Release 0.1.20 (#166)

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

.github/workflows/e2e.yml

@@ -17,7 +17,7 @@ jobs:
       matrix:
         node-version: [18, 20]
         python-version: ["3.11"]
-        os: [macos-latest, windows-latest]
+        os: [macos-latest, windows-latest, ubuntu-22.04]
     defaults:
       run:
         shell: bash
@@ -26,7 +26,7 @@ jobs:
       - uses: actions/checkout@v4
 
       - name: Set up python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5
         with:
           python-version: ${{ matrix.python-version }}
 

.gitignore

@@ -46,5 +46,8 @@ e2e/cache
 # intellij
 **/.idea
 
+# Python
+.mypy_cache/
+
 # build artifacts
 create-llama-*.tgz

CHANGELOG.md

@@ -1,5 +1,124 @@
 # create-llama
 
+## 0.1.20
+
+### Patch Changes
+
+- 624c721: Update to LlamaIndex 0.10.55
+
+## 0.1.19
+
+### Patch Changes
+
+- df96159: Use Qdrant FastEmbed as local embedding provider
+- 32fb32a: Support upload document files: pdf, docx, txt
+
+## 0.1.18
+
+### Patch Changes
+
+- d1026ea: support Mistral as llm and embedding
+- a221cfc: Use LlamaParse for all the file types that it supports (if activated)
+
+## 0.1.17
+
+### Patch Changes
+
+- 9ecd061: Add new template for a multi-agents app
+
+## 0.1.16
+
+### Patch Changes
+
+- a0aab03: Add T-System's LLMHUB as a model provider
+
+## 0.1.15
+
+### Patch Changes
+
+- 64732f0: Fix the issue of images not showing with the sandbox URL from OpenAI's models
+- aeb6fef: use llamacloud for chat
+
+## 0.1.14
+
+### Patch Changes
+
+- f2c3389: chore: update to llamaindex 0.4.3
+- 5093b37: Remove non-working file selectors for Linux
+
+## 0.1.13
+
+### Patch Changes
+
+- b3c969d: Add image generator tool
+
+## 0.1.12
+
+### Patch Changes
+
+- aa69014: Fix NextJS for TS 5.2
+
+## 0.1.11
+
+### Patch Changes
+
+- 48b96ff: Add DuckDuckGo search tool
+- 9c9decb: Reuse function tool instances and improve e2b interpreter tool for Python
+- 02ed277: Add Groq as a model provider
+- 0748f2e: Remove hard-coded Gemini supported models
+
+## 0.1.10
+
+### Patch Changes
+
+- 9112d08: Add OpenAPI tool for Typescript
+- 8f03f8d: Add OLLAMA_REQUEST_TIMEOUT variable to config Ollama timeout (Python)
+- 8f03f8d: Apply nest_asyncio for llama parse
+
+## 0.1.9
+
+### Patch Changes
+
+- a42fa53: Add CSV upload
+- 563b51d: Fix Vercel streaming (python) to stream data events instantly
+- d60b3c5: Add E2B code interpreter tool for FastAPI
+- 956538e: Add OpenAPI action tool for FastAPI
+
+## 0.1.8
+
+### Patch Changes
+
+- cd50a33: Add interpreter tool for TS using e2b.dev
+
+## 0.1.7
+
+### Patch Changes
+
+- 260d37a: Add system prompt env variable for TS
+- bbd5b8d: Fix postgres connection leaking issue
+- bb53425: Support HTTP proxies by setting the GLOBAL_AGENT_HTTP_PROXY env variable
+- 69c2e16: Fix streaming for Express
+- 7873bfb: Update Ollama provider to run with the base URL from the environment variable
+
+## 0.1.6
+
+### Patch Changes
+
+- 56537a1: Display PDF files in source nodes
+
+## 0.1.5
+
+### Patch Changes
+
+- 84db798: feat: support display latex in chat markdown
+
+## 0.1.4
+
+### Patch Changes
+
+- 0bc8e75: Use ingestion pipeline for dedicated vector stores (Python only)
+- cb1001d: Add ChromaDB vector store
+
 ## 0.1.3
 
 ### Patch Changes

create-app.ts

@@ -151,5 +151,19 @@ export async function createApp({
     );
   }
 
+  if (
+    dataSources.some((dataSource) => dataSource.type === "file") &&
+    process.platform === "linux"
+  ) {
+    console.log(
+      yellow(
+        `You can add your own data files to ${terminalLink(
+          "data",
+          `file://${root}/data`,
+        )} folder manually.`,
+      ),
+    );
+  }
+
   console.log();
 }

helpers/env-variables.ts

@@ -1,13 +1,17 @@
 import fs from "fs/promises";
 import path from "path";
+import { TOOL_SYSTEM_PROMPT_ENV_VAR, Tool } from "./tools";
 import {
   ModelConfig,
   TemplateDataSource,
   TemplateFramework,
+  TemplateType,
   TemplateVectorDB,
 } from "./types";
 
-type EnvVar = {
+import { TSYSTEMS_LLMHUB_API_URL } from "./providers/llmhub";
+
+export type EnvVar = {
   name?: string;
   description?: string;
   value?: string;
@@ -29,17 +33,20 @@ const renderEnvVar = (envVars: EnvVar[]): string => {
   );
 };
 
-const getVectorDBEnvs = (vectorDb?: TemplateVectorDB): EnvVar[] => {
-  if (!vectorDb) {
+const getVectorDBEnvs = (
+  vectorDb?: TemplateVectorDB,
+  framework?: TemplateFramework,
+): EnvVar[] => {
+  if (!vectorDb || !framework) {
     return [];
   }
   switch (vectorDb) {
     case "mongo":
       return [
         {
-          name: "MONGO_URI",
+          name: "MONGODB_URI",
           description:
-            "For generating a connection URI, see https://docs.timescale.com/use-timescale/latest/services/create-a-service\nThe MongoDB connection URI.",
+            "For generating a connection URI, see https://www.mongodb.com/docs/manual/reference/connection-string/ \nThe MongoDB connection URI.",
         },
         {
           name: "MONGODB_DATABASE",
@@ -129,6 +136,51 @@ const getVectorDBEnvs = (vectorDb?: TemplateVectorDB): EnvVar[] => {
             "Optional API key for authenticating requests to Qdrant.",
         },
       ];
+    case "llamacloud":
+      return [
+        {
+          name: "LLAMA_CLOUD_INDEX_NAME",
+          description:
+            "The name of the LlamaCloud index to use (part of the LlamaCloud project).",
+          value: "test",
+        },
+        {
+          name: "LLAMA_CLOUD_PROJECT_NAME",
+          description: "The name of the LlamaCloud project.",
+          value: "Default",
+        },
+        {
+          name: "LLAMA_CLOUD_BASE_URL",
+          description:
+            "The base URL for the LlamaCloud API. Only change this for non-production environments",
+          value: "https://api.cloud.llamaindex.ai",
+        },
+      ];
+    case "chroma":
+      const envs = [
+        {
+          name: "CHROMA_COLLECTION",
+          description: "The name of the collection in your Chroma database",
+        },
+        {
+          name: "CHROMA_HOST",
+          description: "The API endpoint for your Chroma database",
+        },
+        {
+          name: "CHROMA_PORT",
+          description: "The port for your Chroma database",
+        },
+      ];
+      // TS Version doesn't support config local storage path
+      if (framework === "fastapi") {
+        envs.push({
+          name: "CHROMA_PATH",
+          description: `The local path to the Chroma database. 
+Specify this if you are using a local Chroma database. 
+Otherwise, use CHROMA_HOST and CHROMA_PORT config above`,
+        });
+      }
+      return envs;
     default:
       return [];
   }
@@ -156,6 +208,10 @@ const getModelEnvs = (modelConfig: ModelConfig): EnvVar[] => {
       description: "Dimension of the embedding model to use.",
       value: modelConfig.dimensions.toString(),
     },
+    {
+      name: "CONVERSATION_STARTERS",
+      description: "The questions to help users get started (multi-line).",
+    },
     ...(modelConfig.provider === "openai"
       ? [
           {
@@ -182,6 +238,15 @@ const getModelEnvs = (modelConfig: ModelConfig): EnvVar[] => {
           },
         ]
       : []),
+    ...(modelConfig.provider === "groq"
+      ? [
+          {
+            name: "GROQ_API_KEY",
+            description: "The Groq API key to use.",
+            value: modelConfig.apiKey,
+          },
+        ]
+      : []),
     ...(modelConfig.provider === "gemini"
       ? [
           {
@@ -191,35 +256,76 @@ const getModelEnvs = (modelConfig: ModelConfig): EnvVar[] => {
           },
         ]
       : []),
+    ...(modelConfig.provider === "ollama"
+      ? [
+          {
+            name: "OLLAMA_BASE_URL",
+            description:
+              "The base URL for the Ollama API. Eg: http://127.0.0.1:11434",
+          },
+        ]
+      : []),
+    ...(modelConfig.provider === "mistral"
+      ? [
+          {
+            name: "MISTRAL_API_KEY",
+            description: "The Mistral API key to use.",
+            value: modelConfig.apiKey,
+          },
+        ]
+      : []),
+    ...(modelConfig.provider === "t-systems"
+      ? [
+          {
+            name: "T_SYSTEMS_LLMHUB_BASE_URL",
+            description:
+              "The base URL for the T-Systems AI Foundation Model API. Eg: http://localhost:11434",
+            value: TSYSTEMS_LLMHUB_API_URL,
+          },
+          {
+            name: "T_SYSTEMS_LLMHUB_API_KEY",
+            description: "API Key for T-System's AI Foundation Model.",
+            value: modelConfig.apiKey,
+          },
+        ]
+      : []),
   ];
 };
 
 const getFrameworkEnvs = (
-  framework?: TemplateFramework,
+  framework: TemplateFramework,
   port?: number,
 ): EnvVar[] => {
-  if (framework !== "fastapi") {
-    return [];
-  }
-  return [
+  const sPort = port?.toString() || "8000";
+  const result: EnvVar[] = [
     {
-      name: "APP_HOST",
-      description: "The address to start the backend app.",
-      value: "0.0.0.0",
-    },
-    {
-      name: "APP_PORT",
-      description: "The port to start the backend app.",
-      value: port?.toString() || "8000",
-    },
-    // TODO: Once LlamaIndexTS supports string templates, move this to `getEngineEnvs`
-    {
-      name: "SYSTEM_PROMPT",
-      description: `Custom system prompt.
-Example:
-SYSTEM_PROMPT="You are a helpful assistant who helps users with their questions."`,
+      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 backend app.",
+          value: "0.0.0.0",
+        },
+        {
+          name: "APP_PORT",
+          description: "The port to start the backend app.",
+          value: sPort,
+        },
+      ],
+    );
+  }
+  return result;
 };
 
 const getEngineEnvs = (): EnvVar[] => {
@@ -230,18 +336,99 @@ const getEngineEnvs = (): EnvVar[] => {
         "The number of similar embeddings to return when retrieving documents.",
       value: "3",
     },
+    {
+      name: "STREAM_TIMEOUT",
+      description:
+        "The time in milliseconds to wait for the stream to return a response.",
+      value: "60000",
+    },
   ];
 };
 
+const getToolEnvs = (tools?: Tool[]): EnvVar[] => {
+  if (!tools?.length) return [];
+  const toolEnvs: EnvVar[] = [];
+  tools.forEach((tool) => {
+    if (tool.envVars?.length) {
+      toolEnvs.push(
+        // Don't include the system prompt env var here
+        // It should be handled separately by merging with the default system prompt
+        ...tool.envVars.filter(
+          (env) => env.name !== TOOL_SYSTEM_PROMPT_ENV_VAR,
+        ),
+      );
+    }
+  });
+  return toolEnvs;
+};
+
+const getSystemPromptEnv = (tools?: Tool[]): EnvVar => {
+  const defaultSystemPrompt =
+    "You are a helpful assistant who helps users with their questions.";
+
+  // build tool system prompt by merging all tool system prompts
+  let toolSystemPrompt = "";
+  tools?.forEach((tool) => {
+    const toolSystemPromptEnv = tool.envVars?.find(
+      (env) => env.name === TOOL_SYSTEM_PROMPT_ENV_VAR,
+    );
+    if (toolSystemPromptEnv) {
+      toolSystemPrompt += toolSystemPromptEnv.value + "\n";
+    }
+  });
+
+  const systemPrompt = toolSystemPrompt
+    ? `\"${toolSystemPrompt}\"`
+    : defaultSystemPrompt;
+
+  return {
+    name: "SYSTEM_PROMPT",
+    description: "The system prompt for the AI model.",
+    value: systemPrompt,
+  };
+};
+
+const getTemplateEnvs = (template?: TemplateType): EnvVar[] => {
+  if (template === "multiagent") {
+    return [
+      {
+        name: "MESSAGE_QUEUE_PORT",
+      },
+      {
+        name: "CONTROL_PLANE_PORT",
+      },
+      {
+        name: "HUMAN_CONSUMER_PORT",
+      },
+      {
+        name: "AGENT_QUERY_ENGINE_PORT",
+        value: "8003",
+      },
+      {
+        name: "AGENT_QUERY_ENGINE_DESCRIPTION",
+        value: "Query information from the provided data",
+      },
+      {
+        name: "AGENT_DUMMY_PORT",
+        value: "8004",
+      },
+    ];
+  } else {
+    return [];
+  }
+};
+
 export const createBackendEnvFile = async (
   root: string,
   opts: {
     llamaCloudKey?: string;
     vectorDb?: TemplateVectorDB;
     modelConfig: ModelConfig;
-    framework?: TemplateFramework;
+    framework: TemplateFramework;
     dataSources?: TemplateDataSource[];
+    template?: TemplateType;
     port?: number;
+    tools?: Tool[];
   },
 ) => {
   // Init env values
@@ -257,8 +444,12 @@ export const createBackendEnvFile = async (
     // Add engine environment variables
     ...getEngineEnvs(),
     // Add vector database environment variables
-    ...getVectorDBEnvs(opts.vectorDb),
+    ...getVectorDBEnvs(opts.vectorDb, opts.framework),
     ...getFrameworkEnvs(opts.framework, opts.port),
+    ...getToolEnvs(opts.tools),
+    // Add template environment variables
+    ...getTemplateEnvs(opts.template),
+    getSystemPromptEnv(opts.tools),
   ];
   // Render and write env file
   const content = renderEnvVar(envVars);

helpers/index.ts

@@ -8,6 +8,7 @@ import { writeLoadersConfig } from "./datasources";
 import { createBackendEnvFile, createFrontendEnvFile } from "./env-variables";
 import { PackageManager } from "./get-pkg-manager";
 import { installLlamapackProject } from "./llama-pack";
+import { makeDir } from "./make-dir";
 import { isHavingPoetryLockFile, tryPoetryRun } from "./poetry";
 import { installPythonTemplate } from "./python";
 import { downloadAndExtractRepo } from "./repo";
@@ -141,14 +142,18 @@ export const installTemplate = async (
     // 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.
-    await createBackendEnvFile(props.root, {
-      modelConfig: props.modelConfig,
-      llamaCloudKey: props.llamaCloudKey,
-      vectorDb: props.vectorDb,
-      framework: props.framework,
-      dataSources: props.dataSources,
-      port: props.externalPort,
-    });
+    if (props.template === "streaming" || props.template === "multiagent") {
+      await createBackendEnvFile(props.root, {
+        modelConfig: props.modelConfig,
+        llamaCloudKey: props.llamaCloudKey,
+        vectorDb: props.vectorDb,
+        framework: props.framework,
+        dataSources: props.dataSources,
+        port: props.externalPort,
+        tools: props.tools,
+        template: props.template,
+      });
+    }
 
     if (props.dataSources.length > 0) {
       console.log("\nGenerating context data...\n");
@@ -170,6 +175,12 @@ export const installTemplate = async (
         );
       }
     }
+
+    // Create outputs directory
+    if (props.tools && props.tools.length > 0) {
+      await makeDir(path.join(props.root, "output/tools"));
+      await makeDir(path.join(props.root, "output/uploaded"));
+    }
   } else {
     // this is a frontend for a full-stack app, create .env file with model information
     await createFrontendEnvFile(props.root, {

helpers/providers/groq.ts

@@ -0,0 +1,99 @@
+import ciInfo from "ci-info";
+import prompts from "prompts";
+import { ModelConfigParams } from ".";
+import { questionHandlers, toChoice } from "../../questions";
+
+const MODELS = ["llama3-8b", "llama3-70b", "mixtral-8x7b"];
+const DEFAULT_MODEL = MODELS[0];
+
+// Use huggingface embedding models for now as Groq doesn't support embedding models
+enum HuggingFaceEmbeddingModelType {
+  XENOVA_ALL_MINILM_L6_V2 = "all-MiniLM-L6-v2",
+  XENOVA_ALL_MPNET_BASE_V2 = "all-mpnet-base-v2",
+}
+type ModelData = {
+  dimensions: number;
+};
+const EMBEDDING_MODELS: Record<HuggingFaceEmbeddingModelType, ModelData> = {
+  [HuggingFaceEmbeddingModelType.XENOVA_ALL_MINILM_L6_V2]: {
+    dimensions: 384,
+  },
+  [HuggingFaceEmbeddingModelType.XENOVA_ALL_MPNET_BASE_V2]: {
+    dimensions: 768,
+  },
+};
+const DEFAULT_EMBEDDING_MODEL = Object.keys(EMBEDDING_MODELS)[0];
+const DEFAULT_DIMENSIONS = Object.values(EMBEDDING_MODELS)[0].dimensions;
+
+type GroqQuestionsParams = {
+  apiKey?: string;
+  askModels: boolean;
+};
+
+export async function askGroqQuestions({
+  askModels,
+  apiKey,
+}: GroqQuestionsParams): Promise<ModelConfigParams> {
+  const config: ModelConfigParams = {
+    apiKey,
+    model: DEFAULT_MODEL,
+    embeddingModel: DEFAULT_EMBEDDING_MODEL,
+    dimensions: DEFAULT_DIMENSIONS,
+    isConfigured(): boolean {
+      if (config.apiKey) {
+        return true;
+      }
+      if (process.env["GROQ_API_KEY"]) {
+        return true;
+      }
+      return false;
+    },
+  };
+
+  if (!config.apiKey) {
+    const { key } = await prompts(
+      {
+        type: "text",
+        name: "key",
+        message:
+          "Please provide your Groq API key (or leave blank to use GROQ_API_KEY env variable):",
+      },
+      questionHandlers,
+    );
+    config.apiKey = key || process.env.GROQ_API_KEY;
+  }
+
+  // use default model values in CI or if user should not be asked
+  const useDefaults = ciInfo.isCI || !askModels;
+  if (!useDefaults) {
+    const { model } = await prompts(
+      {
+        type: "select",
+        name: "model",
+        message: "Which LLM model would you like to use?",
+        choices: MODELS.map(toChoice),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    config.model = model;
+
+    const { embeddingModel } = await prompts(
+      {
+        type: "select",
+        name: "embeddingModel",
+        message: "Which embedding model would you like to use?",
+        choices: Object.keys(EMBEDDING_MODELS).map(toChoice),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    config.embeddingModel = embeddingModel;
+    config.dimensions =
+      EMBEDDING_MODELS[
+        embeddingModel as HuggingFaceEmbeddingModelType
+      ].dimensions;
+  }
+
+  return config;
+}

helpers/providers/index.ts

@@ -1,9 +1,12 @@
 import ciInfo from "ci-info";
 import prompts from "prompts";
 import { questionHandlers } from "../../questions";
-import { ModelConfig, ModelProvider } from "../types";
+import { ModelConfig, ModelProvider, TemplateFramework } from "../types";
 import { askAnthropicQuestions } from "./anthropic";
 import { askGeminiQuestions } from "./gemini";
+import { askGroqQuestions } from "./groq";
+import { askLLMHubQuestions } from "./llmhub";
+import { askMistralQuestions } from "./mistral";
 import { askOllamaQuestions } from "./ollama";
 import { askOpenAIQuestions } from "./openai";
 
@@ -12,6 +15,7 @@ const DEFAULT_MODEL_PROVIDER = "openai";
 export type ModelConfigQuestionsParams = {
   openAiKey?: string;
   askModels: boolean;
+  framework?: TemplateFramework;
 };
 
 export type ModelConfigParams = Omit<ModelConfig, "provider">;
@@ -19,23 +23,28 @@ export type ModelConfigParams = Omit<ModelConfig, "provider">;
 export async function askModelConfig({
   askModels,
   openAiKey,
+  framework,
 }: ModelConfigQuestionsParams): Promise<ModelConfig> {
   let modelProvider: ModelProvider = DEFAULT_MODEL_PROVIDER;
   if (askModels && !ciInfo.isCI) {
+    let choices = [
+      { title: "OpenAI", value: "openai" },
+      { title: "Groq", value: "groq" },
+      { title: "Ollama", value: "ollama" },
+      { title: "Anthropic", value: "anthropic" },
+      { title: "Gemini", value: "gemini" },
+      { title: "Mistral", value: "mistral" },
+    ];
+
+    if (framework === "fastapi") {
+      choices.push({ title: "T-Systems", value: "t-systems" });
+    }
     const { provider } = await prompts(
       {
         type: "select",
         name: "provider",
         message: "Which model provider would you like to use",
-        choices: [
-          {
-            title: "OpenAI",
-            value: "openai",
-          },
-          { title: "Ollama", value: "ollama" },
-          { title: "Anthropic", value: "anthropic" },
-          { title: "Gemini", value: "gemini" },
-        ],
+        choices: choices,
         initial: 0,
       },
       questionHandlers,
@@ -48,12 +57,21 @@ export async function askModelConfig({
     case "ollama":
       modelConfig = await askOllamaQuestions({ askModels });
       break;
+    case "groq":
+      modelConfig = await askGroqQuestions({ askModels });
+      break;
     case "anthropic":
       modelConfig = await askAnthropicQuestions({ askModels });
       break;
     case "gemini":
       modelConfig = await askGeminiQuestions({ askModels });
       break;
+    case "mistral":
+      modelConfig = await askMistralQuestions({ askModels });
+      break;
+    case "t-systems":
+      modelConfig = await askLLMHubQuestions({ askModels });
+      break;
     default:
       modelConfig = await askOpenAIQuestions({
         openAiKey,

helpers/providers/llmhub.ts

@@ -0,0 +1,169 @@
+import ciInfo from "ci-info";
+import got from "got";
+import ora from "ora";
+import { red } from "picocolors";
+import prompts from "prompts";
+import { ModelConfigParams } from ".";
+import { questionHandlers } from "../../questions";
+
+export const TSYSTEMS_LLMHUB_API_URL =
+  "https://llm-server.llmhub.t-systems.net/v2";
+
+const DEFAULT_MODEL = "gpt-3.5-turbo";
+const DEFAULT_EMBEDDING_MODEL = "text-embedding-3-large";
+
+const LLMHUB_MODELS = [
+  "gpt-35-turbo",
+  "gpt-4-32k-1",
+  "gpt-4-32k-canada",
+  "gpt-4-32k-france",
+  "gpt-4-turbo-128k-france",
+  "Llama2-70b-Instruct",
+  "Llama-3-70B-Instruct",
+  "Mixtral-8x7B-Instruct-v0.1",
+  "mistral-large-32k-france",
+  "CodeLlama-2",
+];
+const LLMHUB_EMBEDDING_MODELS = [
+  "text-embedding-ada-002",
+  "text-embedding-ada-002-france",
+  "jina-embeddings-v2-base-de",
+  "jina-embeddings-v2-base-code",
+  "text-embedding-bge-m3",
+];
+
+type LLMHubQuestionsParams = {
+  apiKey?: string;
+  askModels: boolean;
+};
+
+export async function askLLMHubQuestions({
+  askModels,
+  apiKey,
+}: LLMHubQuestionsParams): Promise<ModelConfigParams> {
+  const config: ModelConfigParams = {
+    apiKey,
+    model: DEFAULT_MODEL,
+    embeddingModel: DEFAULT_EMBEDDING_MODEL,
+    dimensions: getDimensions(DEFAULT_EMBEDDING_MODEL),
+    isConfigured(): boolean {
+      if (config.apiKey) {
+        return true;
+      }
+      if (process.env["T_SYSTEMS_LLMHUB_API_KEY"]) {
+        return true;
+      }
+      return false;
+    },
+  };
+
+  if (!config.apiKey) {
+    const { key } = await prompts(
+      {
+        type: "text",
+        name: "key",
+        message: askModels
+          ? "Please provide your LLMHub API key (or leave blank to use T_SYSTEMS_LLMHUB_API_KEY env variable):"
+          : "Please provide your LLMHub API key (leave blank to skip):",
+        validate: (value: string) => {
+          if (askModels && !value) {
+            if (process.env.T_SYSTEMS_LLMHUB_API_KEY) {
+              return true;
+            }
+            return "T_SYSTEMS_LLMHUB_API_KEY env variable is not set - key is required";
+          }
+          return true;
+        },
+      },
+      questionHandlers,
+    );
+    config.apiKey = key || process.env.T_SYSTEMS_LLMHUB_API_KEY;
+  }
+
+  // use default model values in CI or if user should not be asked
+  const useDefaults = ciInfo.isCI || !askModels;
+  if (!useDefaults) {
+    const { model } = await prompts(
+      {
+        type: "select",
+        name: "model",
+        message: "Which LLM model would you like to use?",
+        choices: await getAvailableModelChoices(false, config.apiKey),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    config.model = model;
+
+    const { embeddingModel } = await prompts(
+      {
+        type: "select",
+        name: "embeddingModel",
+        message: "Which embedding model would you like to use?",
+        choices: await getAvailableModelChoices(true, config.apiKey),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    config.embeddingModel = embeddingModel;
+    config.dimensions = getDimensions(embeddingModel);
+  }
+
+  return config;
+}
+
+async function getAvailableModelChoices(
+  selectEmbedding: boolean,
+  apiKey?: string,
+) {
+  if (!apiKey) {
+    throw new Error("Need LLMHub key to retrieve model choices");
+  }
+  const isLLMModel = (modelId: string) => {
+    return LLMHUB_MODELS.includes(modelId);
+  };
+
+  const isEmbeddingModel = (modelId: string) => {
+    return LLMHUB_EMBEDDING_MODELS.includes(modelId);
+  };
+
+  const spinner = ora("Fetching available models").start();
+  try {
+    const response = await got(`${TSYSTEMS_LLMHUB_API_URL}/models`, {
+      headers: {
+        Authorization: "Bearer " + apiKey,
+      },
+      timeout: 5000,
+      responseType: "json",
+    });
+    const data: any = await response.body;
+    spinner.stop();
+    return data.data
+      .filter((model: any) =>
+        selectEmbedding ? isEmbeddingModel(model.id) : isLLMModel(model.id),
+      )
+      .map((el: any) => {
+        return {
+          title: el.id,
+          value: el.id,
+        };
+      });
+  } catch (error) {
+    spinner.stop();
+    if ((error as any).response?.statusCode === 401) {
+      console.log(
+        red(
+          "Invalid LLMHub API key provided! Please provide a valid key and try again!",
+        ),
+      );
+    } else {
+      console.log(red("Request failed: " + error));
+    }
+    process.exit(1);
+  }
+}
+
+function getDimensions(modelName: string) {
+  // Assuming dimensions similar to OpenAI for simplicity. Update if different.
+  return modelName === "text-embedding-004" ? 768 : 1536;
+}

helpers/providers/mistral.ts

@@ -0,0 +1,86 @@
+import ciInfo from "ci-info";
+import prompts from "prompts";
+import { ModelConfigParams } from ".";
+import { questionHandlers, toChoice } from "../../questions";
+
+const MODELS = ["mistral-tiny", "mistral-small", "mistral-medium"];
+type ModelData = {
+  dimensions: number;
+};
+const EMBEDDING_MODELS: Record<string, ModelData> = {
+  "mistral-embed": { dimensions: 1024 },
+};
+
+const DEFAULT_MODEL = MODELS[0];
+const DEFAULT_EMBEDDING_MODEL = Object.keys(EMBEDDING_MODELS)[0];
+const DEFAULT_DIMENSIONS = Object.values(EMBEDDING_MODELS)[0].dimensions;
+
+type MistralQuestionsParams = {
+  apiKey?: string;
+  askModels: boolean;
+};
+
+export async function askMistralQuestions({
+  askModels,
+  apiKey,
+}: MistralQuestionsParams): Promise<ModelConfigParams> {
+  const config: ModelConfigParams = {
+    apiKey,
+    model: DEFAULT_MODEL,
+    embeddingModel: DEFAULT_EMBEDDING_MODEL,
+    dimensions: DEFAULT_DIMENSIONS,
+    isConfigured(): boolean {
+      if (config.apiKey) {
+        return true;
+      }
+      if (process.env["MISTRAL_API_KEY"]) {
+        return true;
+      }
+      return false;
+    },
+  };
+
+  if (!config.apiKey) {
+    const { key } = await prompts(
+      {
+        type: "text",
+        name: "key",
+        message:
+          "Please provide your Mistral API key (or leave blank to use MISTRAL_API_KEY env variable):",
+      },
+      questionHandlers,
+    );
+    config.apiKey = key || process.env.MISTRAL_API_KEY;
+  }
+
+  // use default model values in CI or if user should not be asked
+  const useDefaults = ciInfo.isCI || !askModels;
+  if (!useDefaults) {
+    const { model } = await prompts(
+      {
+        type: "select",
+        name: "model",
+        message: "Which LLM model would you like to use?",
+        choices: MODELS.map(toChoice),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    config.model = model;
+
+    const { embeddingModel } = await prompts(
+      {
+        type: "select",
+        name: "embeddingModel",
+        message: "Which embedding model would you like to use?",
+        choices: Object.keys(EMBEDDING_MODELS).map(toChoice),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    config.embeddingModel = embeddingModel;
+    config.dimensions = EMBEDDING_MODELS[embeddingModel].dimensions;
+  }
+
+  return config;
+}

helpers/proxy.ts

@@ -0,0 +1,8 @@
+/* Function to conditionally load the global-agent/bootstrap module */
+export async function initializeGlobalAgent() {
+  if (process.env.GLOBAL_AGENT_HTTP_PROXY) {
+    /* Dynamically import global-agent/bootstrap */
+    await import("global-agent/bootstrap");
+    console.log("Proxy enabled via global-agent.");
+  }
+}

helpers/python.ts

@@ -55,11 +55,11 @@ const getAdditionalDependencies = (
     case "milvus": {
       dependencies.push({
         name: "llama-index-vector-stores-milvus",
-        version: "^0.1.6",
+        version: "^0.1.20",
       });
       dependencies.push({
         name: "pymilvus",
-        version: "2.3.7",
+        version: "2.4.4",
       });
       break;
     }
@@ -70,6 +70,20 @@ const getAdditionalDependencies = (
       });
       break;
     }
+    case "qdrant": {
+      dependencies.push({
+        name: "llama-index-vector-stores-qdrant",
+        version: "^0.2.8",
+      });
+      break;
+    }
+    case "chroma": {
+      dependencies.push({
+        name: "llama-index-vector-stores-chroma",
+        version: "^0.1.8",
+      });
+      break;
+    }
   }
 
   // Add data source dependencies
@@ -104,6 +118,12 @@ const getAdditionalDependencies = (
             version: "^2.9.9",
           });
           break;
+        case "llamacloud":
+          dependencies.push({
+            name: "llama-index-indices-managed-llama-cloud",
+            version: "^0.2.5",
+          });
+          break;
       }
     }
   }
@@ -130,7 +150,17 @@ const getAdditionalDependencies = (
     case "openai":
       dependencies.push({
         name: "llama-index-agent-openai",
-        version: "0.2.2",
+        version: "0.2.6",
+      });
+      break;
+    case "groq":
+      dependencies.push({
+        name: "llama-index-llms-groq",
+        version: "0.1.4",
+      });
+      dependencies.push({
+        name: "llama-index-embeddings-fastembed",
+        version: "^0.1.4",
       });
       break;
     case "anthropic":
@@ -139,20 +169,40 @@ const getAdditionalDependencies = (
         version: "0.1.10",
       });
       dependencies.push({
-        name: "llama-index-embeddings-huggingface",
-        version: "0.2.0",
+        name: "llama-index-embeddings-fastembed",
+        version: "^0.1.4",
       });
       break;
     case "gemini":
       dependencies.push({
         name: "llama-index-llms-gemini",
-        version: "0.1.7",
+        version: "0.1.10",
       });
       dependencies.push({
         name: "llama-index-embeddings-gemini",
         version: "0.1.6",
       });
       break;
+    case "mistral":
+      dependencies.push({
+        name: "llama-index-llms-mistralai",
+        version: "0.1.17",
+      });
+      dependencies.push({
+        name: "llama-index-embeddings-mistralai",
+        version: "0.1.4",
+      });
+      break;
+    case "t-systems":
+      dependencies.push({
+        name: "llama-index-agent-openai",
+        version: "0.2.2",
+      });
+      dependencies.push({
+        name: "llama-index-llms-openai-like",
+        version: "0.1.3",
+      });
+      break;
   }
 
   return dependencies;
@@ -290,20 +340,27 @@ export const installPythonTemplate = async ({
     cwd: path.join(compPath, "loaders", "python"),
   });
 
-  // Select and copy engine code based on data sources and tools
-  let engine;
-  tools = tools ?? [];
-  if (dataSources.length > 0 && tools.length === 0) {
-    console.log("\nNo tools selected - use optimized context chat engine\n");
-    engine = "chat";
-  } else {
-    engine = "agent";
-  }
-  await copy("**", enginePath, {
-    parents: true,
-    cwd: path.join(compPath, "engines", "python", engine),
+  // Copy settings.py to app
+  await copy("**", path.join(root, "app"), {
+    cwd: path.join(compPath, "settings", "python"),
   });
 
+  if (template === "streaming") {
+    // For the streaming template only:
+    // Select and copy engine code based on data sources and tools
+    let engine;
+    if (dataSources.length > 0 && (!tools || tools.length === 0)) {
+      console.log("\nNo tools selected - use optimized context chat engine\n");
+      engine = "chat";
+    } else {
+      engine = "agent";
+    }
+    await copy("**", enginePath, {
+      parents: true,
+      cwd: path.join(compPath, "engines", "python", engine),
+    });
+  }
+
   console.log("Adding additional dependencies");
 
   const addOnDependencies = getAdditionalDependencies(

helpers/tools.ts

@@ -2,9 +2,12 @@ import fs from "fs/promises";
 import path from "path";
 import { red } from "picocolors";
 import yaml from "yaml";
+import { EnvVar } from "./env-variables";
 import { makeDir } from "./make-dir";
 import { TemplateFramework } from "./types";
 
+export const TOOL_SYSTEM_PROMPT_ENV_VAR = "TOOL_SYSTEM_PROMPT";
+
 export enum ToolType {
   LLAMAHUB = "llamahub",
   LOCAL = "local",
@@ -17,6 +20,7 @@ export type Tool = {
   dependencies?: ToolDependencies[];
   supportedFrameworks?: Array<TemplateFramework>;
   type: ToolType;
+  envVars?: EnvVar[];
 };
 
 export type ToolDependencies = {
@@ -26,7 +30,7 @@ export type ToolDependencies = {
 
 export const supportedTools: Tool[] = [
   {
-    display: "Google Search (configuration required after installation)",
+    display: "Google Search",
     name: "google.GoogleSearchToolSpec",
     config: {
       engine:
@@ -42,6 +46,36 @@ export const supportedTools: Tool[] = [
     ],
     supportedFrameworks: ["fastapi"],
     type: ToolType.LLAMAHUB,
+    envVars: [
+      {
+        name: TOOL_SYSTEM_PROMPT_ENV_VAR,
+        description: "System prompt for google search tool.",
+        value: `You are a Google search agent. You help users to get information from Google search.`,
+      },
+    ],
+  },
+  {
+    // For python app, we will use a local DuckDuckGo search tool (instead of DuckDuckGo search tool in LlamaHub)
+    // to get the same results as the TS app.
+    display: "DuckDuckGo Search",
+    name: "duckduckgo",
+    dependencies: [
+      {
+        name: "duckduckgo-search",
+        version: "6.1.7",
+      },
+    ],
+    supportedFrameworks: ["fastapi", "nextjs", "express"],
+    type: ToolType.LOCAL,
+    envVars: [
+      {
+        name: TOOL_SYSTEM_PROMPT_ENV_VAR,
+        description: "System prompt for DuckDuckGo search tool.",
+        value: `You are a DuckDuckGo search agent. 
+You can use the duckduckgo search tool to get information from the web to answer user questions.
+For better results, you can specify the region parameter to get results from a specific region but it's optional.`,
+      },
+    ],
   },
   {
     display: "Wikipedia",
@@ -54,6 +88,13 @@ export const supportedTools: Tool[] = [
     ],
     supportedFrameworks: ["fastapi", "express", "nextjs"],
     type: ToolType.LLAMAHUB,
+    envVars: [
+      {
+        name: TOOL_SYSTEM_PROMPT_ENV_VAR,
+        description: "System prompt for wiki tool.",
+        value: `You are a Wikipedia agent. You help users to get information from Wikipedia.`,
+      },
+    ],
   },
   {
     display: "Weather",
@@ -61,6 +102,91 @@ export const supportedTools: Tool[] = [
     dependencies: [],
     supportedFrameworks: ["fastapi", "express", "nextjs"],
     type: ToolType.LOCAL,
+    envVars: [
+      {
+        name: TOOL_SYSTEM_PROMPT_ENV_VAR,
+        description: "System prompt for weather tool.",
+        value: `You are a weather forecast agent. You help users to get the weather forecast for a given location.`,
+      },
+    ],
+  },
+  {
+    display: "Code Interpreter",
+    name: "interpreter",
+    dependencies: [
+      {
+        name: "e2b_code_interpreter",
+        version: "0.0.7",
+      },
+    ],
+    supportedFrameworks: ["fastapi", "express", "nextjs"],
+    type: ToolType.LOCAL,
+    envVars: [
+      {
+        name: "E2B_API_KEY",
+        description:
+          "E2B_API_KEY key is required to run code interpreter tool. Get it here: https://e2b.dev/docs/getting-started/api-key",
+      },
+      {
+        name: TOOL_SYSTEM_PROMPT_ENV_VAR,
+        description: "System prompt for code interpreter tool.",
+        value: `-You are a Python interpreter that can run any python code in a secure environment.
+- The python code runs in a Jupyter notebook. Every time you call the 'interpreter' tool, the python code is executed in a separate cell. 
+- You are given tasks to complete and you run python code to solve them.
+- It's okay to make multiple calls to interpreter tool. If you get an error or the result is not what you expected, you can call the tool again. Don't give up too soon!
+- Plot visualizations using matplotlib or any other visualization library directly in the notebook.
+- You can install any pip package (if it exists) by running a cell with pip install.`,
+      },
+    ],
+  },
+  {
+    display: "OpenAPI action",
+    name: "openapi_action.OpenAPIActionToolSpec",
+    dependencies: [
+      {
+        name: "llama-index-tools-openapi",
+        version: "0.1.3",
+      },
+      {
+        name: "jsonschema",
+        version: "^4.22.0",
+      },
+      {
+        name: "llama-index-tools-requests",
+        version: "0.1.3",
+      },
+    ],
+    config: {
+      openapi_uri: "The URL or file path of the OpenAPI schema",
+    },
+    supportedFrameworks: ["fastapi", "express", "nextjs"],
+    type: ToolType.LOCAL,
+    envVars: [
+      {
+        name: TOOL_SYSTEM_PROMPT_ENV_VAR,
+        description: "System prompt for openapi action tool.",
+        value:
+          "You are an OpenAPI action agent. You help users to make requests to the provided OpenAPI schema.",
+      },
+    ],
+  },
+  {
+    display: "Image Generator",
+    name: "img_gen",
+    supportedFrameworks: ["fastapi", "express", "nextjs"],
+    type: ToolType.LOCAL,
+    envVars: [
+      {
+        name: "STABILITY_API_KEY",
+        description:
+          "STABILITY_API_KEY key is required to run image generator. Get it here: https://platform.stability.ai/account/keys",
+      },
+      {
+        name: TOOL_SYSTEM_PROMPT_ENV_VAR,
+        description: "System prompt for image generator tool.",
+        value: `You are an image generator agent. You help users to generate images using the Stability API.`,
+      },
+    ],
   },
 ];
 
@@ -87,9 +213,15 @@ export const getTools = (toolsName: string[]): Tool[] => {
   return tools;
 };
 
+export const toolRequiresConfig = (tool: Tool): boolean => {
+  const hasConfig = Object.keys(tool.config || {}).length > 0;
+  const hasEmptyEnvVar = tool.envVars?.some((envVar) => !envVar.value) ?? false;
+  return hasConfig || hasEmptyEnvVar;
+};
+
 export const toolsRequireConfig = (tools?: Tool[]): boolean => {
   if (tools) {
-    return tools?.some((tool) => Object.keys(tool.config || {}).length > 0);
+    return tools?.some(toolRequiresConfig);
   }
   return false;
 };
@@ -104,7 +236,6 @@ export const writeToolsConfig = async (
   tools: Tool[] = [],
   type: ConfigFileType = ConfigFileType.YAML,
 ) => {
-  if (tools.length === 0) return; // no tools selected, no config need
   const configContent: {
     [key in ToolType]: Record<string, any>;
   } = {

helpers/types.ts

@@ -1,7 +1,14 @@
 import { PackageManager } from "../helpers/get-pkg-manager";
 import { Tool } from "./tools";
 
-export type ModelProvider = "openai" | "ollama" | "anthropic" | "gemini";
+export type ModelProvider =
+  | "openai"
+  | "groq"
+  | "ollama"
+  | "anthropic"
+  | "gemini"
+  | "mistral"
+  | "t-systems";
 export type ModelConfig = {
   provider: ModelProvider;
   apiKey?: string;
@@ -10,7 +17,11 @@ export type ModelConfig = {
   dimensions: number;
   isConfigured(): boolean;
 };
-export type TemplateType = "streaming" | "community" | "llamapack";
+export type TemplateType =
+  | "streaming"
+  | "community"
+  | "llamapack"
+  | "multiagent";
 export type TemplateFramework = "nextjs" | "express" | "fastapi";
 export type TemplateUI = "html" | "shadcn";
 export type TemplateVectorDB =
@@ -20,7 +31,9 @@ export type TemplateVectorDB =
   | "pinecone"
   | "milvus"
   | "astra"
-  | "qdrant";
+  | "qdrant"
+  | "chroma"
+  | "llamacloud";
 export type TemplatePostInstallAction =
   | "none"
   | "VSCode"
@@ -30,7 +43,7 @@ export type TemplateDataSource = {
   type: TemplateDataSourceType;
   config: TemplateDataSourceConfig;
 };
-export type TemplateDataSourceType = "file" | "web" | "db";
+export type TemplateDataSourceType = "file" | "web" | "db" | "llamacloud";
 export type TemplateObservability = "none" | "opentelemetry";
 // Config for both file and folder
 export type FileSourceConfig = {

helpers/typescript.ts

@@ -1,7 +1,7 @@
 import fs from "fs/promises";
 import os from "os";
 import path from "path";
-import { bold, cyan } from "picocolors";
+import { bold, cyan, yellow } from "picocolors";
 import { assetRelocator, copy } from "../helpers/copy";
 import { callPackageManager } from "../helpers/install";
 import { templatesDir } from "./dir";
@@ -104,8 +104,20 @@ export const installTSTemplate = async ({
       : path.join("src", "controllers");
   const enginePath = path.join(root, relativeEngineDestPath, "engine");
 
+  // copy llamaindex code for TS templates
+  await copy("**", path.join(root, relativeEngineDestPath, "llamaindex"), {
+    parents: true,
+    cwd: path.join(compPath, "llamaindex", "typescript"),
+  });
+
   // copy vector db component
-  console.log("\nUsing vector DB:", vectorDb ?? "none", "\n");
+  if (vectorDb === "llamacloud") {
+    console.log(
+      `\nUsing managed index from LlamaCloud. Ensure the ${yellow("LLAMA_CLOUD_* environment variables are set correctly.")}`,
+    );
+  } else {
+    console.log("\nUsing vector DB:", vectorDb ?? "none");
+  }
   await copy("**", enginePath, {
     parents: true,
     cwd: path.join(compPath, "vectordbs", "typescript", vectorDb ?? "none"),

index.ts

@@ -12,12 +12,16 @@ import { createApp } from "./create-app";
 import { getDataSources } from "./helpers/datasources";
 import { getPkgManager } from "./helpers/get-pkg-manager";
 import { isFolderEmpty } from "./helpers/is-folder-empty";
+import { initializeGlobalAgent } from "./helpers/proxy";
 import { runApp } from "./helpers/run-app";
 import { getTools } from "./helpers/tools";
 import { validateNpmName } from "./helpers/validate-pkg";
 import packageJson from "./package.json";
 import { QuestionArgs, askQuestions, onPromptState } from "./questions";
 
+// Run the initialization function
+initializeGlobalAgent();
+
 let projectPath: string = "";
 
 const handleSigTerm = () => process.exit(0);

package.json

@@ -1,6 +1,6 @@
 {
   "name": "create-llama",
-  "version": "0.1.3",
+  "version": "0.1.20",
   "description": "Create LlamaIndex-powered apps with one command",
   "keywords": [
     "rag",
@@ -52,6 +52,7 @@
     "cross-spawn": "7.0.3",
     "fast-glob": "3.3.1",
     "fs-extra": "11.2.0",
+    "global-agent": "^3.0.0",
     "got": "10.7.0",
     "ollama": "^0.5.0",
     "ora": "^8.0.1",

questions.ts

@@ -9,6 +9,7 @@ import {
   TemplateDataSource,
   TemplateDataSourceType,
   TemplateFramework,
+  TemplateType,
 } from "./helpers";
 import { COMMUNITY_OWNER, COMMUNITY_REPO } from "./helpers/constant";
 import { EXAMPLE_FILE } from "./helpers/datasources";
@@ -16,7 +17,11 @@ import { templatesDir } from "./helpers/dir";
 import { getAvailableLlamapackOptions } from "./helpers/llama-pack";
 import { askModelConfig } from "./helpers/providers";
 import { getProjectOptions } from "./helpers/repo";
-import { supportedTools, toolsRequireConfig } from "./helpers/tools";
+import {
+  supportedTools,
+  toolRequiresConfig,
+  toolsRequireConfig,
+} from "./helpers/tools";
 
 export type QuestionArgs = Omit<
   InstallAppArgs,
@@ -97,6 +102,7 @@ const getVectorDbChoices = (framework: TemplateFramework) => {
     { title: "Milvus", value: "milvus" },
     { title: "Astra", value: "astra" },
     { title: "Qdrant", value: "qdrant" },
+    { title: "ChromaDB", value: "chroma" },
   ];
 
   const vectordbLang = framework === "fastapi" ? "python" : "typescript";
@@ -117,8 +123,15 @@ const getVectorDbChoices = (framework: TemplateFramework) => {
 export const getDataSourceChoices = (
   framework: TemplateFramework,
   selectedDataSource: TemplateDataSource[],
+  template?: TemplateType,
 ) => {
+  // If LlamaCloud is already selected, don't show any other options
+  if (selectedDataSource.find((s) => s.type === "llamacloud")) {
+    return [];
+  }
+
   const choices = [];
+
   if (selectedDataSource.length > 0) {
     choices.push({
       title: "No",
@@ -126,29 +139,37 @@ export const getDataSourceChoices = (
     });
   }
   if (selectedDataSource === undefined || selectedDataSource.length === 0) {
+    if (template !== "multiagent") {
+      choices.push({
+        title: "No datasource",
+        value: "none",
+      });
+    }
     choices.push({
-      title: "No data, just a simple chat or agent",
-      value: "none",
-    });
-    choices.push({
-      title: "Use an example PDF",
+      title:
+        process.platform !== "linux"
+          ? "Use an example PDF"
+          : "Use an example PDF (you can add your own data files later)",
       value: "exampleFile",
     });
   }
 
-  choices.push(
-    {
-      title: `Use local files (${supportedContextFileTypes.join(", ")})`,
-      value: "file",
-    },
-    {
-      title:
-        process.platform === "win32"
-          ? "Use a local folder"
-          : "Use local folders",
-      value: "folder",
-    },
-  );
+  // Linux has many distros so we won't support file/folder picker for now
+  if (process.platform !== "linux") {
+    choices.push(
+      {
+        title: `Use local files (${supportedContextFileTypes.join(", ")})`,
+        value: "file",
+      },
+      {
+        title:
+          process.platform === "win32"
+            ? "Use a local folder"
+            : "Use local folders",
+        value: "folder",
+      },
+    );
+  }
 
   if (framework === "fastapi") {
     choices.push({
@@ -160,6 +181,13 @@ export const getDataSourceChoices = (
       value: "db",
     });
   }
+
+  if (!selectedDataSource.length) {
+    choices.push({
+      title: "Use managed index from LlamaCloud",
+      value: "llamacloud",
+    });
+  }
   return choices;
 };
 
@@ -257,25 +285,27 @@ export const askQuestions = async (
           },
         ];
 
-        const modelConfigured =
-          !program.llamapack && program.modelConfig.isConfigured();
-        // If using LlamaParse, require LlamaCloud API key
-        const llamaCloudKeyConfigured = program.useLlamaParse
-          ? program.llamaCloudKey || process.env["LLAMA_CLOUD_API_KEY"]
-          : true;
-        const hasVectorDb = program.vectorDb && program.vectorDb !== "none";
-        // Can run the app if all tools do not require configuration
-        if (
-          !hasVectorDb &&
-          modelConfigured &&
-          llamaCloudKeyConfigured &&
-          !toolsRequireConfig(program.tools)
-        ) {
-          actionChoices.push({
-            title:
-              "Generate code, install dependencies, and run the app (~2 min)",
-            value: "runApp",
-          });
+        if (program.template !== "multiagent") {
+          const modelConfigured =
+            !program.llamapack && program.modelConfig.isConfigured();
+          // If using LlamaParse, require LlamaCloud API key
+          const llamaCloudKeyConfigured = program.useLlamaParse
+            ? program.llamaCloudKey || process.env["LLAMA_CLOUD_API_KEY"]
+            : true;
+          const hasVectorDb = program.vectorDb && program.vectorDb !== "none";
+          // Can run the app if all tools do not require configuration
+          if (
+            !hasVectorDb &&
+            modelConfigured &&
+            llamaCloudKeyConfigured &&
+            !toolsRequireConfig(program.tools)
+          ) {
+            actionChoices.push({
+              title:
+                "Generate code, install dependencies, and run the app (~2 min)",
+              value: "runApp",
+            });
+          }
         }
 
         const { action } = await prompts(
@@ -307,7 +337,11 @@ export const askQuestions = async (
           name: "template",
           message: "Which template would you like to use?",
           choices: [
-            { title: "Chat", value: "streaming" },
+            { title: "Agentic RAG (single agent)", value: "streaming" },
+            {
+              title: "Multi-agent app (using llama-agents)",
+              value: "multiagent",
+            },
             {
               title: `Community template from ${styledRepo}`,
               value: "community",
@@ -371,6 +405,10 @@ export const askQuestions = async (
     return; // early return - no further questions needed for llamapack projects
   }
 
+  if (program.template === "multiagent") {
+    // TODO: multi-agents currently only supports FastAPI
+    program.framework = preferences.framework = "fastapi";
+  }
   if (!program.framework) {
     if (ciInfo.isCI) {
       program.framework = getPrefOrDefault("framework");
@@ -396,7 +434,10 @@ export const askQuestions = async (
     }
   }
 
-  if (program.framework === "express" || program.framework === "fastapi") {
+  if (
+    (program.framework === "express" || program.framework === "fastapi") &&
+    program.template === "streaming"
+  ) {
     // if a backend-only framework is selected, ask whether we should create a frontend
     if (program.frontend === undefined) {
       if (ciInfo.isCI) {
@@ -433,7 +474,7 @@ export const askQuestions = async (
     }
   }
 
-  if (!program.observability) {
+  if (!program.observability && program.template === "streaming") {
     if (ciInfo.isCI) {
       program.observability = getPrefOrDefault("observability");
     } else {
@@ -460,6 +501,7 @@ export const askQuestions = async (
     const modelConfig = await askModelConfig({
       openAiKey,
       askModels: program.askModels ?? false,
+      framework: program.framework,
     });
     program.modelConfig = modelConfig;
     preferences.modelConfig = modelConfig;
@@ -473,6 +515,12 @@ export const askQuestions = async (
       // continue asking user for data sources if none are initially provided
       while (true) {
         const firstQuestion = program.dataSources.length === 0;
+        const choices = getDataSourceChoices(
+          program.framework,
+          program.dataSources,
+          program.template,
+        );
+        if (choices.length === 0) break;
         const { selectedSource } = await prompts(
           {
             type: "select",
@@ -480,10 +528,7 @@ export const askQuestions = async (
             message: firstQuestion
               ? "Which data source would you like to use?"
               : "Would you like to add another data source?",
-            choices: getDataSourceChoices(
-              program.framework,
-              program.dataSources,
-            ),
+            choices,
             initial: firstQuestion ? 1 : 0,
           },
           questionHandlers,
@@ -580,51 +625,76 @@ export const askQuestions = async (
               config: await prompts(dbPrompts, questionHandlers),
             });
           }
+          case "llamacloud": {
+            program.dataSources.push({
+              type: "llamacloud",
+              config: {},
+            });
+            program.dataSources.push(EXAMPLE_FILE);
+            break;
+          }
         }
       }
     }
   }
 
-  // Asking for LlamaParse if user selected file or folder data source
-  if (
-    program.dataSources.some((ds) => ds.type === "file") &&
-    program.useLlamaParse === undefined
-  ) {
-    if (ciInfo.isCI) {
-      program.useLlamaParse = getPrefOrDefault("useLlamaParse");
-      program.llamaCloudKey = getPrefOrDefault("llamaCloudKey");
-    } else {
-      const { useLlamaParse } = await prompts(
-        {
-          type: "toggle",
-          name: "useLlamaParse",
-          message:
-            "Would you like to use LlamaParse (improved parser for RAG - requires API key)?",
-          initial: false,
-          active: "yes",
-          inactive: "no",
-        },
-        questionHandlers,
-      );
-      program.useLlamaParse = useLlamaParse;
+  const isUsingLlamaCloud = program.dataSources.some(
+    (ds) => ds.type === "llamacloud",
+  );
 
-      // Ask for LlamaCloud API key
-      if (useLlamaParse && program.llamaCloudKey === undefined) {
-        const { llamaCloudKey } = await prompts(
+  // Asking for LlamaParse if user selected file data source
+  if (isUsingLlamaCloud) {
+    // default to use LlamaParse if using LlamaCloud
+    program.useLlamaParse = preferences.useLlamaParse = true;
+  } else {
+    if (program.dataSources.some((ds) => ds.type === "file")) {
+      if (ciInfo.isCI) {
+        program.useLlamaParse = getPrefOrDefault("useLlamaParse");
+      } else {
+        const { useLlamaParse } = await prompts(
           {
-            type: "text",
-            name: "llamaCloudKey",
+            type: "toggle",
+            name: "useLlamaParse",
             message:
-              "Please provide your LlamaIndex Cloud API key (leave blank to skip):",
+              "Would you like to use LlamaParse (improved parser for RAG - requires API key)?",
+            initial: false,
+            active: "yes",
+            inactive: "no",
           },
           questionHandlers,
         );
-        program.llamaCloudKey = llamaCloudKey;
+        program.useLlamaParse = useLlamaParse;
+        preferences.useLlamaParse = useLlamaParse;
       }
     }
   }
 
-  if (program.dataSources.length > 0 && !program.vectorDb) {
+  // Ask for LlamaCloud API key when using a LlamaCloud index or LlamaParse
+  if (isUsingLlamaCloud || program.useLlamaParse) {
+    if (ciInfo.isCI) {
+      program.llamaCloudKey = getPrefOrDefault("llamaCloudKey");
+    } else {
+      // Ask for LlamaCloud API key
+      const { llamaCloudKey } = await prompts(
+        {
+          type: "text",
+          name: "llamaCloudKey",
+          message:
+            "Please provide your LlamaCloud API key (leave blank to skip):",
+        },
+        questionHandlers,
+      );
+      program.llamaCloudKey = preferences.llamaCloudKey =
+        llamaCloudKey || process.env.LLAMA_CLOUD_API_KEY;
+    }
+  }
+
+  if (isUsingLlamaCloud) {
+    // When using a LlamaCloud index, don't ask for vector database and use code in `llamacloud` folder for vector database
+    const vectorDb = "llamacloud";
+    program.vectorDb = vectorDb;
+    preferences.vectorDb = vectorDb;
+  } else if (program.dataSources.length > 0 && !program.vectorDb) {
     if (ciInfo.isCI) {
       program.vectorDb = getPrefOrDefault("vectorDb");
     } else {
@@ -643,7 +713,8 @@ export const askQuestions = async (
     }
   }
 
-  if (!program.tools) {
+  if (!program.tools && program.template === "streaming") {
+    // TODO: allow to select tools also for multi-agent framework
     if (ciInfo.isCI) {
       program.tools = getPrefOrDefault("tools");
     } else {
@@ -651,7 +722,7 @@ export const askQuestions = async (
         t.supportedFrameworks?.includes(program.framework),
       );
       const toolChoices = options.map((tool) => ({
-        title: tool.display,
+        title: `${tool.display}${toolRequiresConfig(tool) ? " (needs configuration)" : ""}`,
         value: tool.name,
       }));
       const { toolsName } = await prompts({

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

@@ -6,7 +6,7 @@ from app.engine.tools import ToolFactory
 from app.engine.index import get_index
 
 
-def get_chat_engine():
+def get_chat_engine(filters=None):
     system_prompt = os.getenv("SYSTEM_PROMPT")
     top_k = os.getenv("TOP_K", "3")
     tools = []
@@ -14,7 +14,9 @@ def get_chat_engine():
     # Add query tool if index exists
     index = get_index()
     if index is not None:
-        query_engine = index.as_query_engine(similarity_top_k=int(top_k))
+        query_engine = index.as_query_engine(
+            similarity_top_k=int(top_k), filters=filters
+        )
         query_engine_tool = QueryEngineTool.from_defaults(query_engine=query_engine)
         tools.append(query_engine_tool)
 

templates/components/engines/python/agent/tools/__init__.py

@@ -1,7 +1,8 @@
 import os
 import yaml
+import json
 import importlib
-
+from cachetools import cached, LRUCache
 from llama_index.core.tools.tool_spec.base import BaseToolSpec
 from llama_index.core.tools.function_tool import FunctionTool
 
@@ -18,7 +19,6 @@ class ToolFactory:
         ToolType.LOCAL: "app.engine.tools",
     }
 
-    @staticmethod
     def load_tools(tool_type: str, tool_name: str, config: dict) -> list[FunctionTool]:
         source_package = ToolFactory.TOOL_SOURCE_PACKAGE_MAP[tool_type]
         try:
@@ -31,7 +31,7 @@ class ToolFactory:
                 return tool_spec.to_tool_list()
             else:
                 module = importlib.import_module(f"{source_package}.{tool_name}")
-                tools = getattr(module, "tools")
+                tools = module.get_tools(**config)
                 if not all(isinstance(tool, FunctionTool) for tool in tools):
                     raise ValueError(
                         f"The module {module} does not contain valid tools"

templates/components/engines/python/agent/tools/duckduckgo.py

@@ -0,0 +1,36 @@
+from llama_index.core.tools.function_tool import FunctionTool
+
+
+def duckduckgo_search(
+    query: str,
+    region: str = "wt-wt",
+    max_results: int = 10,
+):
+    """
+    Use this function to search for any query in DuckDuckGo.
+    Args:
+        query (str): The query to search in DuckDuckGo.
+        region Optional(str): The region to be used for the search in [country-language] convention, ex us-en, uk-en, ru-ru, etc...
+        max_results Optional(int): The maximum number of results to be returned. Default is 10.
+    """
+    try:
+        from duckduckgo_search import DDGS
+    except ImportError:
+        raise ImportError(
+            "duckduckgo_search package is required to use this function."
+            "Please install it by running: `poetry add duckduckgo_search` or `pip install duckduckgo_search`"
+        )
+
+    params = {
+        "keywords": query,
+        "region": region,
+        "max_results": max_results,
+    }
+    results = []
+    with DDGS() as ddg:
+        results = list(ddg.text(**params))
+    return results
+
+
+def get_tools(**kwargs):
+    return [FunctionTool.from_defaults(duckduckgo_search)]

templates/components/engines/python/agent/tools/img_gen.py

@@ -0,0 +1,108 @@
+import os
+import uuid
+import logging
+import requests
+from typing import Optional
+from pydantic import BaseModel, Field
+from llama_index.core.tools import FunctionTool
+
+logger = logging.getLogger(__name__)
+
+
+class ImageGeneratorToolOutput(BaseModel):
+    is_success: bool = Field(
+        ...,
+        description="Whether the image generation was successful.",
+    )
+    image_url: Optional[str] = Field(
+        None,
+        description="The URL of the generated image.",
+    )
+    error_message: Optional[str] = Field(
+        None,
+        description="The error message if the image generation failed.",
+    )
+
+
+class ImageGeneratorTool:
+    _IMG_OUTPUT_FORMAT = "webp"
+    _IMG_OUTPUT_DIR = "output/tool"
+    _IMG_GEN_API = "https://api.stability.ai/v2beta/stable-image/generate/core"
+
+    def __init__(self, api_key: str = None):
+        if not api_key:
+            api_key = os.getenv("STABILITY_API_KEY")
+        self._api_key = api_key
+        self.fileserver_url_prefix = os.getenv("FILESERVER_URL_PREFIX")
+        if self._api_key is None:
+            raise ValueError(
+                "STABILITY_API_KEY key is required to run image generator. Get it here: https://platform.stability.ai/account/keys"
+            )
+        if self.fileserver_url_prefix is None:
+            raise ValueError("FILESERVER_URL_PREFIX is required.")
+
+    def _prepare_output_dir(self):
+        """
+        Create the output directory if it doesn't exist
+        """
+        if not os.path.exists(self._IMG_OUTPUT_DIR):
+            os.makedirs(self._IMG_OUTPUT_DIR, exist_ok=True)
+
+    def _save_image(self, image_data: bytes):
+        self._prepare_output_dir()
+        filename = f"{uuid.uuid4()}.{self._IMG_OUTPUT_FORMAT}"
+        output_path = os.path.join(self._IMG_OUTPUT_DIR, filename)
+        with open(output_path, "wb") as f:
+            f.write(image_data)
+        url = f"{os.getenv('FILESERVER_URL_PREFIX')}/{self._IMG_OUTPUT_DIR}/{filename}"
+        logger.info(f"Saved image to {output_path}.\nURL: {url}")
+        return url
+
+    def _call_stability_api(self, prompt: str):
+        headers = {
+            "authorization": f"Bearer {self._api_key}",
+            "accept": "image/*",
+        }
+        data = {
+            "prompt": prompt,
+            "output_format": self._IMG_OUTPUT_FORMAT,
+        }
+
+        response = requests.post(
+            self._IMG_GEN_API,
+            headers=headers,
+            files={"none": ""},
+            data=data,
+        )
+        response.raise_for_status()
+
+        return response
+
+    def generate_image(self, prompt: str) -> ImageGeneratorToolOutput:
+        """
+        Use this tool to generate an image based on the prompt.
+        Args:
+            prompt (str): The prompt to generate the image from.
+        """
+
+        try:
+            # Call the Stability API
+            response = self._call_stability_api(prompt)
+
+            # Save the image and get the URL
+            image_url = self._save_image(response.content)
+
+            return ImageGeneratorToolOutput(
+                is_success=True,
+                image_url=image_url,
+            )
+        except Exception as e:
+            logger.exception(e, exc_info=True)
+            return ImageGeneratorToolOutput(
+                is_success=False,
+                error_message=str(e),
+            )
+
+
+def get_tools(**kwargs):
+    return [FunctionTool.from_defaults(ImageGeneratorTool(**kwargs).generate_image)]

templates/components/engines/python/agent/tools/interpreter.py

@@ -0,0 +1,143 @@
+import os
+import logging
+import base64
+import uuid
+from pydantic import BaseModel
+from typing import List, Tuple, Dict, Optional
+from llama_index.core.tools import FunctionTool
+from e2b_code_interpreter import CodeInterpreter
+from e2b_code_interpreter.models import Logs
+
+
+logger = logging.getLogger(__name__)
+
+
+class InterpreterExtraResult(BaseModel):
+    type: str
+    content: Optional[str] = None
+    filename: Optional[str] = None
+    url: Optional[str] = None
+
+
+class E2BToolOutput(BaseModel):
+    is_error: bool
+    logs: Logs
+    results: List[InterpreterExtraResult] = []
+
+
+class E2BCodeInterpreter:
+
+    output_dir = "output/tool"
+
+    def __init__(self, api_key: str = None):
+        if api_key is None:
+            api_key = os.getenv("E2B_API_KEY")
+        filesever_url_prefix = os.getenv("FILESERVER_URL_PREFIX")
+        if not api_key:
+            raise ValueError(
+                "E2B_API_KEY key is required to run code interpreter. Get it here: https://e2b.dev/docs/getting-started/api-key"
+            )
+        if not filesever_url_prefix:
+            raise ValueError(
+                "FILESERVER_URL_PREFIX is required to display file output from sandbox"
+            )
+
+        self.filesever_url_prefix = filesever_url_prefix
+        self.interpreter = CodeInterpreter(api_key=api_key)
+
+    def __del__(self):
+        self.interpreter.close()
+
+    def get_output_path(self, filename: str) -> str:
+        # if output directory doesn't exist, create it
+        if not os.path.exists(self.output_dir):
+            os.makedirs(self.output_dir, exist_ok=True)
+        return os.path.join(self.output_dir, filename)
+
+    def save_to_disk(self, base64_data: str, ext: str) -> Dict:
+        filename = f"{uuid.uuid4()}.{ext}"  # generate a unique filename
+        buffer = base64.b64decode(base64_data)
+        output_path = self.get_output_path(filename)
+
+        try:
+            with open(output_path, "wb") as file:
+                file.write(buffer)
+        except IOError as e:
+            logger.error(f"Failed to write to file {output_path}: {str(e)}")
+            raise e
+
+        logger.info(f"Saved file to {output_path}")
+
+        return {
+            "outputPath": output_path,
+            "filename": filename,
+        }
+
+    def get_file_url(self, filename: str) -> str:
+        return f"{self.filesever_url_prefix}/{self.output_dir}/{filename}"
+
+    def parse_result(self, result) -> List[InterpreterExtraResult]:
+        """
+        The result could include multiple formats (e.g. png, svg, etc.) but encoded in base64
+        We save each result to disk and return saved file metadata (extension, filename, url)
+        """
+        if not result:
+            return []
+
+        output = []
+
+        try:
+            formats = result.formats()
+            results = [result[format] for format in formats]
+
+            for ext, data in zip(formats, results):
+                match ext:
+                    case "png" | "svg" | "jpeg" | "pdf":
+                        result = self.save_to_disk(data, ext)
+                        filename = result["filename"]
+                        output.append(
+                            InterpreterExtraResult(
+                                type=ext,
+                                filename=filename,
+                                url=self.get_file_url(filename),
+                            )
+                        )
+                    case _:
+                        output.append(
+                            InterpreterExtraResult(
+                                type=ext,
+                                content=data,
+                            )
+                        )
+        except Exception as error:
+            logger.exception(error, exc_info=True)
+            logger.error("Error when parsing output from E2b interpreter tool", error)
+
+        return output
+
+    def interpret(self, code: str) -> E2BToolOutput:
+        """
+        Execute python code in a Jupyter notebook cell, the toll will return result, stdout, stderr, display_data, and error.
+
+        Parameters:
+            code (str): The python code to be executed in a single cell.
+        """
+        logger.info(
+            f"\n{'='*50}\n> Running following AI-generated code:\n{code}\n{'='*50}"
+        )
+        exec = self.interpreter.notebook.exec_cell(code)
+
+        if exec.error:
+            logger.error("Error when executing code", exec.error)
+            output = E2BToolOutput(is_error=True, logs=exec.logs, results=[])
+        else:
+            if len(exec.results) == 0:
+                output = E2BToolOutput(is_error=False, logs=exec.logs, results=[])
+            else:
+                results = self.parse_result(exec.results[0])
+                output = E2BToolOutput(is_error=False, logs=exec.logs, results=results)
+        return output
+
+
+def get_tools(**kwargs):
+    return [FunctionTool.from_defaults(E2BCodeInterpreter(**kwargs).interpret)]

templates/components/engines/python/agent/tools/openapi_action.py

@@ -0,0 +1,78 @@
+from typing import Dict, List, Tuple
+from llama_index.tools.openapi import OpenAPIToolSpec
+from llama_index.tools.requests import RequestsToolSpec
+
+
+class OpenAPIActionToolSpec(OpenAPIToolSpec, RequestsToolSpec):
+    """
+    A combination of OpenAPI and Requests tool specs that can parse OpenAPI specs and make requests.
+
+    openapi_uri: str: The file path or URL to the OpenAPI spec.
+    domain_headers: dict: Whitelist domains and the headers to use.
+    """
+
+    spec_functions = OpenAPIToolSpec.spec_functions + RequestsToolSpec.spec_functions
+    # Cached parsed specs by URI
+    _specs: Dict[str, Tuple[Dict, List[str]]] = {}
+
+    def __init__(self, openapi_uri: str, domain_headers: dict = None, **kwargs):
+        if domain_headers is None:
+            domain_headers = {}
+        if openapi_uri not in self._specs:
+            openapi_spec, servers = self._load_openapi_spec(openapi_uri)
+            self._specs[openapi_uri] = (openapi_spec, servers)
+        else:
+            openapi_spec, servers = self._specs[openapi_uri]
+
+        # Add the servers to the domain headers if they are not already present
+        for server in servers:
+            if server not in domain_headers:
+                domain_headers[server] = {}
+
+        OpenAPIToolSpec.__init__(self, spec=openapi_spec)
+        RequestsToolSpec.__init__(self, domain_headers)
+
+    @staticmethod
+    def _load_openapi_spec(uri: str) -> Tuple[Dict, List[str]]:
+        """
+        Load an OpenAPI spec from a URI.
+
+        Args:
+            uri (str): A file path or URL to the OpenAPI spec.
+
+        Returns:
+            List[Document]: A list of Document objects.
+        """
+        import yaml
+        from urllib.parse import urlparse
+
+        if uri.startswith("http"):
+            import requests
+
+            response = requests.get(uri)
+            if response.status_code != 200:
+                raise ValueError(
+                    "Could not initialize OpenAPIActionToolSpec: "
+                    f"Failed to load OpenAPI spec from {uri}, status code: {response.status_code}"
+                )
+            spec = yaml.safe_load(response.text)
+        elif uri.startswith("file"):
+            filepath = urlparse(uri).path
+            with open(filepath, "r") as file:
+                spec = yaml.safe_load(file)
+        else:
+            raise ValueError(
+                "Could not initialize OpenAPIActionToolSpec: Invalid OpenAPI URI provided. "
+                "Only HTTP and file path are supported."
+            )
+        # Add the servers to the whitelist
+        try:
+            servers = [
+                urlparse(server["url"]).netloc for server in spec.get("servers", [])
+            ]
+        except KeyError as e:
+            raise ValueError(
+                "Could not initialize OpenAPIActionToolSpec: Invalid OpenAPI spec provided. "
+                "Could not get `servers` from the spec."
+            ) from e
+        return spec, servers

templates/components/engines/python/agent/tools/weather.py

@@ -69,4 +69,5 @@ class OpenMeteoWeather:
         return response.json()
 
 
-tools = [FunctionTool.from_defaults(OpenMeteoWeather.get_weather_information)]
+def get_tools(**kwargs):
+    return [FunctionTool.from_defaults(OpenMeteoWeather.get_weather_information)]

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

@@ -3,7 +3,7 @@ from app.engine.index import get_index
 from fastapi import HTTPException
 
 
-def get_chat_engine():
+def get_chat_engine(filters=None):
     system_prompt = os.getenv("SYSTEM_PROMPT")
     top_k = os.getenv("TOP_K", 3)
 
@@ -20,4 +20,5 @@ def get_chat_engine():
         similarity_top_k=int(top_k),
         system_prompt=system_prompt,
         chat_mode="condense_plus_context",
+        filters=filters,
     )

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

@@ -1,12 +1,10 @@
 import { BaseToolWithCall, OpenAIAgent, QueryEngineTool } from "llamaindex";
-import { ToolsFactory } from "llamaindex/tools/ToolsFactory";
 import fs from "node:fs/promises";
 import path from "node:path";
 import { getDataSource } from "./index";
-import { STORAGE_CACHE_DIR } from "./shared";
-import { createLocalTools } from "./tools";
+import { createTools } from "./tools";
 
-export async function createChatEngine() {
+export async function createChatEngine(documentIds?: string[]) {
   const tools: BaseToolWithCall[] = [];
 
   // Add a query engine tool if we have a data source
@@ -15,31 +13,31 @@ export async function createChatEngine() {
   if (index) {
     tools.push(
       new QueryEngineTool({
-        queryEngine: index.asQueryEngine(),
+        queryEngine: index.asQueryEngine({
+          preFilters: undefined, // TODO: Add filters once LITS supports it (getQueryFilters)
+        }),
         metadata: {
           name: "data_query_engine",
-          description: `A query engine for documents in storage folder: ${STORAGE_CACHE_DIR}`,
+          description: `A query engine for documents from your data source.`,
         },
       }),
     );
   }
 
+  const configFile = path.join("config", "tools.json");
+  let toolConfig: any;
   try {
     // add tools from config file if it exists
-    const config = JSON.parse(
-      await fs.readFile(path.join("config", "tools.json"), "utf8"),
-    );
-
-    // add local tools from the 'tools' folder (if configured)
-    const localTools = createLocalTools(config.local);
-    tools.push(...localTools);
-
-    // add tools from LlamaIndexTS (if configured)
-    const llamaTools = await ToolsFactory.createTools(config.llamahub);
-    tools.push(...llamaTools);
-  } catch {}
+    toolConfig = JSON.parse(await fs.readFile(configFile, "utf8"));
+  } catch (e) {
+    console.info(`Could not read ${configFile} file. Using no tools.`);
+  }
+  if (toolConfig) {
+    tools.push(...(await createTools(toolConfig)));
+  }
 
   return new OpenAIAgent({
     tools,
+    systemPrompt: process.env.SYSTEM_PROMPT,
   });
 }

templates/components/engines/typescript/agent/tools/duckduckgo.ts

@@ -0,0 +1,61 @@
+import { JSONSchemaType } from "ajv";
+import { search } from "duck-duck-scrape";
+import { BaseTool, ToolMetadata } from "llamaindex";
+
+export type DuckDuckGoParameter = {
+  query: string;
+  region?: string;
+};
+
+export type DuckDuckGoToolParams = {
+  metadata?: ToolMetadata<JSONSchemaType<DuckDuckGoParameter>>;
+};
+
+const DEFAULT_META_DATA: ToolMetadata<JSONSchemaType<DuckDuckGoParameter>> = {
+  name: "duckduckgo",
+  description: "Use this function to search for any query in DuckDuckGo.",
+  parameters: {
+    type: "object",
+    properties: {
+      query: {
+        type: "string",
+        description: "The query to search in DuckDuckGo.",
+      },
+      region: {
+        type: "string",
+        description:
+          "Optional, The region to be used for the search in [country-language] convention, ex us-en, uk-en, ru-ru, etc...",
+        nullable: true,
+      },
+    },
+    required: ["query"],
+  },
+};
+
+type DuckDuckGoSearchResult = {
+  title: string;
+  description: string;
+  url: string;
+};
+
+export class DuckDuckGoSearchTool implements BaseTool<DuckDuckGoParameter> {
+  metadata: ToolMetadata<JSONSchemaType<DuckDuckGoParameter>>;
+
+  constructor(params: DuckDuckGoToolParams) {
+    this.metadata = params.metadata ?? DEFAULT_META_DATA;
+  }
+
+  async call(input: DuckDuckGoParameter) {
+    const { query, region } = input;
+    const options = region ? { region } : {};
+    const searchResults = await search(query, options);
+
+    return searchResults.results.map((result) => {
+      return {
+        title: result.title,
+        description: result.description,
+        url: result.url,
+      } as DuckDuckGoSearchResult;
+    });
+  }
+}

templates/components/engines/typescript/agent/tools/img-gen.ts

@@ -0,0 +1,112 @@
+import type { JSONSchemaType } from "ajv";
+import { FormData } from "formdata-node";
+import fs from "fs";
+import got from "got";
+import { BaseTool, ToolMetadata } from "llamaindex";
+import path from "node:path";
+import { Readable } from "stream";
+
+export type ImgGeneratorParameter = {
+  prompt: string;
+};
+
+export type ImgGeneratorToolParams = {
+  metadata?: ToolMetadata<JSONSchemaType<ImgGeneratorParameter>>;
+};
+
+export type ImgGeneratorToolOutput = {
+  isSuccess: boolean;
+  imageUrl?: string;
+  errorMessage?: string;
+};
+
+const DEFAULT_META_DATA: ToolMetadata<JSONSchemaType<ImgGeneratorParameter>> = {
+  name: "image_generator",
+  description: `Use this function to generate an image based on the prompt.`,
+  parameters: {
+    type: "object",
+    properties: {
+      prompt: {
+        type: "string",
+        description: "The prompt to generate the image",
+      },
+    },
+    required: ["prompt"],
+  },
+};
+
+export class ImgGeneratorTool implements BaseTool<ImgGeneratorParameter> {
+  readonly IMG_OUTPUT_FORMAT = "webp";
+  readonly IMG_OUTPUT_DIR = "output/tool";
+  readonly IMG_GEN_API =
+    "https://api.stability.ai/v2beta/stable-image/generate/core";
+
+  metadata: ToolMetadata<JSONSchemaType<ImgGeneratorParameter>>;
+
+  constructor(params?: ImgGeneratorToolParams) {
+    this.checkRequiredEnvVars();
+    this.metadata = params?.metadata || DEFAULT_META_DATA;
+  }
+
+  async call(input: ImgGeneratorParameter): Promise<ImgGeneratorToolOutput> {
+    return await this.generateImage(input.prompt);
+  }
+
+  private generateImage = async (
+    prompt: string,
+  ): Promise<ImgGeneratorToolOutput> => {
+    try {
+      const buffer = await this.promptToImgBuffer(prompt);
+      const imageUrl = this.saveImage(buffer);
+      return { isSuccess: true, imageUrl };
+    } catch (error) {
+      console.error(error);
+      return {
+        isSuccess: false,
+        errorMessage: "Failed to generate image. Please try again.",
+      };
+    }
+  };
+
+  private promptToImgBuffer = async (prompt: string) => {
+    const form = new FormData();
+    form.append("prompt", prompt);
+    form.append("output_format", this.IMG_OUTPUT_FORMAT);
+    const buffer = await got
+      .post(this.IMG_GEN_API, {
+        // Not sure why it shows an type error when passing form to body
+        // Although I follow document: https://github.com/sindresorhus/got/blob/main/documentation/2-options.md#body
+        // Tt still works fine, so I make casting to unknown to avoid the typescript warning
+        // Found a similar issue: https://github.com/sindresorhus/got/discussions/1877
+        body: form as unknown as Buffer | Readable | string,
+        headers: {
+          Authorization: `Bearer ${process.env.STABILITY_API_KEY}`,
+          Accept: "image/*",
+        },
+      })
+      .buffer();
+    return buffer;
+  };
+
+  private saveImage = (buffer: Buffer) => {
+    const filename = `${crypto.randomUUID()}.${this.IMG_OUTPUT_FORMAT}`;
+    const outputPath = path.join(this.IMG_OUTPUT_DIR, filename);
+    fs.writeFileSync(outputPath, buffer);
+    const url = `${process.env.FILESERVER_URL_PREFIX}/${this.IMG_OUTPUT_DIR}/${filename}`;
+    console.log(`Saved image to ${outputPath}.\nURL: ${url}`);
+    return url;
+  };
+
+  private checkRequiredEnvVars = () => {
+    if (!process.env.STABILITY_API_KEY) {
+      throw new Error(
+        "STABILITY_API_KEY key is required to run image generator. Get it here: https://platform.stability.ai/account/keys",
+      );
+    }
+    if (!process.env.FILESERVER_URL_PREFIX) {
+      throw new Error(
+        "FILESERVER_URL_PREFIX is required to display file output after generation",
+      );
+    }
+  };
+}

templates/components/engines/typescript/agent/tools/index.ts

@@ -1,26 +1,61 @@
 import { BaseToolWithCall } from "llamaindex";
+import { ToolsFactory } from "llamaindex/tools/ToolsFactory";
+import { DuckDuckGoSearchTool, DuckDuckGoToolParams } from "./duckduckgo";
+import { ImgGeneratorTool, ImgGeneratorToolParams } from "./img-gen";
+import { InterpreterTool, InterpreterToolParams } from "./interpreter";
+import { OpenAPIActionTool } from "./openapi-action";
 import { WeatherTool, WeatherToolParams } from "./weather";
 
-type ToolCreator = (config: unknown) => BaseToolWithCall;
+type ToolCreator = (config: unknown) => Promise<BaseToolWithCall[]>;
+
+export async function createTools(toolConfig: {
+  local: Record<string, unknown>;
+  llamahub: any;
+}): Promise<BaseToolWithCall[]> {
+  // add local tools from the 'tools' folder (if configured)
+  const tools = await createLocalTools(toolConfig.local);
+  // add tools from LlamaIndexTS (if configured)
+  tools.push(...(await ToolsFactory.createTools(toolConfig.llamahub)));
+  return tools;
+}
 
 const toolFactory: Record<string, ToolCreator> = {
-  weather: (config: unknown) => {
-    return new WeatherTool(config as WeatherToolParams);
+  weather: async (config: unknown) => {
+    return [new WeatherTool(config as WeatherToolParams)];
+  },
+  interpreter: async (config: unknown) => {
+    return [new InterpreterTool(config as InterpreterToolParams)];
+  },
+  "openapi_action.OpenAPIActionToolSpec": async (config: unknown) => {
+    const { openapi_uri, domain_headers } = config as {
+      openapi_uri: string;
+      domain_headers: Record<string, Record<string, string>>;
+    };
+    const openAPIActionTool = new OpenAPIActionTool(
+      openapi_uri,
+      domain_headers,
+    );
+    return await openAPIActionTool.toToolFunctions();
+  },
+  duckduckgo: async (config: unknown) => {
+    return [new DuckDuckGoSearchTool(config as DuckDuckGoToolParams)];
+  },
+  img_gen: async (config: unknown) => {
+    return [new ImgGeneratorTool(config as ImgGeneratorToolParams)];
   },
 };
 
-export function createLocalTools(
+async function createLocalTools(
   localConfig: Record<string, unknown>,
-): BaseToolWithCall[] {
+): Promise<BaseToolWithCall[]> {
   const tools: BaseToolWithCall[] = [];
 
-  Object.keys(localConfig).forEach((key) => {
+  for (const [key, toolConfig] of Object.entries(localConfig)) {
     if (key in toolFactory) {
-      const toolConfig = localConfig[key];
-      const tool = toolFactory[key](toolConfig);
-      tools.push(tool);
+      const newTools = await toolFactory[key](toolConfig);
+      tools.push(...newTools);
     }
-  });
+  }
 
   return tools;
 }

templates/components/engines/typescript/agent/tools/interpreter.ts

@@ -0,0 +1,189 @@
+import { CodeInterpreter, Logs, Result } from "@e2b/code-interpreter";
+import type { JSONSchemaType } from "ajv";
+import fs from "fs";
+import { BaseTool, ToolMetadata } from "llamaindex";
+import crypto from "node:crypto";
+import path from "node:path";
+
+export type InterpreterParameter = {
+  code: string;
+};
+
+export type InterpreterToolParams = {
+  metadata?: ToolMetadata<JSONSchemaType<InterpreterParameter>>;
+  apiKey?: string;
+  fileServerURLPrefix?: string;
+};
+
+export type InterpreterToolOutput = {
+  isError: boolean;
+  logs: Logs;
+  extraResult: InterpreterExtraResult[];
+};
+
+type InterpreterExtraType =
+  | "html"
+  | "markdown"
+  | "svg"
+  | "png"
+  | "jpeg"
+  | "pdf"
+  | "latex"
+  | "json"
+  | "javascript";
+
+export type InterpreterExtraResult = {
+  type: InterpreterExtraType;
+  content?: string;
+  filename?: string;
+  url?: string;
+};
+
+const DEFAULT_META_DATA: ToolMetadata<JSONSchemaType<InterpreterParameter>> = {
+  name: "interpreter",
+  description:
+    "Execute python code in a Jupyter notebook cell and return any result, stdout, stderr, display_data, and error.",
+  parameters: {
+    type: "object",
+    properties: {
+      code: {
+        type: "string",
+        description: "The python code to execute in a single cell.",
+      },
+    },
+    required: ["code"],
+  },
+};
+
+export class InterpreterTool implements BaseTool<InterpreterParameter> {
+  private readonly outputDir = "output/tool";
+  private apiKey?: string;
+  private fileServerURLPrefix?: string;
+  metadata: ToolMetadata<JSONSchemaType<InterpreterParameter>>;
+  codeInterpreter?: CodeInterpreter;
+
+  constructor(params?: InterpreterToolParams) {
+    this.metadata = params?.metadata || DEFAULT_META_DATA;
+    this.apiKey = params?.apiKey || process.env.E2B_API_KEY;
+    this.fileServerURLPrefix =
+      params?.fileServerURLPrefix || process.env.FILESERVER_URL_PREFIX;
+
+    if (!this.apiKey) {
+      throw new Error(
+        "E2B_API_KEY key is required to run code interpreter. Get it here: https://e2b.dev/docs/getting-started/api-key",
+      );
+    }
+    if (!this.fileServerURLPrefix) {
+      throw new Error(
+        "FILESERVER_URL_PREFIX is required to display file output from sandbox",
+      );
+    }
+  }
+
+  public async initInterpreter() {
+    if (!this.codeInterpreter) {
+      this.codeInterpreter = await CodeInterpreter.create({
+        apiKey: this.apiKey,
+      });
+    }
+    return this.codeInterpreter;
+  }
+
+  public async codeInterpret(code: string): Promise<InterpreterToolOutput> {
+    console.log(
+      `\n${"=".repeat(50)}\n> Running following AI-generated code:\n${code}\n${"=".repeat(50)}`,
+    );
+    const interpreter = await this.initInterpreter();
+    const exec = await interpreter.notebook.execCell(code);
+    if (exec.error) console.error("[Code Interpreter error]", exec.error);
+    const extraResult = await this.getExtraResult(exec.results[0]);
+    const result: InterpreterToolOutput = {
+      isError: !!exec.error,
+      logs: exec.logs,
+      extraResult,
+    };
+    return result;
+  }
+
+  async call(input: InterpreterParameter): Promise<InterpreterToolOutput> {
+    const result = await this.codeInterpret(input.code);
+    return result;
+  }
+
+  async close() {
+    await this.codeInterpreter?.close();
+  }
+
+  private async getExtraResult(
+    res?: Result,
+  ): Promise<InterpreterExtraResult[]> {
+    if (!res) return [];
+    const output: InterpreterExtraResult[] = [];
+
+    try {
+      const formats = res.formats(); // formats available for the result. Eg: ['png', ...]
+      const results = formats.map((f) => res[f as keyof Result]); // get base64 data for each format
+
+      // save base64 data to file and return the url
+      for (let i = 0; i < formats.length; i++) {
+        const ext = formats[i];
+        const data = results[i];
+        switch (ext) {
+          case "png":
+          case "jpeg":
+          case "svg":
+          case "pdf":
+            const { filename } = this.saveToDisk(data, ext);
+            output.push({
+              type: ext as InterpreterExtraType,
+              filename,
+              url: this.getFileUrl(filename),
+            });
+            break;
+          default:
+            output.push({
+              type: ext as InterpreterExtraType,
+              content: data,
+            });
+            break;
+        }
+      }
+    } catch (error) {
+      console.error("Error when parsing e2b response", error);
+    }
+
+    return output;
+  }
+
+  // Consider saving to cloud storage instead but it may cost more for you
+  // See: https://e2b.dev/docs/sandbox/api/filesystem#write-to-file
+  private saveToDisk(
+    base64Data: string,
+    ext: string,
+  ): {
+    outputPath: string;
+    filename: string;
+  } {
+    const filename = `${crypto.randomUUID()}.${ext}`; // generate a unique filename
+    const buffer = Buffer.from(base64Data, "base64");
+    const outputPath = this.getOutputPath(filename);
+    fs.writeFileSync(outputPath, buffer);
+    console.log(`Saved file to ${outputPath}`);
+    return {
+      outputPath,
+      filename,
+    };
+  }
+
+  private getOutputPath(filename: string): string {
+    // if outputDir doesn't exist, create it
+    if (!fs.existsSync(this.outputDir)) {
+      fs.mkdirSync(this.outputDir, { recursive: true });
+    }
+    return path.join(this.outputDir, filename);
+  }
+
+  private getFileUrl(filename: string): string {
+    return `${this.fileServerURLPrefix}/${this.outputDir}/${filename}`;
+  }
+}

templates/components/engines/typescript/agent/tools/openapi-action.ts

@@ -0,0 +1,164 @@
+import SwaggerParser from "@apidevtools/swagger-parser";
+import { JSONSchemaType } from "ajv";
+import got from "got";
+import { FunctionTool, JSONValue, ToolMetadata } from "llamaindex";
+
+interface DomainHeaders {
+  [key: string]: { [header: string]: string };
+}
+
+type Input = {
+  url: string;
+  params: object;
+};
+
+type APIInfo = {
+  description: string;
+  title: string;
+};
+
+export class OpenAPIActionTool {
+  // cache the loaded specs by URL
+  private static specs: Record<string, any> = {};
+
+  private readonly INVALID_URL_PROMPT =
+    "This url did not include a hostname or scheme. Please determine the complete URL and try again.";
+
+  private createLoadSpecMetaData = (info: APIInfo) => {
+    return {
+      name: "load_openapi_spec",
+      description: `Use this to retrieve the OpenAPI spec for the API named ${info.title} with the following description: ${info.description}. Call it before making any requests to the API.`,
+    };
+  };
+
+  private readonly createMethodCallMetaData = (
+    method: "POST" | "PATCH" | "GET",
+    info: APIInfo,
+  ) => {
+    return {
+      name: `${method.toLowerCase()}_request`,
+      description: `Use this to call the ${method} method on the API named ${info.title}`,
+      parameters: {
+        type: "object",
+        properties: {
+          url: {
+            type: "string",
+            description: `The url to make the ${method} request against`,
+          },
+          params: {
+            type: "object",
+            description:
+              method === "GET"
+                ? "the URL parameters to provide with the get request"
+                : `the key-value pairs to provide with the ${method} request`,
+          },
+        },
+        required: ["url"],
+      },
+    } as ToolMetadata<JSONSchemaType<Input>>;
+  };
+
+  constructor(
+    public openapi_uri: string,
+    public domainHeaders: DomainHeaders = {},
+  ) {}
+
+  async loadOpenapiSpec(url: string): Promise<any> {
+    const api = await SwaggerParser.validate(url);
+    return {
+      servers: "servers" in api ? api.servers : "",
+      info: { description: api.info.description, title: api.info.title },
+      endpoints: api.paths,
+    };
+  }
+
+  async getRequest(input: Input): Promise<JSONValue> {
+    if (!this.validUrl(input.url)) {
+      return this.INVALID_URL_PROMPT;
+    }
+    try {
+      const data = await got
+        .get(input.url, {
+          headers: this.getHeadersForUrl(input.url),
+          searchParams: input.params as URLSearchParams,
+        })
+        .json();
+      return data as JSONValue;
+    } catch (error) {
+      return error as JSONValue;
+    }
+  }
+
+  async postRequest(input: Input): Promise<JSONValue> {
+    if (!this.validUrl(input.url)) {
+      return this.INVALID_URL_PROMPT;
+    }
+    try {
+      const res = await got.post(input.url, {
+        headers: this.getHeadersForUrl(input.url),
+        json: input.params,
+      });
+      return res.body as JSONValue;
+    } catch (error) {
+      return error as JSONValue;
+    }
+  }
+
+  async patchRequest(input: Input): Promise<JSONValue> {
+    if (!this.validUrl(input.url)) {
+      return this.INVALID_URL_PROMPT;
+    }
+    try {
+      const res = await got.patch(input.url, {
+        headers: this.getHeadersForUrl(input.url),
+        json: input.params,
+      });
+      return res.body as JSONValue;
+    } catch (error) {
+      return error as JSONValue;
+    }
+  }
+
+  public async toToolFunctions() {
+    if (!OpenAPIActionTool.specs[this.openapi_uri]) {
+      console.log(`Loading spec for URL: ${this.openapi_uri}`);
+      const spec = await this.loadOpenapiSpec(this.openapi_uri);
+      OpenAPIActionTool.specs[this.openapi_uri] = spec;
+    }
+    const spec = OpenAPIActionTool.specs[this.openapi_uri];
+    // TODO: read endpoints with parameters from spec and create one tool for each endpoint
+    // For now, we just create a tool for each HTTP method which does not work well for passing parameters
+    return [
+      FunctionTool.from(() => {
+        return spec;
+      }, this.createLoadSpecMetaData(spec.info)),
+      FunctionTool.from(
+        this.getRequest.bind(this),
+        this.createMethodCallMetaData("GET", spec.info),
+      ),
+      FunctionTool.from(
+        this.postRequest.bind(this),
+        this.createMethodCallMetaData("POST", spec.info),
+      ),
+      FunctionTool.from(
+        this.patchRequest.bind(this),
+        this.createMethodCallMetaData("PATCH", spec.info),
+      ),
+    ];
+  }
+
+  private validUrl(url: string): boolean {
+    const parsed = new URL(url);
+    return !!parsed.protocol && !!parsed.hostname;
+  }
+
+  private getDomain(url: string): string {
+    const parsed = new URL(url);
+    return parsed.hostname;
+  }
+
+  private getHeadersForUrl(url: string): { [header: string]: string } {
+    const domain = this.getDomain(url);
+    return this.domainHeaders[domain] || {};
+  }
+}

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

@@ -1,20 +1,21 @@
 import { ContextChatEngine, Settings } from "llamaindex";
 import { getDataSource } from "./index";
 
-export async function createChatEngine() {
+export async function createChatEngine(documentIds?: string[]) {
   const index = await getDataSource();
   if (!index) {
     throw new Error(
       `StorageContext is empty - call 'npm run generate' to generate the storage first`,
     );
   }
-  const retriever = index.asRetriever();
-  retriever.similarityTopK = process.env.TOP_K
-    ? parseInt(process.env.TOP_K)
-    : 3;
+  const retriever = index.asRetriever({
+    similarityTopK: process.env.TOP_K ? parseInt(process.env.TOP_K) : 3,
+  });
 
   return new ContextChatEngine({
     chatModel: Settings.llm,
     retriever,
+    // disable as a custom system prompt disables the generated context
+    // systemPrompt: process.env.SYSTEM_PROMPT,
   });
 }

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

@@ -0,0 +1,115 @@
+import fs from "fs";
+import {
+  BaseNode,
+  Document,
+  IngestionPipeline,
+  Metadata,
+  Settings,
+  SimpleNodeParser,
+  storageContextFromDefaults,
+  VectorStoreIndex,
+} from "llamaindex";
+import { DocxReader } from "llamaindex/readers/DocxReader";
+import { PDFReader } from "llamaindex/readers/PDFReader";
+import { TextFileReader } from "llamaindex/readers/TextFileReader";
+import crypto from "node:crypto";
+import { getDataSource } from "../../engine";
+
+const MIME_TYPE_TO_EXT: Record<string, string> = {
+  "application/pdf": "pdf",
+  "text/plain": "txt",
+  "application/vnd.openxmlformats-officedocument.wordprocessingml.document":
+    "docx",
+};
+
+export async function uploadDocument(raw: string): Promise<string[]> {
+  const [header, content] = raw.split(",");
+  const mimeType = header.replace("data:", "").replace(";base64", "");
+  const fileBuffer = Buffer.from(content, "base64");
+  const documents = await loadDocuments(fileBuffer, mimeType);
+  const { filename } = await saveDocument(fileBuffer, mimeType);
+  return await runPipeline(documents, filename);
+}
+
+async function runPipeline(
+  documents: Document[],
+  filename: string,
+): Promise<string[]> {
+  // mark documents to add to the vector store as private
+  for (const document of documents) {
+    document.metadata = {
+      ...document.metadata,
+      file_name: filename,
+      private: true,
+    };
+  }
+  const pipeline = new IngestionPipeline({
+    transformations: [
+      new SimpleNodeParser({
+        chunkSize: Settings.chunkSize,
+        chunkOverlap: Settings.chunkOverlap,
+      }),
+      Settings.embedModel,
+    ],
+  });
+  const nodes = await pipeline.run({ documents });
+  await addNodesToVectorStore(nodes);
+  return documents.map((document) => document.id_);
+}
+
+async function loadDocuments(fileBuffer: Buffer, mimeType: string) {
+  console.log(`Processing uploaded document of type: ${mimeType}`);
+  switch (mimeType) {
+    case "application/pdf": {
+      const pdfReader = new PDFReader();
+      return await pdfReader.loadDataAsContent(new Uint8Array(fileBuffer));
+    }
+    case "text/plain": {
+      const textReader = new TextFileReader();
+      return await textReader.loadDataAsContent(fileBuffer);
+    }
+    case "application/vnd.openxmlformats-officedocument.wordprocessingml.document": {
+      const docxReader = new DocxReader();
+      return await docxReader.loadDataAsContent(fileBuffer);
+    }
+    default:
+      throw new Error(`Unsupported document type: ${mimeType}`);
+  }
+}
+
+async function saveDocument(fileBuffer: Buffer, mimeType: string) {
+  const fileExt = MIME_TYPE_TO_EXT[mimeType];
+  if (!fileExt) throw new Error(`Unsupported document type: ${mimeType}`);
+
+  const folder = "output/uploaded";
+  const filename = `${crypto.randomUUID()}.${fileExt}`;
+  const filepath = `${folder}/${filename}`;
+  const fileurl = `${process.env.FILESERVER_URL_PREFIX}/${filepath}`;
+
+  if (!fs.existsSync(folder)) {
+    fs.mkdirSync(folder, { recursive: true });
+  }
+  await fs.promises.writeFile(filepath, fileBuffer);
+
+  console.log(`Saved document file to ${filepath}.\nURL: ${fileurl}`);
+  return {
+    filename,
+    filepath,
+    fileurl,
+  };
+}
+
+async function addNodesToVectorStore(nodes: BaseNode<Metadata>[]) {
+  let currentIndex = await getDataSource(); // always not null with an vectordb
+  if (currentIndex) {
+    await currentIndex.insertNodes(nodes);
+  } else {
+    // Not using vectordb and haven't generated local index yet
+    const storageContext = await storageContextFromDefaults({
+      persistDir: "./cache",
+    });
+    currentIndex = await VectorStoreIndex.init({ nodes, storageContext });
+  }
+  currentIndex.storageContext.docStore.persist();
+  console.log("Added nodes to the vector store.");
+}

templates/components/llamaindex/typescript/streaming/annotations.ts

@@ -0,0 +1,124 @@
+import { JSONValue } from "ai";
+import { MessageContent, MessageContentDetail } from "llamaindex";
+
+export type DocumentFileType = "csv" | "pdf" | "txt" | "docx";
+
+export type DocumentFileContent = {
+  type: "ref" | "text";
+  value: string[] | string;
+};
+
+export type DocumentFile = {
+  id: string;
+  filename: string;
+  filesize: number;
+  filetype: DocumentFileType;
+  content: DocumentFileContent;
+};
+
+type Annotation = {
+  type: string;
+  data: object;
+};
+
+export function retrieveDocumentIds(annotations?: JSONValue[]): string[] {
+  if (!annotations) return [];
+
+  const ids: string[] = [];
+
+  for (const annotation of annotations) {
+    const { type, data } = getValidAnnotation(annotation);
+    if (
+      type === "document_file" &&
+      "files" in data &&
+      Array.isArray(data.files)
+    ) {
+      const files = data.files as DocumentFile[];
+      for (const file of files) {
+        if (Array.isArray(file.content)) {
+          // it's an array, so it's an array of doc IDs
+          for (const id of file.content) {
+            ids.push(id);
+          }
+        }
+      }
+    }
+  }
+
+  return ids;
+}
+
+export function convertMessageContent(
+  content: string,
+  annotations?: JSONValue[],
+): MessageContent {
+  if (!annotations) return content;
+  return [
+    {
+      type: "text",
+      text: content,
+    },
+    ...convertAnnotations(annotations),
+  ];
+}
+
+function convertAnnotations(annotations: JSONValue[]): MessageContentDetail[] {
+  const content: MessageContentDetail[] = [];
+  annotations.forEach((annotation: JSONValue) => {
+    const { type, data } = getValidAnnotation(annotation);
+    // convert image
+    if (type === "image" && "url" in data && typeof data.url === "string") {
+      content.push({
+        type: "image_url",
+        image_url: {
+          url: data.url,
+        },
+      });
+    }
+    // convert the content of files to a text message
+    if (
+      type === "document_file" &&
+      "files" in data &&
+      Array.isArray(data.files)
+    ) {
+      // get all CSV files and convert their whole content to one text message
+      // currently CSV files are the only files where we send the whole content - we don't use an index
+      const csvFiles: DocumentFile[] = data.files.filter(
+        (file: DocumentFile) => file.filetype === "csv",
+      );
+      if (csvFiles && csvFiles.length > 0) {
+        const csvContents = csvFiles.map((file: DocumentFile) => {
+          const fileContent = Array.isArray(file.content.value)
+            ? file.content.value.join("\n")
+            : file.content.value;
+          return "```csv\n" + fileContent + "\n```";
+        });
+        const text =
+          "Use the following CSV content:\n" + csvContents.join("\n\n");
+        content.push({
+          type: "text",
+          text,
+        });
+      }
+    }
+  });
+
+  return content;
+}
+
+function getValidAnnotation(annotation: JSONValue): Annotation {
+  if (
+    !(
+      annotation &&
+      typeof annotation === "object" &&
+      "type" in annotation &&
+      typeof annotation.type === "string" &&
+      "data" in annotation &&
+      annotation.data &&
+      typeof annotation.data === "object"
+    )
+  ) {
+    throw new Error("Client sent invalid annotation. Missing data and type");
+  }
+  return { type: annotation.type, data: annotation.data };
+}

templates/components/llamaindex/typescript/streaming/events.ts

@@ -7,16 +7,6 @@ import {
   ToolOutput,
 } from "llamaindex";
 
-export function appendImageData(data: StreamData, imageUrl?: string) {
-  if (!imageUrl) return;
-  data.appendMessageAnnotation({
-    type: "image",
-    data: {
-      url: imageUrl,
-    },
-  });
-}
-
 export function appendSourceData(
   data: StreamData,
   sourceNodes?: NodeWithScore<Metadata>[],
@@ -29,6 +19,7 @@ export function appendSourceData(
         ...node.node.toMutableJSON(),
         id: node.node.id_,
         score: node.score ?? null,
+        url: getNodeUrl(node.node.metadata),
       })),
     },
   });
@@ -65,11 +56,21 @@ export function appendToolData(
   });
 }
 
+export function createStreamTimeout(stream: StreamData) {
+  const timeout = Number(process.env.STREAM_TIMEOUT ?? 1000 * 60 * 5); // default to 5 minutes
+  const t = setTimeout(() => {
+    appendEventData(stream, `Stream timed out after ${timeout / 1000} seconds`);
+    stream.close();
+  }, timeout);
+  return t;
+}
+
 export function createCallbackManager(stream: StreamData) {
   const callbackManager = new CallbackManager();
 
-  callbackManager.on("retrieve", (data) => {
-    const { nodes, query } = data.detail;
+  callbackManager.on("retrieve-end", (data) => {
+    const { nodes, query } = data.detail.payload;
+    appendSourceData(stream, nodes);
     appendEventData(stream, `Retrieving context for query: '${query}'`);
     appendEventData(
       stream,
@@ -95,3 +96,20 @@ export function createCallbackManager(stream: StreamData) {
 
   return callbackManager;
 }
+
+function getNodeUrl(metadata: Metadata) {
+  const url = metadata["URL"];
+  if (url) return url;
+  const fileName = metadata["file_name"];
+  if (!process.env.FILESERVER_URL_PREFIX) {
+    console.warn(
+      "FILESERVER_URL_PREFIX is not set. File URLs will not be generated.",
+    );
+    return undefined;
+  }
+  if (fileName) {
+    const folder = metadata["private"] ? "output/uploaded" : "data";
+    return `${process.env.FILESERVER_URL_PREFIX}/${folder}/${fileName}`;
+  }
+  return undefined;
+}

templates/components/llamaindex/typescript/streaming/stream.ts

@@ -0,0 +1,40 @@
+import {
+  StreamData,
+  createCallbacksTransformer,
+  createStreamDataTransformer,
+  trimStartOfStreamHelper,
+  type AIStreamCallbacksAndOptions,
+} from "ai";
+import { EngineResponse } from "llamaindex";
+
+export function LlamaIndexStream(
+  response: AsyncIterable<EngineResponse>,
+  data: StreamData,
+  opts?: {
+    callbacks?: AIStreamCallbacksAndOptions;
+  },
+): ReadableStream<Uint8Array> {
+  return createParser(response, data)
+    .pipeThrough(createCallbacksTransformer(opts?.callbacks))
+    .pipeThrough(createStreamDataTransformer());
+}
+
+function createParser(res: AsyncIterable<EngineResponse>, data: StreamData) {
+  const it = res[Symbol.asyncIterator]();
+  const trimStartOfStream = trimStartOfStreamHelper();
+
+  return new ReadableStream<string>({
+    async pull(controller): Promise<void> {
+      const { value, done } = await it.next();
+      if (done) {
+        controller.close();
+        data.close();
+        return;
+      }
+      const text = trimStartOfStream(value.delta ?? "");
+      if (text) {
+        controller.enqueue(text);
+      }
+    },
+  });
+}

templates/components/loaders/python/file.py

@@ -1,7 +1,11 @@
 import os
+import logging
+from typing import Dict
 from llama_parse import LlamaParse
 from pydantic import BaseModel, validator
 
+logger = logging.getLogger(__name__)
+
 
 class FileLoaderConfig(BaseModel):
     data_dir: str = "data"
@@ -20,18 +24,55 @@ def llama_parse_parser():
             "LLAMA_CLOUD_API_KEY environment variable is not set. "
             "Please set it in .env file or in your shell environment then run again!"
         )
-    parser = LlamaParse(result_type="markdown", verbose=True, language="en")
+    parser = LlamaParse(
+        result_type="markdown",
+        verbose=True,
+        language="en",
+        ignore_errors=False,
+    )
     return parser
 
 
+def llama_parse_extractor() -> Dict[str, LlamaParse]:
+    from llama_parse.utils import SUPPORTED_FILE_TYPES
+
+    parser = llama_parse_parser()
+    return {file_type: parser for file_type in SUPPORTED_FILE_TYPES}
+
+
 def get_file_documents(config: FileLoaderConfig):
     from llama_index.core.readers import SimpleDirectoryReader
 
-    reader = SimpleDirectoryReader(
-        config.data_dir,
-        recursive=True,
-    )
-    if config.use_llama_parse:
-        parser = llama_parse_parser()
-        reader.file_extractor = {".pdf": parser}
-    return reader.load_data()
+    try:
+        file_extractor = None
+        if config.use_llama_parse:
+            # LlamaParse is async first,
+            # so we need to use nest_asyncio to run it in sync mode
+            import nest_asyncio
+
+            nest_asyncio.apply()
+
+            file_extractor = llama_parse_extractor()
+        reader = SimpleDirectoryReader(
+            config.data_dir,
+            recursive=True,
+            filename_as_id=True,
+            raise_on_error=True,
+            file_extractor=file_extractor,
+        )
+        return reader.load_data()
+    except Exception as e:
+        import sys, traceback
+
+        # Catch the error if the data dir is empty
+        # and return as empty document list
+        _, _, exc_traceback = sys.exc_info()
+        function_name = traceback.extract_tb(exc_traceback)[-1].name
+        if function_name == "_add_files":
+            logger.warning(
+                f"Failed to load file documents, error message: {e} . Return as empty document list."
+            )
+            return []
+        else:
+            # Raise the error if it is not the case of empty data dir
+            raise e

templates/components/settings/python/llmhub.py

@@ -0,0 +1,61 @@
+from llama_index.embeddings.openai import OpenAIEmbedding
+from llama_index.core.settings import Settings
+from typing import Dict
+import os
+
+DEFAULT_MODEL = "gpt-3.5-turbo"
+DEFAULT_EMBEDDING_MODEL = "text-embedding-3-large"
+
+class TSIEmbedding(OpenAIEmbedding):
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        self._query_engine = self._text_engine = self.model_name
+
+def llm_config_from_env() -> Dict:
+    from llama_index.core.constants import DEFAULT_TEMPERATURE
+
+    model = os.getenv("MODEL", DEFAULT_MODEL)
+    temperature = os.getenv("LLM_TEMPERATURE", DEFAULT_TEMPERATURE)
+    max_tokens = os.getenv("LLM_MAX_TOKENS")
+    api_key = os.getenv("T_SYSTEMS_LLMHUB_API_KEY")
+    api_base = os.getenv("T_SYSTEMS_LLMHUB_BASE_URL")
+
+    config = {
+        "model": model,
+        "api_key": api_key,
+        "api_base": api_base,
+        "temperature": float(temperature),
+        "max_tokens": int(max_tokens) if max_tokens is not None else None,
+    }
+    return config
+
+
+def embedding_config_from_env() -> Dict:
+    from llama_index.core.constants import DEFAULT_EMBEDDING_DIM
+    
+    model = os.getenv("EMBEDDING_MODEL", DEFAULT_EMBEDDING_MODEL)
+    dimension = os.getenv("EMBEDDING_DIM", DEFAULT_EMBEDDING_DIM)
+    api_key = os.getenv("T_SYSTEMS_LLMHUB_API_KEY")
+    api_base = os.getenv("T_SYSTEMS_LLMHUB_BASE_URL")
+
+    config = {
+        "model_name": model,
+        "dimension": int(dimension) if dimension is not None else None,
+        "api_key": api_key,
+        "api_base": api_base,
+    }
+    return config
+
+def init_llmhub():
+    from llama_index.llms.openai_like import OpenAILike
+
+    llm_configs = llm_config_from_env()
+    embedding_configs = embedding_config_from_env()
+
+    Settings.embed_model = TSIEmbedding(**embedding_configs)
+    Settings.llm = OpenAILike(
+        **llm_configs,
+        is_chat_model=True,
+        is_function_calling_model=False,
+        context_window=4096,
+    )

templates/components/settings/python/settings.py

@@ -0,0 +1,166 @@
+import os
+from typing import Dict
+
+from llama_index.core.settings import Settings
+
+
+def init_settings():
+    model_provider = os.getenv("MODEL_PROVIDER")
+    match model_provider:
+        case "openai":
+            init_openai()
+        case "groq":
+            init_groq()
+        case "ollama":
+            init_ollama()
+        case "anthropic":
+            init_anthropic()
+        case "gemini":
+            init_gemini()
+        case "mistral":
+            init_mistral()
+        case "azure-openai":
+            init_azure_openai()
+        case "t-systems":
+            from .llmhub import init_llmhub
+
+            init_llmhub()
+        case _:
+            raise ValueError(f"Invalid model provider: {model_provider}")
+
+    Settings.chunk_size = int(os.getenv("CHUNK_SIZE", "1024"))
+    Settings.chunk_overlap = int(os.getenv("CHUNK_OVERLAP", "20"))
+
+
+def init_ollama():
+    from llama_index.embeddings.ollama import OllamaEmbedding
+    from llama_index.llms.ollama.base import DEFAULT_REQUEST_TIMEOUT, Ollama
+
+    base_url = os.getenv("OLLAMA_BASE_URL") or "http://127.0.0.1:11434"
+    request_timeout = float(
+        os.getenv("OLLAMA_REQUEST_TIMEOUT", DEFAULT_REQUEST_TIMEOUT)
+    )
+    Settings.embed_model = OllamaEmbedding(
+        base_url=base_url,
+        model_name=os.getenv("EMBEDDING_MODEL"),
+    )
+    Settings.llm = Ollama(
+        base_url=base_url, model=os.getenv("MODEL"), request_timeout=request_timeout
+    )
+
+
+def init_openai():
+    from llama_index.core.constants import DEFAULT_TEMPERATURE
+    from llama_index.embeddings.openai import OpenAIEmbedding
+    from llama_index.llms.openai import OpenAI
+
+    max_tokens = os.getenv("LLM_MAX_TOKENS")
+    config = {
+        "model": os.getenv("MODEL"),
+        "temperature": float(os.getenv("LLM_TEMPERATURE", DEFAULT_TEMPERATURE)),
+        "max_tokens": int(max_tokens) if max_tokens is not None else None,
+    }
+    Settings.llm = OpenAI(**config)
+
+    dimensions = os.getenv("EMBEDDING_DIM")
+    config = {
+        "model": os.getenv("EMBEDDING_MODEL"),
+        "dimensions": int(dimensions) if dimensions is not None else None,
+    }
+    Settings.embed_model = OpenAIEmbedding(**config)
+
+
+def init_azure_openai():
+    from llama_index.core.constants import DEFAULT_TEMPERATURE
+    from llama_index.embeddings.azure_openai import AzureOpenAIEmbedding
+    from llama_index.llms.azure_openai import AzureOpenAI
+
+    llm_deployment = os.getenv("AZURE_OPENAI_LLM_DEPLOYMENT")
+    embedding_deployment = os.getenv("AZURE_OPENAI_EMBEDDING_DEPLOYMENT")
+    max_tokens = os.getenv("LLM_MAX_TOKENS")
+    api_key = os.getenv("AZURE_OPENAI_API_KEY")
+    llm_config = {
+        "api_key": api_key,
+        "deployment_name": llm_deployment,
+        "model": os.getenv("MODEL"),
+        "temperature": float(os.getenv("LLM_TEMPERATURE", DEFAULT_TEMPERATURE)),
+        "max_tokens": int(max_tokens) if max_tokens is not None else None,
+    }
+    Settings.llm = AzureOpenAI(**llm_config)
+
+    dimensions = os.getenv("EMBEDDING_DIM")
+    embedding_config = {
+        "api_key": api_key,
+        "deployment_name": embedding_deployment,
+        "model": os.getenv("EMBEDDING_MODEL"),
+        "dimensions": int(dimensions) if dimensions is not None else None,
+    }
+    Settings.embed_model = AzureOpenAIEmbedding(**embedding_config)
+
+
+def init_fastembed():
+    """
+    Use Qdrant Fastembed as the local embedding provider.
+    """
+    from llama_index.embeddings.fastembed import FastEmbedEmbedding
+
+    embed_model_map: Dict[str, str] = {
+        # Small and multilingual
+        "all-MiniLM-L6-v2": "sentence-transformers/all-MiniLM-L6-v2",
+        # Large and multilingual
+        "paraphrase-multilingual-mpnet-base-v2": "sentence-transformers/paraphrase-multilingual-mpnet-base-v2",   # noqa: E501
+    }
+
+    # This will download the model automatically if it is not already downloaded
+    Settings.embed_model = FastEmbedEmbedding(
+        model_name=embed_model_map[os.getenv("EMBEDDING_MODEL")]
+    )
+
+def init_groq():
+    from llama_index.llms.groq import Groq
+
+    model_map: Dict[str, str] = {
+        "llama3-8b": "llama3-8b-8192",
+        "llama3-70b": "llama3-70b-8192",
+        "mixtral-8x7b": "mixtral-8x7b-32768",
+    }
+
+
+    Settings.llm = Groq(model=model_map[os.getenv("MODEL")])
+    # Groq does not provide embeddings, so we use FastEmbed instead
+    init_fastembed()
+
+
+def init_anthropic():
+    from llama_index.llms.anthropic import Anthropic
+
+    model_map: Dict[str, str] = {
+        "claude-3-opus": "claude-3-opus-20240229",
+        "claude-3-sonnet": "claude-3-sonnet-20240229",
+        "claude-3-haiku": "claude-3-haiku-20240307",
+        "claude-2.1": "claude-2.1",
+        "claude-instant-1.2": "claude-instant-1.2",
+    }
+
+    Settings.llm = Anthropic(model=model_map[os.getenv("MODEL")])
+    # Anthropic does not provide embeddings, so we use FastEmbed instead
+    init_fastembed()
+
+
+def init_gemini():
+    from llama_index.embeddings.gemini import GeminiEmbedding
+    from llama_index.llms.gemini import Gemini
+
+    model_name = f"models/{os.getenv('MODEL')}"
+    embed_model_name = f"models/{os.getenv('EMBEDDING_MODEL')}"
+
+    Settings.llm = Gemini(model=model_name)
+    Settings.embed_model = GeminiEmbedding(model_name=embed_model_name)
+
+
+def init_mistral():
+    from llama_index.embeddings.mistralai import MistralAIEmbedding
+    from llama_index.llms.mistralai import MistralAI
+
+    Settings.llm = MistralAI(model=os.getenv("MODEL"))
+    Settings.embed_model = MistralAIEmbedding(model_name=os.getenv("EMBEDDING_MODEL"))

templates/components/ui/html/chat/chat-input.tsx

@@ -1,5 +1,7 @@
 "use client";
 
+import { Message } from "./chat-messages";
+
 export interface ChatInputProps {
   /** The current value of the input */
   input?: string;
@@ -12,7 +14,8 @@ export interface ChatInputProps {
   /** Form submission handler to automatically reset input and append a user message  */
   handleSubmit: (e: React.FormEvent<HTMLFormElement>) => void;
   isLoading: boolean;
-  multiModal?: boolean;
+  messages: Message[];
+  setInput?: (input: string) => void;
 }
 
 export default function ChatInput(props: ChatInputProps) {

templates/components/ui/html/chat/chat-messages.tsx

@@ -19,8 +19,12 @@ export default function ChatMessages({
   isLoading?: boolean;
   stop?: () => void;
   reload?: () => void;
+  append?: (
+    message: Message | Omit<Message, "id">,
+  ) => Promise<string | null | undefined>;
 }) {
   const scrollableChatContainerRef = useRef<HTMLDivElement>(null);
+  const lastMessage = messages[messages.length - 1];
 
   const scrollToBottom = () => {
     if (scrollableChatContainerRef.current) {
@@ -31,14 +35,14 @@ export default function ChatMessages({
 
   useEffect(() => {
     scrollToBottom();
-  }, [messages.length]);
+  }, [messages.length, lastMessage]);
 
   return (
-    <div className="w-full max-w-5xl p-4 bg-white rounded-xl shadow-xl">
-      <div
-        className="flex flex-col gap-5 divide-y h-[50vh] overflow-auto"
-        ref={scrollableChatContainerRef}
-      >
+    <div
+      className="flex-1 w-full max-w-5xl p-4 bg-white rounded-xl shadow-xl overflow-auto"
+      ref={scrollableChatContainerRef}
+    >
+      <div className="flex flex-col gap-5 divide-y">
         {messages.map((m: Message) => (
           <ChatItem key={m.id} {...m} />
         ))}

templates/components/ui/html/chat/hooks/use-config.ts

@@ -0,0 +1,31 @@
+"use client";
+
+import { useEffect, useMemo, useState } from "react";
+
+export interface ChatConfig {
+  backend?: string;
+  starterQuestions?: string[];
+}
+
+export function useClientConfig(): ChatConfig {
+  const chatAPI = process.env.NEXT_PUBLIC_CHAT_API;
+  const [config, setConfig] = useState<ChatConfig>();
+
+  const backendOrigin = useMemo(() => {
+    return chatAPI ? new URL(chatAPI).origin : "";
+  }, [chatAPI]);
+
+  const configAPI = `${backendOrigin}/api/chat/config`;
+
+  useEffect(() => {
+    fetch(configAPI)
+      .then((response) => response.json())
+      .then((data) => setConfig({ ...data, chatAPI }))
+      .catch((error) => console.error("Error fetching config", error));
+  }, [chatAPI, configAPI]);
+
+  return {
+    backend: backendOrigin,
+    starterQuestions: config?.starterQuestions,
+  };
+}

templates/components/vectordbs/python/astra/generate.py

@@ -1,37 +0,0 @@
-from dotenv import load_dotenv
-
-load_dotenv()
-
-import os
-import logging
-from llama_index.core.storage import StorageContext
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.vector_stores.astra_db import AstraDBVectorStore
-from app.settings import init_settings
-from app.engine.loaders import get_documents
-
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger()
-
-
-def generate_datasource():
-    init_settings()
-    logger.info("Creating new index")
-    documents = get_documents()
-    store = AstraDBVectorStore(
-        token=os.environ["ASTRA_DB_APPLICATION_TOKEN"],
-        api_endpoint=os.environ["ASTRA_DB_ENDPOINT"],
-        collection_name=os.environ["ASTRA_DB_COLLECTION"],
-        embedding_dimension=int(os.environ["EMBEDDING_DIM"]),
-    )
-    storage_context = StorageContext.from_defaults(vector_store=store)
-    VectorStoreIndex.from_documents(
-        documents,
-        storage_context=storage_context,
-        show_progress=True,  # this will show you a progress bar as the embeddings are created
-    )
-    logger.info(f"Successfully created embeddings in the AstraDB")
-
-
-if __name__ == "__main__":
-    generate_datasource()

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

@@ -1,21 +0,0 @@
-import logging
-import os
-
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.vector_stores.astra_db import AstraDBVectorStore
-
-
-logger = logging.getLogger("uvicorn")
-
-
-def get_index():
-    logger.info("Connecting to index from AstraDB...")
-    store = AstraDBVectorStore(
-        token=os.environ["ASTRA_DB_APPLICATION_TOKEN"],
-        api_endpoint=os.environ["ASTRA_DB_ENDPOINT"],
-        collection_name=os.environ["ASTRA_DB_COLLECTION"],
-        embedding_dimension=int(os.environ["EMBEDDING_DIM"]),
-    )
-    index = VectorStoreIndex.from_vector_store(store)
-    logger.info("Finished connecting to index from AstraDB.")
-    return index

templates/components/vectordbs/python/astra/vectordb.py

@@ -0,0 +1,20 @@
+import os
+from llama_index.vector_stores.astra_db import AstraDBVectorStore
+
+
+def get_vector_store():
+    endpoint = os.getenv("ASTRA_DB_ENDPOINT")
+    token = os.getenv("ASTRA_DB_APPLICATION_TOKEN")
+    collection = os.getenv("ASTRA_DB_COLLECTION")
+    if not endpoint or not token or not collection:
+        raise ValueError(
+            "Please config ASTRA_DB_ENDPOINT, ASTRA_DB_APPLICATION_TOKEN and ASTRA_DB_COLLECTION"
+            " to your environment variables or config them in the .env file"
+        )
+    store = AstraDBVectorStore(
+        token=token,
+        api_endpoint=endpoint,
+        collection_name=collection,
+        embedding_dimension=int(os.getenv("EMBEDDING_DIM")),
+    )
+    return store

templates/components/vectordbs/python/chroma/vectordb.py

@@ -0,0 +1,24 @@
+import os
+from llama_index.vector_stores.chroma import ChromaVectorStore
+
+
+def get_vector_store():
+    collection_name = os.getenv("CHROMA_COLLECTION", "default")
+    chroma_path = os.getenv("CHROMA_PATH")
+    # if CHROMA_PATH is set, use a local ChromaVectorStore from the path
+    # otherwise, use a remote ChromaVectorStore (ChromaDB Cloud is not supported yet)
+    if chroma_path:
+        store = ChromaVectorStore.from_params(
+            persist_dir=chroma_path, collection_name=collection_name
+        )
+    else:
+        if not os.getenv("CHROMA_HOST") or not os.getenv("CHROMA_PORT"):
+            raise ValueError(
+                "Please provide either CHROMA_PATH or CHROMA_HOST and CHROMA_PORT"
+            )
+        store = ChromaVectorStore.from_params(
+            host=os.getenv("CHROMA_HOST"),
+            port=int(os.getenv("CHROMA_PORT")),
+            collection_name=collection_name,
+        )
+    return store

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

@@ -0,0 +1,45 @@
+from dotenv import load_dotenv
+
+load_dotenv()
+
+import os
+import logging
+from app.settings import init_settings
+from app.engine.loaders import get_documents
+from llama_index.indices.managed.llama_cloud import LlamaCloudIndex
+
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger()
+
+
+def generate_datasource():
+    init_settings()
+    logger.info("Generate index for the provided data")
+
+    name = os.getenv("LLAMA_CLOUD_INDEX_NAME")
+    project_name = os.getenv("LLAMA_CLOUD_PROJECT_NAME")
+    api_key = os.getenv("LLAMA_CLOUD_API_KEY")
+    base_url = os.getenv("LLAMA_CLOUD_BASE_URL")
+
+    if name is None or project_name is None or api_key is None:
+        raise ValueError(
+            "Please set LLAMA_CLOUD_INDEX_NAME, LLAMA_CLOUD_PROJECT_NAME and LLAMA_CLOUD_API_KEY"
+            " to your environment variables or config them in .env file"
+        )
+
+    documents = get_documents()
+
+    LlamaCloudIndex.from_documents(
+        documents=documents,
+        name=name,
+        project_name=project_name,
+        api_key=api_key,
+        base_url=base_url,
+    )
+
+    logger.info("Finished generating the index")
+
+
+if __name__ == "__main__":
+    generate_datasource()

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

@@ -0,0 +1,28 @@
+import logging
+import os
+from llama_index.indices.managed.llama_cloud import LlamaCloudIndex
+
+
+logger = logging.getLogger("uvicorn")
+
+
+def get_index():
+    name = os.getenv("LLAMA_CLOUD_INDEX_NAME")
+    project_name = os.getenv("LLAMA_CLOUD_PROJECT_NAME")
+    api_key = os.getenv("LLAMA_CLOUD_API_KEY")
+    base_url = os.getenv("LLAMA_CLOUD_BASE_URL")
+
+    if name is None or project_name is None or api_key is None:
+        raise ValueError(
+            "Please set LLAMA_CLOUD_INDEX_NAME, LLAMA_CLOUD_PROJECT_NAME and LLAMA_CLOUD_API_KEY"
+            " to your environment variables or config them in .env file"
+        )
+
+    index = LlamaCloudIndex(
+        name=name,
+        project_name=project_name,
+        api_key=api_key,
+        base_url=base_url,
+    )
+
+    return index

templates/components/vectordbs/python/milvus/generate.py

@@ -1,39 +0,0 @@
-from dotenv import load_dotenv
-
-load_dotenv()
-
-import os
-import logging
-from llama_index.core.storage import StorageContext
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.vector_stores.milvus import MilvusVectorStore
-from app.settings import init_settings
-from app.engine.loaders import get_documents
-
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger()
-
-
-def generate_datasource():
-    init_settings()
-    logger.info("Creating new index")
-    # load the documents and create the index
-    documents = get_documents()
-    store = MilvusVectorStore(
-        uri=os.environ["MILVUS_ADDRESS"],
-        user=os.getenv("MILVUS_USERNAME"),
-        password=os.getenv("MILVUS_PASSWORD"),
-        collection_name=os.getenv("MILVUS_COLLECTION"),
-        dim=int(os.getenv("EMBEDDING_DIM")),
-    )
-    storage_context = StorageContext.from_defaults(vector_store=store)
-    VectorStoreIndex.from_documents(
-        documents,
-        storage_context=storage_context,
-        show_progress=True,  # this will show you a progress bar as the embeddings are created
-    )
-    logger.info(f"Successfully created embeddings in the Milvus")
-
-
-if __name__ == "__main__":
-    generate_datasource()

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

@@ -1,22 +0,0 @@
-import logging
-import os
-
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.vector_stores.milvus import MilvusVectorStore
-
-
-logger = logging.getLogger("uvicorn")
-
-
-def get_index():
-    logger.info("Connecting to index from Milvus...")
-    store = MilvusVectorStore(
-        uri=os.getenv("MILVUS_ADDRESS"),
-        user=os.getenv("MILVUS_USERNAME"),
-        password=os.getenv("MILVUS_PASSWORD"),
-        collection_name=os.getenv("MILVUS_COLLECTION"),
-        dim=int(os.getenv("EMBEDDING_DIM")),
-    )
-    index = VectorStoreIndex.from_vector_store(store)
-    logger.info("Finished connecting to index from Milvus.")
-    return index

templates/components/vectordbs/python/milvus/vectordb.py

@@ -0,0 +1,20 @@
+import os
+from llama_index.vector_stores.milvus import MilvusVectorStore
+
+
+def get_vector_store():
+    address = os.getenv("MILVUS_ADDRESS")
+    collection = os.getenv("MILVUS_COLLECTION")
+    if not address or not collection:
+        raise ValueError(
+            "Please set MILVUS_ADDRESS and MILVUS_COLLECTION to your environment variables"
+            " or config them in the .env file"
+        )
+    store = MilvusVectorStore(
+        uri=address,
+        user=os.getenv("MILVUS_USERNAME"),
+        password=os.getenv("MILVUS_PASSWORD"),
+        collection_name=collection,
+        dim=int(os.getenv("EMBEDDING_DIM")),
+    )
+    return store

templates/components/vectordbs/python/mongo/generate.py

@@ -1,43 +0,0 @@
-from dotenv import load_dotenv
-
-load_dotenv()
-
-import os
-import logging
-from llama_index.core.storage import StorageContext
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.vector_stores.mongodb import MongoDBAtlasVectorSearch
-from app.settings import init_settings
-from app.engine.loaders import get_documents
-
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger()
-
-
-def generate_datasource():
-    init_settings()
-    logger.info("Creating new index")
-    # load the documents and create the index
-    documents = get_documents()
-    store = MongoDBAtlasVectorSearch(
-        db_name=os.environ["MONGODB_DATABASE"],
-        collection_name=os.environ["MONGODB_VECTORS"],
-        index_name=os.environ["MONGODB_VECTOR_INDEX"],
-    )
-    storage_context = StorageContext.from_defaults(vector_store=store)
-    VectorStoreIndex.from_documents(
-        documents,
-        storage_context=storage_context,
-        show_progress=True,  # this will show you a progress bar as the embeddings are created
-    )
-    logger.info(
-        f"Successfully created embeddings in the MongoDB collection {os.environ['MONGODB_VECTORS']}"
-    )
-    logger.info(
-        """IMPORTANT: You can't query your index yet because you need to create a vector search index in MongoDB's UI now.
-See https://github.com/run-llama/mongodb-demo/tree/main?tab=readme-ov-file#create-a-vector-search-index"""
-    )
-
-
-if __name__ == "__main__":
-    generate_datasource()

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

@@ -1,20 +0,0 @@
-import logging
-import os
-
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.vector_stores.mongodb import MongoDBAtlasVectorSearch
-
-
-logger = logging.getLogger("uvicorn")
-
-
-def get_index():
-    logger.info("Connecting to index from MongoDB...")
-    store = MongoDBAtlasVectorSearch(
-        db_name=os.environ["MONGODB_DATABASE"],
-        collection_name=os.environ["MONGODB_VECTORS"],
-        index_name=os.environ["MONGODB_VECTOR_INDEX"],
-    )
-    index = VectorStoreIndex.from_vector_store(store)
-    logger.info("Finished connecting to index from MongoDB.")
-    return index

templates/components/vectordbs/python/mongo/vectordb.py

@@ -0,0 +1,20 @@
+import os
+from llama_index.vector_stores.mongodb import MongoDBAtlasVectorSearch
+
+
+def get_vector_store():
+    db_uri = os.getenv("MONGODB_URI")
+    db_name = os.getenv("MONGODB_DATABASE")
+    collection_name = os.getenv("MONGODB_VECTORS")
+    index_name = os.getenv("MONGODB_VECTOR_INDEX")
+    if not db_uri or not db_name or not collection_name or not index_name:
+        raise ValueError(
+            "Please set MONGODB_URI, MONGODB_DATABASE, MONGODB_VECTORS, and MONGODB_VECTOR_INDEX"
+            " to your environment variables or config them in .env file"
+        )
+    store = MongoDBAtlasVectorSearch(
+        db_name=db_name,
+        collection_name=collection_name,
+        index_name=index_name,
+    )
+    return store

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

@@ -1 +0,0 @@
-STORAGE_DIR = "storage"  # directory to cache the generated index

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

@@ -2,11 +2,11 @@ from dotenv import load_dotenv
 
 load_dotenv()
 
+import os
 import logging
 from llama_index.core.indices import (
     VectorStoreIndex,
 )
-from app.engine.constants import STORAGE_DIR
 from app.engine.loaders import get_documents
 from app.settings import init_settings
 
@@ -18,14 +18,18 @@ logger = logging.getLogger()
 def generate_datasource():
     init_settings()
     logger.info("Creating new index")
+    storage_dir = os.environ.get("STORAGE_DIR", "storage")
     # load the documents and create the index
     documents = get_documents()
+    # Set private=false to mark the document as public (required for filtering)
+    for doc in documents:
+        doc.metadata["private"] = "false"
     index = VectorStoreIndex.from_documents(
         documents,
     )
     # store it for later
-    index.storage_context.persist(STORAGE_DIR)
-    logger.info(f"Finished creating new index. Stored in {STORAGE_DIR}")
+    index.storage_context.persist(storage_dir)
+    logger.info(f"Finished creating new index. Stored in {storage_dir}")
 
 
 if __name__ == "__main__":

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

@@ -1,20 +1,30 @@
-import logging
 import os
+import logging
+from datetime import timedelta
 
-from app.engine.constants import STORAGE_DIR
+from cachetools import cached, TTLCache
 from llama_index.core.storage import StorageContext
 from llama_index.core.indices import load_index_from_storage
 
 logger = logging.getLogger("uvicorn")
 
 
+@cached(
+    TTLCache(maxsize=10, ttl=timedelta(minutes=5).total_seconds()),
+    key=lambda *args, **kwargs: "global_storage_context",
+)
+def get_storage_context(persist_dir: str) -> StorageContext:
+    return StorageContext.from_defaults(persist_dir=persist_dir)
+
+
 def get_index():
+    storage_dir = os.getenv("STORAGE_DIR", "storage")
     # check if storage already exists
-    if not os.path.exists(STORAGE_DIR):
+    if not os.path.exists(storage_dir):
         return None
     # load the existing index
-    logger.info(f"Loading index from {STORAGE_DIR}...")
-    storage_context = StorageContext.from_defaults(persist_dir=STORAGE_DIR)
+    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}")
+    logger.info(f"Finished loading index from {storage_dir}")
     return index

templates/components/vectordbs/python/pg/constants.py

@@ -1,2 +0,0 @@
-PGVECTOR_SCHEMA = "public"
-PGVECTOR_TABLE = "llamaindex_embedding"

templates/components/vectordbs/python/pg/generate.py

@@ -1,35 +0,0 @@
-from dotenv import load_dotenv
-
-load_dotenv()
-
-import logging
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.core.storage import StorageContext
-
-from app.engine.loaders import get_documents
-from app.settings import init_settings
-from app.engine.utils import init_pg_vector_store_from_env
-
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger()
-
-
-def generate_datasource():
-    init_settings()
-    logger.info("Creating new index")
-    # load the documents and create the index
-    documents = get_documents()
-    store = init_pg_vector_store_from_env()
-    storage_context = StorageContext.from_defaults(vector_store=store)
-    VectorStoreIndex.from_documents(
-        documents,
-        storage_context=storage_context,
-        show_progress=True,  # this will show you a progress bar as the embeddings are created
-    )
-    logger.info(
-        f"Successfully created embeddings in the PG vector store, schema={store.schema_name} table={store.table_name}"
-    )
-
-
-if __name__ == "__main__":
-    generate_datasource()

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

@@ -1,13 +0,0 @@
-import logging
-from llama_index.core.indices.vector_store import VectorStoreIndex
-from app.engine.utils import init_pg_vector_store_from_env
-
-logger = logging.getLogger("uvicorn")
-
-
-def get_index():
-    logger.info("Connecting to index from PGVector...")
-    store = init_pg_vector_store_from_env()
-    index = VectorStoreIndex.from_vector_store(store)
-    logger.info("Finished connecting to index from PGVector.")
-    return index

templates/components/vectordbs/python/pg/utils.py

@@ -1,27 +0,0 @@
-import os
-from llama_index.vector_stores.postgres import PGVectorStore
-from urllib.parse import urlparse
-from app.engine.constants import PGVECTOR_SCHEMA, PGVECTOR_TABLE
-
-
-def init_pg_vector_store_from_env():
-    original_conn_string = os.environ.get("PG_CONNECTION_STRING")
-    if original_conn_string is None or original_conn_string == "":
-        raise ValueError("PG_CONNECTION_STRING environment variable is not set.")
-
-    # The PGVectorStore requires both two connection strings, one for psycopg2 and one for asyncpg
-    # Update the configured scheme with the psycopg2 and asyncpg schemes
-    original_scheme = urlparse(original_conn_string).scheme + "://"
-    conn_string = original_conn_string.replace(
-        original_scheme, "postgresql+psycopg2://"
-    )
-    async_conn_string = original_conn_string.replace(
-        original_scheme, "postgresql+asyncpg://"
-    )
-
-    return PGVectorStore(
-        connection_string=conn_string,
-        async_connection_string=async_conn_string,
-        schema_name=PGVECTOR_SCHEMA,
-        table_name=PGVECTOR_TABLE,
-    )

templates/components/vectordbs/python/pg/vectordb.py

@@ -0,0 +1,37 @@
+import os
+from llama_index.vector_stores.postgres import PGVectorStore
+from urllib.parse import urlparse
+
+PGVECTOR_SCHEMA = "public"
+PGVECTOR_TABLE = "llamaindex_embedding"
+
+vector_store: PGVectorStore = None
+
+
+def get_vector_store():
+    global vector_store
+
+    if vector_store is None:
+        original_conn_string = os.environ.get("PG_CONNECTION_STRING")
+        if original_conn_string is None or original_conn_string == "":
+            raise ValueError("PG_CONNECTION_STRING environment variable is not set.")
+
+        # The PGVectorStore requires both two connection strings, one for psycopg2 and one for asyncpg
+        # Update the configured scheme with the psycopg2 and asyncpg schemes
+        original_scheme = urlparse(original_conn_string).scheme + "://"
+        conn_string = original_conn_string.replace(
+            original_scheme, "postgresql+psycopg2://"
+        )
+        async_conn_string = original_conn_string.replace(
+            original_scheme, "postgresql+asyncpg://"
+        )
+
+        vector_store = PGVectorStore(
+            connection_string=conn_string,
+            async_connection_string=async_conn_string,
+            schema_name=PGVECTOR_SCHEMA,
+            table_name=PGVECTOR_TABLE,
+            embed_dim=int(os.environ.get("EMBEDDING_DIM", 1024)),
+        )
+
+    return vector_store

templates/components/vectordbs/python/pinecone/generate.py

@@ -1,39 +0,0 @@
-from dotenv import load_dotenv
-
-load_dotenv()
-
-import os
-import logging
-from llama_index.core.storage import StorageContext
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.vector_stores.pinecone import PineconeVectorStore
-from app.settings import init_settings
-from app.engine.loaders import get_documents
-
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger()
-
-
-def generate_datasource():
-    init_settings()
-    logger.info("Creating new index")
-    # load the documents and create the index
-    documents = get_documents()
-    store = PineconeVectorStore(
-        api_key=os.environ["PINECONE_API_KEY"],
-        index_name=os.environ["PINECONE_INDEX_NAME"],
-        environment=os.environ["PINECONE_ENVIRONMENT"],
-    )
-    storage_context = StorageContext.from_defaults(vector_store=store)
-    VectorStoreIndex.from_documents(
-        documents,
-        storage_context=storage_context,
-        show_progress=True,  # this will show you a progress bar as the embeddings are created
-    )
-    logger.info(
-        f"Successfully created embeddings and save to your Pinecone index {os.environ['PINECONE_INDEX_NAME']}"
-    )
-
-
-if __name__ == "__main__":
-    generate_datasource()

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

@@ -1,20 +0,0 @@
-import logging
-import os
-
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.vector_stores.pinecone import PineconeVectorStore
-
-
-logger = logging.getLogger("uvicorn")
-
-
-def get_index():
-    logger.info("Connecting to index from Pinecone...")
-    store = PineconeVectorStore(
-        api_key=os.environ["PINECONE_API_KEY"],
-        index_name=os.environ["PINECONE_INDEX_NAME"],
-        environment=os.environ["PINECONE_ENVIRONMENT"],
-    )
-    index = VectorStoreIndex.from_vector_store(store)
-    logger.info("Finished connecting to index from Pinecone.")
-    return index

templates/components/vectordbs/python/pinecone/vectordb.py

@@ -0,0 +1,19 @@
+import os
+from llama_index.vector_stores.pinecone import PineconeVectorStore
+
+
+def get_vector_store():
+    api_key = os.getenv("PINECONE_API_KEY")
+    index_name = os.getenv("PINECONE_INDEX_NAME")
+    environment = os.getenv("PINECONE_ENVIRONMENT")
+    if not api_key or not index_name or not environment:
+        raise ValueError(
+            "Please set PINECONE_API_KEY, PINECONE_INDEX_NAME, and PINECONE_ENVIRONMENT"
+            " to your environment variables or config them in the .env file"
+        )
+    store = PineconeVectorStore(
+        api_key=api_key,
+        index_name=index_name,
+        environment=environment,
+    )
+    return store

templates/components/vectordbs/python/qdrant/generate.py

@@ -1,37 +0,0 @@
-import logging
-import os
-from app.engine.loaders import get_documents
-from app.settings import init_settings
-from dotenv import load_dotenv
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.core.storage import StorageContext
-from llama_index.vector_stores.qdrant import QdrantVectorStore
-load_dotenv()
-
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger()
-
-
-def generate_datasource():
-    init_settings()
-    logger.info("Creating new index with Qdrant")
-    # load the documents and create the index
-    documents = get_documents()
-    store = QdrantVectorStore(
-        collection_name=os.getenv("QDRANT_COLLECTION"),
-        url=os.getenv("QDRANT_URL"),
-        api_key=os.getenv("QDRANT_API_KEY"),
-    )
-    storage_context = StorageContext.from_defaults(vector_store=store)
-    VectorStoreIndex.from_documents(
-        documents,
-        storage_context=storage_context,
-        show_progress=True,  # this will show you a progress bar as the embeddings are created
-    )
-    logger.info(
-        f"Successfully uploaded documents to the {os.getenv('QDRANT_COLLECTION')} collection."
-    )
-
-
-if __name__ == "__main__":
-    generate_datasource()

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

@@ -1,20 +0,0 @@
-import logging
-import os
-
-from llama_index.core.indices import VectorStoreIndex
-from llama_index.vector_stores.qdrant import QdrantVectorStore
-
-
-logger = logging.getLogger("uvicorn")
-
-
-def get_index():
-    logger.info("Connecting to Qdrant collection..")
-    store = QdrantVectorStore(
-        collection_name=os.getenv("QDRANT_COLLECTION"),
-        url=os.getenv("QDRANT_URL"),
-        api_key=os.getenv("QDRANT_API_KEY"),
-    )
-    index = VectorStoreIndex.from_vector_store(store)
-    logger.info("Finished connecting to Qdrant collection.")
-    return index

templates/components/vectordbs/python/qdrant/vectordb.py

@@ -0,0 +1,19 @@
+import os
+from llama_index.vector_stores.qdrant import QdrantVectorStore
+
+
+def get_vector_store():
+    collection_name = os.getenv("QDRANT_COLLECTION")
+    url = os.getenv("QDRANT_URL")
+    api_key = os.getenv("QDRANT_API_KEY")
+    if not collection_name or not url:
+        raise ValueError(
+            "Please set QDRANT_COLLECTION, QDRANT_URL"
+            " to your environment variables or config them in the .env file"
+        )
+    store = QdrantVectorStore(
+        collection_name=collection_name,
+        url=url,
+        api_key=api_key,
+    )
+    return store

templates/components/vectordbs/typescript/chroma/generate.ts

@@ -0,0 +1,37 @@
+/* eslint-disable turbo/no-undeclared-env-vars */
+import * as dotenv from "dotenv";
+import { VectorStoreIndex, storageContextFromDefaults } from "llamaindex";
+import { ChromaVectorStore } from "llamaindex/storage/vectorStore/ChromaVectorStore";
+import { getDocuments } from "./loader";
+import { initSettings } from "./settings";
+import { checkRequiredEnvVars } from "./shared";
+
+dotenv.config();
+
+async function loadAndIndex() {
+  // load objects from storage and convert them into LlamaIndex Document objects
+  const documents = await getDocuments();
+
+  // create vector store
+  const chromaUri = `http://${process.env.CHROMA_HOST}:${process.env.CHROMA_PORT}`;
+
+  const vectorStore = new ChromaVectorStore({
+    collectionName: process.env.CHROMA_COLLECTION,
+    chromaClientParams: { path: chromaUri },
+  });
+
+  // create index from all the Documentss and store them in Pinecone
+  console.log("Start creating embeddings...");
+  const storageContext = await storageContextFromDefaults({ vectorStore });
+  await VectorStoreIndex.fromDocuments(documents, { storageContext });
+  console.log(
+    "Successfully created embeddings and save to your ChromaDB index.",
+  );
+}
+
+(async () => {
+  checkRequiredEnvVars();
+  initSettings();
+  await loadAndIndex();
+  console.log("Finished generating storage.");
+})();

templates/components/vectordbs/typescript/chroma/index.ts

@@ -0,0 +1,16 @@
+/* eslint-disable turbo/no-undeclared-env-vars */
+import { VectorStoreIndex } from "llamaindex";
+import { ChromaVectorStore } from "llamaindex/storage/vectorStore/ChromaVectorStore";
+import { checkRequiredEnvVars } from "./shared";
+
+export async function getDataSource() {
+  checkRequiredEnvVars();
+  const chromaUri = `http://${process.env.CHROMA_HOST}:${process.env.CHROMA_PORT}`;
+
+  const store = new ChromaVectorStore({
+    collectionName: process.env.CHROMA_COLLECTION,
+    chromaClientParams: { path: chromaUri },
+  });
+
+  return await VectorStoreIndex.fromVectorStore(store);
+}

templates/components/vectordbs/typescript/chroma/shared.ts

@@ -0,0 +1,18 @@
+const REQUIRED_ENV_VARS = ["CHROMA_COLLECTION", "CHROMA_HOST", "CHROMA_PORT"];
+
+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(", ")}`,
+    );
+  }
+}

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

@@ -0,0 +1,26 @@
+import * as dotenv from "dotenv";
+import { LlamaCloudIndex } from "llamaindex";
+import { getDocuments } from "./loader";
+import { initSettings } from "./settings";
+import { checkRequiredEnvVars } from "./shared";
+
+dotenv.config();
+
+async function loadAndIndex() {
+  const documents = await getDocuments();
+  await LlamaCloudIndex.fromDocuments({
+    documents,
+    name: process.env.LLAMA_CLOUD_INDEX_NAME!,
+    projectName: process.env.LLAMA_CLOUD_PROJECT_NAME!,
+    apiKey: process.env.LLAMA_CLOUD_API_KEY,
+    baseUrl: process.env.LLAMA_CLOUD_BASE_URL,
+  });
+  console.log(`Successfully created embeddings!`);
+}
+
+(async () => {
+  checkRequiredEnvVars();
+  initSettings();
+  await loadAndIndex();
+  console.log("Finished generating storage.");
+})();

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

@@ -0,0 +1,13 @@
+import { LlamaCloudIndex } from "llamaindex/cloud/LlamaCloudIndex";
+import { checkRequiredEnvVars } from "./shared";
+
+export async function getDataSource() {
+  checkRequiredEnvVars();
+  const index = new LlamaCloudIndex({
+    name: process.env.LLAMA_CLOUD_INDEX_NAME!,
+    projectName: process.env.LLAMA_CLOUD_PROJECT_NAME!,
+    apiKey: process.env.LLAMA_CLOUD_API_KEY,
+    baseUrl: process.env.LLAMA_CLOUD_BASE_URL,
+  });
+  return index;
+}

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

@@ -0,0 +1,22 @@
+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(", ")}`,
+    );
+  }
+}

templates/components/vectordbs/typescript/mongo/generate.ts

@@ -9,7 +9,7 @@ import { checkRequiredEnvVars } from "./shared";
 
 dotenv.config();
 
-const mongoUri = process.env.MONGO_URI!;
+const mongoUri = process.env.MONGODB_URI!;
 const databaseName = process.env.MONGODB_DATABASE!;
 const vectorCollectionName = process.env.MONGODB_VECTORS!;
 const indexName = process.env.MONGODB_VECTOR_INDEX;

templates/components/vectordbs/typescript/mongo/shared.ts

@@ -1,5 +1,5 @@
 const REQUIRED_ENV_VARS = [
-  "MONGO_URI",
+  "MONGODB_URI",
   "MONGODB_DATABASE",
   "MONGODB_VECTORS",
   "MONGODB_VECTOR_INDEX",

templates/types/multiagent/fastapi/README-template.md

@@ -0,0 +1,50 @@
+This is a [LlamaIndex](https://www.llamaindex.ai/) project using [FastAPI](https://fastapi.tiangolo.com/) bootstrapped with [`create-llama`](https://github.com/run-llama/LlamaIndexTS/tree/main/packages/create-llama).
+
+## Getting Started
+
+First, setup the environment with poetry:
+
+> **_Note:_** This step is not needed if you are using the dev-container.
+
+```shell
+poetry install
+poetry shell
+```
+
+Then check the parameters that have been pre-configured in the `.env` file in this directory. (E.g. you might need to configure an `OPENAI_API_KEY` if you're using OpenAI as model provider).
+
+Second, generate the embeddings of the documents in the `./data` directory (if this folder exists - otherwise, skip this step):
+
+```shell
+poetry run generate
+```
+
+Third, run all the services in one command:
+
+```shell
+poetry run python main.py
+```
+
+You can monitor and test the agent services with `llama-agents` monitor TUI:
+
+```shell
+poetry run llama-agents monitor --control-plane-url http://127.0.0.1:8001
+```
+
+## Services:
+
+- Message queue (port 8000): To exchange the message between services
+- Control plane (port 8001): A gateway to manage the tasks and services.
+- Human consumer (port 8002): To handle result when the task is completed.
+- Agent service `query_engine` (port 8003): Agent that can query information from the configured LlamaIndex index.
+- Agent service `dummy_agent` (port 8004): A dummy agent that does nothing. Good starting point to add more agents.
+
+The ports listed above are set by default, but you can change them in the `.env` file.
+
+## Learn More
+
+To learn more about LlamaIndex, take a look at the following resources:
+
+- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex.
+
+You can check out [the LlamaIndex GitHub repository](https://github.com/run-llama/llama_index) - your feedback and contributions are welcome!

templates/types/multiagent/fastapi/app/agents/dummy/agent.py

@@ -0,0 +1,33 @@
+from llama_agents import AgentService, SimpleMessageQueue
+from llama_index.core.agent import FunctionCallingAgentWorker
+from llama_index.core.tools import FunctionTool
+from llama_index.core.settings import Settings
+from app.utils import load_from_env
+
+
+DEFAULT_DUMMY_AGENT_DESCRIPTION = "I'm a dummy agent which does nothing."
+
+
+def dummy_function():
+    """
+    This function does nothing.
+    """
+    return ""
+
+
+def init_dummy_agent(message_queue: SimpleMessageQueue) -> AgentService:
+    agent = FunctionCallingAgentWorker(
+        tools=[FunctionTool.from_defaults(fn=dummy_function)],
+        llm=Settings.llm,
+        prefix_messages=[],
+    ).as_agent()
+
+    return AgentService(
+        service_name="dummy_agent",
+        agent=agent,
+        message_queue=message_queue.client,
+        description=load_from_env("AGENT_DUMMY_DESCRIPTION", throw_error=False)
+        or DEFAULT_DUMMY_AGENT_DESCRIPTION,
+        host=load_from_env("AGENT_DUMMY_HOST", throw_error=False) or "127.0.0.1",
+        port=int(load_from_env("AGENT_DUMMY_PORT")),
+    )

templates/types/multiagent/fastapi/app/agents/query_engine/agent.py

@@ -0,0 +1,52 @@
+import os
+from llama_agents import AgentService, SimpleMessageQueue
+from llama_index.core.agent import FunctionCallingAgentWorker
+from llama_index.core.tools import QueryEngineTool, ToolMetadata
+from llama_index.core.settings import Settings
+from app.engine.index import get_index
+from app.utils import load_from_env
+
+
+DEFAULT_QUERY_ENGINE_AGENT_DESCRIPTION = (
+    "Used to answer the questions using the provided context data."
+)
+
+
+def get_query_engine_tool() -> QueryEngineTool:
+    """
+    Provide an agent worker that can be used to query the index.
+    """
+    index = get_index()
+    if index is None:
+        raise ValueError("Index not found. Please create an index first.")
+    query_engine = index.as_query_engine(similarity_top_k=int(os.getenv("TOP_K", 3)))
+    return QueryEngineTool(
+        query_engine=query_engine,
+        metadata=ToolMetadata(
+            name="context_data",
+            description="""
+                Provide the provided context information. 
+                Use a detailed plain text question as input to the tool.
+            """,
+        ),
+    )
+
+
+def init_query_engine_agent(
+    message_queue: SimpleMessageQueue,
+) -> AgentService:
+    """
+    Initialize the agent service.
+    """
+    agent = FunctionCallingAgentWorker(
+        tools=[get_query_engine_tool()], llm=Settings.llm, prefix_messages=[]
+    ).as_agent()
+    return AgentService(
+        service_name="context_query_agent",
+        agent=agent,
+        message_queue=message_queue.client,
+        description=load_from_env("AGENT_QUERY_ENGINE_DESCRIPTION", throw_error=False)
+        or DEFAULT_QUERY_ENGINE_AGENT_DESCRIPTION,
+        host=load_from_env("AGENT_QUERY_ENGINE_HOST", throw_error=False) or "127.0.0.1",
+        port=int(load_from_env("AGENT_QUERY_ENGINE_PORT")),
+    )

templates/types/multiagent/fastapi/app/core/control_plane.py

@@ -0,0 +1,19 @@
+from llama_index.llms.openai import OpenAI
+from llama_agents import AgentOrchestrator, ControlPlaneServer
+from app.core.message_queue import message_queue
+from app.utils import load_from_env
+
+
+control_plane_host = (
+    load_from_env("CONTROL_PLANE_HOST", throw_error=False) or "127.0.0.1"
+)
+control_plane_port = load_from_env("CONTROL_PLANE_PORT", throw_error=False) or "8001"
+
+
+# setup control plane
+control_plane = ControlPlaneServer(
+    message_queue=message_queue,
+    orchestrator=AgentOrchestrator(llm=OpenAI()),
+    host=control_plane_host,
+    port=int(control_plane_port) if control_plane_port else None,
+)

templates/types/multiagent/fastapi/app/core/message_queue.py

@@ -0,0 +1,12 @@
+from llama_agents import SimpleMessageQueue
+from app.utils import load_from_env
+
+message_queue_host = (
+    load_from_env("MESSAGE_QUEUE_HOST", throw_error=False) or "127.0.0.1"
+)
+message_queue_port = load_from_env("MESSAGE_QUEUE_PORT", throw_error=False) or "8000"
+
+message_queue = SimpleMessageQueue(
+    host=message_queue_host,
+    port=int(message_queue_port) if message_queue_port else None,
+)

templates/types/multiagent/fastapi/app/core/task_result.py

@@ -0,0 +1,88 @@
+import json
+from logging import getLogger
+from pathlib import Path
+from fastapi import FastAPI
+from typing import Dict, Optional
+from llama_agents import CallableMessageConsumer, QueueMessage
+from llama_agents.message_queues.base import BaseMessageQueue
+from llama_agents.message_consumers.base import BaseMessageQueueConsumer
+from llama_agents.message_consumers.remote import RemoteMessageConsumer
+from app.utils import load_from_env
+from app.core.message_queue import message_queue
+
+
+logger = getLogger(__name__)
+
+
+class TaskResultService:
+    def __init__(
+        self,
+        message_queue: BaseMessageQueue,
+        name: str = "human",
+        host: str = "127.0.0.1",
+        port: Optional[int] = 8002,
+    ) -> None:
+        self.name = name
+        self.host = host
+        self.port = port
+
+        self._message_queue = message_queue
+
+        # app
+        self._app = FastAPI()
+        self._app.add_api_route(
+            "/", self.home, methods=["GET"], tags=["Human Consumer"]
+        )
+        self._app.add_api_route(
+            "/process_message",
+            self.process_message,
+            methods=["POST"],
+            tags=["Human Consumer"],
+        )
+
+    @property
+    def message_queue(self) -> BaseMessageQueue:
+        return self._message_queue
+
+    def as_consumer(self, remote: bool = False) -> BaseMessageQueueConsumer:
+        if remote:
+            return RemoteMessageConsumer(
+                url=(
+                    f"http://{self.host}:{self.port}/process_message"
+                    if self.port
+                    else f"http://{self.host}/process_message"
+                ),
+                message_type=self.name,
+            )
+
+        return CallableMessageConsumer(
+            message_type=self.name,
+            handler=self.process_message,
+        )
+
+    async def process_message(self, message: QueueMessage) -> None:
+        Path("task_results").mkdir(exist_ok=True)
+        with open("task_results/task_results.json", "+a") as f:
+            json.dump(message.model_dump(), f)
+            f.write("\n")
+
+    async def home(self) -> Dict[str, str]:
+        return {"message": "hello, human."}
+
+    async def register_to_message_queue(self) -> None:
+        """Register to the message queue."""
+        await self.message_queue.register_consumer(self.as_consumer(remote=True))
+
+
+human_consumer_host = (
+    load_from_env("HUMAN_CONSUMER_HOST", throw_error=False) or "127.0.0.1"
+)
+human_consumer_port = load_from_env("HUMAN_CONSUMER_PORT", throw_error=False) or "8002"
+
+
+human_consumer_server = TaskResultService(
+    message_queue=message_queue,
+    host=human_consumer_host,
+    port=int(human_consumer_port) if human_consumer_port else None,
+    name="human",
+)

templates/types/multiagent/fastapi/app/utils.py

@@ -0,0 +1,8 @@
+import os
+
+
+def load_from_env(var: str, throw_error: bool = True) -> str:
+    res = os.getenv(var)
+    if res is None and throw_error:
+        raise ValueError(f"Missing environment variable: {var}")
+    return res

templates/types/multiagent/fastapi/main.py

@@ -0,0 +1,27 @@
+from dotenv import load_dotenv
+from app.settings import init_settings
+
+load_dotenv()
+init_settings()
+
+from llama_agents import ServerLauncher
+from app.core.message_queue import message_queue
+from app.core.control_plane import control_plane
+from app.core.task_result import human_consumer_server
+from app.agents.query_engine.agent import init_query_engine_agent
+from app.agents.dummy.agent import init_dummy_agent
+
+agents = [
+    init_query_engine_agent(message_queue),
+    init_dummy_agent(message_queue),
+]
+
+launcher = ServerLauncher(
+    agents,
+    control_plane,
+    message_queue,
+    additional_consumers=[human_consumer_server.as_consumer()],
+)
+
+if __name__ == "__main__":
+    launcher.launch_servers()

templates/types/multiagent/fastapi/pyproject.toml

@@ -0,0 +1,20 @@
+[tool.poetry]
+name = "app"
+version = "0.1.0"
+description = ""
+authors = ["Marcus Schiesser <mail@marcusschiesser.de>"]
+readme = "README.md"
+
+[tool.poetry.scripts]
+generate = "app.engine.generate:generate_datasource"
+
+[tool.poetry.dependencies]
+python = "^3.11"
+llama-agents = "^0.0.3"
+llama-index-agent-openai = "^0.2.7"
+llama-index-embeddings-openai = "^0.1.10"
+llama-index-llms-openai = "^0.1.23"
+
+[build-system]
+requires = ["poetry-core"]
+build-backend = "poetry.core.masonry.api"

templates/types/streaming/express/eslintrc.json

@@ -3,5 +3,8 @@
   "rules": {
     "max-params": ["error", 4],
     "prefer-const": "error"
+  },
+  "parserOptions": {
+    "sourceType": "module"
   }
 }

templates/types/streaming/express/gitignore

@@ -1,3 +1,5 @@
 # local env files
 .env
-node_modules/
+node_modules/
+
+output/

templates/types/streaming/express/index.ts

@@ -14,7 +14,7 @@ const prodCorsOrigin = process.env["PROD_CORS_ORIGIN"];
 
 initObservability();
 
-app.use(express.json());
+app.use(express.json({ limit: "50mb" }));
 
 if (isDevelopment) {
   console.warn("Running in development mode - allowing CORS for all origins");
@@ -31,6 +31,8 @@ if (isDevelopment) {
   console.warn("Production CORS origin not set, defaulting to no CORS.");
 }
 
+app.use("/api/files/data", express.static("data"));
+app.use("/api/files/output", express.static("output"));
 app.use(express.text());
 
 app.get("/", (req: Request, res: Response) => {

templates/types/streaming/express/package.json

@@ -1,22 +1,32 @@
 {
   "name": "llama-index-express-streaming",
   "version": "1.0.0",
-  "main": "dist/index.mjs",
+  "exports": "./index.js",
+  "types": "./index.d.ts",
+  "type": "module",
+  "engines": {
+    "node": ">=18"
+  },
   "scripts": {
     "format": "prettier --ignore-unknown --cache --check .",
     "format:write": "prettier --ignore-unknown --write .",
-    "build": "tsup index.ts --format cjs --dts",
+    "build": "tsup index.ts --format esm --dts",
     "start": "node dist/index.js",
-    "dev": "concurrently \"tsup index.ts --format cjs --dts --watch\" \"nodemon -q dist/index.mjs\""
+    "dev": "concurrently \"tsup index.ts --format esm --dts --watch\" \"nodemon --watch dist/index.js\""
   },
   "dependencies": {
     "ai": "^3.0.21",
     "cors": "^2.8.5",
     "dotenv": "^16.3.1",
+    "duck-duck-scrape": "^2.2.5",
     "express": "^4.18.2",
-    "llamaindex": "0.3.9",
+    "llamaindex": "0.4.6",
     "pdf2json": "3.0.5",
-    "ajv": "^8.12.0"
+    "ajv": "^8.12.0",
+    "@e2b/code-interpreter": "^0.0.5",
+    "got": "^14.4.1",
+    "@apidevtools/swagger-parser": "^10.1.0",
+    "formdata-node": "^6.0.3"
   },
   "devDependencies": {
     "@types/cors": "^2.8.16",

templates/types/streaming/express/src/controllers/chat-config.controller.ts

@@ -0,0 +1,14 @@
+import { Request, Response } from "express";
+
+export const chatConfig = async (_req: Request, res: Response) => {
+  let starterQuestions = undefined;
+  if (
+    process.env.CONVERSATION_STARTERS &&
+    process.env.CONVERSATION_STARTERS.trim()
+  ) {
+    starterQuestions = process.env.CONVERSATION_STARTERS.trim().split("\n");
+  }
+  return res.status(200).json({
+    starterQuestions,
+  });
+};

templates/types/streaming/express/src/controllers/chat-upload.controller.ts

@@ -0,0 +1,12 @@
+import { Request, Response } from "express";
+import { uploadDocument } from "./llamaindex/documents/documents";
+
+export const chatUpload = async (req: Request, res: Response) => {
+  const { base64 }: { base64: string } = req.body;
+  if (!base64) {
+    return res.status(400).json({
+      error: "base64 is required in the request body",
+    });
+  }
+  return res.status(200).json(await uploadDocument(base64));
+};