
-    useCase: answers.appType,
    ui: "shadcn",
    llamaCloudKey: answers.llamaCloudKey,
    useLlamaParse: answers.useLlamaCloud,
+    llamapack: "",
    vectorDb: answers.useLlamaCloud ? "llamacloud" : "none",
+    observability: "none",
    ...results,
    modelConfig:
      results.modelConfig ??
@@ -161,6 +252,6 @@ const convertAnswers = async (
        askModels: args.askModels ?? false,
        framework: answers.language,
      })),
-    frontend: true,
+    frontend: answers.language === "nextjs" ? false : results.frontend,
  };
 };
@@ -1,409 +0,0 @@
-function Component({ events }) {
-  // Aggregate events by type and track their state progression
-  const aggregateEvents = () => {
-    const retrieveEvents = events.filter((e) => e.event === "retrieve");
-    const analyzeEvents = events.filter((e) => e.event === "analyze");
-    const answerEvents = events.filter((e) => e.event === "answer");
-
-    // Get the latest state for retrieve and analyze events
-    const retrieveState =
-      retrieveEvents.length > 0
-        ? retrieveEvents[retrieveEvents.length - 1].state
-        : null;
-    const analyzeState =
-      analyzeEvents.length > 0
-        ? analyzeEvents[analyzeEvents.length - 1].state
-        : null;
-
-    // Group answer events by their ID
-    const answerGroups = {};
-    for (const event of answerEvents) {
-      if (!event.id) continue;
-
-      if (!answerGroups[event.id]) {
-        answerGroups[event.id] = {
-          id: event.id,
-          question: event.question,
-          answer: event.answer,
-          states: [],
-        };
-      }
-
-      const lastState =
-        answerGroups[event.id].states[answerGroups[event.id].states.length - 1];
-      if (lastState !== event.state) {
-        answerGroups[event.id].states.push(event.state);
-      }
-
-      if (event.answer) {
-        answerGroups[event.id].answer = event.answer;
-      }
-    }
-
-    return {
-      retrieveState,
-      analyzeState,
-      answerGroups: Object.values(answerGroups),
-    };
-  };
-
-  const { retrieveState, analyzeState, answerGroups } = aggregateEvents();
-
-  // Styles
-  const styles = {
-    container: {
-      maxWidth: "900px",
-      margin: "0 auto",
-      padding: "20px",
-      backgroundColor: "#FFFFFF",
-      borderRadius: "8px",
-      boxShadow: "0 2px 8px rgba(0, 0, 0, 0.15)",
-    },
-    header: {
-      fontSize: "24px",
-      fontWeight: "bold",
-      textAlign: "center",
-      marginBottom: "20px",
-      color: "#333",
-    },
-    cardsContainer: {
-      display: "flex",
-      gap: "16px",
-      marginBottom: "30px",
-      flexWrap: "wrap",
-    },
-    card: {
-      flex: "1",
-      minWidth: "250px",
-      backgroundColor: "#FFFFFF",
-      borderRadius: "8px",
-      boxShadow: "0 2px 4px rgba(0, 0, 0, 0.1)",
-      padding: "16px",
-      border: "1px solid #E5E7EB",
-      transition: "all 0.3s ease",
-    },
-    cardHeader: {
-      display: "flex",
-      alignItems: "center",
-      marginBottom: "12px",
-    },
-    cardTitle: {
-      fontSize: "18px",
-      fontWeight: "bold",
-      marginLeft: "8px",
-    },
-    cardContent: {
-      display: "flex",
-      justifyContent: "space-between",
-      alignItems: "center",
-    },
-    badge: {
-      padding: "4px 8px",
-      borderRadius: "9999px",
-      fontSize: "12px",
-      fontWeight: "bold",
-      flexShrink: 0,
-    },
-    questionsList: {
-      marginTop: "30px",
-    },
-    questionItem: {
-      backgroundColor: "#F9FAFB",
-      border: "1px solid #E5E7EB",
-      borderRadius: "8px",
-      marginBottom: "12px",
-      overflow: "hidden",
-    },
-    questionHeader: {
-      padding: "12px 16px",
-      cursor: "pointer",
-      display: "flex",
-      justifyContent: "space-between",
-      alignItems: "flex-start",
-      userSelect: "none",
-    },
-    questionTitle: {
-      fontWeight: "medium",
-      marginLeft: "12px",
-    },
-    answerContainer: {
-      padding: "0",
-      backgroundColor: "#F3F4F6",
-      overflow: "hidden",
-      maxHeight: "0",
-      transition: "max-height 0.3s ease-out",
-      whiteSpace: "pre-line",
-      fontSize: "13px",
-    },
-    answerContent: {
-      padding: "16px",
-    },
-    answerContainerExpanded: {
-      maxHeight: "1000px", // Adjust based on content
-    },
-    arrow: {
-      marginLeft: "8px",
-      transition: "transform 0.3s ease",
-      display: "inline-block",
-    },
-    loadingContainer: {
-      display: "flex",
-      justifyContent: "center",
-      alignItems: "center",
-      padding: "24px",
-      color: "#6B7280",
-    },
-    stateIconContainer: {
-      display: "flex",
-      alignItems: "center",
-      gap: "8px",
-    },
-  };
-
-  // Helper function to get color for a state
-  const getStateColor = (state) => {
-    switch (state) {
-      case "inprogress":
-        return "#FCD34D";
-      case "done":
-        return "#34D399";
-      case "pending":
-        return "#9CA3AF";
-      default:
-        return "#D1D5DB";
-    }
-  };
-
-  // Helper function to get badge styles based on state
-  const getBadgeStyles = (state) => {
-    const colors = {
-      inprogress: { background: "#FEF3C7", color: "#92400E" },
-      done: { background: "#D1FAE5", color: "#065F46" },
-      pending: { background: "#F3F4F6", color: "#1F2937" },
-    };
-    const stateColors = colors[state] || colors.pending;
-    return {
-      ...styles.badge,
-      backgroundColor: stateColors.background,
-      color: stateColors.color,
-    };
-  };
-
-  // Helper function to render state icon
-  const renderStateIcon = (state) => {
-    const color = getStateColor(state);
-    if (state === "inprogress") {
-      return (
-        <div style={{ color, animation: "spin 1s linear infinite" }}>⟳</div>
-      );
-    } else if (state === "done") {
-      return <div style={{ color }}>✓</div>;
-    } else if (state === "pending") {
-      return <div style={{ color }}>⏱</div>;
-    }
-    return <div style={{ color }}>?</div>;
-  };
-
-  // State for toggling question answers
-  const [expandedQuestions, setExpandedQuestions] = React.useState({});
-
-  const toggleQuestion = (questionId) => {
-    setExpandedQuestions((prev) => ({
-      ...prev,
-      [questionId]: !prev[questionId],
-    }));
-  };
-
-  return (
-    <div style={styles.container}>
-      <h1 style={styles.header}>Research Progress</h1>
-
-      {/* Status Cards */}
-      <div style={styles.cardsContainer}>
-        {/* Retrieve Card */}
-        <div
-          style={{
-            ...styles.card,
-            borderColor:
-              retrieveState === "done"
-                ? "#A7F3D0"
-                : retrieveState === "inprogress"
-                  ? "#FDE68A"
-                  : "#E5E7EB",
-          }}
-        >
-          <div style={styles.cardHeader}>
-            <span style={{ color: "#8B5CF6" }}>🔍</span>
-            <div style={styles.cardTitle}>Data Retrieval</div>
-          </div>
-          <div style={styles.cardContent}>
-            <span style={{ fontSize: "14px", color: "#6B7280" }}>Status:</span>
-            <div style={styles.stateIconContainer}>
-              {retrieveState && (
-                <span style={getBadgeStyles(retrieveState)}>
-                  {retrieveState === "inprogress"
-                    ? "In Progress"
-                    : retrieveState === "done"
-                      ? "Completed"
-                      : "Pending"}
-                </span>
-              )}
-              {renderStateIcon(retrieveState)}
-            </div>
-          </div>
-        </div>
-
-        {/* Analyze Card */}
-        <div
-          style={{
-            ...styles.card,
-            borderColor:
-              analyzeState === "done"
-                ? "#A7F3D0"
-                : analyzeState === "inprogress"
-                  ? "#FDE68A"
-                  : "#E5E7EB",
-          }}
-        >
-          <div style={styles.cardHeader}>
-            <span style={{ color: "#06B6D4" }}>📊</span>
-            <div style={styles.cardTitle}>Data Analysis</div>
-          </div>
-          <div style={styles.cardContent}>
-            <span style={{ fontSize: "14px", color: "#6B7280" }}>Status:</span>
-            <div style={styles.stateIconContainer}>
-              {analyzeState && (
-                <span style={getBadgeStyles(analyzeState)}>
-                  {analyzeState === "inprogress"
-                    ? "In Progress"
-                    : analyzeState === "done"
-                      ? "Completed"
-                      : "Pending"}
-                </span>
-              )}
-              {renderStateIcon(analyzeState)}
-            </div>
-          </div>
-        </div>
-
-        {/* Questions Card */}
-        <div style={styles.card}>
-          <div style={styles.cardHeader}>
-            <span style={{ color: "#10B981" }}>💬</span>
-            <div style={styles.cardTitle}>Questions</div>
-          </div>
-          <div style={styles.cardContent}>
-            <span style={{ fontSize: "14px", color: "#6B7280" }}>Status:</span>
-            <div style={{ display: "flex", alignItems: "center", gap: "8px" }}>
-              <span
-                style={{
-                  ...styles.badge,
-                  backgroundColor: "#F3F4F6",
-                  color: "#1F2937",
-                }}
-              >
-                {answerGroups.length} Questions
-              </span>
-              <span
-                style={{
-                  fontSize: "14px",
-                  fontWeight: "500",
-                  color: "#059669",
-                }}
-              >
-                {answerGroups.filter((g) => g.states.includes("done")).length}{" "}
-                Answered
-              </span>
-            </div>
-          </div>
-        </div>
-      </div>
-
-      {/* Questions List */}
-      {answerGroups.length > 0 && (
-        <div style={styles.questionsList}>
-          <h2
-            style={{
-              fontSize: "20px",
-              fontWeight: "600",
-              marginBottom: "16px",
-            }}
-          >
-            Research Questions
-          </h2>
-          {answerGroups.map((group, index) => {
-            const latestState = group.states[group.states.length - 1];
-            const questionId = group.id || `question-${index}`;
-            const isExpanded = expandedQuestions[questionId];
-
-            const answerWithoutCitation = group.answer?.replace(
-              /\[citation:[a-f0-9-]+\]/g,
-              "",
-            );
-
-            return (
-              <div key={questionId} style={styles.questionItem}>
-                <div
-                  style={styles.questionHeader}
-                  onClick={() => toggleQuestion(questionId)}
-                >
-                  <div
-                    style={{
-                      display: "flex",
-                      alignItems: "flex-start",
-                      gap: "12px",
-                      fontSize: "14px",
-                    }}
-                  >
-                    {renderStateIcon(latestState)}
-                    <span style={styles.questionTitle}>{group.question}</span>
-                    <span
-                      style={{
-                        ...styles.arrow,
-                        transform: isExpanded
-                          ? "rotate(180deg)"
-                          : "rotate(0deg)",
-                      }}
-                    >
-                      ▼
-                    </span>
-                  </div>
-                  <span style={getBadgeStyles(latestState)}>
-                    {latestState === "inprogress"
-                      ? "In Progress"
-                      : latestState === "done"
-                        ? "Answered"
-                        : "Pending"}
-                  </span>
-                </div>
-                <div
-                  style={{
-                    ...styles.answerContainer,
-                    ...(isExpanded ? styles.answerContainerExpanded : {}),
-                  }}
-                >
-                  {answerWithoutCitation ? (
-                    <div style={styles.answerContent}>
-                      {answerWithoutCitation}
-                    </div>
-                  ) : (
-                    <div
-                      style={{
-                        ...styles.loadingContainer,
-                        ...styles.answerContent,
-                      }}
-                    >
-                      <span style={{ marginLeft: "8px" }}>
-                        Generating answer...
-                      </span>
-                    </div>
-                  )}
-                </div>
-              </div>
-            );
-          })}
-        </div>
-      )}
-    </div>
-  );
-}
@@ -1,31 +0,0 @@
-# flake8: noqa: E402
-
-from dotenv import load_dotenv
-
-load_dotenv()
-
-import logging
-
-from app.index import get_index
-from app.settings import init_settings
-from llama_index.server.services.llamacloud.generate import (
-    load_to_llamacloud,
-)
-
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger()
-
-
-def generate_datasource():
-    init_settings()
-    logger.info("Generate index for the provided data")
-
-    index = get_index(create_if_missing=True)
-    if index is None:
-        raise ValueError("Index not found and could not be created")
-
-    load_to_llamacloud(index, logger=logger)
-
-
-if __name__ == "__main__":
-    generate_datasource()
@@ -1,7 +0,0 @@
-from llama_index.server.services.llamacloud import (
-    LlamaCloudIndex,
-    get_client,
-    get_index,
-)
-
-__all__ = ["LlamaCloudIndex", "get_client", "get_index"]
@@ -1,100 +0,0 @@
-import * as dotenv from "dotenv";
-import "dotenv/config";
-import * as fs from "fs/promises";
-import { LLamaCloudFileService } from "llamaindex";
-import * as path from "path";
-import { getIndex } from "./app/data";
-import { initSettings } from "./app/settings";
-
-dotenv.config();
-
-const REQUIRED_ENV_VARS = [
-  "LLAMA_CLOUD_INDEX_NAME",
-  "LLAMA_CLOUD_PROJECT_NAME",
-  "LLAMA_CLOUD_API_KEY",
-];
-
-export function checkRequiredEnvVars() {
-  const missingEnvVars = REQUIRED_ENV_VARS.filter((envVar) => {
-    return !process.env[envVar];
-  });
-
-  if (missingEnvVars.length > 0) {
-    console.log(
-      `The following environment variables are required but missing: ${missingEnvVars.join(
-        ", ",
-      )}`,
-    );
-    throw new Error(
-      `Missing environment variables: ${missingEnvVars.join(", ")}`,
-    );
-  }
-}
-
-async function* walk(dir: string): AsyncGenerator<string> {
-  const directory = await fs.opendir(dir);
-
-  for await (const dirent of directory) {
-    const entryPath = path.join(dir, dirent.name);
-
-    if (dirent.isDirectory()) {
-      yield* walk(entryPath); // Recursively walk through directories
-    } else if (dirent.isFile()) {
-      yield entryPath; // Yield file paths
-    }
-  }
-}
-
-async function loadAndIndex() {
-  const index = await getIndex();
-  // ensure the index is available or create a new one
-  await index.ensureIndex({
-    verbose: true,
-    embedding: {
-      type: "OPENAI_EMBEDDING",
-      component: {
-        api_key: process.env.OPENAI_API_KEY,
-        model_name: "text-embedding-3-small",
-      },
-    },
-  });
-
-  const projectId = await index.getProjectId();
-  const pipelineId = await index.getPipelineId();
-
-  // walk through the data directory and upload each file to LlamaCloud
-  for await (const filePath of walk("data")) {
-    const buffer = await fs.readFile(filePath);
-    const filename = path.basename(filePath);
-    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!`);
-}
-
-(async () => {
-  try {
-    checkRequiredEnvVars();
-    initSettings();
-    await loadAndIndex();
-    console.log("Finished generating storage.");
-  } catch (error) {
-    console.error("Error generating storage.", error);
-  }
-})();
@@ -1,28 +0,0 @@
-import { LlamaCloudIndex } from "llamaindex/cloud/LlamaCloudIndex";
-
-type LlamaCloudDataSourceParams = {
-  llamaCloudPipeline?: {
-    project: string;
-    pipeline: string;
-  };
-};
-
-export async function getIndex(params?: LlamaCloudDataSourceParams) {
-  const { project, pipeline } = params?.llamaCloudPipeline ?? {};
-  const projectName = project ?? process.env.LLAMA_CLOUD_PROJECT_NAME;
-  const pipelineName = pipeline ?? process.env.LLAMA_CLOUD_INDEX_NAME;
-  const apiKey = process.env.LLAMA_CLOUD_API_KEY;
-  if (!projectName || !pipelineName || !apiKey) {
-    throw new Error(
-      "LlamaCloud is not configured. Please set project, pipeline, and api key in the params or as environment variables.",
-    );
-  }
-  const index = new LlamaCloudIndex({
-    organizationId: process.env.LLAMA_CLOUD_ORGANIZATION_ID,
-    name: pipelineName,
-    projectName,
-    apiKey,
-    baseUrl: process.env.LLAMA_CLOUD_BASE_URL,
-  });
-  return index;
-}
@@ -1,59 +0,0 @@
-This is a [LlamaIndex](https://www.llamaindex.ai/) simple agentic RAG project using [Agent Workflows](https://docs.llamaindex.ai/en/stable/examples/agent/agent_workflow_basic/).
-
-## Getting Started
-
-First, setup the environment with poetry:
-
-> **_Note:_** This step is not needed if you are using the dev-container.
-
-```shell
-poetry install
-```
-
-Then check the parameters that have been pre-configured in the `.env` file in this directory.
-Make sure you have set the `OPENAI_API_KEY` for the LLM.
-
-Second, generate the embeddings of the documents in the `./data` directory:
-
-```shell
-poetry run generate
-```
-
-Third, run the development server:
-
-```shell
-poetry run dev
-```
-
-Then open [http://localhost:8000](http://localhost:8000) with your browser to start the chat UI.
-
-To start the app optimized for **production**, run:
-
-```
-poetry run prod
-```
-
-## Configure LLM and Embedding Model
-
-You can configure [LLM model](https://docs.llamaindex.ai/en/stable/module_guides/models/llms) and [embedding model](https://docs.llamaindex.ai/en/stable/module_guides/models/embeddings) in [settings.py](app/settings.py).
-
-## Use Case
-
-We have prepared an [example workflow](./app/workflow.py) for the agentic RAG use case, where you can ask questions about the example documents in the [./data](./data) directory.
-
-You can start by sending an request on the [chat UI](http://localhost:8000) or you can test the `/api/chat` endpoint with the following curl request:
-
-```
-curl --location 'localhost:8000/api/chat' \
--header 'Content-Type: application/json' \
--data '{ "messages": [{ "role": "user", "content": "What standards for a letter exist?" }] }'
-```
-
-## Learn More
-
-To learn more about LlamaIndex, take a look at the following resources:
-
- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex.
- [Workflows Introduction](https://docs.llamaindex.ai/en/stable/understanding/workflows/) - learn about LlamaIndex workflows.
-
-You can check out [the LlamaIndex GitHub repository](https://github.com/run-llama/llama_index) - your feedback and contributions are welcome!
@@ -1,22 +0,0 @@
-from typing import Optional
-
-from app.index import get_index
-from llama_index.core.agent.workflow import AgentWorkflow
-from llama_index.core.settings import Settings
-from llama_index.llms.openai import OpenAI
-from llama_index.server.api.models import ChatRequest
-from llama_index.server.tools.index import get_query_engine_tool
-
-
-def create_workflow(chat_request: Optional[ChatRequest] = None) -> AgentWorkflow:
-    index = get_index(chat_request=chat_request)
-    if index is None:
-        raise RuntimeError(
-            "Index not found! Please run `poetry run generate` to index the data first."
-        )
-    query_tool = get_query_engine_tool(index=index)
-    return AgentWorkflow.from_tools_or_functions(
-        tools_or_functions=[query_tool],
-        llm=Settings.llm or OpenAI(model="gpt-4o-mini"),
-        system_prompt="You are a helpful assistant.",
-    )
@@ -1,59 +0,0 @@
-This is a [LlamaIndex](https://www.llamaindex.ai/) multi-agents project using [Workflows](https://docs.llamaindex.ai/en/stable/understanding/workflows/).
-
-## Getting Started
-
-First, setup the environment with poetry:
-
-> **_Note:_** This step is not needed if you are using the dev-container.
-
-```shell
-poetry install
-```
-
-Then check the parameters that have been pre-configured in the `.env` file in this directory.
-Make sure you have set the `OPENAI_API_KEY` for the LLM.
-
-Second, generate the embeddings of the documents in the `./data` directory:
-
-```shell
-poetry run generate
-```
-
-Third, run the development server:
-
-```shell
-poetry run dev
-```
-
-Then open [http://localhost:8000](http://localhost:8000) with your browser to start the chat UI.
-
-To start the app optimized for **production**, run:
-
-```
-poetry run prod
-```
-
-## Configure LLM and Embedding Model
-
-You can configure [LLM model](https://docs.llamaindex.ai/en/stable/module_guides/models/llms) and [embedding model](https://docs.llamaindex.ai/en/stable/module_guides/models/embeddings) in [settings.py](app/settings.py).
-
-## Use Case
-
-We have prepared an [example workflow](./app/workflow.py) for the deep research use case, where you can ask questions about the example documents in the [./data](./data) directory.
-
-You can start by sending an request on the [chat UI](http://localhost:8000) or you can test the `/api/chat` endpoint with the following curl request:
-
-```
-curl --location 'localhost:8000/api/chat' \
--header 'Content-Type: application/json' \
--data '{ "messages": [{ "role": "user", "content": "Create a report comparing the finances of Apple and Tesla" }] }'
-```
-
-## Learn More
-
-To learn more about LlamaIndex, take a look at the following resources:
-
- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex.
- [Workflows Introduction](https://docs.llamaindex.ai/en/stable/understanding/workflows/) - learn about LlamaIndex workflows.
-
-You can check out [the LlamaIndex GitHub repository](https://github.com/run-llama/llama_index) - your feedback and contributions are welcome!
@@ -1,534 +0,0 @@
-import logging
-import os
-import uuid
-from typing import List, Literal, Optional
-
-from app.index import get_index
-from llama_index.core.base.llms.types import (
-    CompletionResponse,
-    CompletionResponseAsyncGen,
-)
-from llama_index.core.indices.base import BaseIndex
-from llama_index.core.memory import ChatMemoryBuffer
-from llama_index.core.memory.simple_composable_memory import SimpleComposableMemory
-from llama_index.core.prompts import PromptTemplate
-from llama_index.core.schema import MetadataMode, Node, NodeWithScore
-from llama_index.core.settings import Settings
-from llama_index.core.types import ChatMessage, MessageRole
-from llama_index.core.workflow import (
-    Context,
-    Event,
-    StartEvent,
-    StopEvent,
-    Workflow,
-    step,
-)
-from llama_index.server.api.models import SourceNodesEvent
-from llama_index.server.api.models import ChatRequest
-from pydantic import BaseModel, Field
-
-logger = logging.getLogger("uvicorn")
-logger.setLevel(logging.INFO)
-
-
-def create_workflow(chat_request: Optional[ChatRequest] = None) -> Workflow:
-    index = get_index(chat_request=chat_request)
-    if index is None:
-        raise ValueError(
-            "Index is not found. Try run generation script to create the index first."
-        )
-
-    return DeepResearchWorkflow(
-        index=index,
-        timeout=120.0,
-    )
-
-
-# Workflow events
-class PlanResearchEvent(Event):
-    pass
-
-
-class ResearchEvent(Event):
-    question_id: str
-    question: str
-    context_nodes: List[NodeWithScore]
-
-
-class CollectAnswersEvent(Event):
-    question_id: str
-    question: str
-    answer: str
-
-
-class ReportEvent(Event):
-    pass
-
-
-# Events that are streamed to the frontend and rendered there
-class DeepResearchEventData(BaseModel):
-    event: Literal["retrieve", "analyze", "answer"]
-    state: Literal["pending", "inprogress", "done", "error"]
-    id: Optional[str] = None
-    question: Optional[str] = None
-    answer: Optional[str] = None
-
-
-class DataEvent(Event):
-    type: Literal["deep_research_event"]
-    data: DeepResearchEventData
-
-    def to_response(self):
-        return self.model_dump()
-
-
-class DeepResearchWorkflow(Workflow):
-    """
-    A workflow to research and analyze documents from multiple perspectives and write a comprehensive report.
-
-    Requirements:
-    - An indexed documents containing the knowledge base related to the topic
-
-    Steps:
-    1. Retrieve information from the knowledge base
-    2. Analyze the retrieved information and provide questions for answering
-    3. Answer the questions
-    4. Write the report based on the research results
-    """
-
-    memory: SimpleComposableMemory
-    context_nodes: List[Node]
-    index: BaseIndex
-    user_request: str
-    stream: bool = True
-
-    def __init__(
-        self,
-        index: BaseIndex,
-        **kwargs,
-    ):
-        super().__init__(**kwargs)
-        self.index = index
-        self.context_nodes = []
-        self.memory = SimpleComposableMemory.from_defaults(
-            primary_memory=ChatMemoryBuffer.from_defaults(),
-        )
-
-    @step
-    async def retrieve(self, ctx: Context, ev: StartEvent) -> PlanResearchEvent:
-        """
-        Initiate the workflow: memory, tools, agent
-        """
-        self.stream = ev.get("stream", True)
-        self.user_request = ev.get("user_msg")
-        chat_history = ev.get("chat_history")
-        if chat_history is not None:
-            self.memory.put_messages(chat_history)
-
-        await ctx.set("total_questions", 0)
-
-        # Add user message to memory
-        self.memory.put_messages(
-            messages=[
-                ChatMessage(
-                    role=MessageRole.USER,
-                    content=self.user_request,
-                )
-            ]
-        )
-        ctx.write_event_to_stream(
-            DataEvent(
-                type="deep_research_event",
-                data={
-                    "event": "retrieve",
-                    "state": "inprogress",
-                },
-            )
-        )
-        retriever = self.index.as_retriever(
-            similarity_top_k=int(os.getenv("TOP_K", 10)),
-        )
-        nodes = retriever.retrieve(self.user_request)
-        self.context_nodes.extend(nodes)  # type: ignore
-        ctx.write_event_to_stream(
-            DataEvent(
-                type="deep_research_event",
-                data={
-                    "event": "retrieve",
-                    "state": "done",
-                },
-            )
-        )
-        # Send source nodes to the stream
-        # Use SourceNodesEvent to display source nodes in the UI.
-        ctx.write_event_to_stream(
-            SourceNodesEvent(
-                nodes=nodes,
-            )
-        )
-        return PlanResearchEvent()
-
-    @step
-    async def analyze(
-        self, ctx: Context, ev: PlanResearchEvent
-    ) -> ResearchEvent | ReportEvent | StopEvent:
-        """
-        Analyze the retrieved information
-        """
-        logger.info("Analyzing the retrieved information")
-        ctx.write_event_to_stream(
-            DataEvent(
-                type="deep_research_event",
-                data={
-                    "event": "analyze",
-                    "state": "inprogress",
-                },
-            )
-        )
-        total_questions = await ctx.get("total_questions")
-        res = await plan_research(
-            memory=self.memory,
-            context_nodes=self.context_nodes,
-            user_request=self.user_request,
-            total_questions=total_questions,
-        )
-        if res.decision == "cancel":
-            ctx.write_event_to_stream(
-                DataEvent(
-                    type="deep_research_event",
-                    data={
-                        "event": "analyze",
-                        "state": "done",
-                    },
-                )
-            )
-            return StopEvent(
-                result=res.cancel_reason,
-            )
-        elif res.decision == "write":
-            # Writing a report without any research context is not allowed.
-            # It's a LLM hallucination.
-            if total_questions == 0:
-                ctx.write_event_to_stream(
-                    DataEvent(
-                        type="deep_research_event",
-                        data={
-                            "event": "analyze",
-                            "state": "done",
-                        },
-                    )
-                )
-                return StopEvent(
-                    result="Sorry, I have a problem when analyzing the retrieved information. Please try again.",
-                )
-
-            self.memory.put(
-                message=ChatMessage(
-                    role=MessageRole.ASSISTANT,
-                    content="No more idea to analyze. We should report the answers.",
-                )
-            )
-            ctx.send_event(ReportEvent())
-        else:
-            total_questions += len(res.research_questions)
-            await ctx.set("total_questions", total_questions)  # For tracking
-            await ctx.set(
-                "waiting_questions", len(res.research_questions)
-            )  # For waiting questions to be answered
-            self.memory.put(
-                message=ChatMessage(
-                    role=MessageRole.ASSISTANT,
-                    content="We need to find answers to the following questions:\n"
-                    + "\n".join(res.research_questions),
-                )
-            )
-            for question in res.research_questions:
-                question_id = str(uuid.uuid4())
-                ctx.write_event_to_stream(
-                    DataEvent(
-                        type="deep_research_event",
-                        data={
-                            "event": "answer",
-                            "state": "pending",
-                            "id": question_id,
-                            "question": question,
-                            "answer": None,
-                        },
-                    )
-                )
-                ctx.send_event(
-                    ResearchEvent(
-                        question_id=question_id,
-                        question=question,
-                        context_nodes=self.context_nodes,
-                    )
-                )
-        ctx.write_event_to_stream(
-            DataEvent(
-                type="deep_research_event",
-                data={
-                    "event": "analyze",
-                    "state": "done",
-                },
-            )
-        )
-        return None
-
-    @step(num_workers=2)
-    async def answer(self, ctx: Context, ev: ResearchEvent) -> CollectAnswersEvent:
-        """
-        Answer the question
-        """
-        ctx.write_event_to_stream(
-            DataEvent(
-                type="deep_research_event",
-                data={
-                    "event": "answer",
-                    "state": "inprogress",
-                    "id": ev.question_id,
-                    "question": ev.question,
-                },
-            )
-        )
-        try:
-            answer = await research(
-                context_nodes=ev.context_nodes,
-                question=ev.question,
-            )
-        except Exception as e:
-            logger.error(f"Error answering question {ev.question}: {e}")
-            answer = f"Got error when answering the question: {ev.question}"
-        ctx.write_event_to_stream(
-            DataEvent(
-                type="deep_research_event",
-                data={
-                    "event": "answer",
-                    "state": "done",
-                    "id": ev.question_id,
-                    "question": ev.question,
-                    "answer": answer,
-                },
-            )
-        )
-
-        return CollectAnswersEvent(
-            question_id=ev.question_id,
-            question=ev.question,
-            answer=answer,
-        )
-
-    @step
-    async def collect_answers(
-        self, ctx: Context, ev: CollectAnswersEvent
-    ) -> PlanResearchEvent:
-        """
-        Collect answers to all questions
-        """
-        num_questions = await ctx.get("waiting_questions")
-        results = ctx.collect_events(
-            ev,
-            expected=[CollectAnswersEvent] * num_questions,
-        )
-        if results is None:
-            return None
-        for result in results:
-            self.memory.put(
-                message=ChatMessage(
-                    role=MessageRole.ASSISTANT,
-                    content=f"<Question>{result.question}</Question>\n<Answer>{result.answer}</Answer>",
-                )
-            )
-        await ctx.set("waiting_questions", 0)
-        self.memory.put(
-            message=ChatMessage(
-                role=MessageRole.ASSISTANT,
-                content="Researched all the questions. Now, i need to analyze if it's ready to write a report or need to research more.",
-            )
-        )
-        return PlanResearchEvent()
-
-    @step
-    async def report(self, ctx: Context, ev: ReportEvent) -> StopEvent:
-        """
-        Report the answers
-        """
-        res = await write_report(
-            memory=self.memory,
-            user_request=self.user_request,
-            stream=self.stream,
-        )
-        return StopEvent(
-            result=res,
-        )
-
-
-class AnalysisDecision(BaseModel):
-    decision: Literal["research", "write", "cancel"] = Field(
-        description="Whether to continue research, write a report, or cancel the research after several retries"
-    )
-    research_questions: Optional[List[str]] = Field(
-        description="""
-        If the decision is to research, provide a list of questions to research that related to the user request.
-        Maximum 3 questions. Set to null or empty if writing a report or cancel the research.
-        """,
-        default_factory=list,
-    )
-    cancel_reason: Optional[str] = Field(
-        description="The reason for cancellation if the decision is to cancel research.",
-        default=None,
-    )
-
-
-async def plan_research(
-    memory: SimpleComposableMemory,
-    context_nodes: List[Node],
-    user_request: str,
-    total_questions: int,
-) -> AnalysisDecision:
-    analyze_prompt = """
-      You are a professor who is guiding a researcher to research a specific request/problem.
-      Your task is to decide on a research plan for the researcher.
-
-      The possible actions are:
-      + Provide a list of questions for the researcher to investigate, with the purpose of clarifying the request.
-      + Write a report if the researcher has already gathered enough research on the topic and can resolve the initial request.
-      + Cancel the research if most of the answers from researchers indicate there is insufficient information to research the request. Do not attempt more than 3 research iterations or too many questions.
-
-      The workflow should be:
-      + Always begin by providing some initial questions for the researcher to investigate.
-      + Analyze the provided answers against the initial topic/request. If the answers are insufficient to resolve the initial request, provide additional questions for the researcher to investigate.
-      + If the answers are sufficient to resolve the initial request, instruct the researcher to write a report.
-
-      Here are the context: 
-      <Collected information>
-      {context_str}
-      </Collected information>
-
-      <Conversation context>
-      {conversation_context}
-      </Conversation context>
-
-      {enhanced_prompt}
-
-      Now, provide your decision in the required format for this user request:
-      <User request>
-      {user_request}
-      </User request>
-      """
-    # Manually craft the prompt to avoid LLM hallucination
-    enhanced_prompt = ""
-    if total_questions == 0:
-        # Avoid writing a report without any research context
-        enhanced_prompt = """
-        
-        The student has no questions to research. Let start by asking some questions.
-        """
-    elif total_questions > 6:
-        # Avoid asking too many questions (when the data is not ready for writing a report)
-        enhanced_prompt = f"""
-
-        The student has researched {total_questions} questions. Should cancel the research if the context is not enough to write a report.
-        """
-
-    conversation_context = "\n".join(
-        [f"{message.role}: {message.content}" for message in memory.get_all()]
-    )
-    context_str = "\n".join(
-        [node.get_content(metadata_mode=MetadataMode.LLM) for node in context_nodes]
-    )
-    res = await Settings.llm.astructured_predict(
-        output_cls=AnalysisDecision,
-        prompt=PromptTemplate(template=analyze_prompt),
-        user_request=user_request,
-        context_str=context_str,
-        conversation_context=conversation_context,
-        enhanced_prompt=enhanced_prompt,
-    )
-    return res
-
-
-async def research(
-    question: str,
-    context_nodes: List[NodeWithScore],
-) -> str:
-    prompt = """
-    You are a researcher who is in the process of answering the question.
-    The purpose is to answer the question based on the collected information, without using prior knowledge or making up any new information.
-    Always add citations to the sentence/point/paragraph using the id of the provided content.
-    The citation should follow this format: [citation:id] where id is the id of the content.
-    
-    E.g:
-    If we have a context like this:
-    <Citation id='abc-xyz'>
-    Baby llama is called cria
-    </Citation id='abc-xyz'>
-
-    And your answer uses the content, then the citation should be:
-    - Baby llama is called cria [citation:abc-xyz]
-
-    Here is the provided context for the question:
-    <Collected information>
-    {context_str}
-    </Collected information>`
-
-    No prior knowledge, just use the provided context to answer the question: {question}
-    """
-    context_str = "\n".join(
-        [_get_text_node_content_for_citation(node) for node in context_nodes]
-    )
-    res = await Settings.llm.acomplete(
-        prompt=prompt.format(question=question, context_str=context_str),
-    )
-    return res.text
-
-
-async def write_report(
-    memory: SimpleComposableMemory,
-    user_request: str,
-    stream: bool = False,
-) -> CompletionResponse | CompletionResponseAsyncGen:
-    report_prompt = """
-    You are a researcher writing a report based on a user request and the research context.
-    You have researched various perspectives related to the user request.
-    The report should provide a comprehensive outline covering all important points from the researched perspectives.
-    Create a well-structured outline for the research report that covers all the answers.
-
-    # IMPORTANT when writing in markdown format:
-    + Use tables or figures where appropriate to enhance presentation.
-    + Preserve all citation syntax (the `[citation:id]()` parts in the provided context). Keep these citations in the final report - no separate reference section is needed.
-    + Do not add links, a table of contents, or a references section to the report.
-
-    <User request>
-    {user_request}
-    </User request>
-
-    <Research context>
-    {research_context}
-    </Research context>
-
-    Now, write a report addressing the user request based on the research provided following the format and guidelines above.
-    """
-    research_context = "\n".join(
-        [f"{message.role}: {message.content}" for message in memory.get_all()]
-    )
-
-    llm_complete_func = (
-        Settings.llm.astream_complete if stream else Settings.llm.acomplete
-    )
-
-    res = await llm_complete_func(
-        prompt=report_prompt.format(
-            user_request=user_request,
-            research_context=research_context,
-        ),
-    )
-    return res
-
-
-def _get_text_node_content_for_citation(node: NodeWithScore) -> str:
-    """
-    Construct node content for LLM with citation flag.
-    """
-    node_id = node.node.node_id
-    content = f"<Citation id='{node_id}'>\n{node.get_content(metadata_mode=MetadataMode.LLM)}</Citation id='{node_id}'>"
-    return content
@@ -1,59 +0,0 @@
-This is a [LlamaIndex](https://www.llamaindex.ai/) multi-agents project using [Workflows](https://docs.llamaindex.ai/en/stable/understanding/workflows/).
-
-## Getting Started
-
-First, setup the environment with poetry:
-
-> **_Note:_** This step is not needed if you are using the dev-container.
-
-```shell
-poetry install
-```
-
-Then check the parameters that have been pre-configured in the `.env` file in this directory.
-Make sure you have set the `OPENAI_API_KEY` for the LLM and the `E2B_API_KEY` for the code interpreter. You can get the E2B API key from [here](https://e2b.dev).
-
-Second, generate the embeddings of the documents in the `./data` directory:
-
-```shell
-poetry run generate
-```
-
-Third, run the development server:
-
-```shell
-poetry run dev
-```
-
-Then open [http://localhost:8000](http://localhost:8000) with your browser to start the chat UI.
-
-To start the app optimized for **production**, run:
-
-```
-poetry run prod
-```
-
-## Configure LLM and Embedding Model
-
-You can configure [LLM model](https://docs.llamaindex.ai/en/stable/module_guides/models/llms) and [embedding model](https://docs.llamaindex.ai/en/stable/module_guides/models/embeddings) in [settings.py](app/settings.py).
-
-## Use Case
-
-We have prepared an [example workflow](./app/workflow.py) for the financial report use case, where you can ask questions about the example documents in the [./data](./data) directory.
-
-You can start by sending an request on the [chat UI](http://localhost:8000) or you can test the `/api/chat` endpoint with the following curl request:
-
-```
-curl --location 'localhost:8000/api/chat' \
--header 'Content-Type: application/json' \
--data '{ "messages": [{ "role": "user", "content": "Create a report comparing the finances of Apple and Tesla" }] }'
-```
-
-## Learn More
-
-To learn more about LlamaIndex, take a look at the following resources:
-
- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex.
- [Workflows Introduction](https://docs.llamaindex.ai/en/stable/understanding/workflows/) - learn about LlamaIndex workflows.
-
-You can check out [the LlamaIndex GitHub repository](https://github.com/run-llama/llama_index) - your feedback and contributions are welcome!
@@ -1,333 +0,0 @@
-import os
-from typing import List, Optional
-
-from app.index import get_index
-from llama_index.core import Settings
-from llama_index.core.base.llms.types import ChatMessage, MessageRole
-from llama_index.core.llms.function_calling import FunctionCallingLLM
-from llama_index.core.memory import ChatMemoryBuffer
-from llama_index.core.tools import FunctionTool, QueryEngineTool, ToolSelection
-from llama_index.core.workflow import (
-    Context,
-    Event,
-    StartEvent,
-    StopEvent,
-    Workflow,
-    step,
-)
-from llama_index.server.api.models import AgentRunEvent, ChatRequest
-from llama_index.server.settings import server_settings
-from llama_index.server.tools.document_generator import DocumentGenerator
-from llama_index.server.tools.index import get_query_engine_tool
-from llama_index.server.tools.interpreter import E2BCodeInterpreter
-from llama_index.server.utils.agent_tool import (
-    call_tools,
-    chat_with_tools,
-)
-
-
-def create_workflow(chat_request: Optional[ChatRequest] = None) -> Workflow:
-    index = get_index(chat_request=chat_request)
-    if index is None:
-        raise ValueError(
-            "Index is not found. Try run generation script to create the index first."
-        )
-    query_engine_tool = get_query_engine_tool(index=index)
-    e2b_api_key = os.getenv("E2B_API_KEY")
-    if e2b_api_key is None:
-        raise ValueError(
-            "E2B_API_KEY is required to use the code interpreter tool. Please check README.md to know how to get the key."
-        )
-    code_interpreter_tool = E2BCodeInterpreter(api_key=e2b_api_key).to_tool()
-    document_generator_tool = DocumentGenerator(
-        file_server_url_prefix=server_settings.file_server_url_prefix,
-    ).to_tool()
-
-    return FinancialReportWorkflow(
-        query_engine_tool=query_engine_tool,
-        code_interpreter_tool=code_interpreter_tool,
-        document_generator_tool=document_generator_tool,
-        timeout=180,
-    )
-
-
-class InputEvent(Event):
-    input: List[ChatMessage]
-    response: bool = False
-
-
-class ResearchEvent(Event):
-    input: list[ToolSelection]
-
-
-class AnalyzeEvent(Event):
-    input: list[ToolSelection] | ChatMessage
-
-
-class ReportEvent(Event):
-    input: list[ToolSelection]
-
-
-class FinancialReportWorkflow(Workflow):
-    """
-    A workflow to generate a financial report using indexed documents.
-
-    Requirements:
-    - Indexed documents containing financial data and a query engine tool to search them
-    - A code interpreter tool to analyze data and generate reports
-    - A document generator tool to create report files
-
-    Steps:
-    1. LLM Input: The LLM determines the next step based on function calling.
-       For example, if the model requests the query engine tool, it returns a ResearchEvent;
-       if it requests document generation, it returns a ReportEvent.
-    2. Research: Uses the query engine to find relevant chunks from indexed documents.
-       After gathering information, it requests analysis (step 3).
-    3. Analyze: Uses a custom prompt to analyze research results and can call the code
-       interpreter tool for visualization or calculation. Returns results to the LLM.
-    4. Report: Uses the document generator tool to create a report. Returns results to the LLM.
-    """
-
-    _default_system_prompt = """
-    You are a financial analyst who are given a set of tools to help you.
-    It's good to using appropriate tools for the user request and always use the information from the tools, don't make up anything yourself.
-    For the query engine tool, you should break down the user request into a list of queries and call the tool with the queries.
-    """
-    stream: bool = True
-
-    def __init__(
-        self,
-        query_engine_tool: QueryEngineTool,
-        code_interpreter_tool: FunctionTool,
-        document_generator_tool: FunctionTool,
-        llm: Optional[FunctionCallingLLM] = None,
-        system_prompt: Optional[str] = None,
-        **kwargs,
-    ):
-        super().__init__(**kwargs)
-        self.system_prompt = system_prompt or self._default_system_prompt
-        self.query_engine_tool = query_engine_tool
-        self.code_interpreter_tool = code_interpreter_tool
-        self.document_generator_tool = document_generator_tool
-        assert query_engine_tool is not None, (
-            "Query engine tool is not found. Try run generation script or upload a document file first."
-        )
-        assert code_interpreter_tool is not None, "Code interpreter tool is required"
-        assert document_generator_tool is not None, (
-            "Document generator tool is required"
-        )
-        self.tools = [
-            self.query_engine_tool,
-            self.code_interpreter_tool,
-            self.document_generator_tool,
-        ]
-        self.llm: FunctionCallingLLM = llm or Settings.llm  # type: ignore
-        assert isinstance(self.llm, FunctionCallingLLM)
-        self.memory = ChatMemoryBuffer.from_defaults(llm=self.llm)
-
-    @step()
-    async def prepare_chat_history(self, ctx: Context, ev: StartEvent) -> InputEvent:
-        self.stream = ev.get("stream", True)
-        user_msg = ev.get("user_msg")
-        chat_history = ev.get("chat_history")
-
-        if chat_history is not None:
-            self.memory.put_messages(chat_history)
-
-        # Add user message to memory
-        self.memory.put(ChatMessage(role=MessageRole.USER, content=user_msg))
-
-        if self.system_prompt:
-            system_msg = ChatMessage(
-                role=MessageRole.SYSTEM, content=self.system_prompt
-            )
-            self.memory.put(system_msg)
-
-        return InputEvent(input=self.memory.get())
-
-    @step()
-    async def handle_llm_input(  # type: ignore
-        self,
-        ctx: Context,
-        ev: InputEvent,
-    ) -> ResearchEvent | AnalyzeEvent | ReportEvent | StopEvent:
-        """
-        Handle an LLM input and decide the next step.
-        """
-        # Always use the latest chat history from the input
-        chat_history: list[ChatMessage] = ev.input
-        # Get tool calls
-        response = await chat_with_tools(
-            self.llm,
-            self.tools,  # type: ignore
-            chat_history,
-        )
-        if not response.has_tool_calls():
-            if self.stream:
-                return StopEvent(result=response.generator)
-            else:
-                return StopEvent(result=await response.full_response())
-        # calling different tools at the same time is not supported at the moment
-        # add an error message to tell the AI to process step by step
-        if response.is_calling_different_tools():
-            self.memory.put(
-                ChatMessage(
-                    role=MessageRole.ASSISTANT,
-                    content="Cannot call different tools at the same time. Try calling one tool at a time.",
-                )
-            )
-            return InputEvent(input=self.memory.get())  # type: ignore
-        self.memory.put(response.tool_call_message)
-        match response.tool_name():
-            case self.code_interpreter_tool.metadata.name:
-                return AnalyzeEvent(input=response.tool_calls)
-            case self.document_generator_tool.metadata.name:
-                return ReportEvent(input=response.tool_calls)
-            case self.query_engine_tool.metadata.name:
-                return ResearchEvent(input=response.tool_calls)
-            case _:
-                raise ValueError(f"Unknown tool: {response.tool_name()}")
-
-    @step()
-    async def research(self, ctx: Context, ev: ResearchEvent) -> AnalyzeEvent:
-        """
-        Do a research to gather information for the user's request.
-        A researcher should have these tools: query engine, search engine, etc.
-        """
-        ctx.write_event_to_stream(
-            AgentRunEvent(
-                name="Researcher",
-                msg="Starting research",
-            )
-        )
-        tool_calls = ev.input
-
-        tool_call_outputs = await call_tools(
-            ctx=ctx,
-            agent_name="Researcher",
-            tools=[self.query_engine_tool],
-            tool_calls=tool_calls,
-        )
-        for tool_call_output in tool_call_outputs:
-            self.memory.put(
-                ChatMessage(
-                    role=MessageRole.TOOL,
-                    content=tool_call_output.tool_output.content,
-                    additional_kwargs={
-                        "name": tool_call_output.tool_output.tool_name,
-                        "tool_call_id": tool_call_output.tool_call_id,
-                    },
-                )
-            )
-        return AnalyzeEvent(
-            input=ChatMessage(
-                role=MessageRole.ASSISTANT,
-                content="Researcher: I've finished the research. Please analyze the result.",
-            ),
-        )
-
-    @step()
-    async def analyze(self, ctx: Context, ev: AnalyzeEvent) -> InputEvent:
-        """
-        Analyze the research result.
-        """
-        ctx.write_event_to_stream(
-            AgentRunEvent(
-                name="Analyst",
-                msg="Starting analysis",
-            )
-        )
-        event_requested_by_workflow_llm = isinstance(ev.input, list)
-        # Requested by the workflow LLM Input step, it's a tool call
-        if event_requested_by_workflow_llm:
-            # Set the tool calls
-            tool_calls = ev.input
-        else:
-            # Otherwise, it's triggered by the research step
-            # Use a custom prompt and independent memory for the analyst agent
-            analysis_prompt = """
-            You are a financial analyst, you are given a research result and a set of tools to help you.
-            Always use the given information, don't make up anything yourself. If there is not enough information, you can asking for more information.
-            If you have enough numerical information, it's good to include some charts/visualizations to the report so you can use the code interpreter tool to generate a report.
-            """
-            # This is handled by analyst agent
-            # Clone the shared memory to avoid conflicting with the workflow.
-            chat_history = self.memory.get()
-            chat_history.append(
-                ChatMessage(role=MessageRole.SYSTEM, content=analysis_prompt)
-            )
-            chat_history.append(ev.input)  # type: ignore
-            # Check if the analyst agent needs to call tools
-            response = await chat_with_tools(
-                self.llm,
-                [self.code_interpreter_tool],
-                chat_history,
-            )
-            if not response.has_tool_calls():
-                # If no tool call, fallback analyst message to the workflow
-                msg_content = await response.full_response()
-                analyst_msg = ChatMessage(
-                    role=MessageRole.ASSISTANT,
-                    content=f"Analyst: "
-                    f"\nHere is the analysis result: {msg_content}"
-                    "\nUse it for other steps or response the content to the user.",
-                )
-                self.memory.put(analyst_msg)
-                return InputEvent(input=self.memory.get())
-            else:
-                tool_calls = response.tool_calls
-                self.memory.put(response.tool_call_message)
-
-        # Call tools
-        tool_call_outputs = await call_tools(
-            ctx=ctx,
-            agent_name="Analyst",
-            tools=[self.code_interpreter_tool],
-            tool_calls=tool_calls,  # type: ignore
-        )
-        for tool_call_output in tool_call_outputs:
-            self.memory.put(
-                ChatMessage(
-                    role=MessageRole.TOOL,
-                    content=tool_call_output.tool_output.content,
-                    additional_kwargs={
-                        "name": tool_call_output.tool_output.tool_name,
-                        "tool_call_id": tool_call_output.tool_call_id,
-                    },
-                )
-            )
-
-        # Fallback to the input with the latest chat history
-        return InputEvent(input=self.memory.get())
-
-    @step()
-    async def report(self, ctx: Context, ev: ReportEvent) -> InputEvent:
-        """
-        Generate a report based on the analysis result.
-        """
-        ctx.write_event_to_stream(
-            AgentRunEvent(
-                name="Reporter",
-                msg="Starting report generation",
-            )
-        )
-        tool_calls = ev.input
-        tool_call_outputs = await call_tools(
-            ctx=ctx,
-            agent_name="Reporter",
-            tools=[self.document_generator_tool],
-            tool_calls=tool_calls,
-        )
-        for tool_call_output in tool_call_outputs:
-            self.memory.put(
-                ChatMessage(
-                    role=MessageRole.TOOL,
-                    content=tool_call_output.tool_output.content,
-                    additional_kwargs={
-                        "name": tool_call_output.tool_output.tool_name,
-                        "tool_call_id": tool_call_output.tool_call_id,
-                    },
-                )
-            )
-        # After the tool calls, fallback to the input with the latest chat history
-        return InputEvent(input=self.memory.get())
@@ -1,52 +0,0 @@
-This is a [LlamaIndex](https://www.llamaindex.ai/) project bootstrapped with [`create-llama`](https://github.com/run-llama/LlamaIndexTS/tree/main/packages/create-llama).
-
-## Getting Started
-
-First, install the dependencies:
-
-```
-npm install
-```
-
-Then check the parameters that have been pre-configured in the `.env` file in this directory.
-Make sure you have set the `OPENAI_API_KEY` for the LLM.
-
-Second, generate the embeddings of the example documents in the `./data` directory:
-
-```
-npm run generate
-```
-
-Third, run the development server:
-
-```
-npm run dev
-```
-
-Open [http://localhost:3000](http://localhost:3000) with your browser to see the chat UI.
-
-## Configure LLM and Embedding Model
-
-You can configure [LLM model](https://ts.llamaindex.ai/docs/llamaindex/modules/llms) and [embedding model](https://ts.llamaindex.ai/docs/llamaindex/modules/embeddings) in the [settings file](src/app/settings.ts).
-
-## Use Case
-
-We have prepared an [example workflow](./src/app/workflow.ts) for the agentic RAG use case, where you can ask questions about the example documents in the [./data](./data) directory.
-
-You can start by sending an request on the [chat UI](http://localhost:3000) or you can test the `/api/chat` endpoint with the following curl request:
-
-```shell
-curl --location 'localhost:3000/api/chat' \
--header 'Content-Type: application/json' \
--data '{ "messages": [{ "role": "user", "content": "What standards for a letter exist?" }] }'
-```
-
-## Learn More
-
-To learn more about LlamaIndex, take a look at the following resources:
-
- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex (Python features).
- [LlamaIndexTS Documentation](https://ts.llamaindex.ai/docs/llamaindex) - learn about LlamaIndex (Typescript features).
- [Agent Workflows Introduction](https://ts.llamaindex.ai/docs/llamaindex/modules/agent_workflow) - learn about LlamaIndexTS Agent Workflows.
-
-You can check out [the LlamaIndexTS GitHub repository](https://github.com/run-llama/LlamaIndexTS) - your feedback and contributions are welcome!
@@ -1,16 +0,0 @@
-import { agent } from "llamaindex";
-import { getIndex } from "./data";
-
-export const workflowFactory = async (reqBody: any) => {
-  const index = await getIndex(reqBody?.data);
-
-  const queryEngineTool = index.queryTool({
-    metadata: {
-      name: "query_document",
-      description: `This tool can retrieve information about Apple and Tesla financial data`,
-    },
-    includeSourceNodes: true,
-  });
-
-  return agent({ tools: [queryEngineTool] });
-};
@@ -1,56 +0,0 @@
-This is a [LlamaIndex](https://www.llamaindex.ai/) project bootstrapped with [`create-llama`](https://github.com/run-llama/LlamaIndexTS/tree/main/packages/create-llama).
-
-## Getting Started
-
-First, install the dependencies:
-
-```
-npm install
-```
-
-Then check the parameters that have been pre-configured in the `.env` file in this directory.
-Make sure you have set the `OPENAI_API_KEY` for the LLM.
-
-Second, generate the embeddings of the example documents in the `./data` directory:
-
-```
-npm run generate
-```
-
-Third, run the development server:
-
-```
-npm run dev
-```
-
-Open [http://localhost:3000](http://localhost:3000) with your browser to see the chat UI.
-
-## Configure LLM and Embedding Model
-
-You can configure [LLM model](https://ts.llamaindex.ai/docs/llamaindex/modules/llms) and [embedding model](https://ts.llamaindex.ai/docs/llamaindex/modules/embeddings) in the [settings file](src/app/settings.ts).
-
-## Custom UI Components
-
-For Deep Research, we have a custom component located in `components/deep_research_event.jsx`. This is used to display the results of the deep research workflow in a more user-friendly way
-
-## Use Case
-
-We have prepared an [example workflow](./src/app/workflow.ts) for the Deep Research use case, where you can request a detailed answer about the example documents in the [./data](./data) directory.
-
-You can start by sending an request on the [chat UI](http://localhost:3000) or you can test the `/api/chat` endpoint with the following curl request:
-
-```shell
-curl --location 'localhost:3000/api/chat' \
--header 'Content-Type: application/json' \
--data '{ "messages": [{ "role": "user", "content": "Compare the financial performance of Apple and Tesla" }] }'
-```
-
-## Learn More
-
-To learn more about LlamaIndex, take a look at the following resources:
-
- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex (Python features).
- [LlamaIndexTS Documentation](https://ts.llamaindex.ai/docs/llamaindex) - learn about LlamaIndex (Typescript features).
- [Workflows Introduction](https://ts.llamaindex.ai/docs/llamaindex/modules/workflows) - learn about LlamaIndexTS workflows.
-
-You can check out [the LlamaIndexTS GitHub repository](https://github.com/run-llama/LlamaIndexTS) - your feedback and contributions are welcome!
@@ -1,425 +0,0 @@
-import {
-  DeepResearchEvent,
-  toSourceEvent,
-  toStreamGenerator,
-} from "@llamaindex/server";
-import {
-  AgentInputData,
-  AgentWorkflowContext,
-  ChatMemoryBuffer,
-  ChatResponseChunk,
-  HandlerContext,
-  LlamaCloudIndex,
-  Metadata,
-  MetadataMode,
-  NodeWithScore,
-  PromptTemplate,
-  Settings,
-  StartEvent,
-  StopEvent as StopEventBase,
-  ToolCallLLM,
-  VectorStoreIndex,
-  Workflow,
-  WorkflowEvent,
-} from "llamaindex";
-import { randomUUID } from "node:crypto";
-import { z } from "zod";
-import { getIndex } from "./data";
-
-// workflow factory
-export const workflowFactory = async (reqBody: any) => {
-  const index = await getIndex(reqBody?.data);
-  return new DeepResearchWorkflow(index);
-};
-
-// workflow configs
-const MAX_QUESTIONS = 6; // max number of questions to research, research will stop when this number is reached
-const TIMEOUT = 360; // timeout in seconds
-const TOP_K = 10; // number of nodes to retrieve from the vector store
-
-const createPlanResearchPrompt = new PromptTemplate({
-  template: `
-You are a professor who is guiding a researcher to research a specific request/problem.
-Your task is to decide on a research plan for the researcher.
-
-The possible actions are:
-+ Provide a list of questions for the researcher to investigate, with the purpose of clarifying the request.
-+ Write a report if the researcher has already gathered enough research on the topic and can resolve the initial request.
-+ Cancel the research if most of the answers from researchers indicate there is insufficient information to research the request. Do not attempt more than 3 research iterations or too many questions.
-
-The workflow should be:
-+ Always begin by providing some initial questions for the researcher to investigate.
-+ Analyze the provided answers against the initial topic/request. If the answers are insufficient to resolve the initial request, provide additional questions for the researcher to investigate.
-+ If the answers are sufficient to resolve the initial request, instruct the researcher to write a report.
-
-Here are the context: 
-<Collected information>
-{context_str}
-</Collected information>
-
-<Conversation context>
-{conversation_context}
-</Conversation context>
-
-{enhanced_prompt}
-
-Now, provide your decision in the required format for this user request:
-<User request>
-{user_request}
-</User request>
-`,
-  templateVars: [
-    "context_str",
-    "conversation_context",
-    "enhanced_prompt",
-    "user_request",
-  ],
-});
-
-const researchPrompt = new PromptTemplate({
-  template: `
-You are a researcher who is in the process of answering the question.
-The purpose is to answer the question based on the collected information, without using prior knowledge or making up any new information.
-Always add citations to the sentence/point/paragraph using the id of the provided content.
-The citation should follow this format: [citation:id] where id is the id of the content.
-
-E.g:
-If we have a context like this:
-<Citation id='abc-xyz'>
-Baby llama is called cria
-</Citation id='abc-xyz'>
-
-And your answer uses the content, then the citation should be:
- Baby llama is called cria [citation:abc-xyz]
-
- Here is the provided context for the question:
-<Collected information>
-{context_str}
-</Collected information>
-
-No prior knowledge, just use the provided context to answer the question: {question}
-`,
-  templateVars: ["context_str", "question"],
-});
-
-const WRITE_REPORT_PROMPT = `
-You are a researcher writing a report based on a user request and the research context.
-You have researched various perspectives related to the user request.
-The report should provide a comprehensive outline covering all important points from the researched perspectives.
-Create a well-structured outline for the research report that covers all the answers.
-
-# IMPORTANT when writing in markdown format:
-+ Use tables or figures where appropriate to enhance presentation.
-+ Preserve all citation syntax (the \`[citation:id]()\` parts in the provided context). Keep these citations in the final report - no separate reference section is needed.
-+ Do not add links, a table of contents, or a references section to the report.
-`;
-
-// workflow events
-type ResearchQuestion = { questionId: string; question: string };
-type ResearchResult = ResearchQuestion & { answer: string };
-
-class PlanResearchEvent extends WorkflowEvent<{}> {}
-class ResearchEvent extends WorkflowEvent<ResearchQuestion[]> {}
-class ReportEvent extends WorkflowEvent<{}> {}
-class StopEvent extends StopEventBase<AsyncGenerator<ChatResponseChunk>> {}
-
-// workflow definition
-class DeepResearchWorkflow extends Workflow<
-  AgentWorkflowContext,
-  AgentInputData,
-  string
-> {
-  #llm = Settings.llm as ToolCallLLM;
-  #index?: VectorStoreIndex | LlamaCloudIndex;
-
-  userRequest: string = "";
-  totalQuestions: number = 0;
-  contextNodes: NodeWithScore<Metadata>[] = [];
-  memory: ChatMemoryBuffer = new ChatMemoryBuffer({ llm: Settings.llm });
-
-  constructor(index: VectorStoreIndex | LlamaCloudIndex) {
-    super({ timeout: TIMEOUT });
-    this.#index = index;
-    this.addWorkflowSteps();
-  }
-
-  addWorkflowSteps() {
-    this.addStep(
-      {
-        inputs: [StartEvent<AgentInputData>],
-        outputs: [PlanResearchEvent],
-      },
-      this.handleStartWorkflow,
-    );
-    this.addStep(
-      {
-        inputs: [PlanResearchEvent],
-        outputs: [ResearchEvent, ReportEvent, StopEvent],
-      },
-      this.handlePlanResearch,
-    );
-    this.addStep(
-      {
-        inputs: [ResearchEvent],
-        outputs: [PlanResearchEvent],
-      },
-      this.handleResearch,
-    );
-    this.addStep(
-      {
-        inputs: [ReportEvent],
-        outputs: [StopEvent],
-      },
-      this.handleReport,
-    );
-  }
-
-  async initWorkflow(data: AgentInputData) {
-    const { userInput, chatHistory = [] } = data;
-    if (!userInput) throw new Error("Invalid input");
-
-    this.userRequest = userInput;
-
-    await this.memory.set(chatHistory);
-    await this.memory.put({ role: "user", content: userInput });
-  }
-
-  handleStartWorkflow = async (
-    ctx: HandlerContext<AgentWorkflowContext>,
-    ev: StartEvent<AgentInputData>,
-  ): Promise<PlanResearchEvent> => {
-    await this.initWorkflow(ev.data);
-
-    ctx.sendEvent(
-      new DeepResearchEvent({
-        type: "deep_research_event",
-        data: { event: "retrieve", state: "inprogress" },
-      }),
-    );
-
-    const retrievedNodes = await this.retriever.retrieve(this.userRequest);
-
-    ctx.sendEvent(toSourceEvent(retrievedNodes));
-
-    ctx.sendEvent(
-      new DeepResearchEvent({
-        type: "deep_research_event",
-        data: { event: "retrieve", state: "done" },
-      }),
-    );
-
-    this.contextNodes = retrievedNodes;
-
-    return new PlanResearchEvent({});
-  };
-
-  handlePlanResearch = async (
-    ctx: HandlerContext<AgentWorkflowContext>,
-    ev: PlanResearchEvent,
-  ): Promise<ResearchEvent | ReportEvent | StopEvent> => {
-    ctx.sendEvent(
-      new DeepResearchEvent({
-        type: "deep_research_event",
-        data: { event: "analyze", state: "inprogress" },
-      }),
-    );
-
-    const { decision, researchQuestions, cancelReason } =
-      await this.createResearchPlan();
-
-    // Stop workflow due to decision from LLM
-    if (decision === "cancel") {
-      ctx.sendEvent(
-        new DeepResearchEvent({
-          type: "deep_research_event",
-          data: { event: "analyze", state: "done" },
-        }),
-      );
-      return new StopEvent(
-        toStreamGenerator(
-          cancelReason ?? "Research cancelled without any reason.",
-        ),
-      );
-    }
-
-    // Trigger research from generated questions
-    if (decision === "research") {
-      this.memory.put({
-        role: "assistant",
-        content:
-          "We need to find answers to the following questions:\n" +
-          researchQuestions.join("\n"),
-      });
-
-      researchQuestions.forEach(({ questionId: id, question }) => {
-        ctx.sendEvent(
-          new DeepResearchEvent({
-            type: "deep_research_event",
-            data: { event: "answer", state: "pending", id, question },
-          }),
-        );
-      });
-
-      return new ResearchEvent(researchQuestions);
-    }
-
-    // Resarch done, start writing report
-    this.memory.put({
-      role: "assistant",
-      content: "No more idea to analyze. We should report the answers.",
-    });
-
-    ctx.sendEvent(
-      new DeepResearchEvent({
-        type: "deep_research_event",
-        data: { event: "analyze", state: "done" },
-      }),
-    );
-
-    return new ReportEvent({});
-  };
-
-  handleResearch = async (
-    ctx: HandlerContext<AgentWorkflowContext>,
-    ev: ResearchEvent,
-  ): Promise<PlanResearchEvent> => {
-    const researchQuestions = ev.data;
-
-    // Answer questions in parallel
-    const researchResults: ResearchResult[] = await Promise.all(
-      researchQuestions.map(async ({ questionId: id, question }) => {
-        ctx.sendEvent(
-          new DeepResearchEvent({
-            type: "deep_research_event",
-            data: { event: "answer", state: "inprogress", id, question },
-          }),
-        );
-
-        const answer = await this.answerQuestion(question);
-
-        ctx.sendEvent(
-          new DeepResearchEvent({
-            type: "deep_research_event",
-            data: { event: "answer", state: "done", id, question, answer },
-          }),
-        );
-
-        return { questionId: id, question, answer };
-      }),
-    );
-
-    // Save answers to memory
-    researchResults.forEach(({ question, answer }) => {
-      this.memory.put({
-        role: "assistant",
-        content: `<Question>${question}</Question>\n<Answer>${answer}</Answer>`,
-      });
-    });
-
-    this.memory.put({
-      role: "assistant",
-      content:
-        "Researched all the questions. Now, I need to analyze if it's ready to write a report or need to research more.",
-    });
-
-    this.totalQuestions += researchResults.length;
-
-    return new PlanResearchEvent({});
-  };
-
-  handleReport = async (
-    ctx: HandlerContext<AgentWorkflowContext>,
-    ev: ReportEvent,
-  ): Promise<StopEvent> => {
-    const chatHistory = await this.memory.getAllMessages();
-
-    const messages = chatHistory.concat([
-      {
-        role: "system",
-        content: WRITE_REPORT_PROMPT,
-      },
-      {
-        role: "user",
-        content:
-          "Write a report addressing the user request based on the research provided the context",
-      },
-    ]);
-
-    const stream = await this.llm.chat({ messages, stream: true });
-
-    return new StopEvent(toStreamGenerator(stream));
-  };
-
-  get llm() {
-    if (!this.#llm.supportToolCall) throw new Error("LLM is not a ToolCallLLM");
-    return this.#llm;
-  }
-
-  get retriever() {
-    if (!this.#index) throw new Error("Index is not initialized");
-    return this.#index.asRetriever({ similarityTopK: TOP_K });
-  }
-
-  get contextStr() {
-    return this.contextNodes
-      .map((node) => {
-        const nodeId = node.node.id_;
-        const nodeContent = node.node.getContent(MetadataMode.NONE);
-        return `<Citation id='${nodeId}'>\n${nodeContent}</Citation id='${nodeId}'>`;
-      })
-      .join("\n");
-  }
-
-  get enhancedPrompt() {
-    if (this.totalQuestions === 0) {
-      return "The student has no questions to research. Let start by asking some questions.";
-    }
-
-    if (this.totalQuestions > MAX_QUESTIONS) {
-      return `The student has researched ${this.totalQuestions} questions. Should cancel the research if the context is not enough to write a report.`;
-    }
-
-    return "";
-  }
-
-  async createResearchPlan() {
-    const chatHistory = await this.memory.getMessages();
-
-    const conversationContext = chatHistory
-      .map((message) => `${message.role}: ${message.content}`)
-      .join("\n");
-
-    const prompt = createPlanResearchPrompt.format({
-      context_str: this.contextStr,
-      conversation_context: conversationContext,
-      enhanced_prompt: this.enhancedPrompt,
-      user_request: this.userRequest,
-    });
-
-    const responseFormat = z.object({
-      decision: z.enum(["research", "write", "cancel"]),
-      researchQuestions: z.array(z.string()),
-      cancelReason: z.string().optional(),
-    });
-
-    const result = await this.llm.complete({ prompt, responseFormat });
-    const plan = JSON.parse(result.text) as z.infer<typeof responseFormat>;
-
-    return {
-      ...plan,
-      researchQuestions: plan.researchQuestions.map((question) => ({
-        questionId: randomUUID(),
-        question,
-      })),
-    };
-  }
-
-  async answerQuestion(question: string) {
-    const prompt = researchPrompt.format({
-      context_str: this.contextStr,
-      question,
-    });
-    const result = await this.llm.complete({ prompt });
-    return result.text;
-  }
-}
@@ -1,52 +0,0 @@
-This is a [LlamaIndex](https://www.llamaindex.ai/) project bootstrapped with [`create-llama`](https://github.com/run-llama/LlamaIndexTS/tree/main/packages/create-llama).
-
-## Getting Started
-
-First, install the dependencies:
-
-```
-npm install
-```
-
-Then check the parameters that have been pre-configured in the `.env` file in this directory.
-Make sure you have set the `OPENAI_API_KEY` for the LLM and the `E2B_API_KEY` for the code interpreter. You can get the E2B API key from [here](https://e2b.dev).
-
-Second, generate the embeddings of the example documents in the `./data` directory:
-
-```
-npm run generate
-```
-
-Third, run the development server:
-
-```
-npm run dev
-```
-
-Open [http://localhost:3000](http://localhost:3000) with your browser to see the chat UI.
-
-## Configure LLM and Embedding Model
-
-You can configure [LLM model](https://ts.llamaindex.ai/docs/llamaindex/modules/llms) and [embedding model](https://ts.llamaindex.ai/docs/llamaindex/modules/embeddings) in the [settings file](src/app/settings.ts).
-
-## Use Case
-
-We have prepared an [example workflow](./src/app/workflow.ts) for the financial report generation use case, where you can request for the app to generate a financial report using the example documents in the [./data](./data).
-
-You can start by sending an request on the [chat UI](http://localhost:3000) or you can test the `/api/chat` endpoint with the following curl request:
-
-```shell
-curl --location 'localhost:3000/api/chat' \
--header 'Content-Type: application/json' \
--data '{ "messages": [{ "role": "user", "content": "Generate a financial report that compares the financial performance of Apple and Tesla" }] }'
-```
-
-## Learn More
-
-To learn more about LlamaIndex, take a look at the following resources:
-
- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex (Python features).
- [LlamaIndexTS Documentation](https://ts.llamaindex.ai/docs/llamaindex) - learn about LlamaIndex (Typescript features).
- [Workflows Introduction](https://ts.llamaindex.ai/docs/llamaindex/modules/workflows) - learn about LlamaIndexTS workflows.
-
-You can check out [the LlamaIndexTS GitHub repository](https://github.com/run-llama/LlamaIndexTS) - your feedback and contributions are welcome!
@@ -1,396 +0,0 @@
-import { toAgentRunEvent, toSourceEvent } from "@llamaindex/server";
-import {
-  callTools,
-  chatWithTools,
-  documentGenerator,
-  interpreter,
-} from "@llamaindex/tools";
-import {
-  AgentInputData,
-  AgentWorkflowContext,
-  BaseToolWithCall,
-  ChatMemoryBuffer,
-  ChatMessage,
-  ChatResponseChunk,
-  HandlerContext,
-  Metadata,
-  NodeWithScore,
-  Settings,
-  StartEvent,
-  StopEvent,
-  ToolCall,
-  ToolCallLLM,
-  Workflow,
-  WorkflowEvent,
-} from "llamaindex";
-import { getIndex } from "./data";
-
-const TIMEOUT = 360 * 1000;
-
-export async function workflowFactory(reqBody: any) {
-  const index = await getIndex(reqBody?.data);
-
-  const queryEngineTool = index.queryTool({
-    metadata: {
-      name: "query_document",
-      description: `This tool can retrieve information about Apple and Tesla financial data`,
-    },
-    includeSourceNodes: true,
-  });
-
-  if (!process.env.E2B_API_KEY) {
-    throw new Error("E2B_API_KEY is required to use the code interpreter tool");
-  }
-
-  const codeInterpreterTool = interpreter({
-    apiKey: process.env.E2B_API_KEY!,
-  });
-  const documentGeneratorTool = documentGenerator();
-
-  return new FinancialReportWorkflow({
-    queryEngineTool,
-    codeInterpreterTool,
-    documentGeneratorTool,
-    timeout: TIMEOUT,
-  });
-}
-
-// Create a custom event type
-class InputEvent extends WorkflowEvent<{ input: ChatMessage[] }> {}
-
-class ResearchEvent extends WorkflowEvent<{
-  toolCalls: ToolCall[];
-}> {}
-
-class AnalyzeEvent extends WorkflowEvent<{
-  input: ChatMessage | ToolCall[];
-}> {}
-
-class ReportGenerationEvent extends WorkflowEvent<{
-  toolCalls: ToolCall[];
-}> {}
-
-const DEFAULT_SYSTEM_PROMPT = `
-You are a financial analyst who are given a set of tools to help you.
-It's good to using appropriate tools for the user request and always use the information from the tools, don't make up anything yourself.
-For the query engine tool, you should break down the user request into a list of queries and call the tool with the queries.
-`;
-
-class FinancialReportWorkflow extends Workflow<
-  AgentWorkflowContext,
-  AgentInputData,
-  string
-> {
-  llm: ToolCallLLM;
-  memory: ChatMemoryBuffer;
-  queryEngineTool: BaseToolWithCall;
-  codeInterpreterTool: BaseToolWithCall;
-  documentGeneratorTool: BaseToolWithCall;
-  systemPrompt?: string;
-
-  constructor(options: {
-    queryEngineTool: BaseToolWithCall;
-    codeInterpreterTool: BaseToolWithCall;
-    documentGeneratorTool: BaseToolWithCall;
-    systemPrompt?: string;
-    verbose?: boolean;
-    timeout?: number;
-  }) {
-    super({
-      verbose: options?.verbose ?? false,
-      timeout: options?.timeout ?? 360,
-    });
-
-    this.llm = Settings.llm as ToolCallLLM;
-    if (!this.llm.supportToolCall) {
-      throw new Error("LLM is not a ToolCallLLM");
-    }
-    this.systemPrompt = options.systemPrompt ?? DEFAULT_SYSTEM_PROMPT;
-    this.queryEngineTool = options.queryEngineTool;
-    this.codeInterpreterTool = options.codeInterpreterTool;
-
-    this.documentGeneratorTool = options.documentGeneratorTool;
-    this.memory = new ChatMemoryBuffer({ llm: this.llm, chatHistory: [] });
-
-    // Add steps
-    this.addStep(
-      {
-        inputs: [StartEvent<AgentInputData>],
-        outputs: [InputEvent],
-      },
-      this.prepareChatHistory,
-    );
-
-    this.addStep(
-      {
-        inputs: [InputEvent],
-        outputs: [
-          InputEvent,
-          ResearchEvent,
-          AnalyzeEvent,
-          ReportGenerationEvent,
-          StopEvent,
-        ],
-      },
-      this.handleLLMInput,
-    );
-
-    this.addStep(
-      {
-        inputs: [ResearchEvent],
-        outputs: [AnalyzeEvent],
-      },
-      this.handleResearch,
-    );
-
-    this.addStep(
-      {
-        inputs: [AnalyzeEvent],
-        outputs: [InputEvent],
-      },
-      this.handleAnalyze,
-    );
-
-    this.addStep(
-      {
-        inputs: [ReportGenerationEvent],
-        outputs: [InputEvent],
-      },
-      this.handleReportGeneration,
-    );
-  }
-
-  prepareChatHistory = async (
-    ctx: HandlerContext<AgentWorkflowContext>,
-    ev: StartEvent<AgentInputData>,
-  ): Promise<InputEvent> => {
-    const { userInput, chatHistory = [] } = ev.data;
-    if (!userInput) throw new Error("Invalid input");
-
-    this.memory.set(chatHistory);
-
-    if (this.systemPrompt) {
-      this.memory.put({ role: "system", content: this.systemPrompt });
-    }
-
-    this.memory.put({ role: "user", content: userInput });
-
-    const messages = await this.memory.getMessages();
-    return new InputEvent({ input: messages });
-  };
-
-  handleLLMInput = async (
-    ctx: HandlerContext<AgentWorkflowContext>,
-    ev: InputEvent,
-  ): Promise<
-    | InputEvent
-    | ResearchEvent
-    | AnalyzeEvent
-    | ReportGenerationEvent
-    | StopEvent<AsyncGenerator<ChatResponseChunk, any, any> | undefined>
-  > => {
-    const chatHistory = ev.data.input;
-
-    const tools = [
-      this.codeInterpreterTool,
-      this.documentGeneratorTool,
-      this.queryEngineTool,
-    ];
-
-    const toolCallResponse = await chatWithTools(this.llm, tools, chatHistory);
-
-    if (!toolCallResponse.hasToolCall()) {
-      return new StopEvent(toolCallResponse.responseGenerator);
-    }
-
-    if (toolCallResponse.hasMultipleTools()) {
-      this.memory.put({
-        role: "assistant",
-        content:
-          "Calling different tools is not allowed. Please only use multiple calls of the same tool.",
-      });
-      const chatHistory = await this.memory.getMessages();
-      return new InputEvent({ input: chatHistory });
-    }
-
-    // Put the LLM tool call message into the memory
-    // And trigger the next step according to the tool call
-    if (toolCallResponse.toolCallMessage) {
-      this.memory.put(toolCallResponse.toolCallMessage);
-    }
-    const toolName = toolCallResponse.getToolNames()[0];
-    switch (toolName) {
-      case this.codeInterpreterTool.metadata.name:
-        return new AnalyzeEvent({
-          input: toolCallResponse.toolCalls,
-        });
-      case this.documentGeneratorTool.metadata.name:
-        return new ReportGenerationEvent({
-          toolCalls: toolCallResponse.toolCalls,
-        });
-      default:
-        if (this.queryEngineTool.metadata.name === toolName) {
-          return new ResearchEvent({
-            toolCalls: toolCallResponse.toolCalls,
-          });
-        }
-        throw new Error(`Unknown tool: ${toolName}`);
-    }
-  };
-
-  handleResearch = async (
-    ctx: HandlerContext<AgentWorkflowContext>,
-    ev: ResearchEvent,
-  ): Promise<AnalyzeEvent> => {
-    ctx.sendEvent(
-      toAgentRunEvent({
-        agent: "Researcher",
-        text: "Researching data",
-        type: "text",
-      }),
-    );
-
-    const { toolCalls } = ev.data;
-
-    const toolMsgs = await callTools({
-      tools: [this.queryEngineTool],
-      toolCalls,
-      writeEvent: (text, step) => {
-        ctx.sendEvent(
-          toAgentRunEvent({
-            agent: "Researcher",
-            text,
-            type: toolCalls.length > 1 ? "progress" : "text",
-            current: step,
-            total: toolCalls.length,
-          }),
-        );
-      },
-    });
-    for (const toolMsg of toolMsgs) {
-      this.memory.put(toolMsg);
-    }
-
-    const sourcesNodes: NodeWithScore<Metadata>[] = toolMsgs
-      .map((msg) => (msg.options as any)?.toolResult?.result?.sourceNodes)
-      .flat()
-      .filter(Boolean);
-
-    if (sourcesNodes.length > 0) {
-      ctx.sendEvent(toSourceEvent(sourcesNodes));
-    }
-
-    return new AnalyzeEvent({
-      input: {
-        role: "assistant",
-        content:
-          "I have finished researching the data, please analyze the data.",
-      },
-    });
-  };
-
-  /**
-   * Analyze a research result or a tool call for code interpreter from the LLM
-   */
-  handleAnalyze = async (
-    ctx: HandlerContext<AgentWorkflowContext>,
-    ev: AnalyzeEvent,
-  ): Promise<InputEvent> => {
-    ctx.sendEvent(
-      toAgentRunEvent({
-        agent: "Analyst",
-        text: "Analyzing data",
-        type: "text",
-      }),
-    );
-    // Request by workflow LLM, input is a list of tool calls
-    let toolCalls: ToolCall[] = [];
-    if (Array.isArray(ev.data.input)) {
-      toolCalls = ev.data.input;
-    } else {
-      // Requested by Researcher, input is a ChatMessage
-      // We start new LLM chat specifically for analyzing the data
-      const analysisPrompt = `
-      You are an expert in analyzing financial data.
-      You are given a set of financial data to analyze. Your task is to analyze the financial data and return a report.
-      Your response should include a detailed analysis of the financial data, including any trends, patterns, or insights that you find.
-      Construct the analysis in textual format; including tables would be great!
-      Don't need to synthesize the data, just analyze and provide your findings.
-      `;
-
-      // Clone the current chat history
-      // Add the analysis system prompt and the message from the researcher
-      const chatHistory = await this.memory.getMessages();
-      const newChatHistory = [
-        ...chatHistory,
-        { role: "system", content: analysisPrompt },
-        ev.data.input,
-      ];
-      const toolCallResponse = await chatWithTools(
-        this.llm,
-        [this.codeInterpreterTool],
-        newChatHistory as ChatMessage[],
-      );
-
-      if (!toolCallResponse.hasToolCall()) {
-        this.memory.put(await toolCallResponse.asFullResponse());
-        const chatHistory = await this.memory.getMessages();
-        return new InputEvent({ input: chatHistory });
-      } else {
-        this.memory.put(toolCallResponse.toolCallMessage!);
-        toolCalls = toolCallResponse.toolCalls;
-      }
-    }
-
-    // Call the tools
-    const toolMsgs = await callTools({
-      tools: [this.codeInterpreterTool],
-      toolCalls,
-      writeEvent: (text, step) => {
-        ctx.sendEvent(
-          toAgentRunEvent({
-            agent: "Analyst",
-            text,
-            type: toolCalls.length > 1 ? "progress" : "text",
-            current: step,
-            total: toolCalls.length,
-          }),
-        );
-      },
-    });
-    for (const toolMsg of toolMsgs) {
-      this.memory.put(toolMsg);
-    }
-
-    const chatHistory = await this.memory.getMessages();
-    return new InputEvent({ input: chatHistory });
-  };
-
-  handleReportGeneration = async (
-    ctx: HandlerContext<AgentWorkflowContext>,
-    ev: ReportGenerationEvent,
-  ): Promise<InputEvent> => {
-    const { toolCalls } = ev.data;
-
-    const toolMsgs = await callTools({
-      tools: [this.documentGeneratorTool],
-      toolCalls,
-      writeEvent: (text, step) => {
-        ctx.sendEvent(
-          toAgentRunEvent({
-            agent: "Reporter",
-            text,
-            type: toolCalls.length > 1 ? "progress" : "text",
-            current: step,
-            total: toolCalls.length,
-          }),
-        );
-      },
-    });
-    for (const toolMsg of toolMsgs) {
-      this.memory.put(toolMsg);
-    }
-    const chatHistory = await this.memory.getMessages();
-    return new InputEvent({ input: chatHistory });
-  };
-}
@@ -1,23 +0,0 @@
-import logging
-import os
-from typing import Optional
-
-from llama_index.core.indices import load_index_from_storage
-from llama_index.server.api.models import ChatRequest
-from llama_index.server.tools.index.utils import get_storage_context
-
-logger = logging.getLogger("uvicorn")
-
-STORAGE_DIR = "storage"
-
-
-def get_index(chat_request: Optional[ChatRequest] = None):
-    # check if storage already exists
-    if not os.path.exists(STORAGE_DIR):
-        return None
-    # load the existing index
-    logger.info(f"Loading index from {STORAGE_DIR}...")
-    storage_context = get_storage_context(STORAGE_DIR)
-    index = load_index_from_storage(storage_context)
-    logger.info(f"Finished loading index from {STORAGE_DIR}")
-    return index
@@ -1,8 +0,0 @@
-from llama_index.core import Settings
-from llama_index.embeddings.openai import OpenAIEmbedding
-from llama_index.llms.openai import OpenAI
-
-
-def init_settings():
-    Settings.llm = OpenAI(model="gpt-4o-mini")
-    Settings.embed_model = OpenAIEmbedding(model="text-embedding-3-small")
@@ -1,33 +0,0 @@
-import logging
-import os
-
-from app.index import STORAGE_DIR
-from app.settings import init_settings
-from dotenv import load_dotenv
-from llama_index.core.indices import (
-    VectorStoreIndex,
-)
-from llama_index.core.readers import SimpleDirectoryReader
-
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger()
-
-
-def generate_datasource():
-    load_dotenv()
-    init_settings()
-
-    logger.info("Creating new index")
-    # load the documents and create the index
-    reader = SimpleDirectoryReader(
-        os.environ.get("DATA_DIR", "data"),
-        recursive=True,
-    )
-    documents = reader.load_data()
-    index = VectorStoreIndex.from_documents(
-        documents,
-        show_progress=True,
-    )
-    # store it for later
-    index.storage_context.persist(STORAGE_DIR)
-    logger.info(f"Finished creating new index. Stored in {STORAGE_DIR}")
@@ -1,5 +0,0 @@
-__pycache__
-storage
-.env
-output
-.ui/
@@ -1,40 +0,0 @@
-import logging
-import os
-import subprocess
-
-from app.settings import init_settings
-from app.workflow import create_workflow
-from dotenv import load_dotenv
-from llama_index.server import LlamaIndexServer
-
-logger = logging.getLogger("uvicorn")
-
-
-def create_app():
-    env = os.environ.get("APP_ENV")
-
-    app = LlamaIndexServer(
-        workflow_factory=create_workflow,  # A factory function that creates a new workflow for each request
-        component_dir="components",
-        env=env,
-        logger=logger,
-    )
-    # You can also add custom routes to the app
-    app.add_api_route("/api/health", lambda: {"message": "OK"}, status_code=200)
-    return app
-
-
-load_dotenv()
-init_settings()
-app = create_app()
-
-
-def run(env: str):
-    os.environ["APP_ENV"] = env
-    app_host = os.getenv("APP_HOST", "0.0.0.0")
-    app_port = os.getenv("APP_PORT", "8000")
-
-    if env == "dev":
-        subprocess.run(["fastapi", "dev", "--host", app_host, "--port", app_port])
-    else:
-        subprocess.run(["fastapi", "run", "--host", app_host, "--port", app_port])
@@ -1,47 +0,0 @@
-[tool]
-[tool.poetry]
-name = "app"
-version = "0.1.0"
-description = ""
-authors = ["Marcus Schiesser <mail@marcusschiesser.de>"]
-readme = "README.md"
-
-[tool.poetry.scripts]
-generate = "generate:generate_datasource"
-dev = "main:run('dev')"
-prod = "main:run('prod')"
-
-[tool.poetry.dependencies]
-python = ">=3.11,<3.14"
-python-dotenv = "^1.0.0"
-pydantic = "<2.10"
-aiostream = "^0.5.2"
-llama-index-core = "^0.12.28"
-llama-index-server = "^0.1.8"
-
-[tool.poetry.group.dev.dependencies]
-mypy = "^1.8.0"
-pytest = "^8.3.5"
-pytest-asyncio = "^0.25.3"
-
-[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", "assignment"]
-
-[[tool.mypy.overrides]]
-module = "app.*"
-ignore_missing_imports = false
-
-[build-system]
-requires = ["poetry-core"]
-build-backend = "poetry.core.masonry.api"
@@ -1,40 +0,0 @@
-# See https://help.github.com/articles/ignoring-files/ for more about ignoring files.
-
-# dependencies
-/node_modules
-/.pnp
-.pnp.js
-
-# testing
-/coverage
-
-# next.js
-/.next/
-/out/
-
-# production
-/build
-
-# misc
-.DS_Store
-*.pem
-
-# debug
-npm-debug.log*
-yarn-debug.log*
-yarn-error.log*
-
-# local env files
-.env*.local
-
-# vercel
-.vercel
-
-# typescript
-*.tsbuildinfo
-next-env.d.ts
-
-output/
-storage/
-
-.env
@@ -1,23 +0,0 @@
-{
-  "name": "llamaindex-server",
-  "version": "0.1.0",
-  "scripts": {
-    "generate": "tsx src/generate.ts",
-    "dev": "tsx watch src/index.ts",
-    "start": "tsx src/index.ts"
-  },
-  "dependencies": {
-    "@llamaindex/openai": "0.2.0",
-    "@llamaindex/readers": "^2.0.0",
-    "@llamaindex/server": "0.0.9",
-    "@llamaindex/tools": "0.0.4",
-    "dotenv": "^16.4.7",
-    "llamaindex": "0.9.17",
-    "zod": "^3.23.8"
-  },
-  "devDependencies": {
-    "@types/node": "^20.10.3",
-    "tsx": "^4.7.2",
-    "typescript": "^5.3.2"
-  }
-}
@@ -1,3 +0,0 @@
-{
-  "trailingComma": "all"
-}
@@ -1,22 +0,0 @@
-import {
-  SimpleDocumentStore,
-  storageContextFromDefaults,
-  VectorStoreIndex,
-} from "llamaindex";
-
-export async function getIndex(params?: any) {
-  const storageContext = await storageContextFromDefaults({
-    persistDir: "storage",
-  });
-
-  const numberOfDocs = Object.keys(
-    (storageContext.docStore as SimpleDocumentStore).toDict(),
-  ).length;
-  if (numberOfDocs === 0) {
-    throw new Error(
-      "Index not found. Please run `pnpm run generate` to generate the embeddings of the documents",
-    );
-  }
-
-  return await VectorStoreIndex.init({ storageContext });
-}
@@ -1,11 +0,0 @@
-import { OpenAI, OpenAIEmbedding } from "@llamaindex/openai";
-import { Settings } from "llamaindex";
-
-export function initSettings() {
-  Settings.llm = new OpenAI({
-    model: "gpt-4o-mini",
-  });
-  Settings.embedModel = new OpenAIEmbedding({
-    model: "text-embedding-3-small",
-  });
-}
@@ -1,25 +0,0 @@
-import "dotenv/config";
-import { SimpleDirectoryReader } from "@llamaindex/readers/directory";
-import { storageContextFromDefaults, VectorStoreIndex } from "llamaindex";
-import { initSettings } from "./app/settings";
-
-async function generateDatasource() {
-  console.log(`Generating storage context...`);
-  // Split documents, create embeddings and store them in the storage context
-  const storageContext = await storageContextFromDefaults({
-    persistDir: "storage",
-  });
-  // load documents from current directoy into an index
-  const reader = new SimpleDirectoryReader();
-  const documents = await reader.loadData("data");
-
-  await VectorStoreIndex.fromDocuments(documents, {
-    storageContext,
-  });
-  console.log("Storage context successfully generated.");
-}
-
-(async () => {
-  initSettings();
-  await generateDatasource();
-})();
@@ -1,12 +0,0 @@
-import { LlamaIndexServer } from "@llamaindex/server";
-import "dotenv/config";
-import { initSettings } from "./app/settings";
-import { workflowFactory } from "./app/workflow";
-
-initSettings();
-
-new LlamaIndexServer({
-  workflow: workflowFactory,
-  appTitle: "LlamaIndex App",
-  componentsDir: "components",
-}).start();
@@ -1,14 +0,0 @@
-{
-  "compilerOptions": {
-    "target": "ES2022",
-    "module": "ES2022",
-    "moduleResolution": "bundler",
-    "esModuleInterop": true,
-    "forceConsistentCasingInFileNames": true,
-    "strict": true,
-    "skipLibCheck": true,
-    "outDir": "dist"
-  },
-  "include": ["src/**/*"],
-  "exclude": ["node_modules", "dist"]
-}
Author	SHA1	Message	Date
leehuwuj	670e2631f3	bump @llamaindex/server	2025-04-02 17:20:32 +07:00
leehuwuj	afb519d231	fix llamacloud api and markdown issue	2025-04-02 16:51:21 +07:00