Release 0.5.2 (#551)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

.changeset/itchy-ads-travel.md

@@ -1,5 +0,0 @@
----
-"create-llama": patch
----
-
-Add support E2B code interpreter tool for FastAPI

.coderabbit.yaml

@@ -0,0 +1,6 @@
+# coderabbit.yml
+reviews:
+  path_instructions:
+    - path: "templates/**"
+      instructions: |
+        For files under the `templates` folder, do not report 'Missing Dependencies Detected' errors.

.github/workflows/e2e.yml

@@ -9,15 +9,17 @@ env:
   POETRY_VERSION: "1.6.1"
 
 jobs:
-  e2e:
-    name: create-llama
+  e2e-python:
+    name: python
     timeout-minutes: 60
     strategy:
       fail-fast: true
       matrix:
-        node-version: [18, 20]
+        node-version: [20]
         python-version: ["3.11"]
-        os: [macos-latest, windows-latest]
+        os: [macos-latest, windows-latest, ubuntu-22.04]
+        frameworks: ["fastapi"]
+        datasources: ["--no-files", "--example-file", "--llamacloud"]
     defaults:
       run:
         shell: bash
@@ -58,15 +60,89 @@ jobs:
         run: pnpm run pack-install
         working-directory: .
 
-      - name: Run Playwright tests
-        run: pnpm run e2e
+      - name: Run Playwright tests for Python
+        run: pnpm run e2e:python
         env:
           OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+          LLAMA_CLOUD_API_KEY: ${{ secrets.LLAMA_CLOUD_API_KEY }}
+          FRAMEWORK: ${{ matrix.frameworks }}
+          DATASOURCE: ${{ matrix.datasources }}
+          PYTHONIOENCODING: utf-8
+          PYTHONLEGACYWINDOWSSTDIO: utf-8
         working-directory: .
 
-      - uses: actions/upload-artifact@v3
+      - uses: actions/upload-artifact@v4
         if: always()
         with:
-          name: playwright-report
+          name: playwright-report-python-${{ matrix.os }}-${{ matrix.frameworks }}-${{ matrix.datasources }}
           path: ./playwright-report/
+          overwrite: true
+          retention-days: 30
+
+  e2e-typescript:
+    name: typescript
+    timeout-minutes: 60
+    strategy:
+      fail-fast: true
+      matrix:
+        node-version: [18, 20]
+        python-version: ["3.11"]
+        os: [macos-latest, windows-latest, ubuntu-22.04]
+        frameworks: ["nextjs"]
+        datasources: ["--no-files", "--example-file", "--llamacloud"]
+    defaults:
+      run:
+        shell: bash
+    runs-on: ${{ matrix.os }}
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Install Poetry
+        uses: snok/install-poetry@v1
+        with:
+          version: ${{ env.POETRY_VERSION }}
+
+      - uses: pnpm/action-setup@v3
+
+      - name: Setup Node.js ${{ matrix.node-version }}
+        uses: actions/setup-node@v4
+        with:
+          node-version: ${{ matrix.node-version }}
+          cache: "pnpm"
+
+      - name: Install dependencies
+        run: pnpm install
+
+      - name: Install Playwright Browsers
+        run: pnpm exec playwright install --with-deps
+        working-directory: .
+
+      - name: Build create-llama
+        run: pnpm run build
+        working-directory: .
+
+      - name: Install
+        run: pnpm run pack-install
+        working-directory: .
+
+      - name: Run Playwright tests for TypeScript
+        run: pnpm run e2e:typescript
+        env:
+          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+          LLAMA_CLOUD_API_KEY: ${{ secrets.LLAMA_CLOUD_API_KEY }}
+          FRAMEWORK: ${{ matrix.frameworks }}
+          DATASOURCE: ${{ matrix.datasources }}
+        working-directory: .
+
+      - uses: actions/upload-artifact@v4
+        if: always()
+        with:
+          name: playwright-report-typescript-${{ matrix.os }}-${{ matrix.frameworks }}-${{ matrix.datasources }}-node${{ matrix.node-version }}
+          path: ./playwright-report/
+          overwrite: true
           retention-days: 30

.github/workflows/lint_on_push_or_pull.yml

@@ -30,3 +30,13 @@ jobs:
 
       - name: Run Prettier
         run: pnpm run format
+
+      - name: Run Python format check
+        uses: chartboost/ruff-action@v1
+        with:
+          args: "format --check"
+
+      - name: Run Python lint
+        uses: chartboost/ruff-action@v1
+        with:
+          args: "check"

.github/workflows/release.yml

@@ -17,6 +17,9 @@ jobs:
 
       - uses: pnpm/action-setup@v3
 
+      - name: Install uv
+        uses: astral-sh/setup-uv@v3
+
       - name: Setup Node.js
         uses: actions/setup-node@v4
         with:

.github/workflows/release_llama_index_server.yml

@@ -0,0 +1,130 @@
+name: Release llama-index-server
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - "llama-index-server/**"
+      - ".github/workflows/release_llama_index_server.yml"
+  pull_request:
+    types:
+      - closed
+
+concurrency: ${{ github.workflow }}-${{ github.ref }}
+
+jobs:
+  release:
+    name: Create Release PR
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: ./llama-index-server
+    if: |
+      github.event_name == 'push' && 
+      !startsWith(github.ref, 'refs/heads/release/llama-index-server-v')
+
+    steps:
+      - name: Checkout Repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+
+      - name: Install Poetry
+        run: |
+          curl -sSL https://install.python-poetry.org | python3 -
+
+      - name: Install dependencies
+        run: poetry install
+
+      - name: Setup Git
+        run: |
+          git config --global user.email "github-actions[bot]@users.noreply.github.com"
+          git config --global user.name "github-actions[bot]"
+
+      - name: Bump patch version
+        run: |
+          poetry version patch
+          git add pyproject.toml
+          git commit -m "chore(release): bump version to $(poetry version -s)"
+
+      - name: Get current version
+        id: get_version
+        run: |
+          version=$(poetry version -s)
+          echo "current_version=${version}" >> "$GITHUB_OUTPUT"
+
+      - name: Create Release PR
+        uses: peter-evans/create-pull-request@v6
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          commit-message: "Release: llama-index-server v${{ steps.get_version.outputs.current_version }}"
+          title: "Release: llama-index-server v${{ steps.get_version.outputs.current_version }}"
+          body: |
+            This PR was automatically created to release a new version of the llama-index-server package.
+
+            Version: ${{ steps.get_version.outputs.current_version }}
+
+            Please review the changes and merge to trigger the release.
+          branch: release/llama-index-server-v${{ steps.get_version.outputs.current_version }}
+          base: main
+          labels: release, llama-index-server
+
+  publish:
+    name: Publish to PyPI
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: ./llama-index-server
+    if: |
+      github.event_name == 'pull_request' && 
+      github.event.pull_request.merged == true && 
+      startsWith(github.event.pull_request.title, 'Release: llama-index-server') &&
+      startsWith(github.event.pull_request.head.ref, 'release/llama-index-server-v')
+
+    steps:
+      - name: Checkout Repository
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+
+      - name: Install Poetry
+        run: |
+          curl -sSL https://install.python-poetry.org | python3 -
+
+      - name: Install dependencies
+        run: poetry install
+
+      - name: Get current version
+        id: get_version
+        run: |
+          version=$(poetry version -s)
+          echo "current_version=${version}" >> "$GITHUB_OUTPUT"
+
+      - name: Build and publish to PyPI
+        uses: JRubics/poetry-publish@v2.1
+        with:
+          python_version: "3.11"
+          pypi_token: ${{ secrets.PYPI_TOKEN }}
+          package_directory: "llama-index-server"
+          poetry_install_options: "--without dev"
+
+      - name: Create GitHub Release
+        uses: softprops/action-gh-release@v2
+        with:
+          tag_name: llama-index-server-v${{ steps.get_version.outputs.current_version }}
+          name: "llama-index-server v${{ steps.get_version.outputs.current_version }}"
+          body: |
+            Release of llama-index-server v${{ steps.get_version.outputs.current_version }}
+          draft: false
+          prerelease: false
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/test_llama_index_server.yml

@@ -0,0 +1,111 @@
+name: Build Package
+
+on:
+  pull_request:
+
+env:
+  POETRY_VERSION: "1.8.3"
+  PYTHON_VERSION: "3.9"
+
+jobs:
+  unit-test:
+    name: Unit Tests
+    runs-on: ${{ matrix.os }}
+    defaults:
+      run:
+        working-directory: llama-index-server
+    strategy:
+      matrix:
+        os: [ubuntu-latest, windows-latest]
+        python-version: ["3.9"]
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install Poetry
+        run: pipx install poetry==${{ env.POETRY_VERSION }}
+
+      - name: Set up python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: "poetry"
+
+      - name: Configure Poetry
+        run: |
+          poetry config virtualenvs.create true
+          poetry config virtualenvs.in-project true
+          poetry env use python
+
+      - name: Install dependencies
+        shell: bash
+        run: poetry install --with dev
+
+      - name: Run unit tests
+        shell: bash
+        run: |
+          poetry run pytest tests
+
+  type-check:
+    name: Type Check
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: llama-index-server
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install Poetry
+        run: pipx install poetry==${{ env.POETRY_VERSION }}
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ env.PYTHON_VERSION }}
+          cache: "poetry"
+
+      - name: Configure Poetry
+        run: |
+          poetry config virtualenvs.create true
+          poetry config virtualenvs.in-project true
+          poetry env use python
+
+      - name: Install dependencies
+        shell: bash
+        run: poetry install --with dev
+
+      - name: Run mypy
+        shell: bash
+        run: poetry run mypy llama_index
+
+  build:
+    needs: [unit-test, type-check]
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: llama-index-server
+    steps:
+      - uses: actions/checkout@v4
+      - name: Install Poetry
+        run: pipx install poetry==${{ env.POETRY_VERSION }}
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ env.PYTHON_VERSION }}
+      - name: Clear python cache
+        shell: bash
+        run: poetry cache clear --all pypi
+      - name: Build package
+        shell: bash
+        run: poetry build
+      - name: Test installing built package
+        shell: bash
+        run: python -m pip install .
+      - name: Test import
+        shell: bash
+        working-directory: ${{ vars.RUNNER_TEMP }}
+        run: python -c "from llama_index.server import LlamaIndexServer"
+      - name: Upload artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: llama-index-server
+          path: llama-index-server/dist/

.gitignore

@@ -8,6 +8,7 @@ node_modules
 
 # testing
 coverage
+.coverage
 
 # next.js
 .next/
@@ -46,5 +47,19 @@ e2e/cache
 # intellij
 **/.idea
 
+# Python
+.mypy_cache/
+venv/
+.venv/
+dist/
+.__pycache__
+__pycache__
+.python-version
+.ui
+
 # build artifacts
 create-llama-*.tgz
+
+# vscode
+.vscode
+!.vscode/settings.json

.husky/pre-commit

@@ -1,2 +1,3 @@
 pnpm format
 pnpm lint
+uvx ruff format --check templates/

CHANGELOG.md

@@ -1,5 +1,615 @@
 # create-llama
 
+## 0.5.2
+
+### Patch Changes
+
+- c9f8f8d: Use custom component for deep research use case
+
+## 0.5.1
+
+### Patch Changes
+
+- 08b3e07: Simplify the local index code.
+
+## 0.5.0
+
+### Minor Changes
+
+- 54c9e2f: Simplified generated code using LlamaIndexServer
+
+### Patch Changes
+
+- 0e4ecfa: fix: add trycatch for generating error
+- ee69ce7: bump: chat-ui and tailwind v4
+
+## 0.4.0
+
+### Minor Changes
+
+- 61204a1: chore: bump LITS 0.9
+
+### Patch Changes
+
+- 9e723c3: Standardize the code of the workflow use case (Python)
+- d5da55b: feat: add components.json to use CLI
+- c1552eb: chore: move wikipedia tool to create-llama
+
+## 0.3.28
+
+### Patch Changes
+
+- 4e06714: Fix the error: Unable to view file sources due to CORS.
+
+## 0.3.27
+
+### Patch Changes
+
+- b4e41aa: Add deep research over own documents use case (Python)
+
+## 0.3.26
+
+### Patch Changes
+
+- f73d46b: Fix missing copy of the multiagent code
+
+## 0.3.25
+
+### Patch Changes
+
+- 5450096: bump: react 19 stable
+
+## 0.3.24
+
+### Patch Changes
+
+- a84743c: Change --agents paramameter to --use-case
+- a84743c: Add LlamaCloud support for Reflex templates
+- a7a6592: Fix the npm issue on the full-stack Python template
+- fc5e56e: bump: code interpreter v1
+
+## 0.3.23
+
+### Patch Changes
+
+- 9077cae: Add contract review use case (Python)
+
+## 0.3.22
+
+### Patch Changes
+
+- 25667d4: Make OpenAPI spec usable by custom GPTs
+
+## 0.3.21
+
+### Patch Changes
+
+- 95227a7: Add query endpoint
+
+## 0.3.20
+
+### Patch Changes
+
+- 27d2499: Bump the LlamaCloud library and fix breaking changes (Python).
+
+## 0.3.19
+
+### Patch Changes
+
+- f9a057d: Add support multimodal indexes (e.g. from LlamaCloud)
+- aedd73d: bump: chat-ui
+
+## 0.3.18
+
+### Patch Changes
+
+- fe90a7e: chore: bump ai v4
+- 02b2473: Show streaming errors in Python, optimize system prompts for tool usage and set the weather tool as default for the Agentic RAG use case
+- 63e961e: Use auto_routed retriever mode for LlamaCloudIndex
+
+## 0.3.17
+
+### Patch Changes
+
+- 28c8808: Add fly.io deployment
+- 0a7dfcf: Generate NEXT_PUBLIC_CHAT_API for NextJS backend to specify alternative backend
+
+## 0.3.16
+
+### Patch Changes
+
+- 8b371d8: Set pydantic version to <2.10 to avoid incompatibility with llama-index.
+- 30fe269: Deactive duckduckgo tool for TS
+- 30fe269: Replace DuckDuckGo by Wikipedia tool for agentic template
+
+## 0.3.15
+
+### Patch Changes
+
+- fc5b266: Improve DX for Python template (use one deployment instead of two)
+- f8f97d2: Add support for python 3.13
+
+## 0.3.14
+
+### Patch Changes
+
+- 00f0b3a: fix: dont include user message in chat history
+- 4663dec: chore: bump react19 rc
+- 44b34fb: chore: update eslint 9, nextjs 15, react 19
+- 6925676: feat: use latest chat UI
+
+## 0.3.13
+
+### Patch Changes
+
+- 282eaa0: Ensure that the index and document store are created when uploading a file with no available index.
+
+## 0.3.12
+
+### Patch Changes
+
+- 6edea6a: Optimize generated workflow code for Python
+- 8431b78: Optimize Typescript multi-agent code
+- 8431b78: Add form filling use case (Typescript)
+
+## 0.3.11
+
+### Patch Changes
+
+- 2b8aaa8: Add support for local models via Hugging Face
+- b9570b2: Fix: use generic LLMAgent instead of OpenAIAgent (adds support for Gemini and Anthropic for Agentic RAG)
+- 1fe21f8: Fix the highlight.js issue with the Next.js static build
+- 00009ae: feat: import pdf css
+
+## 0.3.10
+
+### Patch Changes
+
+- 9172fed: feat: bump LITS 0.8.2
+- 78ccde7: feat: use llamaindex chat-ui for nextjs frontend
+
+## 0.3.9
+
+### Patch Changes
+
+- ed59927: Add form filling use case (Python)
+
+## 0.3.8
+
+### Patch Changes
+
+- 4a83469: Add multi-agent financial report for Typescript (and update LITS to 0.7.10)
+
+## 0.3.7
+
+### Patch Changes
+
+- fa80378: DocumentInfo working with relative URLs
+
+## 0.3.6
+
+### Patch Changes
+
+- 0182368: Fix the streaming issue to prevent the UI from hanging.
+
+## 0.3.5
+
+### Patch Changes
+
+- 2209409: Add financial report as the default use case in the multi-agent template (Python).
+
+## 0.3.4
+
+### Patch Changes
+
+- 384a136: Fix import error if the artifact tool is selected
+
+## 0.3.3
+
+### Patch Changes
+
+- 99b8247: Simplify and unify handling file uploads
+
+## 0.3.2
+
+### Patch Changes
+
+- 6d1b6b9: Update README.md for pro mode
+
+## 0.3.1
+
+### Patch Changes
+
+- f3577c5: Fix event streaming is blocked
+- f3577c5: Add upload file to sandbox (artifact and code interpreter)
+
+## 0.3.0
+
+### Minor Changes
+
+- 7562cb4: Simplified default questions and added pro mode
+
+### Patch Changes
+
+- 0a69fe0: fix: missing params when init Astra vectorstore
+- 98a82b0: docs: chroma env variables
+
+## 0.2.19
+
+### Patch Changes
+
+- 3d41488: feat: use selected llamacloud for multiagent
+
+## 0.2.18
+
+### Patch Changes
+
+- 75e1f61: Fix cannot query public document from llamacloud
+- 88220f1: fix workflow doesn't stop when user presses stop generation button
+- 75e1f61: Fix typescript templates cannot upload file to llamacloud
+- 88220f1: Bump llama_index@0.11.17
+
+## 0.2.17
+
+### Patch Changes
+
+- cd3fcd0: bump: use LlamaIndexTS 0.6.18
+- 6335de1: Fix using LlamaCloud selector does not use the configured values in the environment (Python)
+
+## 0.2.16
+
+### Patch Changes
+
+- 0e78ba4: Fix: programmatically ensure index for LlamaCloud
+- 0e78ba4: Fix .env not loaded on poetry run generate
+- 7f4ac22: Don't need to run generate script for LlamaCloud
+- 5263bde: Use selected LlamaCloud index in multi-agent template
+
+## 0.2.15
+
+### Patch Changes
+
+- 16e6124: Bump package for llamatrace observability
+- 3790ca0: Add multi-agent task selector for TS template
+- d18f039: Add e2b code artifact tool for the FastAPI template
+
+## 0.2.14
+
+### Patch Changes
+
+- 5a7216e: feat: implement artifact tool in TS
+
+## 0.2.13
+
+### Patch Changes
+
+- 04ddebc: Add publisher agent to multi-agents for generating documents (PDF and HTML)
+- 04ddebc: Allow tool selection for multi-agents (Python and TS)
+
+## 0.2.12
+
+### Patch Changes
+
+- 70f7dca: feat: add test deps for llamaparse
+- ef070c0: Add multi agents template for Typescript
+
+## 0.2.11
+
+### Patch Changes
+
+- 7c2a3f6: fix: postgres import
+
+## 0.2.10
+
+### Patch Changes
+
+- cb8d535: Fix only produces one agent event
+
+## 0.2.9
+
+### Patch Changes
+
+- 0213fe0: Update dependencies for vector stores and add e2e test to ensure that they work as expected.
+
+## 0.2.8
+
+### Patch Changes
+
+- 0031e67: Bump llama-index to 0.11.11 for the multi-agent template
+
+## 0.2.7
+
+### Patch Changes
+
+- 505b8e9: bump: use latest ai package version
+- cf3ec97: Dynamically select model for Groq
+- 8c1087f: feat: enhance style for markdown
+
+## 0.2.6
+
+### Patch Changes
+
+- adc40cf: fix: vercel ai update crash sending annotations
+
+## 0.2.5
+
+### Patch Changes
+
+- 38a8be8: fix: filter in mongo vector store
+
+## 0.2.4
+
+### Patch Changes
+
+- 917e862: Fix errors in building the frontend
+
+## 0.2.3
+
+### Patch Changes
+
+- b6da3c2: Ensure the generation script always works
+
+## 0.2.2
+
+### Patch Changes
+
+- 8105c5c: Add env config for next questions feature
+
+## 0.2.1
+
+### Patch Changes
+
+- 6a409cb: Bump web and database reader packages
+
+## 0.2.0
+
+### Minor Changes
+
+- 435109f: Add multi-agents template based on workflows
+
+## 0.1.44
+
+### Patch Changes
+
+- bedde2b: Change metadata filters to use already existing documents in LlamaCloud Index
+- 5cd12fa: Use one callback manager per request
+- 5cd12fa: Bump llama_index version to 0.11.1
+- fd4abb3: Fix to use filename for uploaded documents in NextJS
+- 2f8feab: Simplify CLI interface
+
+## 0.1.43
+
+### Patch Changes
+
+- 4fa2b76: feat: implement citation for TS
+
+## 0.1.42
+
+### Patch Changes
+
+- 8f670a9: Allow relative URL in documents
+
+## 0.1.41
+
+### Patch Changes
+
+- 57e7638: Use the retrieval defaults from LlamaCloud
+
+## 0.1.40
+
+### Patch Changes
+
+- 8ce4a85: Add UI for extractor template
+
+## 0.1.39
+
+### Patch Changes
+
+- 3fb93c7: Use LlamaCloud pipeline for data ingestion in TS (private file uploads and generate script)
+
+## 0.1.38
+
+### Patch Changes
+
+- bd5e39a: Fix error that files in sub folders of 'data' are not displayed
+
+## 0.1.37
+
+### Patch Changes
+
+- 9fd832c: Add in-text citation references
+
+## 0.1.36
+
+### Patch Changes
+
+- 2b7a5d8: Fix: private file upload not working in Python without LlamaCloud
+
+## 0.1.35
+
+### Patch Changes
+
+- 81ef7f0: Use LlamaCloud pipeline for data ingestion (private file uploads and generate script)
+
+## 0.1.34
+
+### Patch Changes
+
+- c49a5e1: Add error handling for generating the next question
+- c49a5e1: Fix wrong api key variable in Azure OpenAI provider
+
+## 0.1.33
+
+### Patch Changes
+
+- d746c75: Add Weaviate vector store (Typescript)
+
+## 0.1.32
+
+### Patch Changes
+
+- 3ec5163: Add Weaviate vector database support (Python)
+
+## 0.1.31
+
+### Patch Changes
+
+- 04a9c71: Cluster nodes by document
+
+## 0.1.30
+
+### Patch Changes
+
+- 09e3022: Add support for LlamaTrace (Python)
+- c06ec4f: Fix imports for MongoDB
+- b6dd7a9: Always send chat data when submit message
+
+## 0.1.29
+
+### Patch Changes
+
+- 8890e27: Let user change indexes in LlamaCloud projects
+
+## 0.1.28
+
+### Patch Changes
+
+- 9a09e8c: Fix Vercel deployment
+
+## 0.1.27
+
+### Patch Changes
+
+- c5c7eee: Make components reusable for chat-llamaindex
+
+## 0.1.26
+
+### Patch Changes
+
+- f43399c: Add metadatafilters to context chat engine (Typescript)
+
+## 0.1.25
+
+### Patch Changes
+
+- c67daeb: fix: missing set private to false for default generate.py
+
+## 0.1.24
+
+### Patch Changes
+
+- 43474a5: Configure LlamaCloud organization ID for Python
+- cf11b23: Add Azure code interpreter for Python and TS
+- fd9fb42: Add Azure OpenAI as model provider
+- 5c13646: Fix starter questions not working in python backend
+
+## 0.1.23
+
+### Patch Changes
+
+- 6bd76fb: Add template for structured extraction
+
+## 0.1.22
+
+### Patch Changes
+
+- b0becaa: Add e2e testing for llamacloud datasource
+- df9cca5: Upgrade pdf viewer
+
+## 0.1.21
+
+### Patch Changes
+
+- bd4714c: Filter private documents for Typescript (Using MetadataFilters) and update to LlamaIndexTS 0.5.7
+- 58e6c15: Add using LlamaParse for private file uploader
+- 455ab68: Display files in sources using LlamaCloud indexes.
+- 23b7357: Use gpt-4o-mini as default model
+- 0900413: Add suggestions for next questions.
+
+## 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

README.md

@@ -1,14 +1,20 @@
-# Create LlamaIndex App
+# Create Llama
 
 The easiest way to get started with [LlamaIndex](https://www.llamaindex.ai/) is by using `create-llama`. This CLI tool enables you to quickly start building a new LlamaIndex application, with everything set up for you.
 
+## Get started
+
 Just run
 
 ```bash
 npx create-llama@latest
 ```
 
-to get started, or see below for more options. Once your app is generated, run
+to get started, or watch this video for a demo session:
+
+<img src="https://github.com/user-attachments/assets/c4a7fe18-8e30-498a-96f8-78127dd706b9" width="100%">
+
+Once your app is generated, run
 
 ```bash
 npm run dev
@@ -18,21 +24,25 @@ to start the development server. You can then visit [http://localhost:3000](http
 
 ## What you'll get
 
-- A Next.js-powered front-end using components from [shadcn/ui](https://ui.shadcn.com/). The app is set up as a chat interface that can answer questions about your data (see below)
-- Your choice of 3 back-ends:
+- A set of pre-configured use cases to get you started, e.g. Agentic RAG, Data Analysis, Report Generation, etc.
+- A Next.js-powered front-end using components from [shadcn/ui](https://ui.shadcn.com/). The app is set up as a chat interface that can answer questions about your data or interact with your agent
+- Your choice of two back-ends:
   - **Next.js**: if you select this option, you’ll have a full-stack Next.js application that you can deploy to a host like [Vercel](https://vercel.com/) in just a few clicks. This uses [LlamaIndex.TS](https://www.npmjs.com/package/llamaindex), our TypeScript library.
-  - **Express**: if you want a more traditional Node.js application you can generate an Express backend. This also uses LlamaIndex.TS.
-  - **Python FastAPI**: if you select this option, you’ll get a backend powered by the [llama-index python package](https://pypi.org/project/llama-index/), which you can deploy to a service like Render or fly.io.
-- The back-end has two endpoints (one streaming, the other one non-streaming) that allow you to send the state of your chat and receive additional responses
-- You add arbitrary data sources to your chat, like local files, websites, or data retrieved from a database.
-- Turn your chat into an AI agent by adding tools (functions called by the LLM).
+  - **Python FastAPI**: if you select this option, you’ll get a separate backend powered by the [llama-index Python package](https://pypi.org/project/llama-index/), which you can deploy to a service like [Render](https://render.com/) or [fly.io](https://fly.io/). The separate Next.js front-end will connect to this backend.
+- Each back-end has two endpoints:
+  - One streaming chat endpoint, that allow you to send the state of your chat and receive additional responses
+  - One endpoint to upload private files which can be used in your chat
 - The app uses OpenAI by default, so you'll need an OpenAI API key, or you can customize it to use any of the dozens of LLMs we support.
 
+Here's how it looks like:
+
+https://github.com/user-attachments/assets/d57af1a1-d99b-4e9c-98d9-4cbd1327eff8
+
 ## Using your data
 
-You can supply your own data; the app will index it and answer questions. Your generated app will have a folder called `data` (If you're using Express or Python and generate a frontend, it will be `./backend/data`).
+Optionally, you can supply your own data; the app will index it and make use of it, e.g. to answer questions. Your generated app will have a folder called `data` (If you're using Express or Python and generate a frontend, it will be `./backend/data`).
 
-The app will ingest any supported files you put in this directory. Your Next.js and Express apps use LlamaIndex.TS so they will be able to ingest any PDF, text, CSV, Markdown, Word and HTML files. The Python backend can read even more types, including video and audio files.
+The app will ingest any supported files you put in this directory. Your Next.js and Express apps use LlamaIndex.TS, so they will be able to ingest any PDF, text, CSV, Markdown, Word and HTML files. The Python backend can read even more types, including video and audio files.
 
 Before you can use your data, you need to index it. If you're using the Next.js or Express apps, run:
 
@@ -48,13 +58,9 @@ If you're using the Python backend, you can trigger indexing of your data by cal
 poetry run generate
 ```
 
-## Want a front-end?
-
-Optionally generate a frontend if you've selected the Python or Express back-ends. If you do so, `create-llama` will generate two folders: `frontend`, for your Next.js-based frontend code, and `backend` containing your API.
-
 ## Customizing the AI models
 
-The app will default to OpenAI's `gpt-4-turbo` LLM and `text-embedding-3-large` embedding model.
+The app will default to OpenAI's `gpt-4o-mini` LLM and `text-embedding-3-large` embedding model.
 
 If you want to use different OpenAI models, add the `--ask-models` CLI parameter.
 
@@ -84,45 +90,40 @@ Need to install the following packages:
   create-llama@latest
 Ok to proceed? (y) y
 ✔ What is your project named? … my-app
-✔ Which template would you like to use? › Chat
-✔ Which framework would you like to use? › NextJS
-✔ Would you like to set up observability? › No
+✔ What app do you want to build? › Agentic RAG
+✔ What language do you want to use? › Python (FastAPI)
+✔ Do you want to use LlamaCloud services? … No / Yes
+✔ Please provide your LlamaCloud API key (leave blank to skip): …
 ✔ Please provide your OpenAI API key (leave blank to skip): …
-✔ Which data source would you like to use? › Use an example PDF
-✔ Would you like to add another data source? › No
-✔ Would you like to use LlamaParse (improved parser for RAG - requires API key)? … no / yes
-✔ Would you like to use a vector database? › No, just store the data in the file system
 ? How would you like to proceed? › - Use arrow-keys. Return to submit.
-   Just generate code (~1 sec)
-❯  Start in VSCode (~1 sec)
-   Generate code and install dependencies (~2 min)
-   Generate code, install dependencies, and run the app (~2 min)
+    Just generate code (~1 sec)
+❯   Start in VSCode (~1 sec)
+    Generate code and install dependencies (~2 min)
 ```
 
 ### Running non-interactively
 
 You can also pass command line arguments to set up a new project
-non-interactively. See `create-llama --help`:
+non-interactively. For a list of the latest options, call `create-llama --help`.
 
-```bash
-create-llama <project-directory> [options]
+### Running in pro mode
 
-Options:
-  -V, --version                      output the version number
+If you prefer more advanced customization options, you can run `create-llama` in pro mode using the `--pro` flag.
 
-  --use-npm
+In pro mode, instead of selecting a predefined use case, you'll be prompted to select each technical component of your project. This allows for greater flexibility in customizing your project, including:
 
-    Explicitly tell the CLI to bootstrap the app using npm
+- **Vector Store**: Choose from a variety of vector stores for keeping your documents, including MongoDB, Pinecone, Weaviate, Qdrant and Chroma.
+- **Tools**: Choose from a variety of agent tools (functions called by the LLM), such as:
+  - Code Interpreter: Executes Python code in a secure Jupyter notebook environment
+  - Artifact Code Generator: Generates code artifacts that can be run in a sandbox
+  - OpenAPI Action: Facilitates requests to a provided OpenAPI schema
+  - Image Generator: Creates images based on text descriptions
+  - Web Search: Performs web searches to retrieve up-to-date information
+- **Data Sources**: Integrate various data sources into your chat application, including local files, websites, or database-retrieved data.
+- **Backend Options**: Besides using Next.js or FastAPI, you can also select to use Express for a more traditional Node.js application.
+- **Observability**: Choose from a variety of LLM observability tools, including LlamaTrace and Traceloop.
 
-  --use-pnpm
-
-    Explicitly tell the CLI to bootstrap the app using pnpm
-
-  --use-yarn
-
-    Explicitly tell the CLI to bootstrap the app using Yarn
-
-```
+Pro mode is ideal for developers who want fine-grained control over their project's configuration and are comfortable with more technical setup options.
 
 ## LlamaIndex Documentation
 

create-app.ts

@@ -7,17 +7,16 @@ import { getOnline } from "./helpers/is-online";
 import { isWriteable } from "./helpers/is-writeable";
 import { makeDir } from "./helpers/make-dir";
 
-import fs from "fs";
 import terminalLink from "terminal-link";
-import type { InstallTemplateArgs } from "./helpers";
+import type { InstallTemplateArgs, TemplateObservability } from "./helpers";
 import { installTemplate } from "./helpers";
-import { writeDevcontainer } from "./helpers/devcontainer";
 import { templatesDir } from "./helpers/dir";
 import { toolsRequireConfig } from "./helpers/tools";
+import { configVSCode } from "./helpers/vscode";
 
 export type InstallAppArgs = Omit<
   InstallTemplateArgs,
-  "appName" | "root" | "isOnline" | "customApiPath"
+  "appName" | "root" | "isOnline" | "port"
 > & {
   appPath: string;
   frontend: boolean;
@@ -35,12 +34,12 @@ export async function createApp({
   communityProjectConfig,
   llamapack,
   vectorDb,
-  externalPort,
   postInstallAction,
   dataSources,
   tools,
   useLlamaParse,
   observability,
+  useCase,
 }: InstallAppArgs): Promise<void> {
   const root = path.resolve(appPath);
 
@@ -80,39 +79,30 @@ export async function createApp({
     communityProjectConfig,
     llamapack,
     vectorDb,
-    externalPort,
     postInstallAction,
     dataSources,
     tools,
     useLlamaParse,
     observability,
+    useCase,
   };
 
-  if (frontend) {
-    // install backend
-    const backendRoot = path.join(root, "backend");
-    await makeDir(backendRoot);
-    await installTemplate({ ...args, root: backendRoot, backend: true });
+  // Install backend
+  await installTemplate({ ...args, backend: true });
+
+  if (frontend && framework === "fastapi" && template !== "llamaindexserver") {
     // install frontend
-    const frontendRoot = path.join(root, "frontend");
+    const frontendRoot = path.join(root, ".frontend");
     await makeDir(frontendRoot);
     await installTemplate({
       ...args,
       root: frontendRoot,
       framework: "nextjs",
-      customApiPath: `http://localhost:${externalPort ?? 8000}/api/chat`,
       backend: false,
     });
-    // copy readme for fullstack
-    await fs.promises.copyFile(
-      path.join(templatesDir, "README-fullstack.md"),
-      path.join(root, "README.md"),
-    );
-  } else {
-    await installTemplate({ ...args, backend: true });
   }
 
-  await writeDevcontainer(root, templatesDir, framework, frontend);
+  await configVSCode(root, templatesDir, framework);
 
   process.chdir(root);
   if (tryGitInit(root)) {
@@ -120,7 +110,7 @@ export async function createApp({
     console.log();
   }
 
-  if (toolsRequireConfig(tools)) {
+  if (toolsRequireConfig(tools) && template !== "llamaindexserver") {
     const configFile =
       framework === "fastapi" ? "config/tools.yaml" : "config/tools.json";
     console.log(
@@ -142,14 +132,42 @@ export async function createApp({
     )} and learn how to get started.`,
   );
 
-  if (args.observability === "opentelemetry") {
+  outputObservability(args.observability);
+
+  if (
+    dataSources.some((dataSource) => dataSource.type === "file") &&
+    process.platform === "linux"
+  ) {
     console.log(
-      `\n${yellow("Observability")}: Visit the ${terminalLink(
-        "documentation",
-        "https://traceloop.com/docs/openllmetry/integrations",
-      )} to set up the environment variables and start seeing execution traces.`,
+      yellow(
+        `You can add your own data files to ${terminalLink(
+          "data",
+          `file://${root}/data`,
+        )} folder manually.`,
+      ),
     );
   }
 
   console.log();
 }
+
+function outputObservability(observability?: TemplateObservability) {
+  switch (observability) {
+    case "traceloop":
+      console.log(
+        `\n${yellow("Observability")}: Visit the ${terminalLink(
+          "documentation",
+          "https://traceloop.com/docs/openllmetry/integrations",
+        )} to set up the environment variables and start seeing execution traces.`,
+      );
+      break;
+    case "llamatrace":
+      console.log(
+        `\n${yellow("Observability")}: LlamaTrace has been configured for your project. Visit the ${terminalLink(
+          "LlamaTrace dashboard",
+          "https://llamatrace.com/login",
+        )} to view your traces and monitor your application.`,
+      );
+      break;
+  }
+}

e2e/basic.spec.ts

@@ -1,129 +0,0 @@
-/* eslint-disable turbo/no-undeclared-env-vars */
-import { expect, test } from "@playwright/test";
-import { ChildProcess } from "child_process";
-import fs from "fs";
-import path from "path";
-import type {
-  TemplateFramework,
-  TemplatePostInstallAction,
-  TemplateType,
-  TemplateUI,
-} from "../helpers";
-import { createTestDir, runCreateLlama, type AppType } from "./utils";
-
-const templateTypes: TemplateType[] = ["streaming"];
-const templateFrameworks: TemplateFramework[] = [
-  "nextjs",
-  "express",
-  "fastapi",
-];
-const dataSources: string[] = ["--no-files", "--example-file"];
-const templateUIs: TemplateUI[] = ["shadcn", "html"];
-const templatePostInstallActions: TemplatePostInstallAction[] = [
-  "none",
-  "runApp",
-];
-
-for (const templateType of templateTypes) {
-  for (const templateFramework of templateFrameworks) {
-    for (const dataSource of dataSources) {
-      for (const templateUI of templateUIs) {
-        for (const templatePostInstallAction of templatePostInstallActions) {
-          const appType: AppType =
-            templateFramework === "nextjs" ? "" : "--frontend";
-          test.describe(`try create-llama ${templateType} ${templateFramework} ${dataSource} ${templateUI} ${appType} ${templatePostInstallAction}`, async () => {
-            let port: number;
-            let externalPort: number;
-            let cwd: string;
-            let name: string;
-            let appProcess: ChildProcess;
-            // Only test without using vector db for now
-            const vectorDb = "none";
-
-            test.beforeAll(async () => {
-              port = Math.floor(Math.random() * 10000) + 10000;
-              externalPort = port + 1;
-              cwd = await createTestDir();
-              const result = await runCreateLlama(
-                cwd,
-                templateType,
-                templateFramework,
-                dataSource,
-                templateUI,
-                vectorDb,
-                appType,
-                port,
-                externalPort,
-                templatePostInstallAction,
-              );
-              name = result.projectName;
-              appProcess = result.appProcess;
-            });
-
-            test("App folder should exist", async () => {
-              const dirExists = fs.existsSync(path.join(cwd, name));
-              expect(dirExists).toBeTruthy();
-            });
-            test("Frontend should have a title", async ({ page }) => {
-              test.skip(templatePostInstallAction !== "runApp");
-              await page.goto(`http://localhost:${port}`);
-              await expect(page.getByText("Built by LlamaIndex")).toBeVisible();
-            });
-
-            test("Frontend should be able to submit a message and receive a response", async ({
-              page,
-            }) => {
-              test.skip(templatePostInstallAction !== "runApp");
-              await page.goto(`http://localhost:${port}`);
-              await page.fill("form input", "hello");
-              const [response] = await Promise.all([
-                page.waitForResponse(
-                  (res) => {
-                    return (
-                      res.url().includes("/api/chat") && res.status() === 200
-                    );
-                  },
-                  {
-                    timeout: 1000 * 60,
-                  },
-                ),
-                page.click("form button[type=submit]"),
-              ]);
-              const text = await response.text();
-              console.log("AI response when submitting message: ", text);
-              expect(response.ok()).toBeTruthy();
-            });
-
-            test("Backend frameworks should response when calling non-streaming chat API", async ({
-              request,
-            }) => {
-              test.skip(templatePostInstallAction !== "runApp");
-              test.skip(templateFramework === "nextjs");
-              const response = await request.post(
-                `http://localhost:${externalPort}/api/chat/request`,
-                {
-                  data: {
-                    messages: [
-                      {
-                        role: "user",
-                        content: "Hello",
-                      },
-                    ],
-                  },
-                },
-              );
-              const text = await response.text();
-              console.log("AI response when calling API: ", text);
-              expect(response.ok()).toBeTruthy();
-            });
-
-            // clean processes
-            test.afterAll(async () => {
-              appProcess?.kill();
-            });
-          });
-        }
-      }
-    }
-  }
-}

e2e/python/resolve_dependencies.spec.ts

@@ -0,0 +1,233 @@
+import { expect, test } from "@playwright/test";
+import { exec } from "child_process";
+import fs from "fs";
+import path from "path";
+import util from "util";
+import { TemplateFramework, TemplateVectorDB } from "../../helpers/types";
+import { RunCreateLlamaOptions, createTestDir, runCreateLlama } from "../utils";
+
+const execAsync = util.promisify(exec);
+
+const templateFramework: TemplateFramework = process.env.FRAMEWORK
+  ? (process.env.FRAMEWORK as TemplateFramework)
+  : "fastapi";
+const dataSource: string = process.env.DATASOURCE
+  ? process.env.DATASOURCE
+  : "--example-file";
+
+// TODO: add support for other templates
+
+if (
+  dataSource === "--example-file" // XXX: this test provides its own data source - only trigger it on one data source (usually the CI matrix will trigger multiple data sources)
+) {
+  // vectorDBs, tools, and data source combinations to test
+  const vectorDbs: TemplateVectorDB[] = [
+    "mongo",
+    "pg",
+    "pinecone",
+    "milvus",
+    "astra",
+    "qdrant",
+    "chroma",
+    "weaviate",
+  ];
+
+  const toolOptions = [
+    "wikipedia.WikipediaToolSpec",
+    "google.GoogleSearchToolSpec",
+    "document_generator",
+    "artifact",
+  ];
+
+  const dataSources = [
+    "--example-file",
+    "--web-source https://www.example.com",
+    "--db-source mysql+pymysql://user:pass@localhost:3306/mydb",
+  ];
+
+  const observabilityOptions = ["llamatrace", "traceloop"];
+
+  test.describe("Mypy check", () => {
+    test.describe.configure({ retries: 0 });
+
+    // Test vector databases
+    for (const vectorDb of vectorDbs) {
+      test(`Mypy check for vectorDB: ${vectorDb}`, async () => {
+        const cwd = await createTestDir();
+        const { pyprojectPath } = await createAndCheckLlamaProject({
+          options: {
+            cwd,
+            templateType: "streaming",
+            templateFramework,
+            dataSource: "--example-file",
+            vectorDb,
+            tools: "none",
+            port: 3000,
+            postInstallAction: "none",
+            templateUI: undefined,
+            appType: "--no-frontend",
+            llamaCloudProjectName: undefined,
+            llamaCloudIndexName: undefined,
+            observability: undefined,
+          },
+        });
+
+        const pyprojectContent = fs.readFileSync(pyprojectPath, "utf-8");
+        if (vectorDb !== "none") {
+          if (vectorDb === "pg") {
+            expect(pyprojectContent).toContain(
+              "llama-index-vector-stores-postgres",
+            );
+          } else {
+            expect(pyprojectContent).toContain(
+              `llama-index-vector-stores-${vectorDb}`,
+            );
+          }
+        }
+      });
+    }
+
+    // Test tools
+    for (const tool of toolOptions) {
+      test(`Mypy check for tool: ${tool}`, async () => {
+        const cwd = await createTestDir();
+        const { pyprojectPath } = await createAndCheckLlamaProject({
+          options: {
+            cwd,
+            templateType: "streaming",
+            templateFramework,
+            dataSource: "--example-file",
+            vectorDb: "none",
+            tools: tool,
+            port: 3000,
+            postInstallAction: "none",
+            templateUI: undefined,
+            appType: "--no-frontend",
+            llamaCloudProjectName: undefined,
+            llamaCloudIndexName: undefined,
+            observability: undefined,
+          },
+        });
+
+        const pyprojectContent = fs.readFileSync(pyprojectPath, "utf-8");
+        if (tool === "wikipedia.WikipediaToolSpec") {
+          expect(pyprojectContent).toContain("wikipedia");
+        }
+        if (tool === "google.GoogleSearchToolSpec") {
+          expect(pyprojectContent).toContain("google");
+        }
+      });
+    }
+
+    // Test data sources
+    for (const dataSource of dataSources) {
+      const dataSourceType = dataSource.split(" ")[0];
+      test(`Mypy check for data source: ${dataSourceType}`, async () => {
+        const cwd = await createTestDir();
+        const { pyprojectPath } = await createAndCheckLlamaProject({
+          options: {
+            cwd,
+            templateType: "streaming",
+            templateFramework,
+            dataSource,
+            vectorDb: "none",
+            tools: "none",
+            port: 3000,
+            postInstallAction: "none",
+            templateUI: undefined,
+            appType: "--no-frontend",
+            llamaCloudProjectName: undefined,
+            llamaCloudIndexName: undefined,
+            observability: undefined,
+          },
+        });
+
+        const pyprojectContent = fs.readFileSync(pyprojectPath, "utf-8");
+        if (dataSource.includes("--web-source")) {
+          expect(pyprojectContent).toContain("llama-index-readers-web");
+        }
+        if (dataSource.includes("--db-source")) {
+          expect(pyprojectContent).toContain("llama-index-readers-database");
+        }
+      });
+    }
+
+    // Test observability options
+    for (const observability of observabilityOptions) {
+      test(`Mypy check for observability: ${observability}`, async () => {
+        const cwd = await createTestDir();
+
+        const { pyprojectPath } = await createAndCheckLlamaProject({
+          options: {
+            cwd,
+            templateType: "streaming",
+            templateFramework,
+            dataSource: "--example-file",
+            vectorDb: "none",
+            tools: "none",
+            port: 3000,
+            postInstallAction: "none",
+            templateUI: undefined,
+            appType: "--no-frontend",
+            llamaCloudProjectName: undefined,
+            llamaCloudIndexName: undefined,
+            observability,
+          },
+        });
+      });
+    }
+  });
+}
+
+async function createAndCheckLlamaProject({
+  options,
+}: {
+  options: RunCreateLlamaOptions;
+}): Promise<{ pyprojectPath: string; projectPath: string }> {
+  const result = await runCreateLlama(options);
+  const name = result.projectName;
+  const projectPath = path.join(options.cwd, name);
+
+  // Check if the app folder exists
+  expect(fs.existsSync(projectPath)).toBeTruthy();
+
+  // Check if pyproject.toml exists
+  const pyprojectPath = path.join(projectPath, "pyproject.toml");
+  expect(fs.existsSync(pyprojectPath)).toBeTruthy();
+
+  const env = {
+    ...process.env,
+    POETRY_VIRTUALENVS_IN_PROJECT: "true",
+  };
+
+  // Run poetry install
+  try {
+    const { stdout: installStdout, stderr: installStderr } = await execAsync(
+      "poetry install",
+      { cwd: projectPath, env },
+    );
+    console.log("poetry install stdout:", installStdout);
+    console.error("poetry install stderr:", installStderr);
+  } catch (error) {
+    console.error("Error running poetry install:", error);
+    throw error;
+  }
+
+  // Run poetry run mypy
+  try {
+    const { stdout: mypyStdout, stderr: mypyStderr } = await execAsync(
+      "poetry run mypy .",
+      { cwd: projectPath, env },
+    );
+    console.log("poetry run mypy stdout:", mypyStdout);
+    console.error("poetry run mypy stderr:", mypyStderr);
+  } catch (error) {
+    console.error("Error running mypy:", error);
+    throw error;
+  }
+
+  // If we reach this point without throwing an error, the test passes
+  expect(true).toBeTruthy();
+
+  return { pyprojectPath, projectPath };
+}

e2e/shared/llamaindexserver_template.spec.ts

@@ -0,0 +1,105 @@
+/* eslint-disable turbo/no-undeclared-env-vars */
+import { expect, test } from "@playwright/test";
+import { ChildProcess } from "child_process";
+import fs from "fs";
+import path from "path";
+import type {
+  TemplateFramework,
+  TemplatePostInstallAction,
+  TemplateUI,
+} from "../../helpers";
+import { createTestDir, runCreateLlama, type AppType } from "../utils";
+
+const templateFramework: TemplateFramework = process.env.FRAMEWORK
+  ? (process.env.FRAMEWORK as TemplateFramework)
+  : "fastapi";
+const dataSource: string = "--example-file";
+const templateUI: TemplateUI = "shadcn";
+const templatePostInstallAction: TemplatePostInstallAction = "runApp";
+const appType: AppType = "--frontend";
+const userMessage = "Write a blog post about physical standards for letters";
+const templateUseCases = ["financial_report", "agentic_rag", "deep_research"];
+
+for (const useCase of templateUseCases) {
+  test.describe(`Test use case ${useCase} ${templateFramework} ${dataSource} ${templateUI} ${appType} ${templatePostInstallAction}`, async () => {
+    test.skip(
+      process.platform !== "linux" ||
+        process.env.DATASOURCE === "--no-files" ||
+        templateFramework === "express",
+      "The llamaindexserver template currently only works with nextjs, fastapi. We also only run on Linux to speed up tests.",
+    );
+    let port: number;
+    let cwd: string;
+    let name: string;
+    let appProcess: ChildProcess;
+    // Only test without using vector db for now
+    const vectorDb = "none";
+
+    test.beforeAll(async () => {
+      port = Math.floor(Math.random() * 10000) + 10000;
+      cwd = await createTestDir();
+      const result = await runCreateLlama({
+        cwd,
+        templateType: "llamaindexserver",
+        templateFramework,
+        dataSource,
+        vectorDb,
+        port,
+        postInstallAction: templatePostInstallAction,
+        templateUI,
+        appType,
+        useCase,
+      });
+      name = result.projectName;
+      appProcess = result.appProcess;
+    });
+
+    test("App folder should exist", async () => {
+      const dirExists = fs.existsSync(path.join(cwd, name));
+      expect(dirExists).toBeTruthy();
+    });
+
+    test("Frontend should have a title", async ({ page }) => {
+      test.skip(
+        templatePostInstallAction !== "runApp" ||
+          templateFramework === "express",
+      );
+      await page.goto(`http://localhost:${port}`);
+      await expect(page.getByText("Built by LlamaIndex")).toBeVisible();
+    });
+
+    test("Frontend should be able to submit a message and receive the start of a streamed response", async ({
+      page,
+    }) => {
+      test.skip(
+        templatePostInstallAction !== "runApp" ||
+          useCase === "financial_report" ||
+          useCase === "deep_research" ||
+          templateFramework === "express",
+        "Skip chat tests for financial report and deep research.",
+      );
+      await page.goto(`http://localhost:${port}`);
+      await page.fill("form textarea", userMessage);
+
+      const responsePromise = page.waitForResponse((res) =>
+        res.url().includes("/api/chat"),
+      );
+
+      await page.click("form button[type=submit]");
+
+      const response = await responsePromise;
+      console.log(`Response status: ${response.status()}`);
+      const responseBody = await response
+        .text()
+        .catch((e) => `Error reading body: ${e}`);
+      console.log(`Response body: ${responseBody}`);
+
+      expect(response.ok()).toBeTruthy();
+    });
+
+    // clean processes
+    test.afterAll(async () => {
+      appProcess?.kill();
+    });
+  });
+}

e2e/shared/reflex_template.spec.ts

@@ -0,0 +1,64 @@
+/* eslint-disable turbo/no-undeclared-env-vars */
+import { expect, test } from "@playwright/test";
+import { ChildProcess } from "child_process";
+import fs from "fs";
+import path from "path";
+import { TemplateFramework, TemplateUseCase } from "../../helpers";
+import { createTestDir, runCreateLlama } from "../utils";
+
+const templateFramework: TemplateFramework = process.env.FRAMEWORK
+  ? (process.env.FRAMEWORK as TemplateFramework)
+  : "fastapi";
+const dataSource: string = process.env.DATASOURCE
+  ? process.env.DATASOURCE
+  : "--example-file";
+const templateUseCases: TemplateUseCase[] = ["extractor", "contract_review"];
+
+// The reflex template currently only works with FastAPI and files (and not on Windows)
+if (
+  process.platform !== "win32" &&
+  templateFramework === "fastapi" &&
+  dataSource === "--example-file"
+) {
+  for (const useCase of templateUseCases) {
+    test.describe(`Test reflex template ${useCase} ${templateFramework} ${dataSource}`, async () => {
+      let appPort: number;
+      let name: string;
+      let appProcess: ChildProcess;
+      let cwd: string;
+
+      // Create reflex app
+      test.beforeAll(async () => {
+        cwd = await createTestDir();
+        appPort = Math.floor(Math.random() * 10000) + 10000;
+        const result = await runCreateLlama({
+          cwd,
+          templateType: "reflex",
+          templateFramework: "fastapi",
+          dataSource: "--example-file",
+          vectorDb: "none",
+          port: appPort,
+          postInstallAction: "runApp",
+          useCase,
+        });
+        name = result.projectName;
+        appProcess = result.appProcess;
+      });
+
+      test.afterAll(async () => {
+        appProcess.kill();
+      });
+
+      test("App folder should exist", async () => {
+        const dirExists = fs.existsSync(path.join(cwd, name));
+        expect(dirExists).toBeTruthy();
+      });
+      test("Frontend should have a title", async ({ page }) => {
+        await page.goto(`http://localhost:${appPort}`);
+        await expect(page.getByText("Built by LlamaIndex")).toBeVisible({
+          timeout: 2000 * 60,
+        });
+      });
+    });
+  }
+}

e2e/shared/streaming_template.spec.ts

@@ -0,0 +1,128 @@
+/* eslint-disable turbo/no-undeclared-env-vars */
+import { expect, test } from "@playwright/test";
+import { ChildProcess } from "child_process";
+import fs from "fs";
+import path from "path";
+import type {
+  TemplateFramework,
+  TemplatePostInstallAction,
+  TemplateUI,
+} from "../../helpers";
+import { createTestDir, runCreateLlama, type AppType } from "../utils";
+
+const templateFramework: TemplateFramework = process.env.FRAMEWORK
+  ? (process.env.FRAMEWORK as TemplateFramework)
+  : "fastapi";
+const dataSource: string = process.env.DATASOURCE
+  ? process.env.DATASOURCE
+  : "--example-file";
+const templateUI: TemplateUI = "shadcn";
+const templatePostInstallAction: TemplatePostInstallAction = "runApp";
+
+const llamaCloudProjectName = "create-llama";
+const llamaCloudIndexName = "e2e-test";
+
+const appType: AppType = templateFramework === "fastapi" ? "--frontend" : "";
+const userMessage =
+  dataSource !== "--no-files" ? "Physical standard for letters" : "Hello";
+
+test.describe(`Test streaming template ${templateFramework} ${dataSource} ${templateUI} ${appType} ${templatePostInstallAction}`, async () => {
+  const isNode18 = process.version.startsWith("v18");
+  const isLlamaCloud = dataSource === "--llamacloud";
+  // llamacloud is using File API which is not supported on node 18
+  if (isNode18 && isLlamaCloud) {
+    test.skip(true, "Skipping tests for Node 18 and LlamaCloud data source");
+  }
+
+  let port: number;
+  let cwd: string;
+  let name: string;
+  let appProcess: ChildProcess;
+  // Only test without using vector db for now
+  const vectorDb = "none";
+
+  test.beforeAll(async () => {
+    port = Math.floor(Math.random() * 10000) + 10000;
+    cwd = await createTestDir();
+    const result = await runCreateLlama({
+      cwd,
+      templateType: "streaming",
+      templateFramework,
+      dataSource,
+      vectorDb,
+      port,
+      postInstallAction: templatePostInstallAction,
+      templateUI,
+      appType,
+      llamaCloudProjectName,
+      llamaCloudIndexName,
+    });
+    name = result.projectName;
+    appProcess = result.appProcess;
+  });
+
+  test("App folder should exist", async () => {
+    const dirExists = fs.existsSync(path.join(cwd, name));
+    expect(dirExists).toBeTruthy();
+  });
+
+  test("Frontend should have a title", async ({ page }) => {
+    test.skip(
+      templatePostInstallAction !== "runApp" || templateFramework === "express",
+    );
+    await page.goto(`http://localhost:${port}`);
+    await expect(page.getByText("Built by LlamaIndex")).toBeVisible();
+  });
+
+  test("Frontend should be able to submit a message and receive a response", async ({
+    page,
+  }) => {
+    test.skip(
+      templatePostInstallAction !== "runApp" || templateFramework === "express",
+    );
+    await page.goto(`http://localhost:${port}`);
+    await page.fill("form textarea", userMessage);
+    const [response] = await Promise.all([
+      page.waitForResponse(
+        (res) => {
+          return res.url().includes("/api/chat") && res.status() === 200;
+        },
+        {
+          timeout: 1000 * 60,
+        },
+      ),
+      page.click("form button[type=submit]"),
+    ]);
+    const text = await response.text();
+    console.log("AI response when submitting message: ", text);
+    expect(response.ok()).toBeTruthy();
+  });
+
+  test("Backend frameworks should response when calling non-streaming chat API", async ({
+    request,
+  }) => {
+    test.skip(templatePostInstallAction !== "runApp");
+    test.skip(templateFramework === "nextjs");
+    const response = await request.post(
+      `http://localhost:${port}/api/chat/request`,
+      {
+        data: {
+          messages: [
+            {
+              role: "user",
+              content: userMessage,
+            },
+          ],
+        },
+      },
+    );
+    const text = await response.text();
+    console.log("AI response when calling API: ", text);
+    expect(response.ok()).toBeTruthy();
+  });
+
+  // clean processes
+  test.afterAll(async () => {
+    appProcess?.kill();
+  });
+});

e2e/typescript/resolve_dependencies.spec.ts

@@ -0,0 +1,105 @@
+import { expect, test } from "@playwright/test";
+import { exec } from "child_process";
+import fs from "fs";
+import path from "path";
+import util from "util";
+import { TemplateFramework, TemplateVectorDB } from "../../helpers/types";
+import { createTestDir, runCreateLlama } from "../utils";
+
+const execAsync = util.promisify(exec);
+
+const templateFramework: TemplateFramework = process.env.FRAMEWORK
+  ? (process.env.FRAMEWORK as TemplateFramework)
+  : "nextjs";
+const dataSource: string = process.env.DATASOURCE
+  ? process.env.DATASOURCE
+  : "--example-file";
+
+// vectorDBs combinations to test
+const vectorDbs: TemplateVectorDB[] = [
+  "mongo",
+  "pg",
+  "qdrant",
+  "pinecone",
+  "milvus",
+  "astra",
+  "chroma",
+  "llamacloud",
+  "weaviate",
+];
+
+test.describe("Test resolve TS dependencies", () => {
+  // Test vector DBs without LlamaParse
+  for (const vectorDb of vectorDbs) {
+    const optionDescription = `vectorDb: ${vectorDb}, dataSource: ${dataSource}`;
+
+    test(`Vector DB test - ${optionDescription}`, async () => {
+      await runTest(vectorDb, false);
+    });
+  }
+
+  // Test LlamaParse with vectorDB 'none'
+  test(`LlamaParse test - vectorDb: none, dataSource: ${dataSource}, llamaParse: true`, async () => {
+    await runTest("none", true);
+  });
+
+  async function runTest(
+    vectorDb: TemplateVectorDB | "none",
+    useLlamaParse: boolean,
+  ) {
+    const cwd = await createTestDir();
+
+    const result = await runCreateLlama({
+      cwd: cwd,
+      templateType: "streaming",
+      templateFramework: templateFramework,
+      dataSource: dataSource,
+      vectorDb: vectorDb,
+      port: 3000,
+      postInstallAction: "none",
+      templateUI: undefined,
+      appType: templateFramework === "nextjs" ? "" : "--no-frontend",
+      llamaCloudProjectName: undefined,
+      llamaCloudIndexName: undefined,
+      tools: undefined,
+      useLlamaParse: useLlamaParse,
+    });
+    const name = result.projectName;
+
+    // Check if the app folder exists
+    const appDir = path.join(cwd, name);
+    const dirExists = fs.existsSync(appDir);
+    expect(dirExists).toBeTruthy();
+
+    // Install dependencies using pnpm
+    try {
+      const { stderr: installStderr } = await execAsync(
+        "pnpm install --prefer-offline",
+        {
+          cwd: appDir,
+        },
+      );
+    } catch (error) {
+      console.error("Error installing dependencies:", error);
+      throw error;
+    }
+
+    // Run tsc type check and capture the output
+    try {
+      const { stdout, stderr } = await execAsync(
+        "pnpm exec tsc -b --diagnostics",
+        {
+          cwd: appDir,
+        },
+      );
+      // Check if there's any error output
+      expect(stderr).toBeFalsy();
+
+      // Log the stdout for debugging purposes
+      console.log("TypeScript type-check output:", stdout);
+    } catch (error) {
+      console.error("Error running tsc:", error);
+      throw error;
+    }
+  }
+});

e2e/utils.ts

@@ -18,141 +18,137 @@ export type CreateLlamaResult = {
   appProcess: ChildProcess;
 };
 
-// eslint-disable-next-line max-params
-export async function checkAppHasStarted(
-  frontend: boolean,
-  framework: TemplateFramework,
-  port: number,
-  externalPort: number,
-  timeout: number,
-) {
-  if (frontend) {
-    await Promise.all([
-      waitPort({
-        host: "localhost",
-        port: port,
-        timeout,
-      }),
-      waitPort({
-        host: "localhost",
-        port: externalPort,
-        timeout,
-      }),
-    ]).catch((err) => {
-      console.error(err);
-      throw err;
-    });
-  } else {
-    let wPort: number;
-    if (framework === "nextjs") {
-      wPort = port;
-    } else {
-      wPort = externalPort;
-    }
-    await waitPort({
-      host: "localhost",
-      port: wPort,
-      timeout,
-    }).catch((err) => {
-      console.error(err);
-      throw err;
-    });
-  }
-}
+export type RunCreateLlamaOptions = {
+  cwd: string;
+  templateType: TemplateType;
+  templateFramework: TemplateFramework;
+  dataSource: string;
+  vectorDb: TemplateVectorDB;
+  port: number;
+  postInstallAction: TemplatePostInstallAction;
+  templateUI?: TemplateUI;
+  appType?: AppType;
+  llamaCloudProjectName?: string;
+  llamaCloudIndexName?: string;
+  tools?: string;
+  useLlamaParse?: boolean;
+  observability?: string;
+  useCase?: string;
+};
 
-// eslint-disable-next-line max-params
-export async function runCreateLlama(
-  cwd: string,
-  templateType: TemplateType,
-  templateFramework: TemplateFramework,
-  dataSource: string,
-  templateUI: TemplateUI,
-  vectorDb: TemplateVectorDB,
-  appType: AppType,
-  port: number,
-  externalPort: number,
-  postInstallAction: TemplatePostInstallAction,
-): Promise<CreateLlamaResult> {
-  if (!process.env.OPENAI_API_KEY) {
-    throw new Error("Setting OPENAI_API_KEY is mandatory to run tests");
+export async function runCreateLlama({
+  cwd,
+  templateType,
+  templateFramework,
+  dataSource,
+  vectorDb,
+  port,
+  postInstallAction,
+  templateUI,
+  appType,
+  llamaCloudProjectName,
+  llamaCloudIndexName,
+  tools,
+  useLlamaParse,
+  observability,
+  useCase,
+}: RunCreateLlamaOptions): Promise<CreateLlamaResult> {
+  if (!process.env.OPENAI_API_KEY || !process.env.LLAMA_CLOUD_API_KEY) {
+    throw new Error(
+      "Setting the OPENAI_API_KEY and LLAMA_CLOUD_API_KEY is mandatory to run tests",
+    );
   }
   const name = [
     templateType,
     templateFramework,
-    dataSource,
+    dataSource.split(" ")[0],
     templateUI,
     appType,
   ].join("-");
-  const command = [
+
+  // Handle different data source types
+  let dataSourceArgs = [];
+  if (dataSource.includes("--web-source" || "--db-source")) {
+    const webSource = dataSource.split(" ")[1];
+    dataSourceArgs.push("--web-source", webSource);
+  } else if (dataSource.includes("--db-source")) {
+    const dbSource = dataSource.split(" ")[1];
+    dataSourceArgs.push("--db-source", dbSource);
+  } else {
+    dataSourceArgs.push(dataSource);
+  }
+
+  const commandArgs = [
     "create-llama",
     name,
     "--template",
     templateType,
     "--framework",
     templateFramework,
-    dataSource,
-    "--ui",
-    templateUI,
+    ...dataSourceArgs,
     "--vector-db",
     vectorDb,
-    "--open-ai-key",
-    process.env.OPENAI_API_KEY,
-    appType,
-    "--use-pnpm",
+    "--use-npm",
     "--port",
     port,
-    "--external-port",
-    externalPort,
     "--post-install-action",
     postInstallAction,
     "--tools",
-    "none",
-    "--no-llama-parse",
+    tools ?? "none",
     "--observability",
     "none",
-  ].join(" ");
+  ];
+
+  if (templateUI) {
+    commandArgs.push("--ui", templateUI);
+  }
+  if (appType) {
+    commandArgs.push(appType);
+  }
+  if (useLlamaParse) {
+    commandArgs.push("--use-llama-parse");
+  } else {
+    commandArgs.push("--no-llama-parse");
+  }
+  if (observability) {
+    commandArgs.push("--observability", observability);
+  }
+  if (
+    (templateType === "multiagent" ||
+      templateType === "reflex" ||
+      templateType === "llamaindexserver") &&
+    useCase
+  ) {
+    commandArgs.push("--use-case", useCase);
+  }
+
+  const command = commandArgs.join(" ");
   console.log(`running command '${command}' in ${cwd}`);
   const appProcess = exec(command, {
     cwd,
     env: {
       ...process.env,
+      LLAMA_CLOUD_PROJECT_NAME: llamaCloudProjectName,
+      LLAMA_CLOUD_INDEX_NAME: llamaCloudIndexName,
     },
   });
   appProcess.stderr?.on("data", (data) => {
-    console.log(data.toString());
+    console.error(data.toString());
   });
   appProcess.on("exit", (code) => {
     if (code !== 0 && code !== null) {
-      throw new Error(`create-llama command was failed!`);
+      throw new Error(`create-llama command failed with exit code ${code}`);
     }
   });
 
   // Wait for app to start
   if (postInstallAction === "runApp") {
-    await checkAppHasStarted(
-      appType === "--frontend",
-      templateFramework,
-      port,
-      externalPort,
-      1000 * 60 * 5,
-    );
+    await waitPorts([port]);
+  } else if (postInstallAction === "dependencies") {
+    await waitForProcess(appProcess, 1000 * 60); // wait 1 min for dependencies to be resolved
   } else {
-    // wait create-llama to exit
-    // we don't test install dependencies for now, so just set timeout for 10 seconds
-    await new Promise((resolve, reject) => {
-      const timeout = setTimeout(() => {
-        reject(new Error("create-llama timeout error"));
-      }, 1000 * 10);
-      appProcess.on("exit", (code) => {
-        if (code !== 0 && code !== null) {
-          clearTimeout(timeout);
-          reject(new Error("create-llama command was failed!"));
-        } else {
-          clearTimeout(timeout);
-          resolve(undefined);
-        }
-      });
-    });
+    // wait 10 seconds for create-llama to exit
+    await waitForProcess(appProcess, 1000 * 10);
   }
 
   return {
@@ -166,3 +162,40 @@ export async function createTestDir() {
   await mkdir(cwd, { recursive: true });
   return cwd;
 }
+
+async function waitPorts(ports: number[]): Promise<void> {
+  const waitForPort = async (port: number): Promise<void> => {
+    await waitPort({
+      host: "localhost",
+      port: port,
+      // wait max. 5 mins for start up of app
+      timeout: 1000 * 60 * 5,
+    });
+  };
+  try {
+    await Promise.all(ports.map(waitForPort));
+  } catch (err) {
+    console.error(err);
+    throw err;
+  }
+}
+
+async function waitForProcess(
+  process: ChildProcess,
+  timeoutMs: number,
+): Promise<void> {
+  return new Promise((resolve, reject) => {
+    const timeout = setTimeout(() => {
+      reject(new Error("Process timeout error"));
+    }, timeoutMs);
+
+    process.on("exit", (code) => {
+      clearTimeout(timeout);
+      if (code !== 0 && code !== null) {
+        reject(new Error("Process exited with non-zero code"));
+      } else {
+        resolve();
+      }
+    });
+  });
+}

helpers/copy.ts

@@ -61,6 +61,9 @@ export const assetRelocator = (name: string) => {
     case "README-template.md": {
       return "README.md";
     }
+    case "vscode_settings.json": {
+      return "settings.json";
+    }
     default: {
       return name;
     }

helpers/datasources.ts

@@ -11,6 +11,47 @@ export const EXAMPLE_FILE: TemplateDataSource = {
   },
 };
 
+export const EXAMPLE_10K_SEC_FILES: TemplateDataSource[] = [
+  {
+    type: "file",
+    config: {
+      url: new URL(
+        "https://s2.q4cdn.com/470004039/files/doc_earnings/2023/q4/filing/_10-K-Q4-2023-As-Filed.pdf",
+      ),
+      filename: "apple_10k_report.pdf",
+    },
+  },
+  {
+    type: "file",
+    config: {
+      url: new URL(
+        "https://ir.tesla.com/_flysystem/s3/sec/000162828024002390/tsla-20231231-gen.pdf",
+      ),
+      filename: "tesla_10k_report.pdf",
+    },
+  },
+];
+
+export const EXAMPLE_GDPR: TemplateDataSource = {
+  type: "file",
+  config: {
+    url: new URL(
+      "https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32016R0679",
+    ),
+    filename: "gdpr.pdf",
+  },
+};
+
+export const AI_REPORTS: TemplateDataSource = {
+  type: "file",
+  config: {
+    url: new URL(
+      "https://www.europarl.europa.eu/RegData/etudes/ATAG/2024/760392/EPRS_ATA(2024)760392_EN.pdf",
+    ),
+    filename: "EPRS_ATA_2024_760392_EN.pdf",
+  },
+};
+
 export function getDataSources(
   files?: string,
   exampleFile?: boolean,
@@ -36,74 +77,66 @@ export async function writeLoadersConfig(
   dataSources: TemplateDataSource[],
   useLlamaParse?: boolean,
 ) {
-  if (dataSources.length === 0) return; // no datasources, no config needed
-  const loaderConfig = new Document({});
-  // Web loader config
+  const loaderConfig: Record<string, any> = {};
+
+  // Always set file loader config
+  loaderConfig.file = createFileLoaderConfig(useLlamaParse);
+
   if (dataSources.some((ds) => ds.type === "web")) {
-    const webLoaderConfig = new Document({});
-
-    // Create config for browser driver arguments
-    const driverArgNodeValue = webLoaderConfig.createNode([
-      "--no-sandbox",
-      "--disable-dev-shm-usage",
-    ]);
-    driverArgNodeValue.commentBefore =
-      " The arguments to pass to the webdriver. E.g.: add --headless to run in headless mode";
-    webLoaderConfig.set("driver_arguments", driverArgNodeValue);
-
-    // Create config for urls
-    const urlConfigs = dataSources
-      .filter((ds) => ds.type === "web")
-      .map((ds) => {
-        const dsConfig = ds.config as WebSourceConfig;
-        return {
-          base_url: dsConfig.baseUrl,
-          prefix: dsConfig.prefix,
-          depth: dsConfig.depth,
-        };
-      });
-    const urlConfigNode = webLoaderConfig.createNode(urlConfigs);
-    urlConfigNode.commentBefore = ` base_url: The URL to start crawling with
- prefix: Only crawl URLs matching the specified prefix
- depth: The maximum depth for BFS traversal
- You can add more websites by adding more entries (don't forget the - prefix from YAML)`;
-    webLoaderConfig.set("urls", urlConfigNode);
-
-    // Add web config to the loaders config
-    loaderConfig.set("web", webLoaderConfig);
+    loaderConfig.web = createWebLoaderConfig(dataSources);
   }
 
-  // File loader config
-  if (dataSources.some((ds) => ds.type === "file")) {
-    // Add documentation to web loader config
-    const node = loaderConfig.createNode({
-      use_llama_parse: useLlamaParse,
-    });
-    node.commentBefore = ` use_llama_parse: Use LlamaParse if \`true\`. Needs a \`LLAMA_CLOUD_API_KEY\` from https://cloud.llamaindex.ai set as environment variable`;
-    loaderConfig.set("file", node);
-  }
-
-  // DB loader config
   const dbLoaders = dataSources.filter((ds) => ds.type === "db");
   if (dbLoaders.length > 0) {
-    const dbLoaderConfig = new Document({});
-    const configEntries = dbLoaders.map((ds) => {
-      const dsConfig = ds.config as DbSourceConfig;
-      return {
-        uri: dsConfig.uri,
-        queries: [dsConfig.queries],
-      };
-    });
-
-    const node = dbLoaderConfig.createNode(configEntries);
-    node.commentBefore = ` The configuration for the database loader, only supports MySQL and PostgreSQL databases for now.
- uri: The URI for the database. E.g.: mysql+pymysql://user:password@localhost:3306/db or postgresql+psycopg2://user:password@localhost:5432/db
- query: The query to fetch data from the database. E.g.: SELECT * FROM table`;
-    loaderConfig.set("db", node);
+    loaderConfig.db = createDbLoaderConfig(dbLoaders);
   }
 
+  // Create a new Document with the loaderConfig
+  const yamlDoc = new Document(loaderConfig);
+
   // Write loaders config
   const loaderConfigPath = path.join(root, "config", "loaders.yaml");
   await fs.mkdir(path.join(root, "config"), { recursive: true });
-  await fs.writeFile(loaderConfigPath, yaml.stringify(loaderConfig));
+  await fs.writeFile(loaderConfigPath, yaml.stringify(yamlDoc));
+}
+
+function createWebLoaderConfig(dataSources: TemplateDataSource[]): any {
+  const webLoaderConfig: Record<string, any> = {};
+
+  // Create config for browser driver arguments
+  webLoaderConfig.driver_arguments = [
+    "--no-sandbox",
+    "--disable-dev-shm-usage",
+  ];
+
+  // Create config for urls
+  const urlConfigs = dataSources
+    .filter((ds) => ds.type === "web")
+    .map((ds) => {
+      const dsConfig = ds.config as WebSourceConfig;
+      return {
+        base_url: dsConfig.baseUrl,
+        prefix: dsConfig.prefix,
+        depth: dsConfig.depth,
+      };
+    });
+  webLoaderConfig.urls = urlConfigs;
+
+  return webLoaderConfig;
+}
+
+function createFileLoaderConfig(useLlamaParse?: boolean): any {
+  return {
+    use_llama_parse: useLlamaParse,
+  };
+}
+
+function createDbLoaderConfig(dbLoaders: TemplateDataSource[]): any {
+  return dbLoaders.map((ds) => {
+    const dsConfig = ds.config as DbSourceConfig;
+    return {
+      uri: dsConfig.uri,
+      queries: [dsConfig.queries],
+    };
+  });
 }

helpers/env-variables.ts

@@ -2,12 +2,23 @@ import fs from "fs/promises";
 import path from "path";
 import { TOOL_SYSTEM_PROMPT_ENV_VAR, Tool } from "./tools";
 import {
+  InstallTemplateArgs,
   ModelConfig,
   TemplateDataSource,
   TemplateFramework,
+  TemplateObservability,
+  TemplateType,
   TemplateVectorDB,
 } from "./types";
 
+import { TSYSTEMS_LLMHUB_API_URL } from "./providers/llmhub";
+
+const DEFAULT_SYSTEM_PROMPT =
+  "You are a helpful assistant who helps users with their questions.";
+
+const DATA_SOURCES_PROMPT =
+  "You have access to a knowledge base including the facts that you should start with to find the answer for the user question. Use the query engine tool to retrieve the facts from the knowledge base.";
+
 export type EnvVar = {
   name?: string;
   description?: string;
@@ -33,6 +44,7 @@ const renderEnvVar = (envVars: EnvVar[]): string => {
 const getVectorDBEnvs = (
   vectorDb?: TemplateVectorDB,
   framework?: TemplateFramework,
+  template?: TemplateType,
 ): EnvVar[] => {
   if (!vectorDb || !framework) {
     return [];
@@ -60,7 +72,7 @@ const getVectorDBEnvs = (
         {
           name: "PG_CONNECTION_STRING",
           description:
-            "For generating a connection URI, see https://docs.timescale.com/use-timescale/latest/services/create-a-service\nThe PostgreSQL connection string.",
+            "For generating a connection URI, see https://supabase.com/vector\nThe PostgreSQL connection string.",
         },
       ];
 
@@ -133,6 +145,42 @@ const getVectorDBEnvs = (
             "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",
+        },
+        {
+          name: "LLAMA_CLOUD_ORGANIZATION_ID",
+          description:
+            "The organization ID for the LlamaCloud project (uses default organization if not specified)",
+        },
+        ...(framework === "nextjs" && template !== "llamaindexserver"
+          ? // activate index selector per default (not needed for non-NextJS backends as it's handled by createFrontendEnvFile)
+            [
+              {
+                name: "NEXT_PUBLIC_USE_LLAMACLOUD",
+                description:
+                  "Let's the user change indexes in LlamaCloud projects",
+                value: "true",
+              },
+            ]
+          : []),
+      ];
     case "chroma":
       const envs = [
         {
@@ -141,11 +189,11 @@ const getVectorDBEnvs = (
         },
         {
           name: "CHROMA_HOST",
-          description: "The API endpoint for your Chroma database",
+          description: "The hostname for your Chroma database. Eg: localhost",
         },
         {
           name: "CHROMA_PORT",
-          description: "The port for your Chroma database",
+          description: "The port for your Chroma database. Eg: 8000",
         },
       ];
       // TS Version doesn't support config local storage path
@@ -158,8 +206,33 @@ Otherwise, use CHROMA_HOST and CHROMA_PORT config above`,
         });
       }
       return envs;
+    case "weaviate":
+      return [
+        {
+          name: "WEAVIATE_CLUSTER_URL",
+          description:
+            "The URL of the Weaviate cloud cluster, see: https://weaviate.io/developers/wcs/connect",
+        },
+        {
+          name: "WEAVIATE_API_KEY",
+          description: "The API key for the Weaviate cloud cluster",
+        },
+        {
+          name: "WEAVIATE_INDEX_NAME",
+          description:
+            "(Optional) The collection name to use, default is LlamaIndex if not specified",
+        },
+      ];
     default:
-      return [];
+      return template !== "llamaindexserver"
+        ? [
+            {
+              name: "STORAGE_CACHE_DIR",
+              description: "The directory to store the local storage cache.",
+              value: ".cache",
+            },
+          ]
+        : [];
   }
 };
 
@@ -185,6 +258,10 @@ 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"
       ? [
           {
@@ -211,6 +288,15 @@ 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"
       ? [
           {
@@ -229,42 +315,118 @@ const getModelEnvs = (modelConfig: ModelConfig): EnvVar[] => {
           },
         ]
       : []),
+    ...(modelConfig.provider === "mistral"
+      ? [
+          {
+            name: "MISTRAL_API_KEY",
+            description: "The Mistral API key to use.",
+            value: modelConfig.apiKey,
+          },
+        ]
+      : []),
+    ...(modelConfig.provider === "azure-openai"
+      ? [
+          {
+            name: "AZURE_OPENAI_API_KEY",
+            description: "The Azure OpenAI key to use.",
+            value: modelConfig.apiKey,
+          },
+          {
+            name: "AZURE_OPENAI_ENDPOINT",
+            description: "The Azure OpenAI endpoint to use.",
+          },
+          {
+            name: "AZURE_OPENAI_API_VERSION",
+            description: "The Azure OpenAI API version to use.",
+          },
+          {
+            name: "AZURE_OPENAI_LLM_DEPLOYMENT",
+            description:
+              "The Azure OpenAI deployment to use for LLM deployment.",
+          },
+          {
+            name: "AZURE_OPENAI_EMBEDDING_DEPLOYMENT",
+            description:
+              "The Azure OpenAI deployment to use for embedding deployment.",
+          },
+        ]
+      : []),
+    ...(modelConfig.provider === "huggingface"
+      ? [
+          {
+            name: "EMBEDDING_BACKEND",
+            description:
+              "The backend to use for the Sentence Transformers embedding model, either 'torch', 'onnx', or 'openvino'. Defaults to 'onnx'.",
+          },
+          {
+            name: "EMBEDDING_TRUST_REMOTE_CODE",
+            description:
+              "Whether to trust remote code for the embedding model, required for some models with custom code.",
+          },
+        ]
+      : []),
+    ...(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,
+  template: TemplateType,
   port?: number,
 ): EnvVar[] => {
   const sPort = port?.toString() || "8000";
-  const result: EnvVar[] = [
-    {
-      name: "FILESERVER_URL_PREFIX",
-      description:
-        "FILESERVER_URL_PREFIX is the URL prefix of the server storing the images generated by the interpreter.",
-      value:
-        framework === "nextjs"
-          ? // FIXME: if we are using nextjs, port should be 3000
-            "http://localhost:3000/api/files"
-          : `http://localhost:${sPort}/api/files`,
-    },
-  ];
+  const result: EnvVar[] =
+    template !== "llamaindexserver"
+      ? [
+          {
+            name: "FILESERVER_URL_PREFIX",
+            description:
+              "FILESERVER_URL_PREFIX is the URL prefix of the server storing the images generated by the interpreter.",
+            value:
+              framework === "nextjs"
+                ? // FIXME: if we are using nextjs, port should be 3000
+                  "http://localhost:3000/api/files"
+                : `http://localhost:${sPort}/api/files`,
+          },
+        ]
+      : [];
   if (framework === "fastapi") {
     result.push(
       ...[
         {
           name: "APP_HOST",
-          description: "The address to start the backend app.",
+          description: "The address to start the FastAPI app.",
           value: "0.0.0.0",
         },
         {
           name: "APP_PORT",
-          description: "The port to start the backend app.",
+          description: "The port to start the FastAPI app.",
           value: sPort,
         },
       ],
     );
   }
+  if (framework === "nextjs" && template !== "llamaindexserver") {
+    result.push({
+      name: "NEXT_PUBLIC_CHAT_API",
+      description:
+        "The API for the chat endpoint. Set when using a custom backend (e.g. Express). Use full URL like http://localhost:8000/api/chat",
+    });
+  }
   return result;
 };
 
@@ -274,7 +436,6 @@ const getEngineEnvs = (): EnvVar[] => {
       name: "TOP_K",
       description:
         "The number of similar embeddings to return when retrieving documents.",
-      value: "3",
     },
   ];
 };
@@ -296,61 +457,160 @@ const getToolEnvs = (tools?: Tool[]): EnvVar[] => {
   return toolEnvs;
 };
 
-const getSystemPromptEnv = (tools?: Tool[]): EnvVar => {
-  const defaultSystemPrompt =
-    "You are a helpful assistant who helps users with their questions.";
-
+const getSystemPromptEnv = (
+  tools?: Tool[],
+  dataSources?: TemplateDataSource[],
+  template?: TemplateType,
+): EnvVar[] => {
+  const systemPromptEnv: EnvVar[] = [];
   // 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";
-    }
-  });
+  // multiagent template doesn't need system prompt
+  if (template !== "multiagent") {
+    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
-    ? `\"${toolSystemPrompt}\"`
-    : defaultSystemPrompt;
+    const systemPrompt =
+      '"' +
+      DEFAULT_SYSTEM_PROMPT +
+      (dataSources?.length ? `\n${DATA_SOURCES_PROMPT}` : "") +
+      (toolSystemPrompt ? `\n${toolSystemPrompt}` : "") +
+      '"';
 
-  return {
-    name: "SYSTEM_PROMPT",
-    description: "The system prompt for the AI model.",
-    value: systemPrompt,
-  };
+    systemPromptEnv.push({
+      name: "SYSTEM_PROMPT",
+      description: "The system prompt for the AI model.",
+      value: systemPrompt,
+    });
+  }
+  if (tools?.length == 0 && (dataSources?.length ?? 0 > 0)) {
+    const citationPrompt = `'You have provided information from a knowledge base that has been passed to you in nodes of information.
+Each node has useful metadata such as node ID, file name, page, etc.
+Please add the citation to the data node for each sentence or paragraph that you reference in the provided information.
+The citation format is: . [citation:<node_id>]()
+Where the <node_id> is the unique identifier of the data node.
+
+Example:
+We have two nodes:
+  node_id: xyz
+  file_name: llama.pdf
+  
+  node_id: abc
+  file_name: animal.pdf
+
+User question: Tell me a fun fact about Llama.
+Your answer:
+A baby llama is called "Cria" [citation:xyz]().
+It often live in desert [citation:abc]().
+It\\'s cute animal.
+'`;
+    systemPromptEnv.push({
+      name: "SYSTEM_CITATION_PROMPT",
+      description:
+        "An additional system prompt to add citation when responding to user questions.",
+      value: citationPrompt,
+    });
+  }
+
+  return systemPromptEnv;
+};
+
+const getTemplateEnvs = (template?: TemplateType): EnvVar[] => {
+  const nextQuestionEnvs: EnvVar[] = [
+    {
+      name: "NEXT_QUESTION_PROMPT",
+      description: `Customize prompt to generate the next question suggestions based on the conversation history.
+Disable this prompt to disable the next question suggestions feature.`,
+      value: `"You're a helpful assistant! Your task is to suggest the next question that user might ask. 
+Here is the conversation history
+---------------------
+{conversation}
+---------------------
+Given the conversation history, please give me 3 questions that user might ask next!
+Your answer should be wrapped in three sticks which follows the following format:
+\`\`\`
+<question 1>
+<question 2>
+<question 3>
+\`\`\`"`,
+    },
+  ];
+
+  if (template === "multiagent" || template === "streaming") {
+    return nextQuestionEnvs;
+  }
+  return [];
+};
+
+const getObservabilityEnvs = (
+  observability?: TemplateObservability,
+): EnvVar[] => {
+  if (observability === "llamatrace") {
+    return [
+      {
+        name: "PHOENIX_API_KEY",
+        description:
+          "API key for LlamaTrace observability. Retrieve from https://llamatrace.com/login",
+      },
+    ];
+  }
+  return [];
 };
 
 export const createBackendEnvFile = async (
   root: string,
-  opts: {
-    llamaCloudKey?: string;
-    vectorDb?: TemplateVectorDB;
-    modelConfig: ModelConfig;
-    framework: TemplateFramework;
-    dataSources?: TemplateDataSource[];
-    port?: number;
-    tools?: Tool[];
-  },
+  opts: Pick<
+    InstallTemplateArgs,
+    | "llamaCloudKey"
+    | "vectorDb"
+    | "modelConfig"
+    | "framework"
+    | "dataSources"
+    | "template"
+    | "port"
+    | "tools"
+    | "observability"
+    | "useLlamaParse"
+  >,
 ) => {
   // Init env values
   const envFileName = ".env";
   const envVars: EnvVar[] = [
-    {
-      name: "LLAMA_CLOUD_API_KEY",
-      description: `The Llama Cloud API key.`,
-      value: opts.llamaCloudKey,
-    },
-    // Add model environment variables
-    ...getModelEnvs(opts.modelConfig),
-    // Add engine environment variables
-    ...getEngineEnvs(),
-    // Add vector database environment variables
-    ...getVectorDBEnvs(opts.vectorDb, opts.framework),
-    ...getFrameworkEnvs(opts.framework, opts.port),
+    ...(opts.useLlamaParse
+      ? [
+          {
+            name: "LLAMA_CLOUD_API_KEY",
+            description: `The Llama Cloud API key.`,
+            value: opts.llamaCloudKey,
+          },
+        ]
+      : []),
+    ...getVectorDBEnvs(opts.vectorDb, opts.framework, opts.template),
     ...getToolEnvs(opts.tools),
-    getSystemPromptEnv(opts.tools),
+    ...getFrameworkEnvs(opts.framework, opts.template, opts.port),
+    // Add environment variables of each component
+    ...(opts.template === "llamaindexserver"
+      ? [
+          {
+            name: "OPENAI_API_KEY",
+            description: "The OpenAI API key to use.",
+            value: opts.modelConfig.apiKey,
+          },
+        ]
+      : [
+          // don't use this stuff for llama-indexserver
+          ...getModelEnvs(opts.modelConfig),
+          ...getEngineEnvs(),
+          ...getTemplateEnvs(opts.template),
+          ...getObservabilityEnvs(opts.observability),
+          ...getSystemPromptEnv(opts.tools, opts.dataSources, opts.template),
+        ]),
   ];
   // Render and write env file
   const content = renderEnvVar(envVars);
@@ -361,16 +621,14 @@ export const createBackendEnvFile = async (
 export const createFrontendEnvFile = async (
   root: string,
   opts: {
-    customApiPath?: string;
+    vectorDb?: TemplateVectorDB;
   },
 ) => {
   const defaultFrontendEnvs = [
     {
-      name: "NEXT_PUBLIC_CHAT_API",
-      description: "The backend API for chat endpoint.",
-      value: opts.customApiPath
-        ? opts.customApiPath
-        : "http://localhost:8000/api/chat",
+      name: "NEXT_PUBLIC_USE_LLAMACLOUD",
+      description: "Let's the user change indexes in LlamaCloud projects",
+      value: opts.vectorDb === "llamacloud" ? "true" : "false",
     },
   ];
   const content = renderEnvVar(defaultFrontendEnvs);

helpers/index.ts

@@ -1,13 +1,14 @@
 import { callPackageManager } from "./install";
 
 import path from "path";
-import { cyan } from "picocolors";
+import picocolors, { cyan } from "picocolors";
 
 import fsExtra from "fs-extra";
 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";
@@ -22,6 +23,35 @@ import {
 } from "./types";
 import { installTSTemplate } from "./typescript";
 
+const checkForGenerateScript = (
+  modelConfig: ModelConfig,
+  vectorDb?: TemplateVectorDB,
+  llamaCloudKey?: string,
+  useLlamaParse?: boolean,
+) => {
+  const missingSettings = [];
+
+  if (!modelConfig.isConfigured()) {
+    missingSettings.push("your model provider API key");
+  }
+
+  const llamaCloudApiKey = llamaCloudKey ?? process.env["LLAMA_CLOUD_API_KEY"];
+  const isRequiredLlamaCloudKey = useLlamaParse || vectorDb === "llamacloud";
+  if (isRequiredLlamaCloudKey && !llamaCloudApiKey) {
+    missingSettings.push("your LLAMA_CLOUD_API_KEY");
+  }
+
+  if (
+    vectorDb !== undefined &&
+    vectorDb !== "none" &&
+    vectorDb !== "llamacloud"
+  ) {
+    missingSettings.push("your Vector DB environment variables");
+  }
+
+  return missingSettings;
+};
+
 // eslint-disable-next-line max-params
 async function generateContextData(
   framework: TemplateFramework,
@@ -37,12 +67,15 @@ async function generateContextData(
         ? "poetry run generate"
         : `${packageManager} run generate`,
     )}`;
-    const modelConfigured = modelConfig.isConfigured();
-    const llamaCloudKeyConfigured = useLlamaParse
-      ? llamaCloudKey || process.env["LLAMA_CLOUD_API_KEY"]
-      : true;
-    const hasVectorDb = vectorDb && vectorDb !== "none";
-    if (modelConfigured && llamaCloudKeyConfigured && !hasVectorDb) {
+
+    const missingSettings = checkForGenerateScript(
+      modelConfig,
+      vectorDb,
+      llamaCloudKey,
+      useLlamaParse,
+    );
+
+    if (!missingSettings.length) {
       // If all the required environment variables are set, run the generate script
       if (framework === "fastapi") {
         if (isHavingPoetryLockFile()) {
@@ -62,30 +95,47 @@ async function generateContextData(
       }
     }
 
-    // generate the message of what to do to run the generate script manually
-    const settings = [];
-    if (!modelConfigured) settings.push("your model provider API key");
-    if (!llamaCloudKeyConfigured) settings.push("your Llama Cloud key");
-    if (hasVectorDb) settings.push("your Vector DB environment variables");
-    const settingsMessage =
-      settings.length > 0 ? `After setting ${settings.join(" and ")}, ` : "";
-    const generateMessage = `run ${runGenerate} to generate the context data.`;
-    console.log(`\n${settingsMessage}${generateMessage}\n\n`);
+    const settingsMessage = `After setting ${missingSettings.join(" and ")}, run ${runGenerate} to generate the context data.`;
+    console.log(picocolors.yellow(`\n${settingsMessage}\n\n`));
   }
 }
 
-const copyContextData = async (
+const downloadFile = async (url: string, destPath: string) => {
+  const response = await fetch(url);
+  const fileBuffer = await response.arrayBuffer();
+  await fsExtra.writeFile(destPath, Buffer.from(fileBuffer));
+};
+
+const prepareContextData = async (
   root: string,
   dataSources: TemplateDataSource[],
 ) => {
+  await makeDir(path.join(root, "data"));
   for (const dataSource of dataSources) {
     const dataSourceConfig = dataSource?.config as FileSourceConfig;
-    // Copy local data
-    const dataPath = dataSourceConfig.path;
-
-    const destPath = path.join(root, "data", path.basename(dataPath));
-    console.log("Copying data from path:", dataPath);
-    await fsExtra.copy(dataPath, destPath);
+    // If the path is URLs, download the data and save it to the data directory
+    if ("url" in dataSourceConfig) {
+      console.log(
+        "Downloading file from URL:",
+        dataSourceConfig.url.toString(),
+      );
+      const destPath = path.join(
+        root,
+        "data",
+        dataSourceConfig.filename ??
+          path.basename(dataSourceConfig.url.toString()),
+      );
+      await downloadFile(dataSourceConfig.url.toString(), destPath);
+    } else {
+      // Copy local data
+      console.log("Copying data from path:", dataSourceConfig.path);
+      const destPath = path.join(
+        root,
+        "data",
+        path.basename(dataSourceConfig.path),
+      );
+      await fsExtra.copy(dataSourceConfig.path, destPath);
+    }
   }
 };
 
@@ -120,66 +170,65 @@ export const installTemplate = async (
 
   if (props.framework === "fastapi") {
     await installPythonTemplate(props);
-    // write loaders configuration (currently Python only)
-    await writeLoadersConfig(
-      props.root,
-      props.dataSources,
-      props.useLlamaParse,
-    );
   } else {
     await installTSTemplate(props);
   }
 
-  // write tools configuration
-  await writeToolsConfig(
-    props.root,
-    props.tools,
-    props.framework === "fastapi" ? ConfigFileType.YAML : ConfigFileType.JSON,
-  );
+  // write configurations
+  if (props.template !== "llamaindexserver") {
+    await writeToolsConfig(
+      props.root,
+      props.tools,
+      props.framework === "fastapi" ? ConfigFileType.YAML : ConfigFileType.JSON,
+    );
+    if (props.vectorDb !== "llamacloud") {
+      // write loaders configuration (currently Python only)
+      // not needed for LlamaCloud as it has its own loaders
+      await writeLoadersConfig(
+        props.root,
+        props.dataSources,
+        props.useLlamaParse,
+      );
+    }
+  }
 
   if (props.backend) {
     // This is a backend, so we need to copy the test data and create the env file.
 
     // Copy the environment file to the target directory.
-    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,
-    });
+    if (props.template !== "community" && props.template !== "llamapack") {
+      await createBackendEnvFile(props.root, props);
+    }
 
-    if (props.dataSources.length > 0) {
+    await prepareContextData(
+      props.root,
+      props.dataSources.filter((ds) => ds.type === "file"),
+    );
+
+    if (
+      props.dataSources.length > 0 &&
+      (props.postInstallAction === "runApp" ||
+        props.postInstallAction === "dependencies")
+    ) {
       console.log("\nGenerating context data...\n");
-      await copyContextData(
-        props.root,
-        props.dataSources.filter((ds) => ds.type === "file"),
+      await generateContextData(
+        props.framework,
+        props.modelConfig,
+        props.packageManager,
+        props.vectorDb,
+        props.llamaCloudKey,
+        props.useLlamaParse,
       );
-      if (
-        props.postInstallAction === "runApp" ||
-        props.postInstallAction === "dependencies"
-      ) {
-        await generateContextData(
-          props.framework,
-          props.modelConfig,
-          props.packageManager,
-          props.vectorDb,
-          props.llamaCloudKey,
-          props.useLlamaParse,
-        );
-      }
     }
 
-    // Create tool-output directory
-    if (props.tools && props.tools.length > 0) {
-      await fsExtra.mkdir(path.join(props.root, "tool-output"));
-    }
+    // Create outputs directory
+    await makeDir(path.join(props.root, "output/tools"));
+    await makeDir(path.join(props.root, "output/uploaded"));
+    await makeDir(path.join(props.root, "output/llamacloud"));
   } else {
     // this is a frontend for a full-stack app, create .env file with model information
     await createFrontendEnvFile(props.root, {
-      customApiPath: props.customApiPath,
+      vectorDb: props.vectorDb,
     });
   }
 };

helpers/providers/anthropic.ts

@@ -1,7 +1,6 @@
-import ciInfo from "ci-info";
 import prompts from "prompts";
 import { ModelConfigParams } from ".";
-import { questionHandlers, toChoice } from "../../questions";
+import { questionHandlers, toChoice } from "../../questions/utils";
 
 const MODELS = [
   "claude-3-opus",
@@ -70,9 +69,7 @@ export async function askAnthropicQuestions({
     config.apiKey = key || process.env.ANTHROPIC_API_KEY;
   }
 
-  // use default model values in CI or if user should not be asked
-  const useDefaults = ciInfo.isCI || !askModels;
-  if (!useDefaults) {
+  if (askModels) {
     const { model } = await prompts(
       {
         type: "select",

helpers/providers/azure.ts

@@ -0,0 +1,115 @@
+import prompts from "prompts";
+import { ModelConfigParams, ModelConfigQuestionsParams } from ".";
+import { questionHandlers } from "../../questions/utils";
+
+const ALL_AZURE_OPENAI_CHAT_MODELS: Record<string, { openAIModel: string }> = {
+  "gpt-35-turbo": { openAIModel: "gpt-3.5-turbo" },
+  "gpt-35-turbo-16k": {
+    openAIModel: "gpt-3.5-turbo-16k",
+  },
+  "gpt-4o": { openAIModel: "gpt-4o" },
+  "gpt-4o-mini": { openAIModel: "gpt-4o-mini" },
+  "gpt-4": { openAIModel: "gpt-4" },
+  "gpt-4-32k": { openAIModel: "gpt-4-32k" },
+  "gpt-4-turbo": {
+    openAIModel: "gpt-4-turbo",
+  },
+  "gpt-4-turbo-2024-04-09": {
+    openAIModel: "gpt-4-turbo",
+  },
+  "gpt-4-vision-preview": {
+    openAIModel: "gpt-4-vision-preview",
+  },
+  "gpt-4-1106-preview": {
+    openAIModel: "gpt-4-1106-preview",
+  },
+  "gpt-4o-2024-05-13": {
+    openAIModel: "gpt-4o-2024-05-13",
+  },
+  "gpt-4o-mini-2024-07-18": {
+    openAIModel: "gpt-4o-mini-2024-07-18",
+  },
+};
+
+const ALL_AZURE_OPENAI_EMBEDDING_MODELS: Record<
+  string,
+  {
+    dimensions: number;
+    openAIModel: string;
+  }
+> = {
+  "text-embedding-3-small": {
+    dimensions: 1536,
+    openAIModel: "text-embedding-3-small",
+  },
+  "text-embedding-3-large": {
+    dimensions: 3072,
+    openAIModel: "text-embedding-3-large",
+  },
+};
+
+const DEFAULT_MODEL = "gpt-4o";
+const DEFAULT_EMBEDDING_MODEL = "text-embedding-3-large";
+
+export async function askAzureQuestions({
+  openAiKey,
+  askModels,
+}: ModelConfigQuestionsParams): Promise<ModelConfigParams> {
+  const config: ModelConfigParams = {
+    apiKey: openAiKey || process.env.AZURE_OPENAI_KEY,
+    model: DEFAULT_MODEL,
+    embeddingModel: DEFAULT_EMBEDDING_MODEL,
+    dimensions: getDimensions(DEFAULT_EMBEDDING_MODEL),
+    isConfigured(): boolean {
+      // the Azure model provider can't be fully configured as endpoint and deployment names have to be configured with env variables
+      return false;
+    },
+  };
+
+  if (askModels) {
+    const { model } = await prompts(
+      {
+        type: "select",
+        name: "model",
+        message: "Which LLM model would you like to use?",
+        choices: getAvailableModelChoices(),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    config.model = model;
+
+    const { embeddingModel } = await prompts(
+      {
+        type: "select",
+        name: "embeddingModel",
+        message: "Which embedding model would you like to use?",
+        choices: getAvailableEmbeddingModelChoices(),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    config.embeddingModel = embeddingModel;
+    config.dimensions = getDimensions(embeddingModel);
+  }
+
+  return config;
+}
+
+function getAvailableModelChoices() {
+  return Object.keys(ALL_AZURE_OPENAI_CHAT_MODELS).map((key) => ({
+    title: key,
+    value: key,
+  }));
+}
+
+function getAvailableEmbeddingModelChoices() {
+  return Object.keys(ALL_AZURE_OPENAI_EMBEDDING_MODELS).map((key) => ({
+    title: key,
+    value: key,
+  }));
+}
+
+function getDimensions(modelName: string) {
+  return ALL_AZURE_OPENAI_EMBEDDING_MODELS[modelName].dimensions;
+}

helpers/providers/gemini.ts

@@ -1,7 +1,6 @@
-import ciInfo from "ci-info";
 import prompts from "prompts";
 import { ModelConfigParams } from ".";
-import { questionHandlers, toChoice } from "../../questions";
+import { questionHandlers, toChoice } from "../../questions/utils";
 
 const MODELS = ["gemini-1.5-pro-latest", "gemini-pro", "gemini-pro-vision"];
 type ModelData = {
@@ -54,9 +53,7 @@ export async function askGeminiQuestions({
     config.apiKey = key || process.env.GOOGLE_API_KEY;
   }
 
-  // use default model values in CI or if user should not be asked
-  const useDefaults = ciInfo.isCI || !askModels;
-  if (!useDefaults) {
+  if (askModels) {
     const { model } = await prompts(
       {
         type: "select",

helpers/providers/groq.ts

@@ -0,0 +1,145 @@
+import prompts from "prompts";
+import { ModelConfigParams } from ".";
+import { questionHandlers, toChoice } from "../../questions/utils";
+
+import got from "got";
+import ora from "ora";
+import { red } from "picocolors";
+
+const GROQ_API_URL = "https://api.groq.com/openai/v1";
+
+async function getAvailableModelChoicesGroq(apiKey: string) {
+  if (!apiKey) {
+    throw new Error("Need Groq API key to retrieve model choices");
+  }
+
+  const spinner = ora("Fetching available models from Groq").start();
+  try {
+    const response = await got(`${GROQ_API_URL}/models`, {
+      headers: {
+        Authorization: `Bearer ${apiKey}`,
+      },
+      timeout: 5000,
+      responseType: "json",
+    });
+    const data: any = await response.body;
+    spinner.stop();
+
+    // Filter out the Whisper models
+    return data.data
+      .filter((model: any) => !model.id.toLowerCase().includes("whisper"))
+      .map((el: any) => {
+        return {
+          title: el.id,
+          value: el.id,
+        };
+      });
+  } catch (error: unknown) {
+    spinner.stop();
+    console.log(error);
+    if ((error as any).response?.statusCode === 401) {
+      console.log(
+        red(
+          "Invalid Groq API key provided! Please provide a valid key and try again!",
+        ),
+      );
+    } else {
+      console.log(red("Request failed: " + error));
+    }
+    process.exit(1);
+  }
+}
+
+const DEFAULT_MODEL = "llama3-70b-8192";
+
+// 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;
+  }
+
+  if (askModels) {
+    const modelChoices = await getAvailableModelChoicesGroq(config.apiKey!);
+
+    const { model } = await prompts(
+      {
+        type: "select",
+        name: "model",
+        message: "Which LLM model would you like to use?",
+        choices: modelChoices,
+        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/huggingface.ts

@@ -0,0 +1,68 @@
+import prompts from "prompts";
+import { ModelConfigParams } from ".";
+import { questionHandlers, toChoice } from "../../questions/utils";
+
+const MODELS = ["HuggingFaceH4/zephyr-7b-alpha"];
+type ModelData = {
+  dimensions: number;
+};
+const EMBEDDING_MODELS: Record<string, ModelData> = {
+  "BAAI/bge-small-en-v1.5": { dimensions: 384 },
+  "BAAI/bge-base-en-v1.5": { dimensions: 768 },
+  "BAAI/bge-large-en-v1.5": { dimensions: 1024 },
+  "sentence-transformers/all-MiniLM-L6-v2": { dimensions: 384 },
+  "sentence-transformers/all-mpnet-base-v2": { dimensions: 768 },
+  "intfloat/multilingual-e5-large": { dimensions: 1024 },
+  "mixedbread-ai/mxbai-embed-large-v1": { dimensions: 1024 },
+  "nomic-ai/nomic-embed-text-v1.5": { dimensions: 768 },
+};
+
+const DEFAULT_MODEL = MODELS[0];
+const DEFAULT_EMBEDDING_MODEL = Object.keys(EMBEDDING_MODELS)[0];
+const DEFAULT_DIMENSIONS = Object.values(EMBEDDING_MODELS)[0].dimensions;
+
+type HuggingfaceQuestionsParams = {
+  askModels: boolean;
+};
+
+export async function askHuggingfaceQuestions({
+  askModels,
+}: HuggingfaceQuestionsParams): Promise<ModelConfigParams> {
+  const config: ModelConfigParams = {
+    model: DEFAULT_MODEL,
+    embeddingModel: DEFAULT_EMBEDDING_MODEL,
+    dimensions: DEFAULT_DIMENSIONS,
+    isConfigured(): boolean {
+      return true;
+    },
+  };
+
+  if (askModels) {
+    const { model } = await prompts(
+      {
+        type: "select",
+        name: "model",
+        message: "Which Hugging Face 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/providers/index.ts

@@ -1,9 +1,13 @@
-import ciInfo from "ci-info";
 import prompts from "prompts";
-import { questionHandlers } from "../../questions";
-import { ModelConfig, ModelProvider } from "../types";
+import { questionHandlers } from "../../questions/utils";
+import { ModelConfig, ModelProvider, TemplateFramework } from "../types";
 import { askAnthropicQuestions } from "./anthropic";
+import { askAzureQuestions } from "./azure";
 import { askGeminiQuestions } from "./gemini";
+import { askGroqQuestions } from "./groq";
+import { askHuggingfaceQuestions } from "./huggingface";
+import { askLLMHubQuestions } from "./llmhub";
+import { askMistralQuestions } from "./mistral";
 import { askOllamaQuestions } from "./ollama";
 import { askOpenAIQuestions } from "./openai";
 
@@ -12,6 +16,7 @@ const DEFAULT_MODEL_PROVIDER = "openai";
 export type ModelConfigQuestionsParams = {
   openAiKey?: string;
   askModels: boolean;
+  framework?: TemplateFramework;
 };
 
 export type ModelConfigParams = Omit<ModelConfig, "provider">;
@@ -19,23 +24,30 @@ 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) {
+  if (askModels) {
+    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" },
+      { title: "AzureOpenAI", value: "azure-openai" },
+    ];
+
+    if (framework === "fastapi") {
+      choices.push({ title: "T-Systems", value: "t-systems" });
+      choices.push({ title: "Huggingface", value: "huggingface" });
+    }
     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 +60,27 @@ 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 "azure-openai":
+      modelConfig = await askAzureQuestions({ askModels });
+      break;
+    case "t-systems":
+      modelConfig = await askLLMHubQuestions({ askModels });
+      break;
+    case "huggingface":
+      modelConfig = await askHuggingfaceQuestions({ askModels });
+      break;
     default:
       modelConfig = await askOpenAIQuestions({
         openAiKey,

helpers/providers/llmhub.ts

@@ -0,0 +1,166 @@
+import got from "got";
+import ora from "ora";
+import { red } from "picocolors";
+import prompts from "prompts";
+import { ModelConfigParams } from ".";
+import { questionHandlers } from "../../questions/utils";
+
+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;
+  }
+
+  if (askModels) {
+    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 +1,83 @@
+import prompts from "prompts";
+import { ModelConfigParams } from ".";
+import { questionHandlers, toChoice } from "../../questions/utils";
+
+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;
+  }
+
+  if (askModels) {
+    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/providers/ollama.ts

@@ -1,9 +1,8 @@
-import ciInfo from "ci-info";
 import ollama, { type ModelResponse } from "ollama";
 import { red } from "picocolors";
 import prompts from "prompts";
 import { ModelConfigParams } from ".";
-import { questionHandlers, toChoice } from "../../questions";
+import { questionHandlers, toChoice } from "../../questions/utils";
 
 type ModelData = {
   dimensions: number;
@@ -34,9 +33,7 @@ export async function askOllamaQuestions({
     },
   };
 
-  // use default model values in CI or if user should not be asked
-  const useDefaults = ciInfo.isCI || !askModels;
-  if (!useDefaults) {
+  if (askModels) {
     const { model } = await prompts(
       {
         type: "select",

helpers/providers/openai.ts

@@ -1,14 +1,14 @@
-import ciInfo from "ci-info";
 import got from "got";
 import ora from "ora";
 import { red } from "picocolors";
 import prompts from "prompts";
 import { ModelConfigParams, ModelConfigQuestionsParams } from ".";
-import { questionHandlers } from "../../questions";
+import { isCI } from "../../questions";
+import { questionHandlers } from "../../questions/utils";
 
 const OPENAI_API_URL = "https://api.openai.com/v1";
 
-const DEFAULT_MODEL = "gpt-3.5-turbo";
+const DEFAULT_MODEL = "gpt-4o-mini";
 const DEFAULT_EMBEDDING_MODEL = "text-embedding-3-large";
 
 export async function askOpenAIQuestions({
@@ -31,7 +31,7 @@ export async function askOpenAIQuestions({
     },
   };
 
-  if (!config.apiKey) {
+  if (!config.apiKey && !isCI) {
     const { key } = await prompts(
       {
         type: "text",
@@ -54,9 +54,7 @@ export async function askOpenAIQuestions({
     config.apiKey = key || process.env.OPENAI_API_KEY;
   }
 
-  // use default model values in CI or if user should not be asked
-  const useDefaults = ciInfo.isCI || !askModels;
-  if (!useDefaults) {
+  if (askModels) {
     const { model } = await prompts(
       {
         type: "select",

helpers/python.ts

@@ -12,6 +12,8 @@ import {
   InstallTemplateArgs,
   ModelConfig,
   TemplateDataSource,
+  TemplateObservability,
+  TemplateType,
   TemplateVectorDB,
 } from "./types";
 
@@ -19,6 +21,7 @@ interface Dependency {
   name: string;
   version?: string;
   extras?: string[];
+  constraints?: Record<string, string>;
 }
 
 const getAdditionalDependencies = (
@@ -26,6 +29,8 @@ const getAdditionalDependencies = (
   vectorDb?: TemplateVectorDB,
   dataSources?: TemplateDataSource[],
   tools?: Tool[],
+  templateType?: TemplateType,
+  observability?: TemplateObservability,
 ) => {
   const dependencies: Dependency[] = [];
 
@@ -34,56 +39,75 @@ const getAdditionalDependencies = (
     case "mongo": {
       dependencies.push({
         name: "llama-index-vector-stores-mongodb",
-        version: "^0.1.3",
+        version: "^0.6.0",
       });
       break;
     }
     case "pg": {
       dependencies.push({
         name: "llama-index-vector-stores-postgres",
-        version: "^0.1.1",
+        version: "^0.3.2",
       });
       break;
     }
     case "pinecone": {
       dependencies.push({
         name: "llama-index-vector-stores-pinecone",
-        version: "^0.1.3",
+        version: "^0.4.1",
+        constraints: {
+          python: ">=3.11,<3.13",
+        },
       });
       break;
     }
     case "milvus": {
       dependencies.push({
         name: "llama-index-vector-stores-milvus",
-        version: "^0.1.6",
+        version: "^0.3.0",
       });
       dependencies.push({
         name: "pymilvus",
-        version: "2.3.7",
+        version: "2.4.4",
       });
       break;
     }
     case "astra": {
       dependencies.push({
         name: "llama-index-vector-stores-astra-db",
-        version: "^0.1.5",
+        version: "^0.4.0",
       });
       break;
     }
     case "qdrant": {
       dependencies.push({
         name: "llama-index-vector-stores-qdrant",
-        version: "^0.2.8",
+        version: "^0.4.0",
+        constraints: {
+          python: ">=3.11,<3.13",
+        },
       });
       break;
     }
     case "chroma": {
       dependencies.push({
         name: "llama-index-vector-stores-chroma",
-        version: "^0.1.8",
+        version: "^0.4.0",
       });
       break;
     }
+    case "weaviate": {
+      dependencies.push({
+        name: "llama-index-vector-stores-weaviate",
+        version: "^1.2.3",
+      });
+      break;
+    }
+    case "llamacloud":
+      dependencies.push({
+        name: "llama-index-indices-managed-llama-cloud",
+        version: "0.6.3",
+      });
+      break;
   }
 
   // Add data source dependencies
@@ -100,13 +124,13 @@ const getAdditionalDependencies = (
         case "web":
           dependencies.push({
             name: "llama-index-readers-web",
-            version: "^0.1.6",
+            version: "^0.3.0",
           });
           break;
         case "db":
           dependencies.push({
             name: "llama-index-readers-database",
-            version: "^0.1.3",
+            version: "^0.3.0",
           });
           dependencies.push({
             name: "pymysql",
@@ -114,7 +138,7 @@ const getAdditionalDependencies = (
             extras: ["rsa"],
           });
           dependencies.push({
-            name: "psycopg2",
+            name: "psycopg2-binary",
             version: "^2.9.9",
           });
           break;
@@ -134,39 +158,131 @@ const getAdditionalDependencies = (
     case "ollama":
       dependencies.push({
         name: "llama-index-llms-ollama",
-        version: "0.1.2",
+        version: "0.3.0",
       });
       dependencies.push({
         name: "llama-index-embeddings-ollama",
-        version: "0.1.2",
+        version: "0.3.0",
       });
       break;
     case "openai":
+      if (templateType !== "multiagent") {
+        dependencies.push({
+          name: "llama-index-llms-openai",
+          version: "^0.3.2",
+        });
+        dependencies.push({
+          name: "llama-index-embeddings-openai",
+          version: "^0.3.1",
+        });
+        dependencies.push({
+          name: "llama-index-agent-openai",
+          version: "^0.4.0",
+        });
+      }
+      break;
+    case "groq":
+      // Fastembed==0.2.0 does not support python3.13 at the moment
+      // Fixed the python version less than 3.13
       dependencies.push({
-        name: "llama-index-agent-openai",
-        version: "0.2.2",
+        name: "python",
+        version: "^3.11,<3.13",
+      });
+      dependencies.push({
+        name: "llama-index-llms-groq",
+        version: "0.2.0",
+      });
+      dependencies.push({
+        name: "llama-index-embeddings-fastembed",
+        version: "^0.2.0",
       });
       break;
     case "anthropic":
+      // Fastembed==0.2.0 does not support python3.13 at the moment
+      // Fixed the python version less than 3.13
       dependencies.push({
-        name: "llama-index-llms-anthropic",
-        version: "0.1.10",
+        name: "python",
+        version: "^3.11,<3.13",
       });
       dependencies.push({
-        name: "llama-index-embeddings-huggingface",
-        version: "0.2.0",
+        name: "llama-index-llms-anthropic",
+        version: "0.3.0",
+      });
+      dependencies.push({
+        name: "llama-index-embeddings-fastembed",
+        version: "^0.2.0",
       });
       break;
     case "gemini":
       dependencies.push({
         name: "llama-index-llms-gemini",
-        version: "0.1.7",
+        version: "0.3.4",
       });
       dependencies.push({
         name: "llama-index-embeddings-gemini",
-        version: "0.1.6",
+        version: "^0.2.0",
       });
       break;
+    case "mistral":
+      dependencies.push({
+        name: "llama-index-llms-mistralai",
+        version: "0.2.1",
+      });
+      dependencies.push({
+        name: "llama-index-embeddings-mistralai",
+        version: "0.2.0",
+      });
+      break;
+    case "azure-openai":
+      dependencies.push({
+        name: "llama-index-llms-azure-openai",
+        version: "0.2.0",
+      });
+      dependencies.push({
+        name: "llama-index-embeddings-azure-openai",
+        version: "0.2.4",
+      });
+      break;
+    case "huggingface":
+      dependencies.push({
+        name: "llama-index-llms-huggingface",
+        version: "^0.3.5",
+      });
+      dependencies.push({
+        name: "llama-index-embeddings-huggingface",
+        version: "^0.3.1",
+      });
+      dependencies.push({
+        name: "optimum",
+        version: "^1.23.3",
+        extras: ["onnxruntime"],
+      });
+      break;
+    case "t-systems":
+      dependencies.push({
+        name: "llama-index-agent-openai",
+        version: "0.3.0",
+      });
+      dependencies.push({
+        name: "llama-index-llms-openai-like",
+        version: "0.2.0",
+      });
+      break;
+  }
+
+  if (observability && observability !== "none") {
+    if (observability === "traceloop") {
+      dependencies.push({
+        name: "traceloop-sdk",
+        version: "^0.15.11",
+      });
+    }
+    if (observability === "llamatrace") {
+      dependencies.push({
+        name: "llama-index-callbacks-arize-phoenix",
+        version: "^0.3.0",
+      });
+    }
   }
 
   return dependencies;
@@ -174,7 +290,7 @@ const getAdditionalDependencies = (
 
 const mergePoetryDependencies = (
   dependencies: Dependency[],
-  existingDependencies: Record<string, Omit<Dependency, "name">>,
+  existingDependencies: Record<string, Omit<Dependency, "name"> | string>,
 ) => {
   for (const dependency of dependencies) {
     let value = existingDependencies[dependency.name] ?? {};
@@ -187,13 +303,35 @@ const mergePoetryDependencies = (
     value.version = dependency.version ?? value.version;
     value.extras = dependency.extras ?? value.extras;
 
+    // Merge constraints if they exist
+    if (dependency.constraints) {
+      value = { ...value, ...dependency.constraints };
+    }
+
     if (value.version === undefined) {
       throw new Error(
         `Dependency "${dependency.name}" is missing attribute "version"!`,
       );
     }
 
-    existingDependencies[dependency.name] = value;
+    // Serialize as object if there are any additional properties
+    if (Object.keys(value).length > 1) {
+      existingDependencies[dependency.name] = value;
+    } else {
+      // Otherwise, serialize just the version string
+      existingDependencies[dependency.name] = value.version;
+    }
+  }
+};
+
+const copyRouterCode = async (root: string, tools: Tool[]) => {
+  // Copy sandbox router if the artifact tool is selected
+  if (tools?.some((t) => t.name === "artifact")) {
+    await copy("sandbox.py", path.join(root, "app", "api", "routers"), {
+      parents: true,
+      cwd: path.join(templatesDir, "components", "routers", "python"),
+      rename: assetRelocator,
+    });
   }
 };
 
@@ -258,36 +396,24 @@ export const installPythonDependencies = (
   }
 };
 
-export const installPythonTemplate = async ({
+const installLegacyPythonTemplate = async ({
   root,
   template,
-  framework,
   vectorDb,
   dataSources,
   tools,
-  postInstallAction,
+  useCase,
   observability,
-  modelConfig,
 }: Pick<
   InstallTemplateArgs,
   | "root"
-  | "framework"
   | "template"
   | "vectorDb"
   | "dataSources"
   | "tools"
-  | "postInstallAction"
+  | "useCase"
   | "observability"
-  | "modelConfig"
 >) => {
-  console.log("\nInitializing Python project with template:", template, "\n");
-  const templatePath = path.join(templatesDir, "types", template, framework);
-  await copy("**", root, {
-    parents: true,
-    cwd: templatePath,
-    rename: assetRelocator,
-  });
-
   const compPath = path.join(templatesDir, "components");
   const enginePath = path.join(root, "app", "engine");
 
@@ -297,62 +423,229 @@ export const installPythonTemplate = async ({
     cwd: path.join(compPath, "vectordbs", "python", vectorDb ?? "none"),
   });
 
-  // Copy all loaders to enginePath
-  const loaderPath = path.join(enginePath, "loaders");
-  await copy("**", loaderPath, {
-    parents: true,
-    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";
+  if (vectorDb !== "llamacloud") {
+    // Copy all loaders to enginePath
+    // Not needed for LlamaCloud as it has its own loaders
+    const loaderPath = path.join(enginePath, "loaders");
+    await copy("**", loaderPath, {
+      parents: true,
+      cwd: path.join(compPath, "loaders", "python"),
+    });
   }
-  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"),
   });
 
-  console.log("Adding additional dependencies");
+  // Copy services
+  if (template == "streaming" || template == "multiagent") {
+    await copy("**", path.join(root, "app", "api", "services"), {
+      cwd: path.join(compPath, "services", "python"),
+    });
+  }
+  // Copy engine code
+  if (template === "streaming" || template === "multiagent") {
+    // Select and copy engine code based on data sources and tools
+    let engine;
+    // Multiagent always uses agent engine
+    if (template === "multiagent") {
+      engine = "agent";
+    } else {
+      // For streaming, use chat engine by default
+      // Unless tools are selected, in which case use agent 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";
+      }
+    }
 
-  const addOnDependencies = getAdditionalDependencies(
-    modelConfig,
-    vectorDb,
-    dataSources,
-    tools,
-  );
-
-  if (observability === "opentelemetry") {
-    addOnDependencies.push({
-      name: "traceloop-sdk",
-      version: "^0.15.11",
+    // Copy engine code
+    await copy("**", enginePath, {
+      parents: true,
+      cwd: path.join(compPath, "engines", "python", engine),
     });
 
+    // Copy router code
+    await copyRouterCode(root, tools ?? []);
+  }
+
+  // Copy multiagents overrides
+  if (template === "multiagent") {
+    await copy("**", path.join(root), {
+      cwd: path.join(compPath, "multiagent", "python"),
+    });
+  }
+
+  if (template === "multiagent" || template === "reflex") {
+    if (useCase) {
+      const sourcePath =
+        template === "multiagent"
+          ? path.join(compPath, "agents", "python", useCase)
+          : path.join(compPath, "reflex", useCase);
+
+      await copy("**", path.join(root), {
+        parents: true,
+        cwd: sourcePath,
+        rename: assetRelocator,
+      });
+    } else {
+      console.log(
+        red(
+          `There is no use case selected for ${template} template. Please pick a use case to use via --use-case flag.`,
+        ),
+      );
+      process.exit(1);
+    }
+  }
+
+  if (observability && observability !== "none") {
     const templateObservabilityPath = path.join(
       templatesDir,
       "components",
       "observability",
       "python",
-      "opentelemetry",
+      observability,
     );
     await copy("**", path.join(root, "app"), {
       cwd: templateObservabilityPath,
     });
   }
+};
+
+const installLlamaIndexServerTemplate = async ({
+  root,
+  useCase,
+  useLlamaParse,
+}: Pick<InstallTemplateArgs, "root" | "useCase" | "useLlamaParse">) => {
+  if (!useCase) {
+    console.log(
+      red(
+        `There is no use case selected. Please pick a use case to use via --use-case flag.`,
+      ),
+    );
+    process.exit(1);
+  }
+
+  await copy("workflow.py", path.join(root, "app"), {
+    parents: true,
+    cwd: path.join(templatesDir, "components", "workflows", "python", useCase),
+  });
+
+  // Copy custom UI component code
+  await copy(`*`, path.join(root, "components"), {
+    parents: true,
+    cwd: path.join(templatesDir, "components", "ui", "workflows", useCase),
+  });
+
+  if (useLlamaParse) {
+    await copy("index.py", path.join(root, "app"), {
+      parents: true,
+      cwd: path.join(
+        templatesDir,
+        "components",
+        "vectordbs",
+        "llamaindexserver",
+        "llamacloud",
+        "python",
+      ),
+    });
+    // TODO: Consider moving generate.py to app folder.
+    await copy("generate.py", path.join(root), {
+      parents: true,
+      cwd: path.join(
+        templatesDir,
+        "components",
+        "vectordbs",
+        "llamaindexserver",
+        "llamacloud",
+        "python",
+      ),
+    });
+  }
+  // Copy README.md
+  await copy("README-template.md", path.join(root), {
+    parents: true,
+    cwd: path.join(templatesDir, "components", "workflows", "python", useCase),
+    rename: assetRelocator,
+  });
+};
+
+export const installPythonTemplate = async ({
+  appName,
+  root,
+  template,
+  framework,
+  vectorDb,
+  postInstallAction,
+  modelConfig,
+  dataSources,
+  tools,
+  useLlamaParse,
+  useCase,
+  observability,
+}: Pick<
+  InstallTemplateArgs,
+  | "appName"
+  | "root"
+  | "template"
+  | "framework"
+  | "vectorDb"
+  | "postInstallAction"
+  | "modelConfig"
+  | "dataSources"
+  | "tools"
+  | "useLlamaParse"
+  | "useCase"
+  | "observability"
+>) => {
+  console.log("\nInitializing Python project with template:", template, "\n");
+  let templatePath;
+  if (template === "reflex") {
+    templatePath = path.join(templatesDir, "types", "reflex");
+  } else {
+    templatePath = path.join(templatesDir, "types", template, framework);
+  }
+  await copy("**", root, {
+    parents: true,
+    cwd: templatePath,
+    rename: assetRelocator,
+  });
+
+  if (template === "llamaindexserver") {
+    await installLlamaIndexServerTemplate({
+      root,
+      useCase,
+      useLlamaParse,
+    });
+  } else {
+    await installLegacyPythonTemplate({
+      root,
+      template,
+      vectorDb,
+      dataSources,
+      tools,
+      useCase,
+      observability,
+    });
+  }
+
+  console.log("Adding additional dependencies");
+  const addOnDependencies = getAdditionalDependencies(
+    modelConfig,
+    vectorDb,
+    dataSources,
+    tools,
+    template,
+  );
 
   await addDependencies(root, addOnDependencies);
 
   if (postInstallAction === "runApp" || postInstallAction === "dependencies") {
     installPythonDependencies();
   }
-
-  // Copy deployment files for python
-  await copy("**", root, {
-    cwd: path.join(compPath, "deployments", "python"),
-  });
 };

helpers/run-app.ts

@@ -1,88 +1,81 @@
-import { ChildProcess, SpawnOptions, spawn } from "child_process";
-import path from "path";
-import { TemplateFramework } from "./types";
+import { SpawnOptions, spawn } from "child_process";
+import { TemplateFramework, TemplateType } from "./types";
 
 const createProcess = (
   command: string,
   args: string[],
   options: SpawnOptions,
-) => {
-  return spawn(command, args, {
-    ...options,
-    shell: true,
-  })
-    .on("exit", function (code) {
-      if (code !== 0) {
-        console.log(`Child process exited with code=${code}`);
-        process.exit(1);
-      }
+): Promise<void> => {
+  return new Promise((resolve, reject) => {
+    spawn(command, args, {
+      ...options,
+      shell: true,
     })
-    .on("error", function (err) {
-      console.log("Error when running chill process: ", err);
-      process.exit(1);
-    });
+      .on("exit", function (code) {
+        if (code !== 0) {
+          console.log(`Child process exited with code=${code}`);
+          reject(code);
+        } else {
+          resolve();
+        }
+      })
+      .on("error", function (err) {
+        console.log("Error when running child process: ", err);
+        reject(err);
+      });
+  });
 };
 
-// eslint-disable-next-line max-params
+export function runReflexApp(appPath: string, port: number) {
+  const commandArgs = [
+    "run",
+    "reflex",
+    "run",
+    "--frontend-port",
+    port.toString(),
+  ];
+  return createProcess("poetry", commandArgs, {
+    stdio: "inherit",
+    cwd: appPath,
+  });
+}
+
+export function runFastAPIApp(appPath: string, port: number) {
+  return createProcess("poetry", ["run", "dev"], {
+    stdio: "inherit",
+    cwd: appPath,
+    env: { ...process.env, APP_PORT: `${port}` },
+  });
+}
+
+export function runTSApp(appPath: string, port: number) {
+  return createProcess("npm", ["run", "dev"], {
+    stdio: "inherit",
+    cwd: appPath,
+    env: { ...process.env, PORT: `${port}` },
+  });
+}
+
 export async function runApp(
   appPath: string,
-  frontend: boolean,
+  template: TemplateType,
   framework: TemplateFramework,
   port?: number,
-  externalPort?: number,
-): Promise<any> {
-  let backendAppProcess: ChildProcess;
-  let frontendAppProcess: ChildProcess | undefined;
-  const frontendPort = port || 3000;
-  let backendPort = externalPort || 8000;
+): Promise<void> {
+  try {
+    // Start the app
+    const defaultPort =
+      framework === "nextjs" || template === "reflex" ? 3000 : 8000;
 
-  // Callback to kill app processes
-  process.on("exit", () => {
-    console.log("Killing app processes...");
-    backendAppProcess.kill();
-    frontendAppProcess?.kill();
-  });
-
-  let backendCommand = "";
-  let backendArgs: string[];
-  if (framework === "fastapi") {
-    backendCommand = "poetry";
-    backendArgs = [
-      "run",
-      "uvicorn",
-      "main:app",
-      "--host=0.0.0.0",
-      "--port=" + backendPort,
-    ];
-  } else if (framework === "nextjs") {
-    backendCommand = "npm";
-    backendArgs = ["run", "dev"];
-    backendPort = frontendPort;
-  } else {
-    backendCommand = "npm";
-    backendArgs = ["run", "dev"];
-  }
-
-  if (frontend) {
-    return new Promise((resolve, reject) => {
-      backendAppProcess = createProcess(backendCommand, backendArgs, {
-        stdio: "inherit",
-        cwd: path.join(appPath, "backend"),
-        env: { ...process.env, PORT: `${backendPort}` },
-      });
-      frontendAppProcess = createProcess("npm", ["run", "dev"], {
-        stdio: "inherit",
-        cwd: path.join(appPath, "frontend"),
-        env: { ...process.env, PORT: `${frontendPort}` },
-      });
-    });
-  } else {
-    return new Promise((resolve, reject) => {
-      backendAppProcess = createProcess(backendCommand, backendArgs, {
-        stdio: "inherit",
-        cwd: path.join(appPath),
-        env: { ...process.env, PORT: `${backendPort}` },
-      });
-    });
+    const appRunner =
+      template === "reflex"
+        ? runReflexApp
+        : framework === "fastapi"
+          ? runFastAPIApp
+          : runTSApp;
+    await appRunner(appPath, port || defaultPort);
+  } catch (error) {
+    console.error("Failed to run app:", error);
+    throw error;
   }
 }

helpers/tools.ts

@@ -30,7 +30,7 @@ export type ToolDependencies = {
 
 export const supportedTools: Tool[] = [
   {
-    display: "Google Search (configuration required after installation)",
+    display: "Google Search",
     name: "google.GoogleSearchToolSpec",
     config: {
       engine:
@@ -41,7 +41,7 @@ export const supportedTools: Tool[] = [
     dependencies: [
       {
         name: "llama-index-tools-google",
-        version: "0.1.2",
+        version: "^0.3.0",
       },
     ],
     supportedFrameworks: ["fastapi"],
@@ -54,24 +54,39 @@ export const supportedTools: Tool[] = [
       },
     ],
   },
+  {
+    // 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: "^6.3.5",
+      },
+    ],
+    supportedFrameworks: ["fastapi"], // TODO: Re-enable this tool once the duck-duck-scrape TypeScript library works again
+    type: ToolType.LOCAL,
+    envVars: [
+      {
+        name: TOOL_SYSTEM_PROMPT_ENV_VAR,
+        description: "System prompt for DuckDuckGo search tool.",
+        value: `You have access to the duckduckgo search tool. Use it 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",
     name: "wikipedia.WikipediaToolSpec",
     dependencies: [
       {
         name: "llama-index-tools-wikipedia",
-        version: "0.1.2",
+        version: "^0.3.0",
       },
     ],
     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",
@@ -79,11 +94,27 @@ export const supportedTools: Tool[] = [
     dependencies: [],
     supportedFrameworks: ["fastapi", "express", "nextjs"],
     type: ToolType.LOCAL,
+  },
+  {
+    display: "Document generator",
+    name: "document_generator",
+    supportedFrameworks: ["fastapi", "nextjs", "express"],
+    dependencies: [
+      {
+        name: "xhtml2pdf",
+        version: "^0.2.14",
+      },
+      {
+        name: "markdown",
+        version: "^3.7",
+      },
+    ],
+    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.`,
+        description: "System prompt for document generator tool.",
+        value: `If user request for a report or a post, use document generator tool to create a file and reply with the link to the file.`,
       },
     ],
   },
@@ -93,7 +124,7 @@ export const supportedTools: Tool[] = [
     dependencies: [
       {
         name: "e2b_code_interpreter",
-        version: "0.0.7",
+        version: "1.1.1",
       },
     ],
     supportedFrameworks: ["fastapi", "express", "nextjs"],
@@ -107,13 +138,120 @@ export const supportedTools: Tool[] = [
       {
         name: TOOL_SYSTEM_PROMPT_ENV_VAR,
         description: "System prompt for code interpreter tool.",
-        value: `You are a Python interpreter.
-        - You are given tasks to complete and you run python code to solve them.
-        - The python code runs in a Jupyter notebook. Every time you call \`interpreter\` tool, the python code is executed in a separate cell. It's okay to make multiple calls to \`interpreter\`.
-        - Display visualizations using matplotlib or any other visualization library directly in the notebook. Shouldn't save the visualizations to a file, just return the base64 encoded data.
-        - You can install any pip package (if it exists) if you need to but the usual packages for data analysis are already preinstalled.
-        - You can run any python code you want in a secure environment.
-        - Use absolute url from result to display images or any other media.`,
+        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: "Artifact Code Generator",
+    name: "artifact",
+    // Using pre-release version of e2b_code_interpreter
+    // TODO: Update to stable version when 0.0.11 is released
+    dependencies: [
+      {
+        name: "e2b_code_interpreter",
+        version: "1.1.1",
+      },
+    ],
+    supportedFrameworks: ["fastapi", "express", "nextjs"],
+    type: ToolType.LOCAL,
+    envVars: [
+      {
+        name: "E2B_API_KEY",
+        description:
+          "E2B_API_KEY key is required to run artifact code generator tool. Get it here: https://e2b.dev/docs/getting-started/api-key",
+      },
+      {
+        name: TOOL_SYSTEM_PROMPT_ENV_VAR,
+        description: "System prompt for artifact code generator tool.",
+        value:
+          "You are a code assistant that can generate and execute code using its tools. Don't generate code yourself, use the provided tools instead. Do not show the code or sandbox url in chat, just describe the steps to build the application based on the code that is generated by your tools. Do not describe how to run the code, just the steps to build the application.",
+      },
+    ],
+  },
+  {
+    display: "OpenAPI action",
+    name: "openapi_action.OpenAPIActionToolSpec",
+    dependencies: [
+      {
+        name: "llama-index-tools-openapi",
+        version: "0.2.0",
+      },
+      {
+        name: "jsonschema",
+        version: "^4.22.0",
+      },
+      {
+        name: "llama-index-tools-requests",
+        version: "0.2.0",
+      },
+    ],
+    config: {
+      openapi_uri: "The URL or file path of the OpenAPI schema",
+    },
+    supportedFrameworks: ["fastapi", "express", "nextjs"],
+    type: ToolType.LOCAL,
+  },
+  {
+    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",
+      },
+    ],
+  },
+  {
+    display: "Azure Code Interpreter",
+    name: "azure_code_interpreter.AzureCodeInterpreterToolSpec",
+    supportedFrameworks: ["fastapi", "nextjs", "express"],
+    type: ToolType.LLAMAHUB,
+    dependencies: [
+      {
+        name: "llama-index-tools-azure-code-interpreter",
+        version: "0.2.0",
+      },
+    ],
+    envVars: [
+      {
+        name: "AZURE_POOL_MANAGEMENT_ENDPOINT",
+        description:
+          "Please follow this guideline to create and get the pool management endpoint: https://learn.microsoft.com/azure/container-apps/sessions?tabs=azure-cli",
+      },
+      {
+        name: TOOL_SYSTEM_PROMPT_ENV_VAR,
+        description: "System prompt for Azure 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: "Form Filling",
+    name: "form_filling",
+    supportedFrameworks: ["fastapi"],
+    type: ToolType.LOCAL,
+    dependencies: [
+      {
+        name: "pandas",
+        version: "^2.2.3",
+      },
+      {
+        name: "tabulate",
+        version: "^0.9.0",
       },
     ],
   },
@@ -142,9 +280,15 @@ 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;
 };
@@ -159,7 +303,6 @@ 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>;
   } = {
@@ -182,9 +325,16 @@ export const writeToolsConfig = async (
       yaml.stringify(configContent),
     );
   } else {
+    // For Typescript, we treat llamahub tools as local tools
+    const tsConfigContent = {
+      local: {
+        ...configContent.local,
+        ...configContent.llamahub,
+      },
+    };
     await fs.writeFile(
       path.join(configPath, "tools.json"),
-      JSON.stringify(configContent, null, 2),
+      JSON.stringify(tsConfigContent, null, 2),
     );
   }
 };

helpers/types.ts

@@ -1,7 +1,16 @@
 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"
+  | "azure-openai"
+  | "huggingface"
+  | "t-systems";
 export type ModelConfig = {
   provider: ModelProvider;
   apiKey?: string;
@@ -10,7 +19,13 @@ export type ModelConfig = {
   dimensions: number;
   isConfigured(): boolean;
 };
-export type TemplateType = "streaming" | "community" | "llamapack";
+export type TemplateType =
+  | "streaming"
+  | "community"
+  | "llamapack"
+  | "multiagent"
+  | "reflex"
+  | "llamaindexserver";
 export type TemplateFramework = "nextjs" | "express" | "fastapi";
 export type TemplateUI = "html" | "shadcn";
 export type TemplateVectorDB =
@@ -21,7 +36,9 @@ export type TemplateVectorDB =
   | "milvus"
   | "astra"
   | "qdrant"
-  | "chroma";
+  | "chroma"
+  | "llamacloud"
+  | "weaviate";
 export type TemplatePostInstallAction =
   | "none"
   | "VSCode"
@@ -32,11 +49,25 @@ export type TemplateDataSource = {
   config: TemplateDataSourceConfig;
 };
 export type TemplateDataSourceType = "file" | "web" | "db";
-export type TemplateObservability = "none" | "opentelemetry";
+export type TemplateObservability = "none" | "traceloop" | "llamatrace";
+export type TemplateUseCase =
+  | "financial_report"
+  | "blog"
+  | "deep_research"
+  | "form_filling"
+  | "extractor"
+  | "contract_review"
+  | "agentic_rag";
 // Config for both file and folder
-export type FileSourceConfig = {
-  path: string;
-};
+export type FileSourceConfig =
+  | {
+      path: string;
+      filename?: string;
+    }
+  | {
+      url: URL;
+      filename?: string;
+    };
 export type WebSourceConfig = {
   baseUrl?: string;
   prefix?: string;
@@ -68,15 +99,15 @@ export interface InstallTemplateArgs {
   framework: TemplateFramework;
   ui: TemplateUI;
   dataSources: TemplateDataSource[];
-  customApiPath?: string;
   modelConfig: ModelConfig;
   llamaCloudKey?: string;
   useLlamaParse?: boolean;
   communityProjectConfig?: CommunityProjectConfig;
   llamapack?: string;
   vectorDb?: TemplateVectorDB;
-  externalPort?: number;
+  port?: number;
   postInstallAction?: TemplatePostInstallAction;
   tools?: Tool[];
   observability?: TemplateObservability;
+  useCase?: TemplateUseCase;
 }

helpers/typescript.ts

@@ -1,47 +1,111 @@
 import fs from "fs/promises";
 import os from "os";
 import path from "path";
-import { bold, cyan } from "picocolors";
+import { bold, cyan, red, yellow } from "picocolors";
 import { assetRelocator, copy } from "../helpers/copy";
 import { callPackageManager } from "../helpers/install";
 import { templatesDir } from "./dir";
 import { PackageManager } from "./get-pkg-manager";
-import { InstallTemplateArgs } from "./types";
+import { InstallTemplateArgs, ModelProvider, TemplateVectorDB } from "./types";
 
-/**
- * Install a LlamaIndex internal template to a given `root` directory.
- */
-export const installTSTemplate = async ({
-  appName,
+const installLlamaIndexServerTemplate = async ({
+  root,
+  useCase,
+  vectorDb,
+}: Pick<InstallTemplateArgs, "root" | "useCase" | "vectorDb">) => {
+  if (!useCase) {
+    console.log(
+      red(
+        `There is no use case selected. Please pick a use case to use via --use-case flag.`,
+      ),
+    );
+    process.exit(1);
+  }
+
+  if (!vectorDb) {
+    console.log(
+      red(
+        `There is no vector db selected. Please pick a vector db to use via --vector-db flag.`,
+      ),
+    );
+    process.exit(1);
+  }
+
+  await copy("workflow.ts", path.join(root, "src", "app"), {
+    parents: true,
+    cwd: path.join(
+      templatesDir,
+      "components",
+      "workflows",
+      "typescript",
+      useCase,
+    ),
+  });
+
+  // copy workflow UI components to output/components folder
+  await copy("*", path.join(root, "components"), {
+    parents: true,
+    cwd: path.join(templatesDir, "components", "ui", "workflows", useCase),
+  });
+
+  if (vectorDb === "llamacloud") {
+    await copy("generate.ts", path.join(root, "src"), {
+      parents: true,
+      cwd: path.join(
+        templatesDir,
+        "components",
+        "vectordbs",
+        "llamaindexserver",
+        "llamacloud",
+        "typescript",
+      ),
+    });
+
+    await copy("index.ts", path.join(root, "src", "app"), {
+      parents: true,
+      cwd: path.join(
+        templatesDir,
+        "components",
+        "vectordbs",
+        "llamaindexserver",
+        "llamacloud",
+        "typescript",
+      ),
+      rename: () => "data.ts",
+    });
+  }
+  // Copy README.md
+  await copy("README-template.md", path.join(root), {
+    parents: true,
+    cwd: path.join(
+      templatesDir,
+      "components",
+      "workflows",
+      "typescript",
+      useCase,
+    ),
+    rename: assetRelocator,
+  });
+};
+
+const installLegacyTSTemplate = async ({
   root,
-  packageManager,
-  isOnline,
   template,
+  backend,
   framework,
   ui,
   vectorDb,
-  postInstallAction,
-  backend,
   observability,
   tools,
   dataSources,
   useLlamaParse,
-}: InstallTemplateArgs & { backend: boolean }) => {
-  console.log(bold(`Using ${packageManager}.`));
-
-  /**
-   * Copy the template files to the target directory.
-   */
-  console.log("\nInitializing project with template:", template, "\n");
-  const templatePath = path.join(templatesDir, "types", template, framework);
-  const copySource = ["**"];
-
-  await copy(copySource, root, {
-    parents: true,
-    cwd: templatePath,
-    rename: assetRelocator,
-  });
-
+  useCase,
+  modelConfig,
+  relativeEngineDestPath,
+}: InstallTemplateArgs & {
+  backend: boolean;
+  relativeEngineDestPath: string;
+}) => {
   /**
    * If next.js is used, update its configuration if necessary
    */
@@ -57,11 +121,9 @@ export const installTSTemplate = async ({
       console.log("\nUsing static site generation\n");
     } else {
       if (vectorDb === "milvus") {
-        nextConfigJson.experimental.serverComponentsExternalPackages =
-          nextConfigJson.experimental.serverComponentsExternalPackages ?? [];
-        nextConfigJson.experimental.serverComponentsExternalPackages.push(
-          "@zilliz/milvus2-sdk-node",
-        );
+        nextConfigJson.serverExternalPackages =
+          nextConfigJson.serverExternalPackages ?? [];
+        nextConfigJson.serverExternalPackages.push("@zilliz/milvus2-sdk-node");
       }
     }
     await fs.writeFile(
@@ -70,7 +132,7 @@ export const installTSTemplate = async ({
     );
 
     const webpackConfigOtelFile = path.join(root, "webpack.config.o11y.mjs");
-    if (observability === "opentelemetry") {
+    if (observability === "traceloop") {
       const webpackConfigDefaultFile = path.join(root, "webpack.config.mjs");
       await fs.rm(webpackConfigDefaultFile);
       await fs.rename(webpackConfigOtelFile, webpackConfigDefaultFile);
@@ -98,19 +160,79 @@ export const installTSTemplate = async ({
   }
 
   const compPath = path.join(templatesDir, "components");
-  const relativeEngineDestPath =
-    framework === "nextjs"
-      ? path.join("app", "api", "chat")
-      : path.join("src", "controllers");
   const enginePath = path.join(root, relativeEngineDestPath, "engine");
 
+  // copy llamaindex code for TS templates
+  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"),
   });
 
+  if (template === "multiagent") {
+    const multiagentPath = path.join(compPath, "multiagent", "typescript");
+
+    // copy workflow code for multiagent template
+    await copy("**", path.join(root, relativeEngineDestPath, "workflow"), {
+      parents: true,
+      cwd: path.join(multiagentPath, "workflow"),
+    });
+
+    // Copy use case code for multiagent template
+    if (useCase) {
+      console.log("\nCopying use case:", useCase, "\n");
+      const useCasePath = path.join(compPath, "agents", "typescript", useCase);
+      const useCaseCodePath = path.join(useCasePath, "workflow");
+
+      // Copy use case codes
+      await copy("**", path.join(root, relativeEngineDestPath, "workflow"), {
+        parents: true,
+        cwd: useCaseCodePath,
+        rename: assetRelocator,
+      });
+
+      // Copy use case files to project root
+      await copy("*.*", path.join(root), {
+        parents: true,
+        cwd: useCasePath,
+        rename: assetRelocator,
+      });
+    } else {
+      console.log(
+        red(
+          `There is no use case selected for ${template} template. Please pick a use case to use via --use-case flag.`,
+        ),
+      );
+      process.exit(1);
+    }
+
+    if (framework === "nextjs") {
+      // patch route.ts file
+      await copy("**", path.join(root, relativeEngineDestPath), {
+        parents: true,
+        cwd: path.join(multiagentPath, "nextjs"),
+      });
+    } else if (framework === "express") {
+      // patch chat.controller.ts file
+      await copy("**", path.join(root, relativeEngineDestPath), {
+        parents: true,
+        cwd: path.join(multiagentPath, "express"),
+      });
+    }
+  }
+
   // copy loader component (TS only supports llama_parse and file for now)
   const loaderFolder = useLlamaParse ? "llama_parse" : "file";
   await copy("**", enginePath, {
@@ -118,10 +240,19 @@ export const installTSTemplate = async ({
     cwd: path.join(compPath, "loaders", "typescript", loaderFolder),
   });
 
+  // copy provider settings
+  await copy("**", enginePath, {
+    parents: true,
+    cwd: path.join(compPath, "providers", "typescript", modelConfig.provider),
+  });
+
   // Select and copy engine code based on data sources and tools
   let engine;
   tools = tools ?? [];
-  if (dataSources.length > 0 && tools.length === 0) {
+  // multiagent template always uses agent engine
+  if (template === "multiagent") {
+    engine = "agent";
+  } else if (dataSources.length > 0 && tools.length === 0) {
     console.log("\nNo tools selected - use optimized context chat engine\n");
     engine = "chat";
   } else {
@@ -132,6 +263,11 @@ export const installTSTemplate = async ({
     cwd: path.join(compPath, "engines", "typescript", engine),
   });
 
+  // copy settings to engine folder
+  await copy("**", enginePath, {
+    cwd: path.join(compPath, "settings", "typescript"),
+  });
+
   /**
    * Copy the selected UI files to the target directory and reference it.
    */
@@ -158,6 +294,75 @@ export const installTSTemplate = async ({
     await fs.rm(path.join(root, "app", "api"), { recursive: true });
     await fs.rm(path.join(root, "config"), { recursive: true, force: true });
   }
+};
+
+/**
+ * Install a LlamaIndex internal template to a given `root` directory.
+ */
+export const installTSTemplate = async ({
+  appName,
+  root,
+  packageManager,
+  isOnline,
+  template,
+  framework,
+  ui,
+  vectorDb,
+  postInstallAction,
+  backend,
+  observability,
+  tools,
+  dataSources,
+  useLlamaParse,
+  useCase,
+  modelConfig,
+}: InstallTemplateArgs & { backend: boolean }) => {
+  console.log(bold(`Using ${packageManager}.`));
+
+  /**
+   * Copy the template files to the target directory.
+   */
+  console.log("\nInitializing project with template:", template, "\n");
+  const templatePath = path.join(templatesDir, "types", template, framework);
+  const copySource = ["**"];
+
+  await copy(copySource, root, {
+    parents: true,
+    cwd: templatePath,
+    rename: assetRelocator,
+  });
+
+  const relativeEngineDestPath =
+    framework === "nextjs"
+      ? path.join("app", "api", "chat")
+      : path.join("src", "controllers");
+
+  if (template === "llamaindexserver") {
+    await installLlamaIndexServerTemplate({
+      root,
+      useCase,
+      vectorDb,
+    });
+  } else {
+    await installLegacyTSTemplate({
+      appName,
+      root,
+      packageManager,
+      isOnline,
+      template,
+      backend,
+      framework,
+      ui,
+      vectorDb,
+      observability,
+      tools,
+      dataSources,
+      useLlamaParse,
+      useCase,
+      modelConfig,
+      relativeEngineDestPath,
+    });
+  }
 
   const packageJson = await updatePackageJson({
     root,
@@ -167,16 +372,80 @@ export const installTSTemplate = async ({
     framework,
     ui,
     observability,
+    vectorDb,
+    backend,
+    modelConfig,
+    template,
   });
 
-  if (postInstallAction === "runApp" || postInstallAction === "dependencies") {
+  if (
+    backend &&
+    (postInstallAction === "runApp" || postInstallAction === "dependencies")
+  ) {
     await installTSDependencies(packageJson, packageManager, isOnline);
   }
+};
 
-  // Copy deployment files for typescript
-  await copy("**", root, {
-    cwd: path.join(compPath, "deployments", "typescript"),
-  });
+const providerDependencies: {
+  [key in ModelProvider]?: Record<string, string>;
+} = {
+  openai: {
+    "@llamaindex/openai": "^0.2.0",
+  },
+  gemini: {
+    "@llamaindex/google": "^0.2.0",
+  },
+  ollama: {
+    "@llamaindex/ollama": "^0.1.0",
+  },
+  mistral: {
+    "@llamaindex/mistral": "^0.2.0",
+  },
+  "azure-openai": {
+    "@llamaindex/openai": "^0.2.0",
+  },
+  groq: {
+    "@llamaindex/groq": "^0.0.61",
+    "@llamaindex/huggingface": "^0.1.0", // groq uses huggingface as default embedding model
+  },
+  anthropic: {
+    "@llamaindex/anthropic": "^0.3.0",
+    "@llamaindex/huggingface": "^0.1.0", // anthropic uses huggingface as default embedding model
+  },
+};
+
+const vectorDbDependencies: Record<TemplateVectorDB, Record<string, string>> = {
+  astra: {
+    "@llamaindex/astra": "^0.0.5",
+  },
+  chroma: {
+    "@llamaindex/chroma": "^0.0.5",
+  },
+  llamacloud: {},
+  milvus: {
+    "@zilliz/milvus2-sdk-node": "^2.4.6",
+    "@llamaindex/milvus": "^0.1.0",
+  },
+  mongo: {
+    mongodb: "6.7.0",
+    "@llamaindex/mongodb": "^0.0.5",
+  },
+  none: {},
+  pg: {
+    pg: "^8.12.0",
+    pgvector: "^0.2.0",
+    "@llamaindex/postgres": "^0.0.33",
+  },
+  pinecone: {
+    "@llamaindex/pinecone": "^0.0.5",
+  },
+  qdrant: {
+    "@qdrant/js-client-rest": "^1.11.0",
+    "@llamaindex/qdrant": "^0.1.0",
+  },
+  weaviate: {
+    "@llamaindex/weaviate": "^0.0.5",
+  },
 };
 
 async function updatePackageJson({
@@ -187,11 +456,24 @@ async function updatePackageJson({
   framework,
   ui,
   observability,
+  vectorDb,
+  backend,
+  modelConfig,
+  template,
 }: Pick<
   InstallTemplateArgs,
-  "root" | "appName" | "dataSources" | "framework" | "ui" | "observability"
+  | "root"
+  | "appName"
+  | "dataSources"
+  | "framework"
+  | "ui"
+  | "observability"
+  | "vectorDb"
+  | "modelConfig"
+  | "template"
 > & {
   relativeEngineDestPath: string;
+  backend: boolean;
 }): Promise<any> {
   const packageJsonFile = path.join(root, "package.json");
   const packageJson: any = JSON.parse(
@@ -200,7 +482,7 @@ async function updatePackageJson({
   packageJson.name = appName;
   packageJson.version = "0.1.0";
 
-  if (relativeEngineDestPath) {
+  if (relativeEngineDestPath && template !== "llamaindexserver") {
     // TODO: move script to {root}/scripts for all frameworks
     // add generate script if using context engine
     packageJson.scripts = {
@@ -227,16 +509,32 @@ async function updatePackageJson({
       "remark-gfm": undefined,
       "remark-math": undefined,
       "react-markdown": undefined,
-      "react-syntax-highlighter": undefined,
-    };
-
-    packageJson.devDependencies = {
-      ...packageJson.devDependencies,
-      "@types/react-syntax-highlighter": undefined,
+      "highlight.js": undefined,
     };
   }
 
-  if (observability === "opentelemetry") {
+  if (backend) {
+    packageJson.dependencies = {
+      ...packageJson.dependencies,
+      "@llamaindex/readers": "^2.0.0",
+    };
+
+    if (vectorDb && vectorDb in vectorDbDependencies) {
+      packageJson.dependencies = {
+        ...packageJson.dependencies,
+        ...vectorDbDependencies[vectorDb],
+      };
+    }
+
+    if (modelConfig.provider && modelConfig.provider in providerDependencies) {
+      packageJson.dependencies = {
+        ...packageJson.dependencies,
+        ...providerDependencies[modelConfig.provider],
+      };
+    }
+  }
+
+  if (observability === "traceloop") {
     packageJson.dependencies = {
       ...packageJson.dependencies,
       "@traceloop/node-server-sdk": "^0.5.19",

helpers/vscode.ts

@@ -1,40 +1,26 @@
 import fs from "fs";
 import path from "path";
+import { assetRelocator, copy } from "./copy";
 import { TemplateFramework } from "./types";
 
 function renderDevcontainerContent(
   templatesDir: string,
   framework: TemplateFramework,
-  frontend: boolean,
 ) {
   const devcontainerJson: any = JSON.parse(
     fs.readFileSync(path.join(templatesDir, "devcontainer.json"), "utf8"),
   );
 
   // Modify postCreateCommand
-  if (frontend) {
-    devcontainerJson.postCreateCommand =
-      framework === "fastapi"
-        ? "cd backend && poetry install && cd ../frontend && npm install"
-        : "cd backend && npm install && cd ../frontend && npm install";
-  } else {
-    devcontainerJson.postCreateCommand =
-      framework === "fastapi" ? "poetry install" : "npm install";
-  }
+  devcontainerJson.postCreateCommand =
+    framework === "fastapi" ? "poetry install" : "npm install";
 
   // Modify containerEnv
   if (framework === "fastapi") {
-    if (frontend) {
-      devcontainerJson.containerEnv = {
-        ...devcontainerJson.containerEnv,
-        PYTHONPATH: "${PYTHONPATH}:${workspaceFolder}/backend",
-      };
-    } else {
-      devcontainerJson.containerEnv = {
-        ...devcontainerJson.containerEnv,
-        PYTHONPATH: "${PYTHONPATH}:${workspaceFolder}",
-      };
-    }
+    devcontainerJson.containerEnv = {
+      ...devcontainerJson.containerEnv,
+      PYTHONPATH: "${PYTHONPATH}:${workspaceFolder}",
+    };
   }
 
   return JSON.stringify(devcontainerJson, null, 2);
@@ -44,7 +30,6 @@ export const writeDevcontainer = async (
   root: string,
   templatesDir: string,
   framework: TemplateFramework,
-  frontend: boolean,
 ) => {
   const devcontainerDir = path.join(root, ".devcontainer");
   if (fs.existsSync(devcontainerDir)) {
@@ -54,7 +39,6 @@ export const writeDevcontainer = async (
   const devcontainerContent = renderDevcontainerContent(
     templatesDir,
     framework,
-    frontend,
   );
   fs.mkdirSync(devcontainerDir);
   await fs.promises.writeFile(
@@ -62,3 +46,25 @@ export const writeDevcontainer = async (
     devcontainerContent,
   );
 };
+
+export const copyVSCodeSettings = async (
+  root: string,
+  templatesDir: string,
+) => {
+  const vscodeDir = path.join(root, ".vscode");
+  await copy("vscode_settings.json", vscodeDir, {
+    cwd: templatesDir,
+    rename: assetRelocator,
+  });
+};
+
+export const configVSCode = async (
+  root: string,
+  templatesDir: string,
+  framework: TemplateFramework,
+) => {
+  await writeDevcontainer(root, templatesDir, framework);
+  if (framework === "fastapi") {
+    await copyVSCodeSettings(root, templatesDir);
+  }
+};

index.ts

@@ -1,7 +1,6 @@
 /* eslint-disable import/no-extraneous-dependencies */
 import { execSync } from "child_process";
-import Commander from "commander";
-import Conf from "conf";
+import { Command } from "commander";
 import fs from "fs";
 import path from "path";
 import { bold, cyan, green, red, yellow } from "picocolors";
@@ -9,7 +8,7 @@ import prompts from "prompts";
 import terminalLink from "terminal-link";
 import checkForUpdate from "update-check";
 import { createApp } from "./create-app";
-import { getDataSources } from "./helpers/datasources";
+import { EXAMPLE_FILE, getDataSources } from "./helpers/datasources";
 import { getPkgManager } from "./helpers/get-pkg-manager";
 import { isFolderEmpty } from "./helpers/is-folder-empty";
 import { initializeGlobalAgent } from "./helpers/proxy";
@@ -17,8 +16,9 @@ import { runApp } from "./helpers/run-app";
 import { getTools } from "./helpers/tools";
 import { validateNpmName } from "./helpers/validate-pkg";
 import packageJson from "./package.json";
-import { QuestionArgs, askQuestions, onPromptState } from "./questions";
-
+import { askQuestions } from "./questions/index";
+import { QuestionArgs } from "./questions/types";
+import { onPromptState } from "./questions/utils";
 // Run the initialization function
 initializeGlobalAgent();
 
@@ -29,12 +29,14 @@ const handleSigTerm = () => process.exit(0);
 process.on("SIGINT", handleSigTerm);
 process.on("SIGTERM", handleSigTerm);
 
-const program = new Commander.Command(packageJson.name)
+const program = new Command(packageJson.name)
   .version(packageJson.version)
-  .arguments("<project-directory>")
-  .usage(`${green("<project-directory>")} [options]`)
+  .arguments("[project-directory]")
+  .usage(`${green("[project-directory]")} [options]`)
   .action((name) => {
-    projectPath = name;
+    if (name) {
+      projectPath = name;
+    }
   })
   .option(
     "--use-npm",
@@ -55,13 +57,6 @@ const program = new Commander.Command(packageJson.name)
     `
 
   Explicitly tell the CLI to bootstrap the application using Yarn
-`,
-  )
-  .option(
-    "--reset-preferences",
-    `
-
-  Explicitly tell the CLI to reset any stored preferences
 `,
   )
   .option(
@@ -90,6 +85,20 @@ const program = new Commander.Command(packageJson.name)
     `
 
   Select to use an example PDF as data source.
+`,
+  )
+  .option(
+    "--web-source <url>",
+    `
+  
+  Specify a website URL to use as a data source.
+`,
+  )
+  .option(
+    "--db-source <connection-string>",
+    `
+  
+  Specify a database connection string to use as a data source.
 `,
   )
   .option(
@@ -110,7 +119,14 @@ const program = new Commander.Command(packageJson.name)
     "--frontend",
     `
 
-  Whether to generate a frontend for your backend.
+  Generate a frontend for your backend.
+`,
+  )
+  .option(
+    "--no-frontend",
+    `
+
+  Do not generate a frontend for your backend.
 `,
   )
   .option(
@@ -118,13 +134,6 @@ const program = new Commander.Command(packageJson.name)
     `
 
   Select UI port.
-`,
-  )
-  .option(
-    "--external-port <external>",
-    `
-
-  Select external port.
 `,
   )
   .option(
@@ -147,6 +156,13 @@ const program = new Commander.Command(packageJson.name)
 
   Specify the tools you want to use by providing a comma-separated list. For example, 'wikipedia.WikipediaToolSpec,google.GoogleSearchToolSpec'. Use 'none' to not using any tools.
 `,
+    (tools, _) => {
+      if (tools === "none") {
+        return [];
+      } else {
+        return getTools(tools.split(","));
+      }
+    },
   )
   .option(
     "--use-llama-parse",
@@ -173,48 +189,75 @@ const program = new Commander.Command(packageJson.name)
     "--ask-models",
     `
 
-  Select LLM and embedding models.
+  Allow interactive selection of LLM and embedding models of different model providers.
+`,
+    false,
+  )
+  .option(
+    "--pro",
+    `
+
+  Allow interactive selection of all features.
+`,
+    false,
+  )
+  .option(
+    "--use-case <useCase>",
+    `
+
+  Select which use case to use for the multi-agent template (e.g: financial_report, blog).
 `,
   )
   .allowUnknownOption()
   .parse(process.argv);
-if (process.argv.includes("--no-frontend")) {
-  program.frontend = false;
+
+const options = program.opts();
+
+if (
+  process.argv.includes("--no-llama-parse") ||
+  options.template === "reflex"
+) {
+  options.useLlamaParse = false;
 }
-if (process.argv.includes("--tools")) {
-  if (program.tools === "none") {
-    program.tools = [];
-  } else {
-    program.tools = getTools(program.tools.split(","));
-  }
-}
-if (process.argv.includes("--no-llama-parse")) {
-  program.useLlamaParse = false;
-}
-program.askModels = process.argv.includes("--ask-models");
 if (process.argv.includes("--no-files")) {
-  program.dataSources = [];
-} else {
-  program.dataSources = getDataSources(program.files, program.exampleFile);
+  options.dataSources = [];
+} else if (process.argv.includes("--example-file")) {
+  options.dataSources = getDataSources(options.files, options.exampleFile);
+} else if (process.argv.includes("--llamacloud")) {
+  options.dataSources = [EXAMPLE_FILE];
+  options.vectorDb = "llamacloud";
+} else if (process.argv.includes("--web-source")) {
+  options.dataSources = [
+    {
+      type: "web",
+      config: {
+        baseUrl: options.webSource,
+        prefix: options.webSource,
+        depth: 1,
+      },
+    },
+  ];
+} else if (process.argv.includes("--db-source")) {
+  options.dataSources = [
+    {
+      type: "db",
+      config: {
+        uri: options.dbSource,
+        queries: options.dbQuery || "SELECT * FROM mytable",
+      },
+    },
+  ];
 }
 
-const packageManager = !!program.useNpm
+const packageManager = !!options.useNpm
   ? "npm"
-  : !!program.usePnpm
+  : !!options.usePnpm
     ? "pnpm"
-    : !!program.useYarn
+    : !!options.useYarn
       ? "yarn"
       : getPkgManager();
 
 async function run(): Promise<void> {
-  const conf = new Conf({ projectName: "create-llama" });
-
-  if (program.resetPreferences) {
-    conf.clear();
-    console.log(`Preferences reset successfully`);
-    return;
-  }
-
   if (typeof projectPath === "string") {
     projectPath = projectPath.trim();
   }
@@ -277,35 +320,15 @@ async function run(): Promise<void> {
     process.exit(1);
   }
 
-  const preferences = (conf.get("preferences") || {}) as QuestionArgs;
-  await askQuestions(
-    program as unknown as QuestionArgs,
-    preferences,
-    program.openAiKey,
-  );
+  const answers = await askQuestions(options as unknown as QuestionArgs);
 
   await createApp({
-    template: program.template,
-    framework: program.framework,
-    ui: program.ui,
+    ...answers,
     appPath: resolvedProjectPath,
     packageManager,
-    frontend: program.frontend,
-    modelConfig: program.modelConfig,
-    llamaCloudKey: program.llamaCloudKey,
-    communityProjectConfig: program.communityProjectConfig,
-    llamapack: program.llamapack,
-    vectorDb: program.vectorDb,
-    externalPort: program.externalPort,
-    postInstallAction: program.postInstallAction,
-    dataSources: program.dataSources,
-    tools: program.tools,
-    useLlamaParse: program.useLlamaParse,
-    observability: program.observability,
   });
-  conf.set("preferences", preferences);
 
-  if (program.postInstallAction === "VSCode") {
+  if (answers.postInstallAction === "VSCode") {
     console.log(`Starting VSCode in ${root}...`);
     try {
       execSync(`code . --new-window --goto README.md`, {
@@ -329,15 +352,9 @@ Please check ${cyan(
         )} for more information.`,
       );
     }
-  } else if (program.postInstallAction === "runApp") {
+  } else if (answers.postInstallAction === "runApp") {
     console.log(`Running app in ${root}...`);
-    await runApp(
-      root,
-      program.frontend,
-      program.framework,
-      program.port,
-      program.externalPort,
-    );
+    await runApp(root, answers.template, answers.framework, options.port);
   }
 }
 

llama-index-server/README.md

@@ -0,0 +1,136 @@
+# LlamaIndex Server
+
+LlamaIndexServer is a FastAPI-based application that allows you to quickly launch your [LlamaIndex Workflows](https://docs.llamaindex.ai/en/stable/module_guides/workflow/#workflows) and [Agent Workflows](https://docs.llamaindex.ai/en/stable/understanding/agent/multi_agent/) as an API server with an optional chat UI. It provides a complete environment for running LlamaIndex workflows with both API endpoints and a user interface for interaction.
+
+## Features
+
+- Serving a workflow as a chatbot
+- Built on FastAPI for high performance and easy API development
+- Optional built-in chat UI with extendable UI components
+- Prebuilt development code
+
+## Installation
+
+```bash
+pip install llama-index-server
+```
+
+## Quick Start
+
+```python
+# main.py
+from llama_index.core.agent.workflow import AgentWorkflow
+from llama_index.core.workflow import Workflow
+from llama_index.core.tools import FunctionTool
+from llama_index.server import LlamaIndexServer
+
+
+# Define a factory function that returns a Workflow or AgentWorkflow
+def create_workflow() -> Workflow:
+    def fetch_weather(city: str) -> str:
+        return f"The weather in {city} is sunny"
+
+    return AgentWorkflow.from_tools(
+        tools=[
+            FunctionTool.from_defaults(
+                fn=fetch_weather,
+            )
+        ]
+    )
+
+
+# Create an API server for the workflow
+app = LlamaIndexServer(
+    workflow_factory=create_workflow,  # Supports Workflow or AgentWorkflow
+    env="dev",  # Enable development mode
+    include_ui=True,  # Include chat UI
+    starter_questions=["What can you do?", "How do I use this?"],
+    verbose=True
+)
+```
+
+## Running the Server
+
+- In the same directory as `main.py`, run the following command to start the server:
+
+  ```bash
+  fastapi dev
+  ```
+
+- Making a request to the server:
+
+  ```bash
+  curl -X POST "http://localhost:8000/api/chat" -H "Content-Type: application/json" -d '{"message": "What is the weather in Tokyo?"}'
+  ```
+
+- See the API documentation at `http://localhost:8000/docs`
+- Access the chat UI at `http://localhost:8000/` (Make sure you set the `env="dev"` or `include_ui=True` in the server configuration)
+
+## Configuration Options
+
+The LlamaIndexServer accepts the following configuration parameters:
+
+- `workflow_factory`: A callable that creates a workflow instance for each request
+- `logger`: Optional logger instance (defaults to uvicorn logger)
+- `use_default_routers`: Whether to include default routers (chat, static file serving)
+- `env`: Environment setting ('dev' enables CORS and UI by default)
+- `include_ui`: Whether to include the chat UI
+- `component_dir`: The directory for custom UI components rendering events emitted by the workflow. The default is None, which does not render custom UI components.
+- `starter_questions`: List of starter questions for the chat UI
+- `verbose`: Enable verbose logging
+- `api_prefix`: API route prefix (default: "/api")
+- `server_url`: The deployment URL of the server (default is None)
+- `ui_path`: Path for downloaded UI static files (default: ".ui")
+
+## Default Routers and Features
+
+### Chat Router
+
+The server includes a default chat router at `/api/chat` for handling chat interactions.
+
+### Static File Serving
+
+- The server automatically mounts the `data` and `output` folders at `{server_url}{api_prefix}/files/data` (default: `/api/files/data`) and `{server_url}{api_prefix}/files/output` (default: `/api/files/output`) respectively.
+- Your workflows can use both folders to store and access files. As a convention, the `data` folder is used for documents that are ingested and the `output` folder is used for documents that are generated by the workflow.
+- The example workflows from `create-llama` (see below) are following this pattern.
+
+### Chat UI
+
+When enabled, the server provides a chat interface at the root path (`/`) with:
+
+- Configurable starter questions
+- Real-time chat interface
+- API endpoint integration
+
+### Custom UI Components
+
+You can add custom UI components for your workflow by providing `component_dir` config and adding custom .jsx or .tsx files to the directory.
+See [Custom UI Components](docs/custom_ui_component.md) for more details.
+
+## Development Mode
+
+In development mode (`env="dev"`), the server:
+
+- Enables CORS for all origins
+- Automatically includes the chat UI
+- Provides more verbose logging
+
+## API Endpoints
+
+The server provides the following default endpoints:
+
+- `/api/chat`: Chat interaction endpoint
+- `/api/files/data/*`: Access to data directory files
+- `/api/files/output/*`: Access to output directory files
+
+## Best Practices
+
+1. Always provide a workflow factory that creates fresh workflow instances
+2. Use environment variables for sensitive configuration
+3. Enable verbose logging during development
+4. Configure CORS appropriately for your deployment environment
+5. Use starter questions to guide users in the chat UI
+
+## Getting Started with a New Project
+
+Want to start a new project with LlamaIndexServer? Check out our [create-llama](https://github.com/run-llama/create-llama) tool to quickly generate a new project with LlamaIndexServer.

llama-index-server/docs/custom_ui_component.md

@@ -0,0 +1,67 @@
+# Custom UI Components
+
+The LlamaIndex server provides support for rendering workflow events using custom UI components, allowing you to extend and customize the chat interface.
+
+## Overview
+
+Custom UI components are a powerful feature that enables you to:
+
+- Add custom interface elements to the chat UI using React JSX or TSX files
+- Extend the default chat interface functionality
+- Create specialized visualizations or interactions
+
+## Configuration
+
+### Workflow events
+
+Your workflow must emit events that fit this structure, allowing the LlamaIndex server to display the right UI components based on the event type.
+
+```json
+{
+    "type": "<event_name>",
+    "data": <data model>
+}
+```
+
+In Pydantic, this is equivalent to:
+
+```python
+from pydantic import BaseModel
+from typing import Literal, Any
+
+class MyCustomEvent(BaseModel):
+    type: Literal["<my_custom_event_name>"]
+    data: dict | Any
+
+    def to_response(self):
+        return self.model_dump()
+```
+
+### Server Setup
+
+1. Initialize the LlamaIndex server with a component directory:
+
+```python
+server = LlamaIndexServer(
+    workflow_factory=your_workflow,
+    component_dir="path/to/components",
+    include_ui=True
+)
+```
+
+2. Add the custom component code to the directory following the naming pattern:
+
+   - File Extension: `.jsx` and `.tsx` for React components
+   - File Name: Should match the event type from your workflow (e.g., `deep_research_event.jsx` for handling `deep_research_event` type that you defined in your workflow). If there are TSX and JSX files with the same name, the TSX file will be used.
+   - Component Name: Export a default React component named `Component` that receives props from the event data
+
+   Example component structure:
+
+   ```jsx
+   function Component({ events }) {
+       // Your component logic here
+       return (
+           // Your UI code here
+       );
+   }
+   ```

llama-index-server/llama_index/server/__init__.py

@@ -0,0 +1,3 @@
+from .server import LlamaIndexServer
+
+__all__ = ["LlamaIndexServer"]

llama-index-server/llama_index/server/api/callbacks/__init__.py

@@ -0,0 +1,13 @@
+from llama_index.server.api.callbacks.base import EventCallback
+from llama_index.server.api.callbacks.llamacloud import LlamaCloudFileDownload
+from llama_index.server.api.callbacks.source_nodes import SourceNodesFromToolCall
+from llama_index.server.api.callbacks.suggest_next_questions import (
+    SuggestNextQuestions,
+)
+
+__all__ = [
+    "EventCallback",
+    "SourceNodesFromToolCall",
+    "SuggestNextQuestions",
+    "LlamaCloudFileDownload",
+]

llama-index-server/llama_index/server/api/callbacks/base.py

@@ -0,0 +1,31 @@
+import logging
+from abc import ABC, abstractmethod
+from typing import Any
+
+logger = logging.getLogger("uvicorn")
+
+
+class EventCallback(ABC):
+    """
+    Base class for event callbacks during event streaming.
+    """
+
+    async def run(self, event: Any) -> Any:
+        """
+        Called for each event in the stream.
+        Default behavior: pass through the event unchanged.
+        """
+        return event
+
+    async def on_complete(self, final_response: str) -> Any:
+        """
+        Called when the stream is complete.
+        Default behavior: return None.
+        """
+        return None
+
+    @abstractmethod
+    def from_default(self, *args: Any, **kwargs: Any) -> "EventCallback":
+        """
+        Create a new instance of the processor from default values.
+        """

llama-index-server/llama_index/server/api/callbacks/llamacloud.py

@@ -0,0 +1,39 @@
+import logging
+from typing import Any, List
+
+from fastapi import BackgroundTasks
+from llama_index.core.schema import NodeWithScore
+from llama_index.server.api.callbacks.base import EventCallback
+from llama_index.server.services.llamacloud.file import LlamaCloudFileService
+
+logger = logging.getLogger("uvicorn")
+
+
+class LlamaCloudFileDownload(EventCallback):
+    """
+    Processor for handling LlamaCloud file downloads from source nodes.
+    """
+
+    def __init__(self, background_tasks: BackgroundTasks) -> None:
+        self.background_tasks = background_tasks
+
+    async def run(self, event: Any) -> Any:
+        if hasattr(event, "to_response"):
+            event_response = event.to_response()
+            if event_response.get("type") == "sources" and hasattr(event, "nodes"):
+                await self._process_response_nodes(event.nodes)
+        return event
+
+    async def _process_response_nodes(self, source_nodes: List[NodeWithScore]) -> None:
+        try:
+            LlamaCloudFileService.download_files_from_nodes(
+                source_nodes, self.background_tasks
+            )
+        except ImportError:
+            pass
+
+    @classmethod
+    def from_default(
+        cls, background_tasks: BackgroundTasks
+    ) -> "LlamaCloudFileDownload":
+        return cls(background_tasks=background_tasks)

llama-index-server/llama_index/server/api/callbacks/source_nodes.py

@@ -0,0 +1,32 @@
+from typing import Any
+
+from llama_index.core.agent.workflow.workflow_events import ToolCallResult
+from llama_index.server.api.callbacks.base import EventCallback
+from llama_index.server.api.models import SourceNodesEvent
+
+
+class SourceNodesFromToolCall(EventCallback):
+    """
+    Extract source nodes from the query tool output.
+
+    Args:
+        query_tool_name: The name of the tool that queries the index.
+                         default is "query_index"
+    """
+
+    def __init__(self, query_tool_name: str = "query_index"):
+        self.query_tool_name = query_tool_name
+
+    def transform_tool_call_result(self, event: ToolCallResult) -> SourceNodesEvent:
+        source_nodes = event.tool_output.raw_output.source_nodes
+        return SourceNodesEvent(nodes=source_nodes)
+
+    async def run(self, event: Any) -> Any:
+        if isinstance(event, ToolCallResult):
+            if event.tool_name == self.query_tool_name:
+                return event, self.transform_tool_call_result(event)
+        return event
+
+    @classmethod
+    def from_default(cls, *args: Any, **kwargs: Any) -> "SourceNodesFromToolCall":
+        return cls()

llama-index-server/llama_index/server/api/callbacks/stream_handler.py

@@ -0,0 +1,69 @@
+import logging
+from typing import Any, AsyncGenerator, List, Optional
+
+from llama_index.core.workflow.handler import WorkflowHandler
+from llama_index.server.api.callbacks.base import EventCallback
+
+logger = logging.getLogger("uvicorn")
+
+
+class StreamHandler:
+    """
+    Streams events from a workflow handler through a chain of callbacks.
+    """
+
+    def __init__(
+        self,
+        workflow_handler: WorkflowHandler,
+        callbacks: Optional[List[EventCallback]] = None,
+    ):
+        self.workflow_handler = workflow_handler
+        self.callbacks = callbacks or []
+        self.accumulated_text = ""
+
+    async def cancel_run(self) -> None:
+        """Cancel the workflow handler."""
+        await self.workflow_handler.cancel_run()
+
+    async def stream_events(self) -> AsyncGenerator[Any, None]:
+        """Stream events through the processor chain."""
+        try:
+            async for event in self.workflow_handler.stream_events():
+                events_to_process = [event]
+                for callback in self.callbacks:
+                    next_events: list[Any] = []
+                    for evt in events_to_process:
+                        callback_output = await callback.run(evt)
+                        if isinstance(callback_output, (list, tuple)):
+                            next_events.extend(callback_output)
+                        elif callback_output is not None:
+                            next_events.append(callback_output)
+                    events_to_process = next_events
+
+                # Yield all processed events
+                for evt in events_to_process:
+                    yield evt
+
+            # After all events are processed, call on_complete for each callback
+            for callback in self.callbacks:
+                result = await callback.on_complete(self.accumulated_text)
+                if result:
+                    yield result
+
+        except Exception:
+            # Make sure to cancel the workflow on error
+            await self.workflow_handler.cancel_run()
+            raise
+
+    def accumulate_text(self, text: str) -> None:
+        """Accumulate text from the workflow handler."""
+        self.accumulated_text += text
+
+    @classmethod
+    def from_default(
+        cls,
+        handler: WorkflowHandler,
+        callbacks: Optional[List[EventCallback]] = None,
+    ) -> "StreamHandler":
+        """Create a new instance with the given workflow handler and callbacks."""
+        return cls(workflow_handler=handler, callbacks=callbacks)

llama-index-server/llama_index/server/api/callbacks/suggest_next_questions.py

@@ -0,0 +1,45 @@
+import logging
+from typing import Any, Optional
+
+from llama_index.server.api.callbacks.base import EventCallback
+from llama_index.server.api.models import ChatRequest
+from llama_index.server.services.suggest_next_question import (
+    SuggestNextQuestionsService,
+)
+
+logger = logging.getLogger("uvicorn")
+
+
+class SuggestNextQuestions(EventCallback):
+    """Processor for generating next question suggestions."""
+
+    def __init__(
+        self, chat_request: ChatRequest, logger: Optional[logging.Logger] = None
+    ):
+        self.chat_request = chat_request
+        self.accumulated_text = ""
+        if logger:
+            self.logger = logger
+        else:
+            self.logger = logging.getLogger("uvicorn")
+
+    async def on_complete(self, final_response: str) -> Any:
+        if final_response == "":
+            self.logger.warning(
+                "SuggestNextQuestions is enabled but final response is empty, make sure your content generator accumulates text"
+            )
+            return None
+
+        questions = await SuggestNextQuestionsService.run(
+            self.chat_request.messages, final_response
+        )
+        if questions:
+            return {
+                "type": "suggested_questions",
+                "data": questions,
+            }
+        return None
+
+    @classmethod
+    def from_default(cls, chat_request: ChatRequest) -> "SuggestNextQuestions":
+        return cls(chat_request=chat_request)

llama-index-server/llama_index/server/api/models.py

@@ -0,0 +1,142 @@
+import logging
+import os
+from enum import Enum
+from typing import Any, Dict, List, Optional
+
+from llama_index.core.schema import NodeWithScore
+from llama_index.core.types import ChatMessage, MessageRole
+from llama_index.core.workflow import Event
+from llama_index.server.settings import server_settings
+from pydantic import BaseModel, Field, field_validator
+
+logger = logging.getLogger("uvicorn")
+
+
+class ChatConfig(BaseModel):
+    next_question_suggestions: bool = Field(
+        default=True,
+        description="Whether to suggest next questions",
+    )
+
+
+class ChatAPIMessage(BaseModel):
+    role: MessageRole
+    content: str
+
+    def to_llamaindex_message(self) -> ChatMessage:
+        return ChatMessage(role=self.role, content=self.content)
+
+
+class ChatRequest(BaseModel):
+    messages: List[ChatAPIMessage]
+    data: Optional[Any] = None
+    config: Optional[ChatConfig] = ChatConfig()
+
+    @field_validator("messages")
+    def validate_messages(cls, v: List[ChatAPIMessage]) -> List[ChatAPIMessage]:
+        if v[-1].role != MessageRole.USER:
+            raise ValueError("Last message must be from user")
+        return v
+
+
+class AgentRunEventType(Enum):
+    TEXT = "text"
+    PROGRESS = "progress"
+
+
+class AgentRunEvent(Event):
+    name: str
+    msg: str
+    event_type: AgentRunEventType = AgentRunEventType.TEXT
+    data: Optional[dict] = None
+
+    def to_response(self) -> dict:
+        return {
+            "type": "agent",
+            "data": {
+                "agent": self.name,
+                "type": self.event_type.value,
+                "text": self.msg,
+                "data": self.data,
+            },
+        }
+
+
+class SourceNodesEvent(Event):
+    nodes: List[NodeWithScore]
+
+    def to_response(self) -> dict:
+        return {
+            "type": "sources",
+            "data": {
+                "nodes": [
+                    SourceNodes.from_source_node(node).model_dump()
+                    for node in self.nodes
+                ]
+            },
+        }
+
+
+class SourceNodes(BaseModel):
+    id: str
+    metadata: Dict[str, Any]
+    score: Optional[float]
+    text: str
+    url: Optional[str]
+
+    @classmethod
+    def from_source_node(cls, source_node: NodeWithScore) -> "SourceNodes":
+        metadata = source_node.node.metadata
+        url = cls.get_url_from_metadata(metadata)
+
+        return cls(
+            id=source_node.node.node_id,
+            metadata=metadata,
+            score=source_node.score,
+            text=source_node.node.text,  # type: ignore
+            url=url,
+        )
+
+    @classmethod
+    def get_url_from_metadata(
+        cls,
+        metadata: Dict[str, Any],
+        data_dir: Optional[str] = None,
+    ) -> Optional[str]:
+        url_prefix = server_settings.file_server_url_prefix
+        if data_dir is None:
+            data_dir = "data"
+        file_name = metadata.get("file_name")
+
+        if file_name and url_prefix:
+            # file_name exists and file server is configured
+            pipeline_id = metadata.get("pipeline_id")
+            if pipeline_id:
+                # file is from LlamaCloud
+                file_name = f"{pipeline_id}${file_name}"
+                return f"{url_prefix}/output/llamacloud/{file_name}"
+            is_private = metadata.get("private", "false") == "true"
+            if is_private:
+                # file is a private upload
+                return f"{url_prefix}/output/uploaded/{file_name}"
+            # file is from calling the 'generate' script
+            # Get the relative path of file_path to data_dir
+            file_path = metadata.get("file_path")
+            data_dir = os.path.abspath(data_dir)
+            if file_path and data_dir:
+                relative_path = os.path.relpath(file_path, data_dir)
+                return f"{url_prefix}/data/{relative_path}"
+        # fallback to URL in metadata (e.g. for websites)
+        return metadata.get("URL")
+
+    @classmethod
+    def from_source_nodes(
+        cls, source_nodes: List[NodeWithScore]
+    ) -> List["SourceNodes"]:
+        return [cls.from_source_node(node) for node in source_nodes]
+
+
+class ComponentDefinition(BaseModel):
+    type: str
+    code: str
+    filename: str

llama-index-server/llama_index/server/api/routers/__init__.py

@@ -0,0 +1,4 @@
+from llama_index.server.api.routers.chat import chat_router
+from llama_index.server.api.routers.ui import custom_components_router
+
+__all__ = ["chat_router", "custom_components_router"]

llama-index-server/llama_index/server/api/routers/chat.py

@@ -0,0 +1,140 @@
+import asyncio
+import inspect
+import logging
+import os
+from typing import AsyncGenerator, Callable, Union
+
+from fastapi import APIRouter, BackgroundTasks, HTTPException
+from fastapi.responses import StreamingResponse
+from llama_index.core.agent.workflow.workflow_events import AgentStream
+from llama_index.core.workflow import StopEvent, Workflow
+from llama_index.server.api.callbacks import (
+    SourceNodesFromToolCall,
+    SuggestNextQuestions,
+)
+from llama_index.server.api.callbacks.base import EventCallback
+from llama_index.server.api.callbacks.llamacloud import LlamaCloudFileDownload
+from llama_index.server.api.callbacks.stream_handler import StreamHandler
+from llama_index.server.api.models import ChatRequest
+from llama_index.server.api.utils.vercel_stream import VercelStreamResponse
+from llama_index.server.services.llamacloud import LlamaCloudFileService
+
+
+def chat_router(
+    workflow_factory: Callable[..., Workflow],
+    logger: logging.Logger,
+) -> APIRouter:
+    router = APIRouter(prefix="/chat")
+
+    @router.post("")
+    async def chat(
+        request: ChatRequest,
+        background_tasks: BackgroundTasks,
+    ) -> StreamingResponse:
+        try:
+            user_message = request.messages[-1].to_llamaindex_message()
+            chat_history = [
+                message.to_llamaindex_message() for message in request.messages[:-1]
+            ]
+            # detect if the workflow factory has chat_request as a parameter
+            factory_sig = inspect.signature(workflow_factory)
+            if "chat_request" in factory_sig.parameters:
+                workflow = workflow_factory(chat_request=request)
+            else:
+                workflow = workflow_factory()
+            workflow_handler = workflow.run(
+                user_msg=user_message.content,
+                chat_history=chat_history,
+            )
+
+            callbacks: list[EventCallback] = [
+                SourceNodesFromToolCall(),
+                LlamaCloudFileDownload(background_tasks),
+            ]
+            if request.config and request.config.next_question_suggestions:
+                callbacks.append(SuggestNextQuestions(request))
+            stream_handler = StreamHandler(
+                workflow_handler=workflow_handler,
+                callbacks=callbacks,
+            )
+
+            return VercelStreamResponse(
+                content_generator=_stream_content(stream_handler, request, logger),
+            )
+        except Exception as e:
+            logger.error(e)
+            raise HTTPException(status_code=500, detail=str(e))
+
+    if LlamaCloudFileService.is_configured():
+
+        @router.get("/config/llamacloud")
+        async def chat_llama_cloud_config() -> dict:
+            if not os.getenv("LLAMA_CLOUD_API_KEY"):
+                raise HTTPException(
+                    status_code=500, detail="LlamaCloud API KEY is not configured"
+                )
+            projects = LlamaCloudFileService.get_all_projects_with_pipelines()
+            pipeline = os.getenv("LLAMA_CLOUD_INDEX_NAME")
+            project = os.getenv("LLAMA_CLOUD_PROJECT_NAME")
+            pipeline_config = None
+            if pipeline and project:
+                pipeline_config = {
+                    "pipeline": pipeline,
+                    "project": project,
+                }
+            return {
+                "projects": projects,
+                "pipeline": pipeline_config,
+            }
+
+    return router
+
+
+async def _stream_content(
+    handler: StreamHandler,
+    request: ChatRequest,
+    logger: logging.Logger,
+) -> AsyncGenerator[str, None]:
+    async def _text_stream(
+        event: Union[AgentStream, StopEvent],
+    ) -> AsyncGenerator[str, None]:
+        if isinstance(event, AgentStream):
+            yield event.delta
+        elif isinstance(event, StopEvent):
+            if isinstance(event.result, str):
+                yield event.result
+            elif isinstance(event.result, AsyncGenerator):
+                async for chunk in event.result:
+                    if isinstance(chunk, str):
+                        yield chunk
+                    elif hasattr(chunk, "delta") and chunk.delta:
+                        yield chunk.delta
+
+    stream_started = False
+    try:
+        async for event in handler.stream_events():
+            if not stream_started:
+                # Start the stream with an empty message
+                stream_started = True
+                yield VercelStreamResponse.convert_text("")
+
+            # Handle different types of events
+            if isinstance(event, (AgentStream, StopEvent)):
+                async for chunk in _text_stream(event):
+                    handler.accumulate_text(chunk)
+                    yield VercelStreamResponse.convert_text(chunk)
+            elif isinstance(event, dict):
+                yield VercelStreamResponse.convert_data(event)
+            elif hasattr(event, "to_response"):
+                event_response = event.to_response()
+                yield VercelStreamResponse.convert_data(event_response)
+            else:
+                yield VercelStreamResponse.convert_data(event.model_dump())
+
+    except asyncio.CancelledError:
+        logger.warning("Client cancelled the request!")
+        await handler.cancel_run()
+    except Exception as e:
+        logger.error(f"Error in stream response: {e}")
+        yield VercelStreamResponse.convert_error(str(e))
+        await handler.cancel_run()

llama-index-server/llama_index/server/api/routers/ui.py

@@ -0,0 +1,20 @@
+import logging
+from typing import List
+
+from fastapi import APIRouter
+from llama_index.server.api.models import ComponentDefinition
+from llama_index.server.services.custom_ui import CustomUI
+
+
+def custom_components_router(
+    component_dir: str,
+    logger: logging.Logger,
+) -> APIRouter:
+    router = APIRouter(prefix="/components")
+
+    @router.get("")
+    async def components() -> List[ComponentDefinition]:
+        custom_ui = CustomUI(component_dir=component_dir, logger=logger)
+        return custom_ui.get_components()
+
+    return router

llama-index-server/llama_index/server/api/utils/vercel_stream.py

@@ -0,0 +1,44 @@
+import json
+import logging
+from typing import Any, AsyncGenerator, Union
+
+from fastapi.responses import StreamingResponse
+
+logger = logging.getLogger("uvicorn")
+
+
+class VercelStreamResponse(StreamingResponse):
+    """
+    Converts preprocessed events into Vercel-compatible streaming response format.
+    """
+
+    TEXT_PREFIX = "0:"
+    DATA_PREFIX = "8:"
+    ERROR_PREFIX = "3:"
+
+    def __init__(
+        self,
+        content_generator: AsyncGenerator[str, None],
+        *args: Any,
+        **kwargs: Any,
+    ):
+        super().__init__(content_generator, *args, **kwargs)
+
+    @classmethod
+    def convert_text(cls, token: str) -> str:
+        """Convert text event to Vercel format."""
+        # Escape newlines and double quotes to avoid breaking the stream
+        token = json.dumps(token)
+        return f"{cls.TEXT_PREFIX}{token}\n"
+
+    @classmethod
+    def convert_data(cls, data: Union[dict, str]) -> str:
+        """Convert data event to Vercel format."""
+        data_str = json.dumps(data) if isinstance(data, dict) else data
+        return f"{cls.DATA_PREFIX}[{data_str}]\n"
+
+    @classmethod
+    def convert_error(cls, error: str) -> str:
+        """Convert error event to Vercel format."""
+        error_str = json.dumps(error)
+        return f"{cls.ERROR_PREFIX}{error_str}\n"

llama-index-server/llama_index/server/chat_ui.py

@@ -0,0 +1,55 @@
+import logging
+import shutil
+from pathlib import Path
+from typing import Optional
+
+import requests
+
+CHAT_UI_VERSION = "0.0.9"
+
+
+def download_chat_ui(
+    logger: Optional[logging.Logger] = None, target_path: str = ".ui"
+) -> None:
+    if logger is None:
+        logger = logging.getLogger("uvicorn")
+    path = Path(target_path)
+    temp_dir = _download_package(_get_download_link(CHAT_UI_VERSION))
+    _copy_ui_files(temp_dir, path)
+    logger.info("Chat UI downloaded and copied to static folder")
+
+
+def _get_download_link(version: str) -> str:
+    """Get the download link for the chat UI from the npm registry."""
+    return f"https://registry.npmjs.org/@llamaindex/server/-/server-{version}.tgz"
+
+
+def _download_package(url: str) -> Path:
+    """Download tar.gz file and extract all files into a temporary directory."""
+    import io
+    import tarfile
+    import tempfile
+
+    response = requests.get(url, headers={"User-Agent": "Mozilla/5.0"})
+    content = response.content
+
+    temp_dir = Path(tempfile.mkdtemp())
+
+    with tarfile.open(fileobj=io.BytesIO(content), mode="r:gz") as tar:
+        tar.extractall(path=temp_dir)
+
+    return temp_dir
+
+
+def _copy_ui_files(temp_dir: Path, target_path: Path) -> None:
+    """Copy files from the .next directory to the static directory."""
+    target_path.mkdir(parents=True, exist_ok=True)
+    next_dir = temp_dir / "package/dist/static"
+
+    if next_dir.exists():
+        for item in next_dir.iterdir():
+            dest = target_path / item.name
+            if item.is_dir():
+                shutil.copytree(item, dest, dirs_exist_ok=True)
+            else:
+                shutil.copy2(item, dest)

llama-index-server/llama_index/server/server.py

@@ -0,0 +1,208 @@
+import json
+import logging
+import os
+from typing import Any, Callable, Optional
+
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.staticfiles import StaticFiles
+from llama_index.core.workflow import Workflow
+from llama_index.server.api.routers import chat_router, custom_components_router
+from llama_index.server.chat_ui import download_chat_ui
+from llama_index.server.settings import server_settings
+
+
+class LlamaIndexServer(FastAPI):
+    workflow_factory: Callable[..., Workflow]
+    include_ui: Optional[bool]
+    starter_questions: Optional[list[str]]
+    verbose: bool = False
+    ui_path: str = ".ui"
+    component_dir: Optional[str] = None
+
+    def __init__(
+        self,
+        workflow_factory: Callable[..., Workflow],
+        logger: Optional[logging.Logger] = None,
+        use_default_routers: Optional[bool] = True,
+        env: Optional[str] = None,
+        include_ui: Optional[bool] = None,
+        component_dir: Optional[str] = None,
+        starter_questions: Optional[list[str]] = None,
+        server_url: Optional[str] = None,
+        api_prefix: Optional[str] = None,
+        verbose: bool = False,
+        *args: Any,
+        **kwargs: Any,
+    ):
+        """
+        Initialize the LlamaIndexServer.
+
+        Args:
+            workflow_factory: A factory function that creates a workflow instance for each request.
+            logger: The logger to use.
+            use_default_routers: Whether to use the default routers (chat, mount `data` and `output` directories).
+            env: The environment to run the server in.
+            include_ui: Whether to show an chat UI in the root path.
+            component_dir: The directory to custom UI components code.
+            starter_questions: A list of starter questions to display in the chat UI.
+            server_url: The URL of the server.
+            api_prefix: The prefix for the API endpoints.
+            verbose: Whether to show verbose logs.
+        """
+        super().__init__(*args, **kwargs)
+
+        self.workflow_factory = workflow_factory
+        self.logger = logger or logging.getLogger("uvicorn")
+        self.verbose = verbose
+        self.include_ui = include_ui  # Store the explicitly passed value first
+        self.starter_questions = starter_questions
+        self.use_default_routers = use_default_routers or True
+        if component_dir:
+            self.component_dir = component_dir
+
+        # Update the settings
+        if server_url:
+            server_settings.set_url(server_url)
+        if api_prefix:
+            server_settings.set_api_prefix(api_prefix)
+
+        if self.use_default_routers:
+            self.add_default_routers()
+
+        if str(env).lower() == "dev":
+            self.allow_cors("*")
+            if self.include_ui is None:
+                self.include_ui = True
+
+        if self.include_ui is None:
+            self.include_ui = False
+
+        if self.include_ui:
+            self.mount_ui()
+
+    @property
+    def _ui_config(self) -> dict:
+        config = {
+            "CHAT_API": f"{server_settings.api_url}/chat",
+            "STARTER_QUESTIONS": self.starter_questions,
+        }
+        is_llamacloud_configured = os.getenv("LLAMA_CLOUD_API_KEY") is not None
+        if is_llamacloud_configured:
+            config["LLAMA_CLOUD_API"] = (
+                f"{server_settings.api_url}/chat/config/llamacloud"
+            )
+        if self.component_dir:
+            config["COMPONENTS_API"] = f"{server_settings.api_url}/components"
+        return config
+
+    # Default routers
+    def add_default_routers(self) -> None:
+        self.add_chat_router()
+        self.mount_data_dir()
+        self.mount_output_dir()
+
+    def add_chat_router(self) -> None:
+        """
+        Add the chat router.
+        """
+        self.include_router(
+            chat_router(
+                self.workflow_factory,
+                self.logger,
+            ),
+            prefix=server_settings.api_prefix,
+        )
+
+    def add_components_router(self) -> None:
+        """
+        Add the UI router.
+        """
+        if self.component_dir is None:
+            raise ValueError("component_dir must be specified to add components router")
+
+        self.include_router(
+            custom_components_router(self.component_dir, self.logger),
+            prefix=server_settings.api_prefix,
+        )
+
+    def mount_ui(self) -> None:
+        """
+        Mount the UI.
+        """
+        if self.include_ui:
+            if self.component_dir:
+                if not os.path.exists(self.component_dir):
+                    os.makedirs(self.component_dir)
+                self.add_components_router()
+            # Check if the static folder exists
+            if not os.path.exists(self.ui_path):
+                self.logger.warning(
+                    f"UI files not found, downloading UI to {self.ui_path}"
+                )
+                download_chat_ui(logger=self.logger, target_path=self.ui_path)
+            self._mount_static_files(directory=self.ui_path, path="/", html=True)
+            self._override_ui_config()
+
+    def _override_ui_config(self) -> None:
+        """
+        Override the UI config by writing a complete configuration file.
+        """
+        try:
+            config_path = os.path.join(self.ui_path, "config.js")
+            if not os.path.exists(config_path):
+                self.logger.error("Config file not found")
+                return
+            config_content = (
+                f"window.LLAMAINDEX = {json.dumps(self._ui_config, indent=2)};"
+            )
+            with open(config_path, "w") as f:
+                f.write(config_content)
+        except Exception as e:
+            self.logger.error(f"Error overriding UI config: {e}")
+
+    def mount_data_dir(self, data_dir: str = "data") -> None:
+        """
+        Mount the data directory.
+        """
+        self._mount_static_files(
+            directory=data_dir,
+            path=f"{server_settings.api_prefix}/files/data",
+            html=True,
+        )
+
+    def mount_output_dir(self, output_dir: str = "output") -> None:
+        """
+        Mount the output directory.
+        """
+        self._mount_static_files(
+            directory=output_dir,
+            path=f"{server_settings.api_prefix}/files/output",
+            html=True,
+        )
+
+    def _mount_static_files(
+        self, directory: str, path: str, html: bool = False
+    ) -> None:
+        """
+        Mount static files from a directory if it exists.
+        """
+        if os.path.exists(directory):
+            self.logger.info(f"Mounting static files '{directory}' at '{path}'")
+            self.mount(
+                path,
+                StaticFiles(directory=directory, check_dir=False, html=html),
+                name=f"{directory}-static",
+            )
+
+    def allow_cors(self, origin: str = "*") -> None:
+        """
+        Allow CORS for a specific origin.
+        """
+        self.add_middleware(
+            CORSMiddleware,
+            allow_origins=[origin],
+            allow_credentials=True,
+            allow_methods=["*"],
+            allow_headers=["*"],
+        )

llama-index-server/llama_index/server/services/custom_ui.py

@@ -0,0 +1,81 @@
+import logging
+import os
+from typing import List, Optional
+
+from llama_index.server.api.models import ComponentDefinition
+
+
+class CustomUI:
+    def __init__(
+        self, component_dir: str, logger: Optional[logging.Logger] = None
+    ) -> None:
+        self.component_dir = component_dir
+        self.logger = logger or logging.getLogger(__name__)
+
+    def get_components(self) -> List[ComponentDefinition]:
+        """
+        List all js files in the component directory and return a list of ComponentDefinition objects.
+        Ignores files that fail to load and logs the error.
+        TSX files take precedence over JSX files when duplicate component names are found.
+        """
+        components_dict: dict[str, ComponentDefinition] = {}
+        if not os.path.exists(self.component_dir):
+            self.logger.warning(
+                f"Component directory {self.component_dir} does not exist"
+            )
+            return []
+        try:
+            for file in os.listdir(self.component_dir):
+                if not file.endswith((".jsx", ".tsx")):
+                    continue
+
+                component_name = file.split(".")[0]
+                file_path = os.path.join(self.component_dir, file)
+                file_ext = os.path.splitext(file)[1]
+
+                try:
+                    with open(file_path, "r") as f:
+                        code = f.read()
+                        new_component = ComponentDefinition(
+                            type=component_name,
+                            code=code,
+                            filename=file,
+                        )
+
+                        if component_name in components_dict:
+                            existing_ext = os.path.splitext(
+                                components_dict[component_name].filename
+                            )[1]
+
+                            # If existing is TSX and new is JSX, skip and warn
+                            if existing_ext == ".tsx" and file_ext == ".jsx":
+                                self.logger.warning(
+                                    f"Skipping duplicate JSX component {file} as TSX version already exists"
+                                )
+                                continue
+
+                            # If both are same extension, warn and skip
+                            if existing_ext == file_ext:
+                                self.logger.warning(
+                                    f"Skipping duplicate component {file} with same extension"
+                                )
+                                continue
+
+                            # If existing is JSX and new is TSX, replace and warn
+                            if existing_ext == ".jsx" and file_ext == ".tsx":
+                                self.logger.warning(
+                                    f"Replacing JSX component {components_dict[component_name].filename} with TSX version {file}"
+                                )
+                                components_dict[component_name] = new_component
+                                continue
+
+                        components_dict[component_name] = new_component
+
+                except Exception as e:
+                    self.logger.error(f"Failed to load component {file}: {str(e)}")
+                    continue
+
+        except Exception as e:
+            self.logger.error(f"Error reading component directory: {str(e)}")
+
+        return list(components_dict.values())

llama-index-server/llama_index/server/services/file.py

@@ -0,0 +1,117 @@
+import logging
+import os
+import re
+import uuid
+from pathlib import Path
+from typing import List, Optional, Union
+
+from llama_index.server.settings import server_settings
+from pydantic import BaseModel, Field
+
+logger = logging.getLogger(__name__)
+
+PRIVATE_STORE_PATH = str(Path("output", "uploaded"))
+TOOL_STORE_PATH = str(Path("output", "tools"))
+LLAMA_CLOUD_STORE_PATH = str(Path("output", "llamacloud"))
+
+
+class DocumentFile(BaseModel):
+    id: str
+    name: str  # Stored file name
+    type: Optional[str] = None
+    size: Optional[int] = None
+    url: Optional[str] = None
+    path: Optional[str] = Field(
+        None,
+        description="The stored file path. Used internally in the server.",
+        exclude=True,
+    )
+    refs: Optional[List[str]] = Field(
+        None, description="The document ids in the index."
+    )
+
+
+class FileService:
+    """
+    To store the files uploaded by the user.
+    """
+
+    @classmethod
+    def save_file(
+        cls,
+        content: Union[bytes, str],
+        file_name: str,
+        save_dir: Optional[str] = None,
+    ) -> DocumentFile:
+        """
+        Save the content to a file in the local file server (accessible via URL).
+
+        Args:
+            content (bytes | str): The content to save, either bytes or string.
+            file_name (str): The original name of the file.
+            save_dir (Optional[str]): The relative path from the current working directory. Defaults to the `output/uploaded` directory.
+
+        Returns:
+            The metadata of the saved file.
+        """
+        if save_dir is None:
+            save_dir = os.path.join("output", "uploaded")
+
+        file_id = str(uuid.uuid4())
+        name, extension = os.path.splitext(file_name)
+        extension = extension.lstrip(".")
+        sanitized_name = _sanitize_file_name(name)
+        if extension == "":
+            raise ValueError("File is not supported!")
+        new_file_name = f"{sanitized_name}_{file_id}.{extension}"
+
+        file_path = os.path.join(save_dir, new_file_name)
+
+        if isinstance(content, str):
+            content = content.encode()
+
+        try:
+            os.makedirs(os.path.dirname(file_path), exist_ok=True)
+            with open(file_path, "wb") as file:
+                file.write(content)
+        except PermissionError as e:
+            logger.error(f"Permission denied when writing to file {file_path}: {e!s}")
+            raise
+        except OSError as e:
+            logger.error(f"IO error occurred when writing to file {file_path}: {e!s}")
+            raise
+        except Exception as e:
+            logger.error(f"Unexpected error when writing to file {file_path}: {e!s}")
+            raise
+
+        logger.info(f"Saved file to {file_path}")
+
+        file_size = os.path.getsize(file_path)
+        file_url = (
+            f"{server_settings.file_server_url_prefix}/{save_dir}/{new_file_name}"
+        )
+        return DocumentFile(
+            id=file_id,
+            name=new_file_name,
+            type=extension,
+            size=file_size,
+            path=file_path,
+            url=file_url,
+            refs=None,
+        )
+
+    @classmethod
+    def get_file_url(cls, file_name: str, save_dir: Optional[str] = None) -> str:
+        """
+        Get the URL of a file.
+        """
+        if save_dir is None:
+            save_dir = os.path.join("output", "uploaded")
+        return f"{server_settings.file_server_url_prefix}/{save_dir}/{file_name}"
+
+
+def _sanitize_file_name(file_name: str) -> str:
+    """
+    Sanitize the file name by replacing all non-alphanumeric characters with underscores.
+    """
+    return re.sub(r"[^a-zA-Z0-9.]", "_", file_name)

llama-index-server/llama_index/server/services/llamacloud/__init__.py

@@ -0,0 +1,11 @@
+from .file import LlamaCloudFileService
+from .generate import load_to_llamacloud
+from .index import LlamaCloudIndex, get_client, get_index
+
+__all__ = [
+    "LlamaCloudFileService",
+    "LlamaCloudIndex",
+    "get_client",
+    "get_index",
+    "load_to_llamacloud",
+]

llama-index-server/llama_index/server/services/llamacloud/file.py

@@ -0,0 +1,184 @@
+import logging
+import os
+import time
+import typing
+from io import BytesIO
+from typing import Any, Dict, List, Optional, Set, Tuple, Union
+
+import requests
+from fastapi import BackgroundTasks
+from llama_cloud import ManagedIngestionStatus, PipelineFileCreateCustomMetadataValue
+from llama_index.core.schema import NodeWithScore
+from llama_index.server.api.models import SourceNodes
+from llama_index.server.services.llamacloud.index import get_client
+from pydantic import BaseModel
+
+logger = logging.getLogger("uvicorn")
+
+
+class LlamaCloudFile(BaseModel):
+    file_name: str
+    pipeline_id: str
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, LlamaCloudFile):
+            return NotImplemented
+        return (
+            self.file_name == other.file_name and self.pipeline_id == other.pipeline_id
+        )
+
+    def __hash__(self) -> int:
+        return hash((self.file_name, self.pipeline_id))
+
+
+class LlamaCloudFileService:
+    LOCAL_STORE_PATH = "output/llamacloud"
+    DOWNLOAD_FILE_NAME_TPL = "{pipeline_id}${filename}"
+
+    @classmethod
+    def get_all_projects_with_pipelines(cls) -> List[Dict[str, Any]]:
+        try:
+            client = get_client()
+            projects = client.projects.list_projects()
+            pipelines = client.pipelines.search_pipelines()
+            return [
+                {
+                    **(project.dict()),
+                    "pipelines": [
+                        {"id": p.id, "name": p.name}
+                        for p in pipelines
+                        if p.project_id == project.id
+                    ],
+                }
+                for project in projects
+            ]
+        except Exception as error:
+            logger.error(f"Error listing projects and pipelines: {error}")
+            return []
+
+    @classmethod
+    def add_file_to_pipeline(
+        cls,
+        project_id: str,
+        pipeline_id: str,
+        upload_file: Union[typing.IO, Tuple[str, BytesIO]],
+        custom_metadata: Optional[Dict[str, PipelineFileCreateCustomMetadataValue]],
+        wait_for_processing: bool = True,
+    ) -> str:
+        client = get_client()
+        file = client.files.upload_file(project_id=project_id, upload_file=upload_file)
+        file_id = file.id
+        files = [
+            {
+                "file_id": file_id,
+                "custom_metadata": {"file_id": file_id, **(custom_metadata or {})},
+            }
+        ]
+        files = client.pipelines.add_files_to_pipeline_api(pipeline_id, request=files)
+
+        if not wait_for_processing:
+            return file_id
+
+        # Wait 2s for the file to be processed
+        max_attempts = 20
+        attempt = 0
+        while attempt < max_attempts:
+            result = client.pipelines.get_pipeline_file_status(
+                file_id=file_id, pipeline_id=pipeline_id
+            )
+            if result.status == ManagedIngestionStatus.ERROR:
+                raise Exception(f"File processing failed: {str(result)}")
+            if result.status == ManagedIngestionStatus.SUCCESS:
+                # File is ingested - return the file id
+                return file_id
+            attempt += 1
+            time.sleep(0.1)  # Sleep for 100ms
+        raise Exception(
+            f"File processing did not complete after {max_attempts} attempts."
+        )
+
+    @classmethod
+    def download_pipeline_file(
+        cls,
+        file: LlamaCloudFile,
+        force_download: bool = False,
+    ) -> None:
+        client = get_client()
+        file_name = file.file_name
+        pipeline_id = file.pipeline_id
+
+        # Check is the file already exists
+        downloaded_file_path = cls._get_file_path(file_name, pipeline_id)
+        if os.path.exists(downloaded_file_path) and not force_download:
+            logger.debug(f"File {file_name} already exists in local storage")
+            return
+        try:
+            logger.info(f"Downloading file {file_name} for pipeline {pipeline_id}")
+            files = client.pipelines.list_pipeline_files(pipeline_id)
+            if not files or not isinstance(files, list):
+                raise Exception("No files found in LlamaCloud")
+            for file_entry in files:
+                if file_entry.name == file_name:
+                    file_id = file_entry.file_id
+                    project_id = file_entry.project_id
+                    file_detail = client.files.read_file_content(
+                        file_id, project_id=project_id
+                    )
+                    cls._download_file(file_detail.url, downloaded_file_path)
+                    break
+        except Exception as error:
+            logger.info(f"Error fetching file from LlamaCloud: {error}")
+
+    @classmethod
+    def download_files_from_nodes(
+        cls, nodes: List[NodeWithScore], background_tasks: BackgroundTasks
+    ) -> None:
+        files = cls._get_files_to_download(nodes)
+        for file in files:
+            logger.info(f"Adding download of {file.file_name} to background tasks")
+            background_tasks.add_task(cls.download_pipeline_file, file)
+
+    @classmethod
+    def _get_files_to_download(cls, nodes: List[NodeWithScore]) -> Set[LlamaCloudFile]:
+        source_nodes = SourceNodes.from_source_nodes(nodes)
+        llama_cloud_files = [
+            LlamaCloudFile(
+                file_name=node.metadata.get("file_name"),  # type: ignore
+                pipeline_id=node.metadata.get("pipeline_id"),  # type: ignore
+            )
+            for node in source_nodes
+            if (
+                node.metadata.get("pipeline_id") is not None
+                and node.metadata.get("file_name") is not None
+            )
+        ]
+        # Remove duplicates and return
+        return set(llama_cloud_files)
+
+    @classmethod
+    def _get_file_name(cls, name: str, pipeline_id: str) -> str:
+        return cls.DOWNLOAD_FILE_NAME_TPL.format(pipeline_id=pipeline_id, filename=name)
+
+    @classmethod
+    def _get_file_path(cls, name: str, pipeline_id: str) -> str:
+        return os.path.join(cls.LOCAL_STORE_PATH, cls._get_file_name(name, pipeline_id))
+
+    @classmethod
+    def _download_file(cls, url: str, local_file_path: str) -> None:
+        logger.info(f"Saving file to {local_file_path}")
+        # Create directory if it doesn't exist
+        os.makedirs(cls.LOCAL_STORE_PATH, exist_ok=True)
+        # Download the file
+        with requests.get(url, stream=True) as r:
+            r.raise_for_status()
+            with open(local_file_path, "wb") as f:
+                for chunk in r.iter_content(chunk_size=8192):
+                    f.write(chunk)
+        logger.info("File downloaded successfully")
+
+    @classmethod
+    def is_configured(cls) -> bool:
+        try:
+            return os.environ.get("LLAMA_CLOUD_API_KEY") is not None
+        except Exception:
+            return False

llama-index-server/llama_index/server/services/llamacloud/generate.py

@@ -0,0 +1,56 @@
+import logging
+from typing import Optional
+
+from tqdm import tqdm
+
+from llama_index.core.readers import SimpleDirectoryReader
+from llama_index.indices.managed.llama_cloud import LlamaCloudIndex
+from llama_index.server.services.llamacloud.file import LlamaCloudFileService
+
+
+def load_to_llamacloud(
+    index: LlamaCloudIndex,
+    data_dir: Optional[str] = None,
+    recursive: Optional[bool] = None,
+    logger: Optional[logging.Logger] = None,
+) -> None:
+    if logger is None:
+        logging.basicConfig(level=logging.INFO)
+        logger = logging.getLogger()
+
+    logger.info("Generate index for the provided data")
+
+    # use SimpleDirectoryReader to retrieve the files to process
+    reader = SimpleDirectoryReader(
+        data_dir or "data",
+        recursive=recursive or True,
+    )
+    files_to_process = reader.input_files
+
+    # add each file to the LlamaCloud pipeline
+    error_files = []
+    for input_file in tqdm(
+        files_to_process,
+        desc="Processing files",
+        unit="file",
+    ):
+        with open(input_file, "rb") as f:
+            logger.debug(
+                f"Adding file {input_file} to pipeline {index.name} in project {index.project_name}"
+            )
+            try:
+                LlamaCloudFileService.add_file_to_pipeline(
+                    index.project.id,
+                    index.pipeline.id,
+                    f,
+                    custom_metadata={},
+                    wait_for_processing=False,
+                )
+            except Exception as e:
+                error_files.append(input_file)
+                logger.error(f"Error adding file {input_file}: {e}")
+
+    if error_files:
+        logger.error(f"Failed to add the following files: {error_files}")
+
+    logger.info("Finished generating the index")

llama-index-server/llama_index/server/services/llamacloud/index.py

@@ -0,0 +1,164 @@
+import logging
+import os
+from typing import TYPE_CHECKING, Any, Optional
+
+from llama_cloud import PipelineType
+from llama_index.core.callbacks import CallbackManager
+from llama_index.core.ingestion.api_utils import (
+    get_client as llama_cloud_get_client,
+)
+from llama_index.core.settings import Settings
+from llama_index.indices.managed.llama_cloud import LlamaCloudIndex
+from llama_index.server.api.models import ChatRequest
+from pydantic import BaseModel, Field, field_validator
+
+if TYPE_CHECKING:
+    from llama_cloud.client import LlamaCloud
+
+logger = logging.getLogger("uvicorn")
+
+
+class LlamaCloudConfig(BaseModel):
+    # Private attributes
+    api_key: str = Field(
+        exclude=True,  # Exclude from the model representation
+    )
+    base_url: Optional[str] = Field(
+        exclude=True,
+    )
+    organization_id: Optional[str] = Field(
+        exclude=True,
+    )
+    # Configuration attributes, can be set by the user
+    pipeline: str = Field(
+        description="The name of the pipeline to use",
+    )
+    project: str = Field(
+        description="The name of the LlamaCloud project",
+    )
+
+    def __init__(self, **kwargs: Any) -> None:
+        if "api_key" not in kwargs:
+            kwargs["api_key"] = os.getenv("LLAMA_CLOUD_API_KEY")
+        if "base_url" not in kwargs:
+            kwargs["base_url"] = os.getenv("LLAMA_CLOUD_BASE_URL")
+        if "organization_id" not in kwargs:
+            kwargs["organization_id"] = os.getenv("LLAMA_CLOUD_ORGANIZATION_ID")
+        if "pipeline" not in kwargs:
+            kwargs["pipeline"] = os.getenv("LLAMA_CLOUD_INDEX_NAME")
+        if "project" not in kwargs:
+            kwargs["project"] = os.getenv("LLAMA_CLOUD_PROJECT_NAME")
+        super().__init__(**kwargs)
+
+    # Validate and throw error if the env variables are not set before starting the app
+    @field_validator("pipeline", "project", "api_key", mode="before")
+    @classmethod
+    def validate_fields(cls, value: Any) -> Any:
+        if value is None:
+            raise ValueError(
+                "Please set LLAMA_CLOUD_INDEX_NAME, LLAMA_CLOUD_PROJECT_NAME and LLAMA_CLOUD_API_KEY"
+                " to your environment variables or config them in .env file"
+            )
+        return value
+
+    def to_client_kwargs(self) -> dict:
+        return {
+            "api_key": self.api_key,
+            "base_url": self.base_url,
+        }
+
+
+class IndexConfig(BaseModel):
+    llama_cloud_pipeline_config: LlamaCloudConfig = Field(
+        default_factory=lambda: LlamaCloudConfig(),
+        alias="llamaCloudPipeline",
+    )
+    callback_manager: Optional[CallbackManager] = Field(
+        default=None,
+    )
+
+    def to_index_kwargs(self) -> dict:
+        return {
+            "name": self.llama_cloud_pipeline_config.pipeline,
+            "project_name": self.llama_cloud_pipeline_config.project,
+            "api_key": self.llama_cloud_pipeline_config.api_key,
+            "base_url": self.llama_cloud_pipeline_config.base_url,
+            "organization_id": self.llama_cloud_pipeline_config.organization_id,
+            "callback_manager": self.callback_manager,
+        }
+
+    @classmethod
+    def from_default(cls, chat_request: Optional[ChatRequest] = None) -> "IndexConfig":
+        default_config = cls()
+        if chat_request is not None and chat_request.data is not None:
+            llamacloud_config = chat_request.data.get("llamaCloudPipeline")
+            if llamacloud_config is not None:
+                default_config.llama_cloud_pipeline_config.pipeline = llamacloud_config[
+                    "pipeline"
+                ]
+                default_config.llama_cloud_pipeline_config.project = llamacloud_config[
+                    "project"
+                ]
+        return default_config
+
+
+def get_index(
+    chat_request: Optional[ChatRequest] = None,
+    create_if_missing: bool = False,
+) -> Optional[LlamaCloudIndex]:
+    config = IndexConfig.from_default(chat_request)
+    # Check whether the index exists
+    try:
+        index = LlamaCloudIndex(**config.to_index_kwargs())
+        return index
+    except ValueError:
+        logger.warning("Index not found")
+        if create_if_missing:
+            logger.info("Creating index")
+            _create_index(config)
+            return LlamaCloudIndex(**config.to_index_kwargs())
+        return None
+
+
+def get_client() -> "LlamaCloud":
+    config = LlamaCloudConfig()
+    return llama_cloud_get_client(**config.to_client_kwargs())
+
+
+def _create_index(
+    config: IndexConfig,
+) -> None:
+    client = get_client()
+    pipeline_name = config.llama_cloud_pipeline_config.pipeline
+
+    pipelines = client.pipelines.search_pipelines(
+        pipeline_name=pipeline_name,
+        pipeline_type=PipelineType.MANAGED.value,
+    )
+    if len(pipelines) == 0:
+        from llama_index.embeddings.openai import OpenAIEmbedding
+
+        if not isinstance(Settings.embed_model, OpenAIEmbedding):
+            raise ValueError(
+                "Creating a new pipeline with a non-OpenAI embedding model is not supported."
+            )
+        client.pipelines.upsert_pipeline(
+            request={
+                "name": pipeline_name,
+                "embedding_config": {
+                    "type": "OPENAI_EMBEDDING",
+                    "component": {
+                        "api_key": os.getenv("OPENAI_API_KEY"),  # editable
+                        "model_name": Settings.embed_model.model_name
+                        or "text-embedding-3-small",
+                    },
+                },
+                "transform_config": {
+                    "mode": "auto",
+                    "config": {
+                        "chunk_size": Settings.chunk_size,  # editable
+                        "chunk_overlap": Settings.chunk_overlap,  # editable
+                    },
+                },
+            },
+        )

llama-index-server/llama_index/server/services/suggest_next_question.py

@@ -0,0 +1,95 @@
+import logging
+import os
+import re
+from typing import List, Optional, Union
+
+from llama_index.core.prompts import PromptTemplate
+from llama_index.core.settings import Settings
+from llama_index.server.api.models import ChatAPIMessage
+
+logger = logging.getLogger("uvicorn")
+
+
+class SuggestNextQuestionsService:
+    """
+    Suggest the next questions that user might ask based on the conversation history.
+    """
+
+    prompt = PromptTemplate(
+        r"""
+You're a helpful assistant! Your task is to suggest the next questions that user might interested in to keep the conversation going.
+Here is the conversation history
+---------------------
+{conversation}
+---------------------
+Given the conversation history, please give me 3 questions that user might ask next!
+Your answer should be wrapped in three sticks without any index numbers and follows the following format:
+\`\`\`
+<question 1>
+<question 2>
+<question 3>
+\`\`\`
+"""
+    )
+
+    @classmethod
+    def get_configured_prompt(cls) -> PromptTemplate:
+        prompt = os.getenv("NEXT_QUESTION_PROMPT", None)
+        if not prompt:
+            return cls.prompt
+        return PromptTemplate(prompt)
+
+    @classmethod
+    async def suggest_next_questions_all_messages(
+        cls,
+        messages: List[ChatAPIMessage],
+    ) -> Optional[List[str]]:
+        """
+        Suggest the next questions that user might ask based on the conversation history.
+        """
+        prompt_template = cls.get_configured_prompt()
+
+        try:
+            # Reduce the cost by only using the last two messages
+            last_user_message = None
+            last_assistant_message = None
+            for message in reversed(messages):
+                if message.role == "user":
+                    last_user_message = f"User: {message.content}"
+                elif message.role == "assistant":
+                    last_assistant_message = f"Assistant: {message.content}"
+                if last_user_message and last_assistant_message:
+                    break
+            conversation: str = f"{last_user_message}\n{last_assistant_message}"
+
+            # Call the LLM and parse questions from the output
+            prompt = prompt_template.format(conversation=conversation)
+            output = await Settings.llm.acomplete(prompt)
+            return cls._extract_questions(output.text)
+
+        except Exception as e:
+            logger.error(f"Error when generating next question: {e}")
+            return None
+
+    @classmethod
+    def _extract_questions(cls, text: str) -> Union[List[str], None]:
+        content_match = re.search(r"```(.*?)```", text, re.DOTALL)
+        content = content_match.group(1) if content_match else None
+        if not content:
+            return None
+        return [q.strip() for q in content.split("\n") if q.strip()]
+
+    @classmethod
+    async def run(
+        cls,
+        chat_history: List[ChatAPIMessage],
+        response: str,
+    ) -> Optional[List[str]]:
+        """
+        Suggest the next questions that user might ask based on the chat history and the last response.
+        """
+        messages = [
+            *chat_history,
+            ChatAPIMessage(role="assistant", content=response),  # type: ignore
+        ]
+        return await cls.suggest_next_questions_all_messages(messages)

llama-index-server/llama_index/server/settings.py

@@ -0,0 +1,47 @@
+from pydantic import Field, validator
+from pydantic_settings import BaseSettings
+
+
+class ServerSettings(BaseSettings):
+    url: str = Field(
+        default="",
+        description="The deployment URL of the server, to be referenced by tools and file services",
+    )
+    api_prefix: str = Field(
+        default="/api",
+        description="The prefix for the API endpoints",
+    )
+
+    @property
+    def file_server_url_prefix(self) -> str:
+        return f"{self.url}{self.api_prefix}/files"
+
+    @property
+    def api_url(self) -> str:
+        return f"{self.url}{self.api_prefix}"
+
+    @validator("url")
+    def validate_url(cls, v: str) -> str:
+        if v.endswith("/"):
+            raise ValueError("URL must not end with a '/'")
+        return v
+
+    @validator("api_prefix")
+    def validate_api_prefix(cls, v: str) -> str:
+        if not v.startswith("/"):
+            raise ValueError("API prefix must start with a '/'")
+        return v
+
+    def set_url(self, v: str) -> None:
+        self.url = v
+        self.validate_url(v)  # type: ignore
+
+    def set_api_prefix(self, v: str) -> None:
+        self.api_prefix = v
+        self.validate_api_prefix(v)  # type: ignore
+
+    class Config:
+        env_file_encoding = "utf-8"
+
+
+server_settings = ServerSettings()

llama-index-server/llama_index/server/tools/document_generator.py

@@ -0,0 +1,242 @@
+import logging
+import os
+import re
+from enum import Enum
+from io import BytesIO
+
+from llama_index.core.tools.function_tool import FunctionTool
+
+OUTPUT_DIR = "output/tools"
+
+
+class DocumentType(Enum):
+    PDF = "pdf"
+    HTML = "html"
+
+
+COMMON_STYLES = """
+body {
+    font-family: Arial, sans-serif;
+    line-height: 1.3;
+    color: #333;
+}
+h1, h2, h3, h4, h5, h6 {
+    margin-top: 1em;
+    margin-bottom: 0.5em;
+}
+p {
+    margin-bottom: 0.7em;
+}
+code {
+    background-color: #f4f4f4;
+    padding: 2px 4px;
+    border-radius: 4px;
+}
+pre {
+    background-color: #f4f4f4;
+    padding: 10px;
+    border-radius: 4px;
+    overflow-x: auto;
+}
+table {
+    border-collapse: collapse;
+    width: 100%;
+    margin-bottom: 1em;
+}
+th, td {
+    border: 1px solid #ddd;
+    padding: 8px;
+    text-align: left;
+}
+th {
+    background-color: #f2f2f2;
+    font-weight: bold;
+}
+"""
+
+HTML_SPECIFIC_STYLES = """
+body {
+    max-width: 800px;
+    margin: 0 auto;
+    padding: 20px;
+}
+"""
+
+PDF_SPECIFIC_STYLES = """
+@page {
+    size: letter;
+    margin: 2cm;
+}
+body {
+    font-size: 11pt;
+}
+h1 { font-size: 18pt; }
+h2 { font-size: 16pt; }
+h3 { font-size: 14pt; }
+h4, h5, h6 { font-size: 12pt; }
+pre, code {
+    font-family: Courier, monospace;
+    font-size: 0.9em;
+}
+"""
+
+HTML_TEMPLATE = """
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <style>
+        {common_styles}
+        {specific_styles}
+    </style>
+</head>
+<body>
+    {content}
+</body>
+</html>
+"""
+
+
+class DocumentGenerator:
+    def __init__(self, file_server_url_prefix: str):
+        if not file_server_url_prefix:
+            raise ValueError("file_server_url_prefix is required")
+        self.file_server_url_prefix = file_server_url_prefix
+
+    @classmethod
+    def _generate_html_content(cls, original_content: str) -> str:
+        """
+        Generate HTML content from the original markdown content.
+        """
+        try:
+            import markdown  # type: ignore
+        except ImportError:
+            raise ImportError(
+                "Failed to import required modules. Please install markdown."
+            )
+
+        # Convert markdown to HTML with fenced code and table extensions
+        return markdown.markdown(original_content, extensions=["fenced_code", "tables"])
+
+    @classmethod
+    def _generate_pdf(cls, html_content: str) -> BytesIO:
+        """
+        Generate a PDF from the HTML content.
+        """
+        try:
+            from xhtml2pdf import pisa
+        except ImportError:
+            raise ImportError(
+                "Failed to import required modules. Please install xhtml2pdf."
+            )
+
+        pdf_html = HTML_TEMPLATE.format(
+            common_styles=COMMON_STYLES,
+            specific_styles=PDF_SPECIFIC_STYLES,
+            content=html_content,
+        )
+
+        buffer = BytesIO()
+        pdf = pisa.pisaDocument(
+            BytesIO(pdf_html.encode("UTF-8")), buffer, encoding="UTF-8"
+        )
+
+        if pdf.err:
+            logging.error(f"PDF generation failed: {pdf.err}")
+            raise ValueError("PDF generation failed")
+
+        buffer.seek(0)
+        return buffer
+
+    @classmethod
+    def _generate_html(cls, html_content: str) -> str:
+        """
+        Generate a complete HTML document with the given HTML content.
+        """
+        return HTML_TEMPLATE.format(
+            common_styles=COMMON_STYLES,
+            specific_styles=HTML_SPECIFIC_STYLES,
+            content=html_content,
+        )
+
+    def generate_document(
+        self, original_content: str, document_type: str, file_name: str
+    ) -> str:
+        """
+        To generate document as PDF or HTML file.
+        Parameters:
+            original_content: str (markdown style)
+            document_type: str (pdf or html) specify the type of the file format based on the use case
+            file_name: str (name of the document file) must be a valid file name, no extensions needed
+        Returns:
+            str (URL to the document file): A file URL ready to serve.
+        """
+        try:
+            doc_type = DocumentType(document_type.lower())
+        except ValueError:
+            raise ValueError(
+                f"Invalid document type: {document_type}. Must be 'pdf' or 'html'."
+            )
+        # Always generate html content first
+        html_content = self._generate_html_content(original_content)
+
+        # Based on the type of document, generate the corresponding file
+        if doc_type == DocumentType.PDF:
+            content = self._generate_pdf(html_content)
+            file_extension = "pdf"
+        elif doc_type == DocumentType.HTML:
+            content = BytesIO(self._generate_html(html_content).encode("utf-8"))
+            file_extension = "html"
+        else:
+            raise ValueError(f"Unexpected document type: {document_type}")
+
+        file_name = self._validate_file_name(file_name)
+        file_path = os.path.join(OUTPUT_DIR, f"{file_name}.{file_extension}")
+
+        self._write_to_file(content, file_path)
+
+        return (
+            f"{self.file_server_url_prefix}/{OUTPUT_DIR}/{file_name}.{file_extension}"
+        )
+
+    @staticmethod
+    def _write_to_file(content: BytesIO, file_path: str) -> None:
+        """
+        Write the content to a file.
+        """
+        try:
+            os.makedirs(os.path.dirname(file_path), exist_ok=True)
+            with open(file_path, "wb") as file:
+                file.write(content.getvalue())
+        except Exception:
+            raise
+
+    @staticmethod
+    def _validate_file_name(file_name: str) -> str:
+        """
+        Validate the file name.
+        """
+        # Don't allow directory traversal
+        if os.path.isabs(file_name):
+            raise ValueError("File name is not allowed.")
+        # Don't allow special characters
+        if re.match(r"^[a-zA-Z0-9_.-]+$", file_name):
+            return file_name
+        else:
+            raise ValueError("File name is not allowed to contain special characters.")
+
+    @classmethod
+    def _validate_packages(cls) -> None:
+        try:
+            import markdown  # noqa: F401
+            import xhtml2pdf  # noqa: F401
+        except ImportError:
+            raise ImportError(
+                "Failed to import required modules. Please install markdown and xhtml2pdf "
+                "using `pip install markdown xhtml2pdf`"
+            )
+
+    def to_tool(self) -> FunctionTool:
+        self._validate_packages()
+        return FunctionTool.from_defaults(self.generate_document)

llama-index-server/llama_index/server/tools/index/__init__.py

@@ -0,0 +1,3 @@
+from .query import get_query_engine_tool
+
+__all__ = ["get_query_engine_tool"]

llama-index-server/llama_index/server/tools/index/query.py

@@ -0,0 +1,49 @@
+import os
+from typing import Any, Optional
+
+from llama_index.core.base.base_query_engine import BaseQueryEngine
+from llama_index.core.tools.query_engine import QueryEngineTool
+from llama_index.core.indices.base import BaseIndex
+
+
+def create_query_engine(index: BaseIndex, **kwargs: Any) -> BaseQueryEngine:
+    """
+    Create a query engine for the given index.
+
+    Args:
+        index: The index to create a query engine for.
+        params (optional): Additional parameters for the query engine, e.g: similarity_top_k
+    """
+    top_k = int(os.getenv("TOP_K", 0))
+    if top_k != 0 and kwargs.get("filters") is None:
+        kwargs["similarity_top_k"] = top_k
+
+    return index.as_query_engine(**kwargs)
+
+
+def get_query_engine_tool(
+    index: BaseIndex,
+    name: Optional[str] = None,
+    description: Optional[str] = None,
+    **kwargs: Any,
+) -> QueryEngineTool:
+    """
+    Get a query engine tool for the given index.
+
+    Args:
+        index: The index to create a query engine for.
+        name (optional): The name of the tool.
+        description (optional): The description of the tool.
+    """
+    if name is None:
+        name = "query_index"
+    if description is None:
+        description = (
+            "Use this tool to retrieve information about the text corpus from an index."
+        )
+    query_engine = create_query_engine(index, **kwargs)
+    return QueryEngineTool.from_defaults(
+        query_engine=query_engine,
+        name=name,
+        description=description,
+    )

llama-index-server/llama_index/server/tools/index/utils.py

@@ -0,0 +1,13 @@
+from datetime import timedelta
+
+from cachetools import TTLCache, cached  # type: ignore
+
+from llama_index.core.storage import StorageContext
+
+
+@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)

llama-index-server/llama_index/server/tools/interpreter.py

@@ -0,0 +1,216 @@
+import base64
+import logging
+import os
+import uuid
+from typing import Any, List, Optional
+
+from llama_index.core.tools import FunctionTool
+from llama_index.server.services.file import DocumentFile, FileService
+from pydantic import BaseModel
+
+logger = logging.getLogger("uvicorn")
+
+
+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"  # type: ignore # noqa: F821
+    error_message: Optional[str] = None
+    results: List[InterpreterExtraResult] = []
+    retry_count: int = 0
+
+
+class E2BCodeInterpreter:
+    output_dir = "output/tools"
+    uploaded_files_dir = "output/uploaded"
+    interpreter: Optional["Sandbox"] = None  # type: ignore # noqa: F821
+
+    def __init__(
+        self,
+        api_key: str,
+        output_dir: Optional[str] = None,
+        uploaded_files_dir: Optional[str] = None,
+    ):
+        """
+        Args:
+            api_key: The API key for the E2B Code Interpreter.
+            output_dir: The directory for the output files. Default is `output/tools`.
+            uploaded_files_dir: The directory for the files to be uploaded to the sandbox. Default is `output/uploaded`.
+        """
+        self._validate_package()
+        if not api_key:
+            raise ValueError(
+                "api_key is required to run code interpreter. Get it here: https://e2b.dev/docs/getting-started/api-key"
+            )
+        self.api_key = api_key
+        self.output_dir = output_dir or "output/tools"
+        self.uploaded_files_dir = uploaded_files_dir or "output/uploaded"
+
+    @classmethod
+    def _validate_package(cls) -> None:
+        try:
+            from e2b_code_interpreter import Sandbox  # noqa: F401
+            from e2b_code_interpreter.models import Logs  # noqa: F401
+        except ImportError:
+            raise ImportError(
+                "e2b_code_interpreter is not installed. Please install it using `pip install e2b-code-interpreter`."
+            )
+
+    def __del__(self) -> None:
+        """
+        Kill the interpreter when the tool is no longer in use.
+        """
+        if self.interpreter is not None:
+            self.interpreter.kill()
+
+    def _init_interpreter(self, sandbox_files: List[str] = []) -> None:
+        """
+        Lazily initialize the interpreter.
+        """
+        from e2b_code_interpreter import Sandbox
+
+        logger.info(f"Initializing interpreter with {len(sandbox_files)} files")
+        self.interpreter = Sandbox(api_key=self.api_key)
+        if len(sandbox_files) > 0:
+            for file_path in sandbox_files:
+                file_name = os.path.basename(file_path)
+                local_file_path = os.path.join(self.uploaded_files_dir, file_name)
+                with open(local_file_path, "rb") as f:
+                    content = f.read()
+                    if self.interpreter and self.interpreter.files:
+                        self.interpreter.files.write(file_path, content)
+            logger.info(f"Uploaded {len(sandbox_files)} files to sandbox")
+
+    def _save_to_disk(self, base64_data: str, ext: str) -> DocumentFile:
+        buffer = base64.b64decode(base64_data)
+
+        # Output from e2b doesn't have a name. Create a random name for it.
+        filename = f"e2b_file_{uuid.uuid4()}.{ext}"
+
+        return FileService.save_file(
+            buffer, file_name=filename, save_dir=self.output_dir
+        )
+
+    def _parse_result(self, result: Any) -> 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):
+                if ext in ["png", "svg", "jpeg", "pdf"]:
+                    document_file = self._save_to_disk(data, ext)
+                    output.append(
+                        InterpreterExtraResult(
+                            type=ext,
+                            filename=document_file.name,
+                            url=document_file.url,
+                        )
+                    )
+                else:
+                    # Try serialize data to string
+                    try:
+                        data = str(data)
+                    except Exception as e:
+                        data = f"Error when serializing data: {e}"
+                    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,
+        sandbox_files: List[str] = [],
+        retry_count: int = 0,
+    ) -> E2BToolOutput:
+        """
+        Execute Python code in a Jupyter notebook cell. The tool will return the result, stdout, stderr, display_data, and error.
+        If the code needs to use a file, ALWAYS pass the file path in the sandbox_files argument.
+        You have a maximum of 3 retries to get the code to run successfully.
+
+        Parameters:
+            code (str): The Python code to be executed in a single cell.
+            sandbox_files (List[str]): List of local file paths to be used by the code. The tool will throw an error if a file is not found.
+            retry_count (int): Number of times the tool has been retried.
+        """
+        from e2b_code_interpreter.models import Logs
+
+        if retry_count > 2:
+            return E2BToolOutput(
+                is_error=True,
+                logs=Logs(
+                    stdout="",
+                    stderr="",
+                    display_data="",
+                    error="",
+                ),
+                error_message="Failed to execute the code after 3 retries. Explain the error to the user and suggest a fix.",
+                retry_count=retry_count,
+            )
+
+        if self.interpreter is None:
+            self._init_interpreter(sandbox_files)
+
+        if self.interpreter:
+            logger.info(
+                f"\n{'=' * 50}\n> Running following AI-generated code:\n{code}\n{'=' * 50}"
+            )
+            exec = self.interpreter.run_code(code)
+
+            if exec.error:
+                error_message = f"The code failed to execute successfully. Error: {exec.error}. Try to fix the code and run again."
+                logger.error(error_message)
+                # Calling the generated code caused an error. Kill the interpreter and return the error to the LLM so it can try to fix the error
+                try:
+                    self.interpreter.kill()  # type: ignore
+                except Exception:
+                    pass
+                finally:
+                    self.interpreter = None
+                output = E2BToolOutput(
+                    is_error=True,
+                    logs=exec.logs,
+                    results=[],
+                    error_message=error_message,
+                    retry_count=retry_count + 1,
+                )
+            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,
+                        retry_count=retry_count + 1,
+                    )
+            return output
+        else:
+            raise ValueError("Interpreter is not initialized.")
+
+    def to_tool(self) -> FunctionTool:
+        self._validate_package()
+        return FunctionTool.from_defaults(self.interpret)

llama-index-server/llama_index/server/utils/agent_tool.py

@@ -0,0 +1,253 @@
+import logging
+import uuid
+from abc import ABC, abstractmethod
+from typing import Any, AsyncGenerator, Optional
+
+from pydantic import BaseModel, ConfigDict
+
+from llama_index.core.base.llms.types import ChatMessage, ChatResponse
+from llama_index.core.llms.function_calling import FunctionCallingLLM
+from llama_index.core.tools import (
+    BaseTool,
+    FunctionTool,
+    ToolOutput,
+    ToolSelection,
+)
+from llama_index.core.workflow import Context
+from llama_index.server.api.models import AgentRunEvent, AgentRunEventType
+from llama_index.core.agent.workflow.workflow_events import ToolCall, ToolCallResult
+
+logger = logging.getLogger("uvicorn")
+
+
+class ToolCallOutput(BaseModel):
+    tool_call_id: str
+    tool_output: ToolOutput
+
+
+class ContextAwareTool(FunctionTool, ABC):
+    @abstractmethod
+    async def acall(self, ctx: Context, input: Any) -> ToolOutput:  # type: ignore
+        pass
+
+
+class ChatWithToolsResponse(BaseModel):
+    """
+    A tool call response from chat_with_tools.
+    """
+
+    tool_calls: Optional[list[ToolSelection]]
+    tool_call_message: Optional[ChatMessage]
+    generator: Optional[AsyncGenerator[ChatResponse | None, None]]
+
+    model_config = ConfigDict(arbitrary_types_allowed=True)
+
+    def is_calling_different_tools(self) -> bool:
+        tool_names = {tool_call.tool_name for tool_call in self.tool_calls or []}
+        return len(tool_names) > 1
+
+    def has_tool_calls(self) -> bool:
+        return self.tool_calls is not None and len(self.tool_calls) > 0
+
+    def tool_name(self) -> str:
+        if not self.has_tool_calls():
+            raise ValueError("No tool calls")
+        if self.is_calling_different_tools():
+            raise ValueError("Calling different tools")
+        return self.tool_calls[0].tool_name  # type: ignore
+
+    async def full_response(self) -> str:
+        assert self.generator is not None
+        full_response = ""
+        async for chunk in self.generator:
+            content = chunk.delta  # type: ignore
+            if content:
+                full_response += content
+        return full_response
+
+
+async def chat_with_tools(  # type: ignore
+    llm: FunctionCallingLLM,
+    tools: list[BaseTool],
+    chat_history: list[ChatMessage],
+) -> ChatWithToolsResponse:
+    """
+    Request LLM to call tools or not.
+    This function doesn't change the memory.
+    """
+    generator = _tool_call_generator(llm, tools, chat_history)
+    is_tool_call = await generator.__anext__()
+    if is_tool_call:
+        # Last chunk is the full response
+        # Wait for the last chunk
+        full_response = None
+        async for chunk in generator:
+            full_response = chunk
+        assert isinstance(full_response, ChatResponse)
+        return ChatWithToolsResponse(
+            tool_calls=llm.get_tool_calls_from_response(full_response),
+            tool_call_message=full_response.message,
+            generator=None,
+        )
+    else:
+        return ChatWithToolsResponse(
+            tool_calls=None,
+            tool_call_message=None,
+            generator=generator,  # type: ignore
+        )
+
+
+async def call_tools(
+    ctx: Context,
+    agent_name: str,
+    tools: list[BaseTool],
+    tool_calls: list[ToolSelection],
+    emit_agent_events: bool = True,
+) -> list[ToolCallOutput]:
+    """
+    Call tools and return the tool call responses.
+    """
+    if len(tool_calls) == 0:
+        return []
+    tools_by_name = {tool.metadata.get_name(): tool for tool in tools}
+    if len(tool_calls) == 1:
+        if emit_agent_events:
+            ctx.write_event_to_stream(
+                AgentRunEvent(
+                    name=agent_name,
+                    msg=f"{tool_calls[0].tool_name}: {tool_calls[0].tool_kwargs}",
+                )
+            )
+        return [
+            await call_tool(ctx, tools_by_name[tool_calls[0].tool_name], tool_calls[0])
+        ]
+    # Multiple tool calls, show progress
+    tool_call_outputs: list[ToolCallOutput] = []
+
+    progress_id = str(uuid.uuid4())
+    total_steps = len(tool_calls)
+    if emit_agent_events:
+        ctx.write_event_to_stream(
+            AgentRunEvent(
+                name=agent_name,
+                msg=f"Making {total_steps} tool calls",
+            )
+        )
+    for i, tool_call in enumerate(tool_calls):
+        tool = tools_by_name.get(tool_call.tool_name)
+        if not tool:
+            tool_call_outputs.append(
+                ToolCallOutput(
+                    tool_call_id=tool_call.tool_id,
+                    tool_output=ToolOutput(
+                        is_error=True,
+                        content=f"Tool {tool_call.tool_name} does not exist",
+                        tool_name=tool_call.tool_name,
+                        raw_input=tool_call.tool_kwargs,
+                        raw_output={
+                            "error": f"Tool {tool_call.tool_name} does not exist",
+                        },
+                    ),
+                )
+            )
+            continue
+
+        tool_call_output = await call_tool(
+            ctx,
+            tool,
+            tool_call,
+        )
+        if emit_agent_events:
+            ctx.write_event_to_stream(
+                AgentRunEvent(
+                    name=agent_name,
+                    msg=f"{tool_call.tool_name}: {tool_call.tool_kwargs}",
+                    event_type=AgentRunEventType.PROGRESS,
+                    data={
+                        "id": progress_id,
+                        "total": total_steps,
+                        "current": i,
+                    },
+                )
+            )
+        tool_call_outputs.append(tool_call_output)
+    return tool_call_outputs
+
+
+async def call_tool(
+    ctx: Context,
+    tool: BaseTool,
+    tool_call: ToolSelection,
+) -> ToolCallOutput:
+    ctx.write_event_to_stream(
+        ToolCall(
+            tool_name=tool_call.tool_name,
+            tool_id=tool_call.tool_id,
+            tool_kwargs=tool_call.tool_kwargs,
+        )
+    )
+    try:
+        if isinstance(tool, ContextAwareTool):
+            if ctx is None:
+                raise ValueError("Context is required for context aware tool")
+            # inject context for calling an context aware tool
+            output = await tool.acall(ctx=ctx, **tool_call.tool_kwargs)
+        else:
+            output = await tool.acall(**tool_call.tool_kwargs)  # type: ignore
+    except Exception as e:
+        logger.error(f"Got error in tool {tool_call.tool_name}: {e!s}")
+        output = ToolOutput(
+            is_error=True,
+            content=f"Error: {e!s}",
+            tool_name=tool.metadata.get_name(),
+            raw_input=tool_call.tool_kwargs,
+            raw_output={
+                "error": str(e),
+            },
+        )
+    ctx.write_event_to_stream(
+        ToolCallResult(
+            tool_name=tool_call.tool_name,
+            tool_kwargs=tool_call.tool_kwargs,
+            tool_id=tool_call.tool_id,
+            tool_output=output,
+            return_direct=False,
+        )
+    )
+    return ToolCallOutput(
+        tool_call_id=tool_call.tool_id,
+        tool_output=output,
+    )
+
+
+async def _tool_call_generator(
+    llm: FunctionCallingLLM,
+    tools: list[BaseTool],
+    chat_history: list[ChatMessage],
+) -> AsyncGenerator[ChatResponse | bool, None]:
+    response_stream = await llm.astream_chat_with_tools(
+        tools,
+        chat_history=chat_history,
+        allow_parallel_tool_calls=False,
+    )
+
+    full_response = None
+    yielded_indicator = False
+    async for chunk in response_stream:
+        if "tool_calls" not in chunk.message.additional_kwargs:
+            # Yield a boolean to indicate whether the response is a tool call
+            if not yielded_indicator:
+                yield False
+                yielded_indicator = True
+
+            # if not a tool call, yield the chunks!
+            yield chunk  # type: ignore
+        elif not yielded_indicator:
+            # Yield the indicator for a tool call
+            yield True
+            yielded_indicator = True
+
+        full_response = chunk
+
+    if full_response:
+        yield full_response  # type: ignore

llama-index-server/pyproject.toml

@@ -0,0 +1,64 @@
+[build-system]
+build-backend = "poetry.core.masonry.api"
+requires = ["poetry-core"]
+
+[tool.codespell]
+check-filenames = true
+check-hidden = true
+# Feel free to un-skip examples, and experimental, you will just need to
+# work through many typos (--write-changes and --interactive will help)
+skip = "*.csv,*.html,*.json,*.jsonl,*.pdf,*.txt,*.ipynb"
+
+[tool.mypy]
+disallow_untyped_defs = true
+# Remove venv skip when integrated with pre-commit
+exclude = ["_static", "build", "examples", "notebooks", "venv"]
+ignore_missing_imports = true
+namespace_packages = true
+explicit_package_bases = true
+python_version = "3.10"
+
+[tool.poetry]
+authors = ["Your Name <you@example.com>"]
+description = "llama-index fastapi server"
+exclude = ["**/BUILD"]
+license = "MIT"
+name = "llama-index-server"
+packages = [{include = "llama_index/"}]
+readme = "README.md"
+version = "0.1.9"
+
+[tool.poetry.dependencies]
+python = ">=3.9,<4.0"
+fastapi = {extras = ["standard"], version = "^0.115.11"}
+cachetools = "^5.5.2"
+requests = "^2.32.3"
+pydantic-settings = "^2.8.1"
+llama-index-core = "0.12.28"
+llama-index-readers-file = "^0.4.6"
+llama-index-indices-managed-llama-cloud = "0.6.3"
+
+[tool.poetry.group.dev.dependencies]
+black = {extras = ["jupyter"], version = "<=23.9.1,>=23.7.0"}
+codespell = {extras = ["toml"], version = ">=v2.2.6"}
+e2b-code-interpreter = "^1.1.1"
+ipython = "8.10.0"
+jupyter = "^1.0.0"
+markdown = "^3.7"
+mypy = "1.15.0"
+pre-commit = "3.2.0"
+pylint = "2.15.10"
+pytest = "^8.3.5"
+pytest-asyncio = "^0.25.3"
+pytest-mock = "3.11.1"
+ruff = "0.0.292"
+tree-sitter-languages = "^1.8.0"
+types-Deprecated = ">=0.1.0"
+types-PyYAML = "^6.0.12.12"
+types-protobuf = "^4.24.0.4"
+types-redis = "4.5.5.0"
+types-requests = "2.28.11.8"  # TODO: unpin when mypy>0.991
+types-setuptools = "67.1.0.0"
+xhtml2pdf = "^0.2.17"
+pytest-cov = "^6.0.0"
+llama-cloud = "^0.1.17"

llama-index-server/tests/api/test_chat_api.py

@@ -0,0 +1,149 @@
+import logging
+from unittest.mock import AsyncMock, MagicMock
+
+import pytest
+from fastapi import FastAPI
+from httpx import ASGITransport, AsyncClient
+
+from llama_index.core.workflow import StopEvent, Workflow
+from llama_index.core.workflow.handler import WorkflowHandler
+from llama_index.server.api.models import ChatAPIMessage, ChatRequest
+from llama_index.server.api.routers.chat import chat_router
+
+
+@pytest.fixture()
+def logger():
+    return logging.getLogger("test")
+
+
+@pytest.fixture()
+def chat_request():
+    """Create a simple chat request with one user message."""
+    return ChatRequest(
+        messages=[ChatAPIMessage(role="user", content="Hello, how are you?")]
+    )
+
+
+@pytest.fixture()
+def mock_workflow():
+    """Create a mock workflow that returns a simple response."""
+    workflow = MagicMock(spec=Workflow)
+    handler = AsyncMock(spec=WorkflowHandler)
+
+    # Setup the handler to stream a simple response event
+    async def mock_stream_events():
+        yield StopEvent(result="I'm doing well, thank you for asking!")
+
+    handler.stream_events.return_value = mock_stream_events()
+    workflow.run.return_value = handler
+
+    return workflow
+
+
+@pytest.fixture()
+def workflow_factory(mock_workflow):
+    """Create a factory function that returns our mock workflow."""
+
+    def factory(verbose=False):
+        return mock_workflow
+
+    return factory
+
+
+@pytest.mark.asyncio()
+async def test_chat_router(chat_request, workflow_factory, logger):
+    """Test that the chat router handles a request correctly."""
+    # Create a FastAPI app and mount our router
+    app = FastAPI()
+    router = chat_router(workflow_factory, logger)
+    app.include_router(router)
+
+    # Make a request to the chat endpoint
+    async with AsyncClient(
+        transport=ASGITransport(app=app), base_url="http://test"
+    ) as client:
+        response = await client.post("/chat", json=chat_request.model_dump())
+
+        # Check response status
+        assert response.status_code == 200
+
+        # For streaming responses we don't check the content-type header directly
+        # Instead, check that we get the expected content in the response body
+
+        # The response is a stream, so we need to collect the chunks
+        content = response.content.decode()
+
+        # Verify content structure follows expected format
+        assert "0:" in content  # Text prefix for VercelStreamResponse
+        # Verify if the response contains the expected message
+        assert "I'm doing well" in content
+
+        # Verify the mock workflow was called correctly
+        mock_workflow = workflow_factory()
+        mock_workflow.run.assert_called_once()
+
+        # Verify the workflow was called with the correct arguments
+        call_args = mock_workflow.run.call_args[1]
+        assert call_args["user_msg"] == "Hello, how are you?"
+        assert isinstance(call_args["chat_history"], list)
+        assert len(call_args["chat_history"]) == 0  # No history for first message
+
+
+@pytest.mark.asyncio()
+async def test_chat_with_agent_workflow(logger):
+    """Test that the chat router works with a workflow that mimics an agent workflow."""
+    # Create a simple workflow that mimics an agent workflow
+    mock_workflow = MagicMock(spec=Workflow)
+    handler = AsyncMock(spec=WorkflowHandler)
+
+    # Setup the handler to stream a simple response about weather
+    async def mock_stream_events():
+        yield StopEvent(
+            result="The weather in New York is sunny. I used the weather tool to get this information."
+        )
+
+    handler.stream_events.return_value = mock_stream_events()
+    mock_workflow.run.return_value = handler
+
+    # Create a factory function that returns our mock workflow
+    def workflow_factory(verbose=False):
+        return mock_workflow
+
+    # Create a FastAPI app and mount our router
+    app = FastAPI()
+    router = chat_router(workflow_factory, logger)
+    app.include_router(router)
+
+    # Create a chat request asking about weather
+    chat_request = ChatRequest(
+        messages=[
+            ChatAPIMessage(role="user", content="What's the weather in New York?")
+        ]
+    )
+
+    # Make a request to the chat endpoint
+    async with AsyncClient(
+        transport=ASGITransport(app=app), base_url="http://test"
+    ) as client:
+        response = await client.post("/chat", json=chat_request.model_dump())
+
+        # Check response status
+        assert response.status_code == 200
+
+        # The response is a stream, so we need to collect the chunks
+        content = response.content.decode()
+
+        # Verify content structure follows expected format
+        assert "0:" in content  # Text prefix for VercelStreamResponse
+
+        # Verify the response content contains expected keywords
+        assert "weather" in content and "New York" in content and "sunny" in content
+
+        # Verify the mock workflow was called correctly
+        mock_workflow.run.assert_called_once()
+
+        # Verify the workflow was called with the correct arguments
+        call_args = mock_workflow.run.call_args[1]
+        assert call_args["user_msg"] == "What's the weather in New York?"
+        assert isinstance(call_args["chat_history"], list)
+        assert len(call_args["chat_history"]) == 0  # No history for first message

llama-index-server/tests/api/test_event_stream.py

@@ -0,0 +1,249 @@
+import asyncio
+import logging
+from unittest.mock import AsyncMock, MagicMock
+
+import pytest
+
+from llama_index.core.agent.workflow.workflow_events import AgentStream
+from llama_index.core.workflow import StopEvent
+from llama_index.core.workflow.handler import WorkflowHandler
+from llama_index.server.api.models import ChatAPIMessage, ChatRequest
+from llama_index.server.api.routers.chat import _stream_content
+from llama_index.server.api.utils.vercel_stream import VercelStreamResponse
+
+
+@pytest.fixture()
+def logger():
+    return logging.getLogger("test")
+
+
+@pytest.fixture()
+def chat_request():
+    return ChatRequest(messages=[ChatAPIMessage(role="user", content="test message")])
+
+
+@pytest.fixture()
+def mock_workflow_handler():
+    handler = AsyncMock(spec=WorkflowHandler)
+    handler.accumulate_text = MagicMock()
+    return handler
+
+
+class TestEventStream:
+    @pytest.mark.asyncio()
+    async def test_stream_content_with_agent_stream(
+        self, mock_workflow_handler, chat_request, logger
+    ):
+        # Setup
+        mock_workflow_handler.stream_events.return_value = (
+            self._mock_agent_stream_events()
+        )
+
+        # Execute
+        result = [
+            chunk
+            async for chunk in _stream_content(
+                mock_workflow_handler, chat_request, logger
+            )
+        ]
+
+        # Assert
+        assert len(result) == 3  # Empty start + 2 text chunks
+        assert result[0] == VercelStreamResponse.convert_text("")
+        assert result[1] == VercelStreamResponse.convert_text("Hello")
+        assert result[2] == VercelStreamResponse.convert_text(" World")
+
+    @pytest.mark.asyncio()
+    async def test_stream_content_with_stop_event_string(
+        self, mock_workflow_handler, chat_request, logger
+    ):
+        # Setup
+        mock_workflow_handler.stream_events.return_value = (
+            self._mock_stop_event_string()
+        )
+
+        # Execute
+        result = [
+            chunk
+            async for chunk in _stream_content(
+                mock_workflow_handler, chat_request, logger
+            )
+        ]
+
+        # Assert
+        assert len(result) == 2  # Empty start + result string
+        assert result[0] == VercelStreamResponse.convert_text("")
+        assert result[1] == VercelStreamResponse.convert_text("Final answer")
+
+    @pytest.mark.asyncio()
+    async def test_stream_content_with_stop_event_delta_objects(
+        self, mock_workflow_handler, chat_request, logger
+    ):
+        # Setup
+        mock_workflow_handler.stream_events.return_value = (
+            self._mock_stop_event_delta_objects()
+        )
+
+        # Execute
+        result = [
+            chunk
+            async for chunk in _stream_content(
+                mock_workflow_handler, chat_request, logger
+            )
+        ]
+
+        # Assert
+        assert len(result) == 3  # Empty start + 2 delta chunks
+        assert result[0] == VercelStreamResponse.convert_text("")
+        assert result[1] == VercelStreamResponse.convert_text("Delta 1")
+        assert result[2] == VercelStreamResponse.convert_text("Delta 2")
+
+    @pytest.mark.asyncio()
+    async def test_stream_content_with_event_with_to_response(
+        self, mock_workflow_handler, chat_request, logger
+    ):
+        # Setup
+        mock_workflow_handler.stream_events.return_value = (
+            self._mock_event_with_to_response()
+        )
+
+        # Execute
+        result = [
+            chunk
+            async for chunk in _stream_content(
+                mock_workflow_handler, chat_request, logger
+            )
+        ]
+
+        # Assert
+        assert len(result) == 2  # Empty start + event with to_response
+        assert result[0] == VercelStreamResponse.convert_text("")
+        assert result[1] == VercelStreamResponse.convert_data({"event_type": "test"})
+
+    @pytest.mark.asyncio()
+    async def test_stream_content_with_event_with_model_dump(
+        self, mock_workflow_handler, chat_request, logger
+    ):
+        # Setup
+        mock_workflow_handler.stream_events.return_value = (
+            self._mock_event_with_model_dump()
+        )
+
+        # Execute
+        result = [
+            chunk
+            async for chunk in _stream_content(
+                mock_workflow_handler, chat_request, logger
+            )
+        ]
+
+        # Assert
+        assert len(result) == 2  # Empty start + event with model_dump
+        assert result[0] == VercelStreamResponse.convert_text("")
+        assert result[1] == VercelStreamResponse.convert_data(None)
+
+    @pytest.mark.asyncio()
+    async def test_stream_content_with_cancelled_error(
+        self, mock_workflow_handler, chat_request, logger
+    ):
+        # Setup
+        mock_workflow_handler.stream_events.side_effect = asyncio.CancelledError()
+        logger.warning = MagicMock()
+
+        # Execute
+        result = [
+            chunk
+            async for chunk in _stream_content(
+                mock_workflow_handler, chat_request, logger
+            )
+        ]
+
+        # Assert
+        assert len(result) == 0
+        mock_workflow_handler.cancel_run.assert_called_once()
+        logger.warning.assert_called_once()
+
+    @pytest.mark.asyncio()
+    async def test_stream_content_with_exception(
+        self, mock_workflow_handler, chat_request, logger
+    ):
+        # Setup
+        error_message = "Test error"
+        mock_workflow_handler.stream_events.side_effect = Exception(error_message)
+        logger.error = MagicMock()
+
+        # Execute
+        result = [
+            chunk
+            async for chunk in _stream_content(
+                mock_workflow_handler, chat_request, logger
+            )
+        ]
+
+        # Assert
+        assert len(result) == 1
+        assert result[0] == VercelStreamResponse.convert_error(error_message)
+        mock_workflow_handler.cancel_run.assert_called_once()
+        logger.error.assert_called_once()
+
+    async def _mock_agent_stream_events(self):
+        yield AgentStream(
+            delta="Hello", response="", current_agent_name="", tool_calls=[], raw=""
+        )
+        yield AgentStream(
+            delta=" World", response="", current_agent_name="", tool_calls=[], raw=""
+        )
+
+    async def _mock_agent_stream_with_empty_deltas(self):
+        yield AgentStream(
+            delta="   ",  # Empty delta with spaces - should be filtered
+            response="",
+            current_agent_name="",
+            tool_calls=[],
+            raw="",
+        )
+        yield AgentStream(
+            delta="Valid delta",
+            response="",
+            current_agent_name="",
+            tool_calls=[],
+            raw="",
+        )
+        yield AgentStream(
+            delta="\n",  # Newline-only delta - should be filtered
+            response="",
+            current_agent_name="",
+            tool_calls=[],
+            raw="",
+        )
+
+    async def _mock_stop_event_string(self):
+        yield StopEvent(result="Final answer")
+
+    async def _mock_stop_event_delta_objects(self):
+        async def generator():
+            # Create proper objects with delta attribute that can be serialized
+            class ObjectWithDelta:
+                def __init__(self, delta_value) -> None:
+                    self.delta = delta_value
+
+            yield ObjectWithDelta("Delta 1")
+            yield ObjectWithDelta("Delta 2")
+
+        yield StopEvent(result=generator())
+
+    async def _mock_dict_event(self):
+        yield {"key": "value"}
+
+    async def _mock_event_with_to_response(self):
+        event = MagicMock()
+        event.to_response.return_value = {"event_type": "test"}
+        yield event
+
+    async def _mock_event_with_model_dump(self):
+        event = MagicMock()
+        event.model_dump.return_value = {"name": "test_event"}
+        # Override to_response to return None - this means convert_data(None) will be called
+        event.to_response = MagicMock(return_value=None)
+        # The model_dump value is ignored when to_response returns None
+        yield event

llama-index-server/tests/services/test_file_service.py

@@ -0,0 +1,205 @@
+import os
+import uuid
+from unittest.mock import mock_open, patch
+
+import pytest
+from llama_index.server.services.file import FileService, _sanitize_file_name
+
+
+class TestFileService:
+    def test_sanitize_file_name(self):
+        # Test with normal alphanumeric name
+        assert _sanitize_file_name("test123") == "test123"
+
+        # Test with spaces
+        assert _sanitize_file_name("test file") == "test_file"
+
+        # Test with special characters
+        assert _sanitize_file_name("test@file!name") == "test_file_name"
+
+        # Test with path-like characters
+        assert _sanitize_file_name("test/file/name") == "test_file_name"
+
+        # Test with dots (should be preserved)
+        assert _sanitize_file_name("test.file.name") == "test.file.name"
+
+    @patch("uuid.uuid4")
+    @patch("os.path.getsize")
+    @patch("builtins.open", new_callable=mock_open)
+    @patch("os.makedirs")
+    def test_save_file_string_content(
+        self, mock_makedirs, mock_file_open, mock_getsize, mock_uuid
+    ):
+        # Setup
+        test_uuid = "12345678-1234-5678-1234-567812345678"
+        mock_uuid.return_value = uuid.UUID(test_uuid)
+        mock_getsize.return_value = 11  # Length of "Hello World"
+
+        # Execute
+        result = FileService.save_file(
+            content="Hello World", file_name="test.txt", save_dir="test_dir"
+        )
+
+        # Assert
+        expected_path = os.path.join("test_dir", f"test_{test_uuid}.txt")
+        mock_makedirs.assert_called_once_with(
+            os.path.dirname(expected_path), exist_ok=True
+        )
+        mock_file_open.assert_called_once_with(expected_path, "wb")
+        mock_file_open().write.assert_called_once_with(b"Hello World")
+
+        assert result.id == test_uuid
+        assert result.name == f"test_{test_uuid}.txt"
+        assert result.type == "txt"
+        assert result.size == 11
+        assert result.path == expected_path
+        assert result.url.endswith(expected_path.replace(os.path.sep, "/"))
+        assert result.refs is None
+
+    @patch("uuid.uuid4")
+    @patch("os.path.getsize")
+    @patch("builtins.open", new_callable=mock_open)
+    @patch("os.makedirs")
+    def test_save_file_bytes_content(
+        self, mock_makedirs, mock_file_open, mock_getsize, mock_uuid
+    ):
+        # Setup
+        test_uuid = "12345678-1234-5678-1234-567812345678"
+        mock_uuid.return_value = uuid.UUID(test_uuid)
+        mock_getsize.return_value = 11  # Length of "Hello World"
+
+        # Execute
+        result = FileService.save_file(
+            content=b"Hello World", file_name="test.txt", save_dir="test_dir"
+        )
+
+        # Assert
+        expected_path = os.path.join("test_dir", f"test_{test_uuid}.txt")
+        mock_makedirs.assert_called_once_with(
+            os.path.dirname(expected_path), exist_ok=True
+        )
+        mock_file_open.assert_called_once_with(expected_path, "wb")
+        mock_file_open().write.assert_called_once_with(b"Hello World")
+        assert result.path == expected_path
+        assert result.type == "txt"
+
+    @patch("uuid.uuid4")
+    @patch("os.path.getsize")
+    @patch("builtins.open", new_callable=mock_open)
+    @patch("os.makedirs")
+    def test_save_file_with_special_characters(
+        self, mock_makedirs, mock_file_open, mock_getsize, mock_uuid
+    ):
+        # Setup
+        test_uuid = "12345678-1234-5678-1234-567812345678"
+        mock_uuid.return_value = uuid.UUID(test_uuid)
+        mock_getsize.return_value = 11
+
+        # Execute
+        result = FileService.save_file(
+            content="Hello World", file_name="test@file!.txt", save_dir="test_dir"
+        )
+
+        # Assert
+        expected_path = os.path.join("test_dir", f"test_file__{test_uuid}.txt")
+        mock_makedirs.assert_called_once_with(
+            os.path.dirname(expected_path), exist_ok=True
+        )
+        mock_file_open.assert_called_once_with(expected_path, "wb")
+        assert result.path == expected_path
+        assert result.name == f"test_file__{test_uuid}.txt"
+
+    @patch("uuid.uuid4")
+    @patch("os.path.getsize")
+    @patch("builtins.open", new_callable=mock_open)
+    @patch("os.makedirs")
+    def test_save_file_default_directory(
+        self, mock_makedirs, mock_file_open, mock_getsize, mock_uuid
+    ):
+        # Setup
+        test_uuid = "12345678-1234-5678-1234-567812345678"
+        mock_uuid.return_value = uuid.UUID(test_uuid)
+        mock_getsize.return_value = 11
+
+        # Execute
+        result = FileService.save_file(content="Hello World", file_name="test.txt")
+
+        # Assert
+        expected_path = os.path.join("output", "uploaded", f"test_{test_uuid}.txt")
+        mock_makedirs.assert_called_once_with(
+            os.path.dirname(expected_path), exist_ok=True
+        )
+        assert result.path == expected_path
+
+    @patch("uuid.uuid4")
+    @patch("os.getenv")
+    @patch("os.path.getsize")
+    @patch("builtins.open", new_callable=mock_open)
+    @patch("os.makedirs")
+    def test_save_file_custom_url_prefix(
+        self, mock_makedirs, mock_file_open, mock_getsize, mock_getenv, mock_uuid
+    ):
+        # Setup
+        test_uuid = "12345678-1234-5678-1234-567812345678"
+        mock_uuid.return_value = uuid.UUID(test_uuid)
+        mock_getsize.return_value = 11
+        mock_getenv.return_value = "/api/files"
+
+        # Execute
+        result = FileService.save_file(
+            content="Hello World", file_name="test.txt", save_dir="test_dir"
+        )
+
+        # Assert
+        expected_path = os.path.join("test_dir", f"test_{test_uuid}.txt")
+        mock_makedirs.assert_called_once_with(
+            os.path.dirname(expected_path), exist_ok=True
+        )
+        mock_file_open.assert_called_once_with(expected_path, "wb")
+        assert result.path == expected_path
+        # URL paths must use forward slashes, even on Windows
+        expected_url = f"/api/files/test_dir/test_{test_uuid}.txt"
+        assert result.url == expected_url
+
+    def test_save_file_no_extension(self):
+        # Test that saving a file without extension raises ValueError
+        with pytest.raises(ValueError, match="File is not supported!"):
+            FileService.save_file(
+                content="Hello World", file_name="test", save_dir="test_dir"
+            )
+
+    @patch("uuid.uuid4")
+    @patch("os.path.getsize")
+    @patch("builtins.open")
+    @patch("os.makedirs")
+    def test_save_file_permission_error(
+        self, mock_makedirs, mock_file_open, mock_getsize, mock_uuid
+    ):
+        # Setup
+        test_uuid = "12345678-1234-5678-1234-567812345678"
+        mock_uuid.return_value = uuid.UUID(test_uuid)
+        mock_file_open.side_effect = PermissionError("Permission denied")
+
+        # Execute and Assert
+        with pytest.raises(PermissionError):
+            FileService.save_file(
+                content="Hello World", file_name="test.txt", save_dir="test_dir"
+            )
+
+    @patch("uuid.uuid4")
+    @patch("os.path.getsize")
+    @patch("builtins.open")
+    @patch("os.makedirs")
+    def test_save_file_io_error(
+        self, mock_makedirs, mock_file_open, mock_getsize, mock_uuid
+    ):
+        # Setup
+        test_uuid = "12345678-1234-5678-1234-567812345678"
+        mock_uuid.return_value = uuid.UUID(test_uuid)
+        mock_file_open.side_effect = OSError("IO Error")
+
+        # Execute and Assert
+        with pytest.raises(IOError):
+            FileService.save_file(
+                content="Hello World", file_name="test.txt", save_dir="test_dir"
+            )

llama-index-server/tests/test_llamaindex_server.py

@@ -0,0 +1,201 @@
+import pytest
+from httpx import ASGITransport, AsyncClient
+from llama_index.core.agent.workflow import AgentWorkflow
+from llama_index.core.llms import MockLLM
+from llama_index.server import LlamaIndexServer
+
+
+def fetch_weather(city: str) -> str:
+    """Fetch the weather for a given city."""
+    return f"The weather in {city} is sunny."
+
+
+def _agent_workflow() -> AgentWorkflow:
+    # Use MockLLM instead of default OpenAI
+    mock_llm = MockLLM()
+    return AgentWorkflow.from_tools_or_functions(
+        tools_or_functions=[fetch_weather],
+        verbose=True,
+        llm=mock_llm,
+    )
+
+
+@pytest.fixture()
+def server() -> LlamaIndexServer:
+    """Fixture to create a LlamaIndexServer instance."""
+    return LlamaIndexServer(
+        workflow_factory=_agent_workflow,
+        verbose=True,
+        use_default_routers=True,
+        mount_ui=False,
+        env="dev",
+    )
+
+
+@pytest.mark.asyncio()
+async def test_server_has_chat_route(server: LlamaIndexServer) -> None:
+    """Test that the server has the chat API route."""
+    chat_route_exists = any(route.path == "/api/chat" for route in server.routes)
+    assert chat_route_exists, "Chat API route not found in server routes"
+
+
+@pytest.mark.asyncio()
+async def test_server_swagger_docs(server: LlamaIndexServer) -> None:
+    """Test that the server serves Swagger UI docs."""
+    async with AsyncClient(
+        transport=ASGITransport(app=server), base_url="http://test"
+    ) as ac:
+        response = await ac.get("/docs")
+        assert response.status_code == 200
+        assert "text/html" in response.headers["content-type"]
+        assert "Swagger UI" in response.text
+
+
+@pytest.mark.asyncio()
+async def test_ui_is_downloaded(server: LlamaIndexServer) -> None:
+    """
+    Test if the UI is downloaded and mounted correctly.
+    """
+    import os
+    import shutil
+
+    # Clean up any existing static directory first
+    if os.path.exists(".ui"):
+        shutil.rmtree(".ui")
+
+    # Create a new server with UI enabled
+    ui_server = LlamaIndexServer(
+        workflow_factory=_agent_workflow,
+        verbose=True,
+        use_default_routers=True,
+        env="dev",
+        include_ui=True,
+    )
+
+    # Verify that static directory was created with index.html
+    assert os.path.exists("./.ui"), "Static directory was not created"
+    assert os.path.isdir("./.ui"), "Static path is not a directory"
+    assert os.path.exists("./.ui/index.html"), "index.html was not downloaded"
+
+    # Check if the UI is mounted and accessible
+    async with AsyncClient(
+        transport=ASGITransport(app=ui_server), base_url="http://test"
+    ) as ac:
+        response = await ac.get("/")
+        assert response.status_code == 200
+        assert "text/html" in response.headers["content-type"]
+
+    # Clean up after test
+    shutil.rmtree("./.ui")
+
+
+@pytest.mark.asyncio()
+async def test_ui_is_accessible(server: LlamaIndexServer) -> None:
+    """
+    Test if the UI is accessible.
+    """
+    # Manually trigger UI mounting
+    server.mount_ui()
+
+    async with AsyncClient(
+        transport=ASGITransport(app=server), base_url="http://test"
+    ) as ac:
+        response = await ac.get("/")
+        assert response.status_code == 200
+        assert "text/html" in response.headers["content-type"]
+
+
+@pytest.mark.asyncio()
+async def test_component_dir_creation(server: LlamaIndexServer) -> None:
+    """
+    Test if the component directory is created when specified and doesn't exist.
+    """
+    import os
+    import shutil
+
+    test_component_dir = "./test_components"
+
+    # Clean up any existing directory
+    if os.path.exists(test_component_dir):
+        shutil.rmtree(test_component_dir)
+
+    # Create server with component directory
+    _ = LlamaIndexServer(
+        workflow_factory=_agent_workflow,
+        verbose=True,
+        component_dir=test_component_dir,
+        include_ui=True,
+    )
+
+    # Verify directory was created
+    assert os.path.exists(test_component_dir), "Component directory was not created"
+    assert os.path.isdir(test_component_dir), "Component path is not a directory"
+
+    # Clean up after test
+    shutil.rmtree(test_component_dir)
+
+
+@pytest.mark.asyncio()
+async def test_component_router_addition(server: LlamaIndexServer, tmp_path) -> None:
+    """
+    Test if the component router is added when component directory is specified.
+    """
+    test_component_dir = tmp_path / "test_components"
+
+    # Create server with component directory
+    component_server = LlamaIndexServer(
+        workflow_factory=_agent_workflow,
+        verbose=True,
+        component_dir=str(test_component_dir),
+        include_ui=True,
+    )
+
+    # Verify component route exists
+    component_route_exists = any(
+        route.path == "/api/components" for route in component_server.routes
+    )
+    assert component_route_exists, "Component API route not found in server routes"
+
+
+@pytest.mark.asyncio()
+async def test_ui_config_includes_components_api(
+    server: LlamaIndexServer, tmp_path
+) -> None:
+    """
+    Test if the UI config includes components API when component directory is set.
+    """
+    test_component_dir = tmp_path / "test_components"
+
+    # Create server with component directory
+    component_server = LlamaIndexServer(
+        workflow_factory=_agent_workflow,
+        verbose=True,
+        component_dir=str(test_component_dir),
+        include_ui=True,
+    )
+
+    # Check if components API is in UI config
+    ui_config = component_server._ui_config
+    assert "COMPONENTS_API" in ui_config, "Components API not found in UI config"
+    assert ui_config["COMPONENTS_API"].endswith("/components"), (
+        "Incorrect components API path"
+    )
+
+
+@pytest.mark.asyncio()
+async def test_component_router_requires_component_dir(
+    server: LlamaIndexServer,
+) -> None:
+    """
+    Test that adding components router without component_dir raises an error.
+    """
+    server_without_component_dir = LlamaIndexServer(
+        workflow_factory=_agent_workflow,
+        verbose=True,
+        include_ui=True,
+    )
+
+    with pytest.raises(
+        ValueError, match="component_dir must be specified to add components router"
+    ):
+        server_without_component_dir.add_components_router()

llama-index-server/tests/tools/test_document_generator.py

@@ -0,0 +1,89 @@
+from io import BytesIO
+from unittest.mock import MagicMock, patch
+
+import pytest
+from llama_index.server.tools.document_generator import (
+    OUTPUT_DIR,
+    DocumentGenerator,
+)
+
+
+class TestDocumentGenerator:
+    def test_validate_file_name(self) -> None:
+        # Valid names
+        assert (
+            DocumentGenerator("/api/files")._validate_file_name("valid-name")
+            == "valid-name"
+        )
+
+        # Invalid names
+        with pytest.raises(ValueError):
+            DocumentGenerator("/api/files")._validate_file_name("/invalid/path")
+
+    @patch("os.makedirs")
+    @patch("builtins.open")
+    def test_write_to_file(self, mock_open, mock_makedirs):  # type: ignore
+        content = BytesIO(b"test")
+        DocumentGenerator("/api/files")._write_to_file(content, "path/file.txt")
+
+        mock_makedirs.assert_called_once()
+        mock_open.assert_called_once()
+        mock_open.return_value.__enter__.return_value.write.assert_called_once_with(
+            b"test"
+        )
+
+    @patch("markdown.markdown")
+    def test_html_generation(self, mock_markdown):  # type: ignore
+        mock_markdown.return_value = "<h1>Test</h1>"
+
+        # Test HTML content generation
+        assert (
+            DocumentGenerator("/api/files")._generate_html_content("# Test")
+            == "<h1>Test</h1>"
+        )
+
+        # Test full HTML generation
+        html = DocumentGenerator("/api/files")._generate_html("<h1>Test</h1>")
+        assert "<!DOCTYPE html>" in html
+        assert "<h1>Test</h1>" in html
+
+    @patch("xhtml2pdf.pisa.pisaDocument")
+    def test_pdf_generation(self, mock_pisa):  # type: ignore
+        # Success case
+        mock_pisa.return_value = MagicMock(err=None)
+        assert isinstance(
+            DocumentGenerator("/api/files")._generate_pdf("test"), BytesIO
+        )
+
+        # Error case
+        mock_pisa.return_value = MagicMock(err="Error")
+        with pytest.raises(ValueError):
+            DocumentGenerator("/api/files")._generate_pdf("test")
+
+    @patch.multiple(
+        DocumentGenerator,
+        _generate_html_content=MagicMock(return_value="<h1>Test</h1>"),
+        _generate_html=MagicMock(
+            return_value="<html><body><h1>Test</h1></body></html>"
+        ),
+        _generate_pdf=MagicMock(return_value=BytesIO(b"pdf")),
+        _write_to_file=MagicMock(),
+    )
+    def test_generate_document(self):  # type: ignore
+        # HTML generation
+        url = DocumentGenerator("/api/files").generate_document(
+            "# Test", "html", "test-doc"
+        )
+        assert url == f"/api/files/{OUTPUT_DIR}/test-doc.html"
+
+        # PDF generation
+        url = DocumentGenerator("/api/files").generate_document(
+            "# Test", "pdf", "test-doc"
+        )
+        assert url == f"/api/files/{OUTPUT_DIR}/test-doc.pdf"
+
+        # Invalid type
+        with pytest.raises(ValueError):
+            DocumentGenerator("/api/files").generate_document(
+                "# Test", "invalid", "test-doc"
+            )

llama-index-server/tests/tools/test_interpreter.py

@@ -0,0 +1,65 @@
+from unittest.mock import MagicMock
+
+import pytest
+from e2b_code_interpreter.models import Execution, Logs
+from llama_index.server.tools.interpreter import E2BCodeInterpreter
+
+
+class TestE2BCodeInterpreter:
+    @pytest.fixture()
+    def sandbox(self):  # type: ignore
+        """Create a mock Sandbox with no API key requirement."""
+        mock_sandbox = MagicMock()
+        mock_sandbox.files = MagicMock()
+        mock_sandbox.files.write = MagicMock()
+        mock_sandbox.run_code = MagicMock()
+        return mock_sandbox
+
+    @pytest.fixture()
+    def code_interpreter(self, sandbox):  # type: ignore
+        """Create E2BCodeInterpreter that uses the mock Sandbox."""
+        interpreter = E2BCodeInterpreter(api_key="dummy_key")
+        interpreter.interpreter = sandbox
+        return interpreter
+
+    def test_interpret_success(self, code_interpreter, sandbox) -> None:  # type: ignore
+        """Test successful code execution."""
+        # Mock execution result
+        mock_execution = Execution()
+        mock_execution.error = None
+        mock_execution.results = []
+        mock_execution.logs = Logs(
+            stdout="stdout", stderr="", display_data="", error=""
+        )
+        sandbox.run_code.return_value = mock_execution
+
+        # Run the code
+        result = code_interpreter.interpret("print('hello')")
+
+        # Verify
+        sandbox.run_code.assert_called_once_with("print('hello')")
+        assert result.is_error is False
+        assert result.logs == mock_execution.logs
+
+    def test_interpret_error(self, code_interpreter, sandbox) -> None:  # type: ignore
+        """Test error in code execution."""
+        # Mock execution result with error
+        mock_execution = Execution()
+        mock_execution.error = "Test error"
+        mock_execution.logs = Logs(
+            stdout="", stderr="error", display_data="", error="Test error"
+        )
+        sandbox.run_code.return_value = mock_execution
+
+        # Run the code
+        result = code_interpreter.interpret("bad code")
+
+        # Verify
+        assert result.is_error is True
+        assert "Error: Test error" in result.error_message
+        sandbox.kill.assert_called_once()
+
+    def test_to_tool(self, code_interpreter) -> None:  # type: ignore
+        """Test tool conversion."""
+        tool = code_interpreter.to_tool()
+        assert tool.fn == code_interpreter.interpret

package.json

@@ -1,6 +1,6 @@
 {
   "name": "create-llama",
-  "version": "0.1.8",
+  "version": "0.5.2",
   "description": "Create LlamaIndex-powered apps with one command",
   "keywords": [
     "rag",
@@ -9,7 +9,7 @@
   ],
   "repository": {
     "type": "git",
-    "url": "https://github.com/run-llama/LlamaIndexTS",
+    "url": "https://github.com/run-llama/create-llama",
     "directory": "packages/create-llama"
   },
   "license": "MIT",
@@ -25,6 +25,8 @@
     "clean": "rimraf --glob ./dist ./templates/**/__pycache__ ./templates/**/node_modules ./templates/**/poetry.lock",
     "dev": "ncc build ./index.ts -w -o dist/",
     "e2e": "playwright test",
+    "e2e:python": "playwright test e2e/shared e2e/python",
+    "e2e:typescript": "playwright test e2e/shared e2e/typescript",
     "format": "prettier --ignore-unknown --cache --check .",
     "format:write": "prettier --ignore-unknown --write .",
     "lint": "eslint . --ignore-pattern dist --ignore-pattern e2e/cache",
@@ -41,14 +43,13 @@
     "@types/cross-spawn": "6.0.0",
     "@types/fs-extra": "11.0.4",
     "@types/node": "^20.11.7",
-    "@types/prompts": "2.0.1",
+    "@types/prompts": "2.4.2",
     "@types/tar": "6.1.5",
     "@types/validate-npm-package-name": "3.0.0",
     "async-retry": "1.3.1",
     "async-sema": "3.0.1",
     "ci-info": "github:watson/ci-info#f43f6a1cefff47fb361c88cf4b943fdbcaafe540",
-    "commander": "2.20.0",
-    "conf": "10.2.0",
+    "commander": "12.1.0",
     "cross-spawn": "7.0.3",
     "fast-glob": "3.3.1",
     "fs-extra": "11.2.0",
@@ -57,7 +58,7 @@
     "ollama": "^0.5.0",
     "ora": "^8.0.1",
     "picocolors": "1.0.0",
-    "prompts": "2.1.0",
+    "prompts": "2.4.2",
     "smol-toml": "^1.1.4",
     "tar": "6.1.15",
     "terminal-link": "^3.0.0",

pnpm-lock.yaml

@@ -24,8 +24,8 @@ importers:
         specifier: ^20.11.7
         version: 20.12.10
       '@types/prompts':
-        specifier: 2.0.1
-        version: 2.0.1
+        specifier: 2.4.2
+        version: 2.4.2
       '@types/tar':
         specifier: 6.1.5
         version: 6.1.5
@@ -42,11 +42,8 @@ importers:
         specifier: github:watson/ci-info#f43f6a1cefff47fb361c88cf4b943fdbcaafe540
         version: https://codeload.github.com/watson/ci-info/tar.gz/f43f6a1cefff47fb361c88cf4b943fdbcaafe540
       commander:
-        specifier: 2.20.0
-        version: 2.20.0
-      conf:
-        specifier: 10.2.0
-        version: 10.2.0
+        specifier: 12.1.0
+        version: 12.1.0
       cross-spawn:
         specifier: 7.0.3
         version: 7.0.3
@@ -72,8 +69,8 @@ importers:
         specifier: 1.0.0
         version: 1.0.0
       prompts:
-        specifier: 2.1.0
-        version: 2.1.0
+        specifier: 2.4.2
+        version: 2.4.2
       smol-toml:
         specifier: ^1.1.4
         version: 1.1.4
@@ -301,8 +298,8 @@ packages:
   '@types/normalize-package-data@2.4.4':
     resolution: {integrity: sha512-37i+OaWTh9qeK4LSHPsyRC7NahnGotNuZvjLSgcPzblpHB3rrCJxAOgI5gCdKm7coonsaX1Of0ILiTcnZjbfxA==}
 
-  '@types/prompts@2.0.1':
-    resolution: {integrity: sha512-AhtMcmETelF8wFDV1ucbChKhLgsc+ytXZXkNz/nnTAMSDeqsjALknEFxi7ZtLgS/G8bV2rp90LhDW5SGACimIQ==}
+  '@types/prompts@2.4.2':
+    resolution: {integrity: sha512-TwNx7qsjvRIUv/BCx583tqF5IINEVjCNqg9ofKHRlSoUHE62WBHrem4B1HGXcIrG511v29d1kJ9a/t2Esz7MIg==}
 
   '@types/responselike@1.0.3':
     resolution: {integrity: sha512-H/+L+UkTV33uf49PH5pCAUBVPNj2nDBXTN+qS1dOwyyg24l3CcicicCA7ca+HMvJBZcFgl5r8e+RR6elsb4Lyw==}
@@ -336,20 +333,9 @@ packages:
     engines: {node: '>=0.4.0'}
     hasBin: true
 
-  ajv-formats@2.1.1:
-    resolution: {integrity: sha512-Wx0Kx52hxE7C18hkMEggYlEifqWZtYaRgouJor+WMdPnQyEK13vgEWyVNup7SoeeoLMsr4kf5h6dOW11I15MUA==}
-    peerDependencies:
-      ajv: ^8.0.0
-    peerDependenciesMeta:
-      ajv:
-        optional: true
-
   ajv@6.12.6:
     resolution: {integrity: sha512-j3fVLgvTo527anyYyJOGTYJbG+vnnQYvE0m5mmkc1TK+nxAppkCLMIL0aZ4dblVCNoGShhm+kzE4ZUykBoMg4g==}
 
-  ajv@8.13.0:
-    resolution: {integrity: sha512-PRA911Blj99jR5RMeTunVbNXMF6Lp4vZXnk5GQjcnUWUTsrXtekg/pnmFFI2u/I36Y/2bITGS30GZCXei6uNkA==}
-
   ansi-colors@4.1.3:
     resolution: {integrity: sha512-/6w/C21Pm1A7aZitlI5Ni/2J6FFQN8i1Cvz3kHABAAbw93v/NlvKdVOqz7CCWz/3iv/JplRSEEZ83XION15ovw==}
     engines: {node: '>=6'}
@@ -410,10 +396,6 @@ packages:
   async-sema@3.0.1:
     resolution: {integrity: sha512-fKT2riE8EHAvJEfLJXZiATQWqZttjx1+tfgnVshCDrH8vlw4YC8aECe0B8MU184g+aVRFVgmfxFlKZKaozSrNw==}
 
-  atomically@1.7.0:
-    resolution: {integrity: sha512-Xcz9l0z7y9yQ9rdDaxlmaI4uJHf/T8g9hOEzJcsEqX2SjCj4J20uK7+ldkDHMbpJDK76wF7xEIgxc/vSlsfw5w==}
-    engines: {node: '>=10.12.0'}
-
   available-typed-arrays@1.0.7:
     resolution: {integrity: sha512-wvUjBtSGN7+7SjNpq/9M2Tg350UZD3q62IFZLbRAR1bSMlCo1ZaeW+BJ+D090e4hIIZLBcTDWe4Mh4jvUDajzQ==}
     engines: {node: '>= 0.4'}
@@ -530,8 +512,9 @@ packages:
   color-name@1.1.4:
     resolution: {integrity: sha512-dOy+3AuW3a2wNbZHIuMZpTcgjGuLU/uBL/ubcZF9OXbDo8ff4O8yVp5Bf0efS8uEoYo5q4Fx7dY9OgQGXgAsQA==}
 
-  commander@2.20.0:
-    resolution: {integrity: sha512-7j2y+40w61zy6YC2iRNpUe/NwhNyoXrYpHMrSunaMG64nRnaf96zO/KMQR4OyN/UnE5KLyEBnKHd4aG3rskjpQ==}
+  commander@12.1.0:
+    resolution: {integrity: sha512-Vw8qHK3bZM9y/P10u3Vib8o/DdkvA2OtPtZvD871QKjy74Wj1WSKFILMPRPSdUSx5RFK1arlJzEtA4PkFgnbuA==}
+    engines: {node: '>=18'}
 
   commander@9.5.0:
     resolution: {integrity: sha512-KRs7WVDKg86PWiuAqhDrAQnTXZKraVcCc6vFdL14qrZ/DcWwuRo7VoiYXalXO7S5GKpqYiVEwCbgFDfxNHKJBQ==}
@@ -540,10 +523,6 @@ packages:
   concat-map@0.0.1:
     resolution: {integrity: sha512-/Srv4dswyQNBfohGpz9o6Yb3Gz3SrUDqBH5rTuhGR7ahtlbYKnVxw2bCFMRljaA7EXHaXZ8wsHdodFvbkhKmqg==}
 
-  conf@10.2.0:
-    resolution: {integrity: sha512-8fLl9F04EJqjSqH+QjITQfJF8BrOVaYr1jewVgSRAEWePfxT0sku4w2hrGQ60BC/TNLGQ2pgxNlTbWQmMPFvXg==}
-    engines: {node: '>=12'}
-
   cross-spawn@5.1.0:
     resolution: {integrity: sha512-pTgQJ5KC0d2hcY8eyL1IzlBPYjTkyH72XRZPnLyKus2mBfNjQs3klqbJU2VILqZryAZUt9JOb3h/mWMy23/f5A==}
 
@@ -576,10 +555,6 @@ packages:
     resolution: {integrity: sha512-t/Ygsytq+R995EJ5PZlD4Cu56sWa8InXySaViRzw9apusqsOO2bQP+SbYzAhR0pFKoB+43lYy8rWban9JSuXnA==}
     engines: {node: '>= 0.4'}
 
-  debounce-fn@4.0.0:
-    resolution: {integrity: sha512-8pYCQiL9Xdcg0UPSD3d+0KMlOjp+KGU5EPwYddgzQ7DATsg4fuUDjQtsYLmWjnk2obnNHgV3vE2Y4jejSOJVBQ==}
-    engines: {node: '>=10'}
-
   debug@4.3.4:
     resolution: {integrity: sha512-PRWFHuSU3eDtQJPvnNY7Jcket1j0t5OuOsFzPPzsekD52Zl8qUfFIPEiswXqIvHWGVHOgX+7G/vCNNhehwxfkQ==}
     engines: {node: '>=6.0'}
@@ -638,10 +613,6 @@ packages:
     resolution: {integrity: sha512-yS+Q5i3hBf7GBkd4KG8a7eBNNWNGLTaEwwYWUijIYM7zrlYDM0BFXHjjPWlWZ1Rg7UaddZeIDmi9jF3HmqiQ2w==}
     engines: {node: '>=6.0.0'}
 
-  dot-prop@6.0.1:
-    resolution: {integrity: sha512-tE7ztYzXHIeyvc7N+hR3oi7FIbf/NIjVP9hmAt3yMXzrQ072/fpjGLx2GxNxGxUl5V73MEqYzioOMoVhGMJ5cA==}
-    engines: {node: '>=10'}
-
   duplexer3@0.1.5:
     resolution: {integrity: sha512-1A8za6ws41LQgv9HrE/66jyC5yuSjQ3L/KOpFtoBilsAK2iA2wuS5rTt1OCzIvtS2V7nVmedsUU+DGRcjBmOYA==}
 
@@ -664,10 +635,6 @@ packages:
     resolution: {integrity: sha512-rRqJg/6gd538VHvR3PSrdRBb/1Vy2YfzHqzvbhGIQpDRKIa4FgV/54b5Q1xYSxOOwKvjXweS26E0Q+nAMwp2pQ==}
     engines: {node: '>=8.6'}
 
-  env-paths@2.2.1:
-    resolution: {integrity: sha512-+h1lkLKhZMTYjog1VEpJNG7NZJWcuc2DDk/qsqSTRRCOXiLjeQ1d1/udrUGhqMxUgAlwKNZ0cf2uqan5GLuS2A==}
-    engines: {node: '>=6'}
-
   error-ex@1.3.2:
     resolution: {integrity: sha512-7dFHNmqeFSEt2ZBsCriorKnn3Z2pj+fd9kmI6QoWw4//DL+icEBfc0U7qJCisqrTsKTjw4fNFy2pW9OqStD84g==}
 
@@ -788,10 +755,6 @@ packages:
     resolution: {integrity: sha512-qOo9F+dMUmC2Lcb4BbVvnKJxTPjCm+RRpe4gDuGrzkL7mEVl/djYSu2OdQ2Pa302N4oqkSg9ir6jaLWJ2USVpQ==}
     engines: {node: '>=8'}
 
-  find-up@3.0.0:
-    resolution: {integrity: sha512-1yD6RmLI1XBfxugvORwlck6f75tYL+iR0jqwsOrOxMZyGYqUuDhJ0l4AXdO1iX/FTs9cBAMEk1gWSEx1kSbylg==}
-    engines: {node: '>=6'}
-
   find-up@4.1.0:
     resolution: {integrity: sha512-PpOwAdQ/YlXQ2vj8a3h8IipDuYRi3wceVQQGYWxNINccq40Anw7BlsEXCMbt1Zt+OLA6Fq9suIpIWD0OsnISlw==}
     engines: {node: '>=8'}
@@ -1057,10 +1020,6 @@ packages:
     resolution: {integrity: sha512-41Cifkg6e8TylSpdtTpeLVMqvSBEVzTttHvERD741+pnZ8ANv0004MRL43QKPDlK9cGvNp6NZWZUBlbGXYxxng==}
     engines: {node: '>=0.12.0'}
 
-  is-obj@2.0.0:
-    resolution: {integrity: sha512-drqDG3cbczxxEJRoOXcOjtdp1J/lyp1mNn0xaznRs8+muBhgQcrnbspox5X5fOw0HnMnbfDzvnEMEtqDEJEo8w==}
-    engines: {node: '>=8'}
-
   is-path-inside@3.0.3:
     resolution: {integrity: sha512-Fd4gABb+ycGAmKou8eMftCupSir5lRxqf4aD/vd0cD2qc4HL07OjCeuHMr8Ro4CoMaeCKDB0/ECBOVWjTwUvPQ==}
     engines: {node: '>=8'}
@@ -1138,12 +1097,6 @@ packages:
   json-schema-traverse@0.4.1:
     resolution: {integrity: sha512-xbbCH5dCYU5T8LcEhhuh7HJ88HXuW3qsI3Y0zOZFKfZEHcpWiHU/Jxzk629Brsab/mMiHQti9wMP+845RPe3Vg==}
 
-  json-schema-traverse@1.0.0:
-    resolution: {integrity: sha512-NM8/P9n3XjXhIZn1lLhkFaACTOURQXjWhV4BA/RnOv8xvgqtqpAX9IO4mRQxSx1Rlo4tqzeqb0sOlruaOy3dug==}
-
-  json-schema-typed@7.0.3:
-    resolution: {integrity: sha512-7DE8mpG+/fVw+dTpjbxnx47TaMnDfOI1jwft9g1VybltZCduyRQPJPvc+zzKY9WPHxhPWczyFuYa6I8Mw4iU5A==}
-
   json-stable-stringify-without-jsonify@1.0.1:
     resolution: {integrity: sha512-Bdboy+l7tA3OGW6FjyFHWkP5LuByj1Tk33Ljyq0axyzdk9//JSi2u3fP1QSmd1KNwq6VOKYGlAu87CisVir6Pw==}
 
@@ -1182,10 +1135,6 @@ packages:
     resolution: {integrity: sha512-OfCBkGEw4nN6JLtgRidPX6QxjBQGQf72q3si2uvqyFEMbycSFFHwAZeXx6cJgFM9wmLrf9zBwCP3Ivqa+LLZPw==}
     engines: {node: '>=6'}
 
-  locate-path@3.0.0:
-    resolution: {integrity: sha512-7AO748wWnIhNqAuaty2ZWHkQHRSNfPVIsPIfwEOWO22AmaoVrWavlOcMR5nzTLNYvp36X220/maaRsrec1G65A==}
-    engines: {node: '>=6'}
-
   locate-path@5.0.0:
     resolution: {integrity: sha512-t7hw9pI+WvuwNJXwk5zVHpyhIqzg2qTlklJOf0mVxGSbe3Fp2VieZcduNYjaLDoy6p9uGpQEGWG87WpMKlNq8g==}
     engines: {node: '>=8'}
@@ -1243,10 +1192,6 @@ packages:
     resolution: {integrity: sha512-OqbOk5oEQeAZ8WXWydlu9HJjz9WVdEIvamMCcXmuqUYjTknH/sqsWvhQ3vgwKFRR1HpjvNBKQ37nbJgYzGqGcg==}
     engines: {node: '>=6'}
 
-  mimic-fn@3.1.0:
-    resolution: {integrity: sha512-Ysbi9uYW9hFyfrThdDEQuykN4Ey6BuwPD2kpI5ES/nFTDn/98yxYNLZJcgUAKPT/mcrLLKaGzJR9YVxJrIdASQ==}
-    engines: {node: '>=8'}
-
   mimic-response@1.0.1:
     resolution: {integrity: sha512-j5EctnkH7amfV/q5Hgmoal1g2QHFJRraOtmx0JpIqkxhBhI/lJSl1nMpQ45hVarwNETOoWEimndZ4QK0RHxuxQ==}
     engines: {node: '>=4'}
@@ -1375,10 +1320,6 @@ packages:
     resolution: {integrity: sha512-TYOanM3wGwNGsZN2cVTYPArw454xnXj5qmWF1bEoAc4+cU/ol7GVh7odevjp1FNHduHc3KZMcFduxU5Xc6uJRQ==}
     engines: {node: '>=10'}
 
-  p-locate@3.0.0:
-    resolution: {integrity: sha512-x+12w/To+4GFfgJhBEpiDcLozRJGegY+Ei7/z0tSLkMmxGZNybVMSfWj9aJn8Z5Fc7dBUNJOOVgPv2H7IwulSQ==}
-    engines: {node: '>=6'}
-
   p-locate@4.1.0:
     resolution: {integrity: sha512-R79ZZ/0wAxKGu3oYMlz8jy/kbhsNrS7SKZ7PxEHBgJ5+F2mtFW2fK2cOtBh1cHYkQsbzFV7I+EoRKe6Yt0oK7A==}
     engines: {node: '>=8'}
@@ -1407,10 +1348,6 @@ packages:
     resolution: {integrity: sha512-ayCKvm/phCGxOkYRSCM82iDwct8/EonSEgCSxWxD7ve6jHggsFl4fZVQBPRNgQoKiuV/odhFrGzQXZwbifC8Rg==}
     engines: {node: '>=8'}
 
-  path-exists@3.0.0:
-    resolution: {integrity: sha512-bpC7GYwiDYQ4wYLe+FA8lhRjhQCMcQGuSgGGqDkg/QerRWw9CmGRT0iSOVRSZJ29NMLZgIzqaljJ63oaL4NIJQ==}
-    engines: {node: '>=4'}
-
   path-exists@4.0.0:
     resolution: {integrity: sha512-ak9Qy5Q7jYb2Wwcey5Fpvg2KoAc/ZIhLSLOSBmRmygPsGwkVVt0fZa0qrtMz+m6tJTAHfZQ8FnmB4MG4LWy7/w==}
     engines: {node: '>=8'}
@@ -1449,10 +1386,6 @@ packages:
     resolution: {integrity: sha512-HRDzbaKjC+AOWVXxAU/x54COGeIv9eb+6CkDSQoNTt4XyWoIJvuPsXizxu/Fr23EiekbtZwmh1IcIG/l/a10GQ==}
     engines: {node: '>=8'}
 
-  pkg-up@3.1.0:
-    resolution: {integrity: sha512-nDywThFk1i4BQK4twPQ6TA4RT8bDY96yeuCVBWL3ePARCiEKDRSrNGbFIgUJpLp+XeIR65v8ra7WuJOFUBtkMA==}
-    engines: {node: '>=8'}
-
   playwright-core@1.44.0:
     resolution: {integrity: sha512-ZTbkNpFfYcGWohvTTl+xewITm7EOuqIqex0c7dNZ+aXsbrLj0qI8XlGKfPpipjm0Wny/4Lt4CJsWJk1stVS5qQ==}
     engines: {node: '>=16'}
@@ -1498,8 +1431,8 @@ packages:
     engines: {node: '>=14'}
     hasBin: true
 
-  prompts@2.1.0:
-    resolution: {integrity: sha512-+x5TozgqYdOwWsQFZizE/Tra3fKvAoy037kOyU6cgz84n8f6zxngLOV4O32kTwt9FcLCxAqw0P/c8rOr9y+Gfg==}
+  prompts@2.4.2:
+    resolution: {integrity: sha512-NxNv/kLguCA7p3jE8oL2aEBsrJWgAakBpgmgK6lpPWV+WuOmY6r2/zbAVnP+T8bQlA0nzHXSJSJW0Hq7ylaD2Q==}
     engines: {node: '>= 6'}
 
   pseudomap@1.0.2:
@@ -1557,10 +1490,6 @@ packages:
     resolution: {integrity: sha512-fGxEI7+wsG9xrvdjsrlmL22OMTTiHRwAMroiEeMgq8gzoLC/PQr7RsRDSTLUg/bZAZtF+TVIkHc6/4RIKrui+Q==}
     engines: {node: '>=0.10.0'}
 
-  require-from-string@2.0.2:
-    resolution: {integrity: sha512-Xf0nWe6RseziFMu+Ap9biiUbmplq6S9/p+7w7YXP/JBHhrUDDUhwa+vANyubuqfZWTveU//DYVGsDG7RKL/vEw==}
-    engines: {node: '>=0.10.0'}
-
   require-main-filename@2.0.0:
     resolution: {integrity: sha512-NKN5kMDylKuldxYLSUfrbo5Tuzh4hd+2E8NPPX02mZtn1VuREQToYe/ZdlJy+J3uCpfaiGF05e7B8W0iXbQHmg==}
 
@@ -2279,7 +2208,10 @@ snapshots:
 
   '@types/normalize-package-data@2.4.4': {}
 
-  '@types/prompts@2.0.1': {}
+  '@types/prompts@2.4.2':
+    dependencies:
+      '@types/node': 20.12.10
+      kleur: 3.0.3
 
   '@types/responselike@1.0.3':
     dependencies:
@@ -2306,10 +2238,6 @@ snapshots:
 
   acorn@8.11.3: {}
 
-  ajv-formats@2.1.1(ajv@8.13.0):
-    optionalDependencies:
-      ajv: 8.13.0
-
   ajv@6.12.6:
     dependencies:
       fast-deep-equal: 3.1.3
@@ -2317,13 +2245,6 @@ snapshots:
       json-schema-traverse: 0.4.1
       uri-js: 4.4.1
 
-  ajv@8.13.0:
-    dependencies:
-      fast-deep-equal: 3.1.3
-      json-schema-traverse: 1.0.0
-      require-from-string: 2.0.2
-      uri-js: 4.4.1
-
   ansi-colors@4.1.3: {}
 
   ansi-escapes@5.0.0:
@@ -2383,8 +2304,6 @@ snapshots:
 
   async-sema@3.0.1: {}
 
-  atomically@1.7.0: {}
-
   available-typed-arrays@1.0.7:
     dependencies:
       possible-typed-array-names: 1.0.0
@@ -2506,25 +2425,12 @@ snapshots:
 
   color-name@1.1.4: {}
 
-  commander@2.20.0: {}
+  commander@12.1.0: {}
 
   commander@9.5.0: {}
 
   concat-map@0.0.1: {}
 
-  conf@10.2.0:
-    dependencies:
-      ajv: 8.13.0
-      ajv-formats: 2.1.1(ajv@8.13.0)
-      atomically: 1.7.0
-      debounce-fn: 4.0.0
-      dot-prop: 6.0.1
-      env-paths: 2.2.1
-      json-schema-typed: 7.0.3
-      onetime: 5.1.2
-      pkg-up: 3.1.0
-      semver: 7.6.1
-
   cross-spawn@5.1.0:
     dependencies:
       lru-cache: 4.1.5
@@ -2568,10 +2474,6 @@ snapshots:
       es-errors: 1.3.0
       is-data-view: 1.0.1
 
-  debounce-fn@4.0.0:
-    dependencies:
-      mimic-fn: 3.1.0
-
   debug@4.3.4:
     dependencies:
       ms: 2.1.2
@@ -2621,10 +2523,6 @@ snapshots:
     dependencies:
       esutils: 2.0.3
 
-  dot-prop@6.0.1:
-    dependencies:
-      is-obj: 2.0.0
-
   duplexer3@0.1.5: {}
 
   eastasianwidth@0.2.0: {}
@@ -2644,8 +2542,6 @@ snapshots:
       ansi-colors: 4.1.3
       strip-ansi: 6.0.1
 
-  env-paths@2.2.1: {}
-
   error-ex@1.3.2:
     dependencies:
       is-arrayish: 0.2.1
@@ -2841,10 +2737,6 @@ snapshots:
     dependencies:
       to-regex-range: 5.0.1
 
-  find-up@3.0.0:
-    dependencies:
-      locate-path: 3.0.0
-
   find-up@4.1.0:
     dependencies:
       locate-path: 5.0.0
@@ -3129,8 +3021,6 @@ snapshots:
 
   is-number@7.0.0: {}
 
-  is-obj@2.0.0: {}
-
   is-path-inside@3.0.3: {}
 
   is-plain-obj@1.1.0: {}
@@ -3197,10 +3087,6 @@ snapshots:
 
   json-schema-traverse@0.4.1: {}
 
-  json-schema-traverse@1.0.0: {}
-
-  json-schema-typed@7.0.3: {}
-
   json-stable-stringify-without-jsonify@1.0.1: {}
 
   json-stringify-safe@5.0.1: {}
@@ -3239,11 +3125,6 @@ snapshots:
       pify: 4.0.1
       strip-bom: 3.0.0
 
-  locate-path@3.0.0:
-    dependencies:
-      p-locate: 3.0.0
-      path-exists: 3.0.0
-
   locate-path@5.0.0:
     dependencies:
       p-locate: 4.1.0
@@ -3301,8 +3182,6 @@ snapshots:
 
   mimic-fn@2.1.0: {}
 
-  mimic-fn@3.1.0: {}
-
   mimic-response@1.0.1: {}
 
   mimic-response@2.1.0: {}
@@ -3425,10 +3304,6 @@ snapshots:
     dependencies:
       yocto-queue: 0.1.0
 
-  p-locate@3.0.0:
-    dependencies:
-      p-limit: 2.3.0
-
   p-locate@4.1.0:
     dependencies:
       p-limit: 2.3.0
@@ -3456,8 +3331,6 @@ snapshots:
       json-parse-even-better-errors: 2.3.1
       lines-and-columns: 1.2.4
 
-  path-exists@3.0.0: {}
-
   path-exists@4.0.0: {}
 
   path-is-absolute@1.0.1: {}
@@ -3483,10 +3356,6 @@ snapshots:
     dependencies:
       find-up: 4.1.0
 
-  pkg-up@3.1.0:
-    dependencies:
-      find-up: 3.0.0
-
   playwright-core@1.44.0: {}
 
   playwright@1.44.0:
@@ -3515,7 +3384,7 @@ snapshots:
 
   prettier@3.2.5: {}
 
-  prompts@2.1.0:
+  prompts@2.4.2:
     dependencies:
       kleur: 3.0.3
       sisteransi: 1.0.5
@@ -3585,8 +3454,6 @@ snapshots:
 
   require-directory@2.1.1: {}
 
-  require-from-string@2.0.2: {}
-
   require-main-filename@2.0.0: {}
 
   resolve-from@4.0.0: {}

questions.ts

@@ -1,678 +0,0 @@
-import { execSync } from "child_process";
-import ciInfo from "ci-info";
-import fs from "fs";
-import path from "path";
-import { blue, green, red } from "picocolors";
-import prompts from "prompts";
-import { InstallAppArgs } from "./create-app";
-import {
-  TemplateDataSource,
-  TemplateDataSourceType,
-  TemplateFramework,
-} from "./helpers";
-import { COMMUNITY_OWNER, COMMUNITY_REPO } from "./helpers/constant";
-import { EXAMPLE_FILE } from "./helpers/datasources";
-import { templatesDir } from "./helpers/dir";
-import { getAvailableLlamapackOptions } from "./helpers/llama-pack";
-import { askModelConfig } from "./helpers/providers";
-import { getProjectOptions } from "./helpers/repo";
-import { supportedTools, toolsRequireConfig } from "./helpers/tools";
-
-export type QuestionArgs = Omit<
-  InstallAppArgs,
-  "appPath" | "packageManager"
-> & {
-  askModels?: boolean;
-};
-const supportedContextFileTypes = [
-  ".pdf",
-  ".doc",
-  ".docx",
-  ".xls",
-  ".xlsx",
-  ".csv",
-];
-const MACOS_FILE_SELECTION_SCRIPT = `
-osascript -l JavaScript -e '
-  a = Application.currentApplication();
-  a.includeStandardAdditions = true;
-  a.chooseFile({ withPrompt: "Please select files to process:", multipleSelectionsAllowed: true }).map(file => file.toString())
-'`;
-const MACOS_FOLDER_SELECTION_SCRIPT = `
-osascript -l JavaScript -e '
-  a = Application.currentApplication();
-  a.includeStandardAdditions = true;
-  a.chooseFolder({ withPrompt: "Please select folders to process:", multipleSelectionsAllowed: true }).map(folder => folder.toString())
-'`;
-const WINDOWS_FILE_SELECTION_SCRIPT = `
-Add-Type -AssemblyName System.Windows.Forms
-$openFileDialog = New-Object System.Windows.Forms.OpenFileDialog
-$openFileDialog.InitialDirectory = [Environment]::GetFolderPath('Desktop')
-$openFileDialog.Multiselect = $true
-$result = $openFileDialog.ShowDialog()
-if ($result -eq 'OK') {
-  $openFileDialog.FileNames
-}
-`;
-const WINDOWS_FOLDER_SELECTION_SCRIPT = `
-Add-Type -AssemblyName System.windows.forms
-$folderBrowser = New-Object System.Windows.Forms.FolderBrowserDialog
-$dialogResult = $folderBrowser.ShowDialog()
-if ($dialogResult -eq [System.Windows.Forms.DialogResult]::OK)
-{
-    $folderBrowser.SelectedPath
-}
-`;
-
-const defaults: Omit<QuestionArgs, "modelConfig"> = {
-  template: "streaming",
-  framework: "nextjs",
-  ui: "shadcn",
-  frontend: false,
-  llamaCloudKey: "",
-  useLlamaParse: false,
-  communityProjectConfig: undefined,
-  llamapack: "",
-  postInstallAction: "dependencies",
-  dataSources: [],
-  tools: [],
-};
-
-export const questionHandlers = {
-  onCancel: () => {
-    console.error("Exiting.");
-    process.exit(1);
-  },
-};
-
-const getVectorDbChoices = (framework: TemplateFramework) => {
-  const choices = [
-    {
-      title: "No, just store the data in the file system",
-      value: "none",
-    },
-    { title: "MongoDB", value: "mongo" },
-    { title: "PostgreSQL", value: "pg" },
-    { title: "Pinecone", value: "pinecone" },
-    { title: "Milvus", value: "milvus" },
-    { title: "Astra", value: "astra" },
-    { title: "Qdrant", value: "qdrant" },
-    { title: "ChromaDB", value: "chroma" },
-  ];
-
-  const vectordbLang = framework === "fastapi" ? "python" : "typescript";
-  const compPath = path.join(templatesDir, "components");
-  const vectordbPath = path.join(compPath, "vectordbs", vectordbLang);
-
-  const availableChoices = fs
-    .readdirSync(vectordbPath)
-    .filter((file) => fs.statSync(path.join(vectordbPath, file)).isDirectory());
-
-  const displayedChoices = choices.filter((choice) =>
-    availableChoices.includes(choice.value),
-  );
-
-  return displayedChoices;
-};
-
-export const getDataSourceChoices = (
-  framework: TemplateFramework,
-  selectedDataSource: TemplateDataSource[],
-) => {
-  const choices = [];
-  if (selectedDataSource.length > 0) {
-    choices.push({
-      title: "No",
-      value: "no",
-    });
-  }
-  if (selectedDataSource === undefined || selectedDataSource.length === 0) {
-    choices.push({
-      title: "No data, just a simple chat or agent",
-      value: "none",
-    });
-    choices.push({
-      title: "Use an example PDF",
-      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",
-    },
-  );
-
-  if (framework === "fastapi") {
-    choices.push({
-      title: "Use website content (requires Chrome)",
-      value: "web",
-    });
-    choices.push({
-      title: "Use data from a database (Mysql, PostgreSQL)",
-      value: "db",
-    });
-  }
-  return choices;
-};
-
-const selectLocalContextData = async (type: TemplateDataSourceType) => {
-  try {
-    let selectedPath: string = "";
-    let execScript: string;
-    let execOpts: any = {};
-    switch (process.platform) {
-      case "win32": // Windows
-        execScript =
-          type === "file"
-            ? WINDOWS_FILE_SELECTION_SCRIPT
-            : WINDOWS_FOLDER_SELECTION_SCRIPT;
-        execOpts = { shell: "powershell.exe" };
-        break;
-      case "darwin": // MacOS
-        execScript =
-          type === "file"
-            ? MACOS_FILE_SELECTION_SCRIPT
-            : MACOS_FOLDER_SELECTION_SCRIPT;
-        break;
-      default: // Unsupported OS
-        console.log(red("Unsupported OS error!"));
-        process.exit(1);
-    }
-    selectedPath = execSync(execScript, execOpts).toString().trim();
-    const paths =
-      process.platform === "win32"
-        ? selectedPath.split("\r\n")
-        : selectedPath.split(", ");
-
-    for (const p of paths) {
-      if (
-        fs.statSync(p).isFile() &&
-        !supportedContextFileTypes.includes(path.extname(p))
-      ) {
-        console.log(
-          red(
-            `Please select a supported file type: ${supportedContextFileTypes}`,
-          ),
-        );
-        process.exit(1);
-      }
-    }
-    return paths;
-  } catch (error) {
-    console.log(
-      red(
-        "Got an error when trying to select local context data! Please try again or select another data source option.",
-      ),
-    );
-    process.exit(1);
-  }
-};
-
-export const onPromptState = (state: any) => {
-  if (state.aborted) {
-    // If we don't re-enable the terminal cursor before exiting
-    // the program, the cursor will remain hidden
-    process.stdout.write("\x1B[?25h");
-    process.stdout.write("\n");
-    process.exit(1);
-  }
-};
-
-export const askQuestions = async (
-  program: QuestionArgs,
-  preferences: QuestionArgs,
-  openAiKey?: string,
-) => {
-  const getPrefOrDefault = <K extends keyof Omit<QuestionArgs, "modelConfig">>(
-    field: K,
-  ): Omit<QuestionArgs, "modelConfig">[K] =>
-    preferences[field] ?? defaults[field];
-
-  // Ask for next action after installation
-  async function askPostInstallAction() {
-    if (program.postInstallAction === undefined) {
-      if (ciInfo.isCI) {
-        program.postInstallAction = getPrefOrDefault("postInstallAction");
-      } else {
-        const actionChoices = [
-          {
-            title: "Just generate code (~1 sec)",
-            value: "none",
-          },
-          {
-            title: "Start in VSCode (~1 sec)",
-            value: "VSCode",
-          },
-          {
-            title: "Generate code and install dependencies (~2 min)",
-            value: "dependencies",
-          },
-        ];
-
-        const modelConfigured =
-          !program.llamapack && program.modelConfig.isConfigured();
-        // If using LlamaParse, require LlamaCloud API key
-        const llamaCloudKeyConfigured = program.useLlamaParse
-          ? program.llamaCloudKey || process.env["LLAMA_CLOUD_API_KEY"]
-          : true;
-        const hasVectorDb = program.vectorDb && program.vectorDb !== "none";
-        // Can run the app if all tools do not require configuration
-        if (
-          !hasVectorDb &&
-          modelConfigured &&
-          llamaCloudKeyConfigured &&
-          !toolsRequireConfig(program.tools)
-        ) {
-          actionChoices.push({
-            title:
-              "Generate code, install dependencies, and run the app (~2 min)",
-            value: "runApp",
-          });
-        }
-
-        const { action } = await prompts(
-          {
-            type: "select",
-            name: "action",
-            message: "How would you like to proceed?",
-            choices: actionChoices,
-            initial: 1,
-          },
-          questionHandlers,
-        );
-
-        program.postInstallAction = action;
-      }
-    }
-  }
-
-  if (!program.template) {
-    if (ciInfo.isCI) {
-      program.template = getPrefOrDefault("template");
-    } else {
-      const styledRepo = blue(
-        `https://github.com/${COMMUNITY_OWNER}/${COMMUNITY_REPO}`,
-      );
-      const { template } = await prompts(
-        {
-          type: "select",
-          name: "template",
-          message: "Which template would you like to use?",
-          choices: [
-            { title: "Chat", value: "streaming" },
-            {
-              title: `Community template from ${styledRepo}`,
-              value: "community",
-            },
-            {
-              title: "Example using a LlamaPack",
-              value: "llamapack",
-            },
-          ],
-          initial: 0,
-        },
-        questionHandlers,
-      );
-      program.template = template;
-      preferences.template = template;
-    }
-  }
-
-  if (program.template === "community") {
-    const projectOptions = await getProjectOptions(
-      COMMUNITY_OWNER,
-      COMMUNITY_REPO,
-    );
-    const { communityProjectConfig } = await prompts(
-      {
-        type: "select",
-        name: "communityProjectConfig",
-        message: "Select community template",
-        choices: projectOptions.map(({ title, value }) => ({
-          title,
-          value: JSON.stringify(value), // serialize value to string in terminal
-        })),
-        initial: 0,
-      },
-      questionHandlers,
-    );
-    const projectConfig = JSON.parse(communityProjectConfig);
-    program.communityProjectConfig = projectConfig;
-    preferences.communityProjectConfig = projectConfig;
-    return; // early return - no further questions needed for community projects
-  }
-
-  if (program.template === "llamapack") {
-    const availableLlamaPacks = await getAvailableLlamapackOptions();
-    const { llamapack } = await prompts(
-      {
-        type: "select",
-        name: "llamapack",
-        message: "Select LlamaPack",
-        choices: availableLlamaPacks.map((pack) => ({
-          title: pack.name,
-          value: pack.folderPath,
-        })),
-        initial: 0,
-      },
-      questionHandlers,
-    );
-    program.llamapack = llamapack;
-    preferences.llamapack = llamapack;
-    await askPostInstallAction();
-    return; // early return - no further questions needed for llamapack projects
-  }
-
-  if (!program.framework) {
-    if (ciInfo.isCI) {
-      program.framework = getPrefOrDefault("framework");
-    } else {
-      const choices = [
-        { title: "NextJS", value: "nextjs" },
-        { title: "Express", value: "express" },
-        { title: "FastAPI (Python)", value: "fastapi" },
-      ];
-
-      const { framework } = await prompts(
-        {
-          type: "select",
-          name: "framework",
-          message: "Which framework would you like to use?",
-          choices,
-          initial: 0,
-        },
-        questionHandlers,
-      );
-      program.framework = framework;
-      preferences.framework = framework;
-    }
-  }
-
-  if (program.framework === "express" || program.framework === "fastapi") {
-    // if a backend-only framework is selected, ask whether we should create a frontend
-    if (program.frontend === undefined) {
-      if (ciInfo.isCI) {
-        program.frontend = getPrefOrDefault("frontend");
-      } else {
-        const styledNextJS = blue("NextJS");
-        const styledBackend = green(
-          program.framework === "express"
-            ? "Express "
-            : program.framework === "fastapi"
-              ? "FastAPI (Python) "
-              : "",
-        );
-        const { frontend } = await prompts({
-          onState: onPromptState,
-          type: "toggle",
-          name: "frontend",
-          message: `Would you like to generate a ${styledNextJS} frontend for your ${styledBackend}backend?`,
-          initial: getPrefOrDefault("frontend"),
-          active: "Yes",
-          inactive: "No",
-        });
-        program.frontend = Boolean(frontend);
-        preferences.frontend = Boolean(frontend);
-      }
-    }
-  } else {
-    program.frontend = false;
-  }
-
-  if (program.framework === "nextjs" || program.frontend) {
-    if (!program.ui) {
-      program.ui = defaults.ui;
-    }
-  }
-
-  if (!program.observability) {
-    if (ciInfo.isCI) {
-      program.observability = getPrefOrDefault("observability");
-    } else {
-      const { observability } = await prompts(
-        {
-          type: "select",
-          name: "observability",
-          message: "Would you like to set up observability?",
-          choices: [
-            { title: "No", value: "none" },
-            { title: "OpenTelemetry", value: "opentelemetry" },
-          ],
-          initial: 0,
-        },
-        questionHandlers,
-      );
-
-      program.observability = observability;
-      preferences.observability = observability;
-    }
-  }
-
-  if (!program.modelConfig) {
-    const modelConfig = await askModelConfig({
-      openAiKey,
-      askModels: program.askModels ?? false,
-    });
-    program.modelConfig = modelConfig;
-    preferences.modelConfig = modelConfig;
-  }
-
-  if (!program.dataSources) {
-    if (ciInfo.isCI) {
-      program.dataSources = getPrefOrDefault("dataSources");
-    } else {
-      program.dataSources = [];
-      // continue asking user for data sources if none are initially provided
-      while (true) {
-        const firstQuestion = program.dataSources.length === 0;
-        const { selectedSource } = await prompts(
-          {
-            type: "select",
-            name: "selectedSource",
-            message: firstQuestion
-              ? "Which data source would you like to use?"
-              : "Would you like to add another data source?",
-            choices: getDataSourceChoices(
-              program.framework,
-              program.dataSources,
-            ),
-            initial: firstQuestion ? 1 : 0,
-          },
-          questionHandlers,
-        );
-
-        if (selectedSource === "no" || selectedSource === "none") {
-          // user doesn't want another data source or any data source
-          break;
-        }
-        switch (selectedSource) {
-          case "exampleFile": {
-            program.dataSources.push(EXAMPLE_FILE);
-            break;
-          }
-          case "file":
-          case "folder": {
-            const selectedPaths = await selectLocalContextData(selectedSource);
-            for (const p of selectedPaths) {
-              program.dataSources.push({
-                type: "file",
-                config: {
-                  path: p,
-                },
-              });
-            }
-            break;
-          }
-          case "web": {
-            const { baseUrl } = await prompts(
-              {
-                type: "text",
-                name: "baseUrl",
-                message: "Please provide base URL of the website: ",
-                initial: "https://www.llamaindex.ai",
-                validate: (value: string) => {
-                  if (!value.includes("://")) {
-                    value = `https://${value}`;
-                  }
-                  const urlObj = new URL(value);
-                  if (
-                    urlObj.protocol !== "https:" &&
-                    urlObj.protocol !== "http:"
-                  ) {
-                    return `URL=${value} has invalid protocol, only allow http or https`;
-                  }
-                  return true;
-                },
-              },
-              questionHandlers,
-            );
-
-            program.dataSources.push({
-              type: "web",
-              config: {
-                baseUrl,
-                prefix: baseUrl,
-                depth: 1,
-              },
-            });
-            break;
-          }
-          case "db": {
-            const dbPrompts: prompts.PromptObject<string>[] = [
-              {
-                type: "text",
-                name: "uri",
-                message:
-                  "Please enter the connection string (URI) for the database.",
-                initial: "mysql+pymysql://user:pass@localhost:3306/mydb",
-                validate: (value: string) => {
-                  if (!value) {
-                    return "Please provide a valid connection string";
-                  } else if (
-                    !(
-                      value.startsWith("mysql+pymysql://") ||
-                      value.startsWith("postgresql+psycopg://")
-                    )
-                  ) {
-                    return "The connection string must start with 'mysql+pymysql://' for MySQL or 'postgresql+psycopg://' for PostgreSQL";
-                  }
-                  return true;
-                },
-              },
-              // Only ask for a query, user can provide more complex queries in the config file later
-              {
-                type: (prev) => (prev ? "text" : null),
-                name: "queries",
-                message: "Please enter the SQL query to fetch data:",
-                initial: "SELECT * FROM mytable",
-              },
-            ];
-            program.dataSources.push({
-              type: "db",
-              config: await prompts(dbPrompts, questionHandlers),
-            });
-          }
-        }
-      }
-    }
-  }
-
-  // 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;
-
-      // Ask for LlamaCloud API key
-      if (useLlamaParse && program.llamaCloudKey === undefined) {
-        const { llamaCloudKey } = await prompts(
-          {
-            type: "text",
-            name: "llamaCloudKey",
-            message:
-              "Please provide your LlamaIndex Cloud API key (leave blank to skip):",
-          },
-          questionHandlers,
-        );
-        program.llamaCloudKey = llamaCloudKey;
-      }
-    }
-  }
-
-  if (program.dataSources.length > 0 && !program.vectorDb) {
-    if (ciInfo.isCI) {
-      program.vectorDb = getPrefOrDefault("vectorDb");
-    } else {
-      const { vectorDb } = await prompts(
-        {
-          type: "select",
-          name: "vectorDb",
-          message: "Would you like to use a vector database?",
-          choices: getVectorDbChoices(program.framework),
-          initial: 0,
-        },
-        questionHandlers,
-      );
-      program.vectorDb = vectorDb;
-      preferences.vectorDb = vectorDb;
-    }
-  }
-
-  if (!program.tools) {
-    if (ciInfo.isCI) {
-      program.tools = getPrefOrDefault("tools");
-    } else {
-      const options = supportedTools.filter((t) =>
-        t.supportedFrameworks?.includes(program.framework),
-      );
-      const toolChoices = options.map((tool) => ({
-        title: tool.display,
-        value: tool.name,
-      }));
-      const { toolsName } = await prompts({
-        type: "multiselect",
-        name: "toolsName",
-        message:
-          "Would you like to build an agent using tools? If so, select the tools here, otherwise just press enter",
-        choices: toolChoices,
-      });
-      const tools = toolsName?.map((tool: string) =>
-        supportedTools.find((t) => t.name === tool),
-      );
-      program.tools = tools;
-      preferences.tools = tools;
-    }
-  }
-
-  await askPostInstallAction();
-};
-
-export const toChoice = (value: string) => {
-  return { title: value, value };
-};

questions/ci.ts

@@ -0,0 +1,30 @@
+import { askModelConfig } from "../helpers/providers";
+import { QuestionArgs, QuestionResults } from "./types";
+
+const defaults: Omit<QuestionArgs, "modelConfig"> = {
+  template: "streaming",
+  framework: "nextjs",
+  ui: "shadcn",
+  frontend: false,
+  llamaCloudKey: "",
+  useLlamaParse: false,
+  communityProjectConfig: undefined,
+  llamapack: "",
+  postInstallAction: "dependencies",
+  dataSources: [],
+  tools: [],
+};
+
+export async function getCIQuestionResults(
+  program: QuestionArgs,
+): Promise<QuestionResults> {
+  return {
+    ...defaults,
+    ...program,
+    modelConfig: await askModelConfig({
+      openAiKey: program.openAiKey,
+      askModels: false,
+      framework: program.framework,
+    }),
+  };
+}

questions/datasources.ts

@@ -0,0 +1,64 @@
+import {
+  TemplateDataSource,
+  TemplateFramework,
+  TemplateType,
+} from "../helpers";
+import { supportedContextFileTypes } from "./utils";
+
+export const getDataSourceChoices = (
+  framework: TemplateFramework,
+  selectedDataSource: TemplateDataSource[],
+  template?: TemplateType,
+) => {
+  const choices = [];
+
+  if (selectedDataSource.length > 0) {
+    choices.push({
+      title: "No",
+      value: "no",
+    });
+  }
+  if (selectedDataSource === undefined || selectedDataSource.length === 0) {
+    choices.push({
+      title: "No datasource",
+      value: "none",
+    });
+    choices.push({
+      title:
+        process.platform !== "linux"
+          ? "Use an example PDF"
+          : "Use an example PDF (you can add your own data files later)",
+      value: "exampleFile",
+    });
+  }
+
+  // Linux has many distros so we won't support file/folder picker for now
+  if (process.platform !== "linux") {
+    choices.push(
+      {
+        title: `Use local files (${supportedContextFileTypes.join(", ")})`,
+        value: "file",
+      },
+      {
+        title:
+          process.platform === "win32"
+            ? "Use a local folder"
+            : "Use local folders",
+        value: "folder",
+      },
+    );
+  }
+
+  if (framework === "fastapi" && template !== "reflex") {
+    choices.push({
+      title: "Use website content (requires Chrome)",
+      value: "web",
+    });
+    choices.push({
+      title: "Use data from a database (Mysql, PostgreSQL)",
+      value: "db",
+    });
+  }
+
+  return choices;
+};

questions/index.ts

@@ -0,0 +1,21 @@
+import ciInfo from "ci-info";
+import { getCIQuestionResults } from "./ci";
+import { askProQuestions } from "./questions";
+import { askSimpleQuestions } from "./simple";
+import { QuestionArgs, QuestionResults } from "./types";
+
+export const isCI = ciInfo.isCI || process.env.PLAYWRIGHT_TEST === "1";
+
+export const askQuestions = async (
+  args: QuestionArgs,
+): Promise<QuestionResults> => {
+  if (isCI) {
+    return await getCIQuestionResults(args);
+  } else if (args.pro) {
+    // TODO: refactor pro questions to return a result object
+    await askProQuestions(args);
+    return args as unknown as QuestionResults;
+  }
+  const results = await askSimpleQuestions(args);
+  return results;
+};

questions/questions.ts

@@ -0,0 +1,459 @@
+import { blue } from "picocolors";
+import prompts from "prompts";
+import { isCI } from ".";
+import { COMMUNITY_OWNER, COMMUNITY_REPO } from "../helpers/constant";
+import { EXAMPLE_FILE, EXAMPLE_GDPR } from "../helpers/datasources";
+import { getAvailableLlamapackOptions } from "../helpers/llama-pack";
+import { askModelConfig } from "../helpers/providers";
+import { getProjectOptions } from "../helpers/repo";
+import { supportedTools, toolRequiresConfig } from "../helpers/tools";
+import { getDataSourceChoices } from "./datasources";
+import { getVectorDbChoices } from "./stores";
+import { QuestionArgs } from "./types";
+import {
+  askPostInstallAction,
+  onPromptState,
+  questionHandlers,
+  selectLocalContextData,
+} from "./utils";
+
+export const askProQuestions = async (program: QuestionArgs) => {
+  if (!program.template) {
+    const styledRepo = blue(
+      `https://github.com/${COMMUNITY_OWNER}/${COMMUNITY_REPO}`,
+    );
+    const { template } = await prompts(
+      {
+        type: "select",
+        name: "template",
+        message: "Which template would you like to use?",
+        choices: [
+          { title: "Agentic RAG (e.g. chat with docs)", value: "streaming" },
+          {
+            title: "Multi-agent app (using workflows)",
+            value: "multiagent",
+          },
+          { title: "Fullstack python template with Reflex", value: "reflex" },
+          {
+            title: `Community template from ${styledRepo}`,
+            value: "community",
+          },
+          {
+            title: "Example using a LlamaPack",
+            value: "llamapack",
+          },
+        ],
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    program.template = template;
+  }
+
+  if (program.template === "community") {
+    const projectOptions = await getProjectOptions(
+      COMMUNITY_OWNER,
+      COMMUNITY_REPO,
+    );
+    const { communityProjectConfig } = await prompts(
+      {
+        type: "select",
+        name: "communityProjectConfig",
+        message: "Select community template",
+        choices: projectOptions.map(({ title, value }) => ({
+          title,
+          value: JSON.stringify(value), // serialize value to string in terminal
+        })),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    const projectConfig = JSON.parse(communityProjectConfig);
+    program.communityProjectConfig = projectConfig;
+    return; // early return - no further questions needed for community projects
+  }
+
+  if (program.template === "llamapack") {
+    const availableLlamaPacks = await getAvailableLlamapackOptions();
+    const { llamapack } = await prompts(
+      {
+        type: "select",
+        name: "llamapack",
+        message: "Select LlamaPack",
+        choices: availableLlamaPacks.map((pack) => ({
+          title: pack.name,
+          value: pack.folderPath,
+        })),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    program.llamapack = llamapack;
+    if (!program.postInstallAction) {
+      program.postInstallAction = await askPostInstallAction(program);
+    }
+    return; // early return - no further questions needed for llamapack projects
+  }
+
+  if (program.template === "reflex") {
+    // Reflex template only supports FastAPI, empty data sources, and llamacloud
+    // So we just use example file for extractor template, this allows user to choose vector database later
+    program.dataSources = [EXAMPLE_FILE];
+    program.framework = "fastapi";
+    // Ask for which Reflex use case to use
+    const { useCase } = await prompts(
+      {
+        type: "select",
+        name: "useCase",
+        message: "Which use case would you like to build?",
+        choices: [
+          { title: "Structured Extractor", value: "extractor" },
+          {
+            title: "Contract review (using Workflow)",
+            value: "contract_review",
+          },
+        ],
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    program.useCase = useCase;
+  }
+
+  if (!program.framework) {
+    const choices = [
+      { title: "NextJS", value: "nextjs" },
+      { title: "Express", value: "express" },
+      { title: "FastAPI (Python)", value: "fastapi" },
+    ];
+
+    const { framework } = await prompts(
+      {
+        type: "select",
+        name: "framework",
+        message: "Which framework would you like to use?",
+        choices,
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    program.framework = framework;
+  }
+
+  if (
+    program.framework === "fastapi" &&
+    (program.template === "streaming" || program.template === "multiagent")
+  ) {
+    // if a backend-only framework is selected, ask whether we should create a frontend
+    if (program.frontend === undefined) {
+      const styledNextJS = blue("NextJS");
+      const { frontend } = await prompts({
+        onState: onPromptState,
+        type: "toggle",
+        name: "frontend",
+        message: `Would you like to generate a ${styledNextJS} frontend for your FastAPI backend?`,
+        initial: false,
+        active: "Yes",
+        inactive: "No",
+      });
+      program.frontend = Boolean(frontend);
+    }
+  } else {
+    program.frontend = false;
+  }
+
+  if (program.framework === "nextjs" || program.frontend) {
+    if (!program.ui) {
+      program.ui = "shadcn";
+    }
+  }
+
+  if (!program.observability && program.template === "streaming") {
+    const { observability } = await prompts(
+      {
+        type: "select",
+        name: "observability",
+        message: "Would you like to set up observability?",
+        choices: [
+          { title: "No", value: "none" },
+          ...(program.framework === "fastapi"
+            ? [{ title: "LlamaTrace", value: "llamatrace" }]
+            : []),
+          { title: "Traceloop", value: "traceloop" },
+        ],
+        initial: 0,
+      },
+      questionHandlers,
+    );
+
+    program.observability = observability;
+  }
+
+  if (
+    (program.template === "reflex" || program.template === "multiagent") &&
+    !program.useCase
+  ) {
+    const choices =
+      program.template === "reflex"
+        ? [
+            { title: "Structured Extractor", value: "extractor" },
+            {
+              title: "Contract review (using Workflow)",
+              value: "contract_review",
+            },
+          ]
+        : [
+            {
+              title: "Financial report (generate a financial report)",
+              value: "financial_report",
+            },
+            {
+              title: "Form filling (fill missing value in a CSV file)",
+              value: "form_filling",
+            },
+            { title: "Blog writer (Write a blog post)", value: "blog" },
+          ];
+
+    const { useCase } = await prompts(
+      {
+        type: "select",
+        name: "useCase",
+        message: "Which use case would you like to use?",
+        choices,
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    program.useCase = useCase;
+  }
+
+  // Configure framework and data sources for Reflex template
+  if (program.template === "reflex") {
+    program.framework = "fastapi";
+
+    program.dataSources =
+      program.useCase === "extractor" ? [EXAMPLE_FILE] : [EXAMPLE_GDPR];
+  }
+
+  if (!program.modelConfig) {
+    const modelConfig = await askModelConfig({
+      openAiKey: program.openAiKey,
+      askModels: program.askModels ?? false,
+      framework: program.framework,
+    });
+    program.modelConfig = modelConfig;
+  }
+
+  if (!program.vectorDb) {
+    const { vectorDb } = await prompts(
+      {
+        type: "select",
+        name: "vectorDb",
+        message: "Would you like to use a vector database?",
+        choices: getVectorDbChoices(program.framework),
+        initial: 0,
+      },
+      questionHandlers,
+    );
+    program.vectorDb = vectorDb;
+  }
+
+  if (program.vectorDb === "llamacloud" && program.dataSources.length === 0) {
+    // When using a LlamaCloud index and no data sources are provided, just copy an example file
+    program.dataSources = [EXAMPLE_FILE];
+  }
+
+  if (!program.dataSources) {
+    program.dataSources = [];
+    // continue asking user for data sources if none are initially provided
+    while (true) {
+      const firstQuestion = program.dataSources.length === 0;
+      const choices = getDataSourceChoices(
+        program.framework,
+        program.dataSources,
+        program.template,
+      );
+      if (choices.length === 0) break;
+      const { selectedSource } = await prompts(
+        {
+          type: "select",
+          name: "selectedSource",
+          message: firstQuestion
+            ? "Which data source would you like to use?"
+            : "Would you like to add another data source?",
+          choices,
+          initial: firstQuestion ? 1 : 0,
+        },
+        questionHandlers,
+      );
+
+      if (selectedSource === "no" || selectedSource === "none") {
+        // user doesn't want another data source or any data source
+        break;
+      }
+      switch (selectedSource) {
+        case "exampleFile": {
+          program.dataSources.push(EXAMPLE_FILE);
+          break;
+        }
+        case "file":
+        case "folder": {
+          const selectedPaths = await selectLocalContextData(selectedSource);
+          for (const p of selectedPaths) {
+            program.dataSources.push({
+              type: "file",
+              config: {
+                path: p,
+              },
+            });
+          }
+          break;
+        }
+        case "web": {
+          const { baseUrl } = await prompts(
+            {
+              type: "text",
+              name: "baseUrl",
+              message: "Please provide base URL of the website: ",
+              initial: "https://www.llamaindex.ai",
+              validate: (value: string) => {
+                if (!value.includes("://")) {
+                  value = `https://${value}`;
+                }
+                const urlObj = new URL(value);
+                if (
+                  urlObj.protocol !== "https:" &&
+                  urlObj.protocol !== "http:"
+                ) {
+                  return `URL=${value} has invalid protocol, only allow http or https`;
+                }
+                return true;
+              },
+            },
+            questionHandlers,
+          );
+
+          program.dataSources.push({
+            type: "web",
+            config: {
+              baseUrl,
+              prefix: baseUrl,
+              depth: 1,
+            },
+          });
+          break;
+        }
+        case "db": {
+          const dbPrompts: prompts.PromptObject<string>[] = [
+            {
+              type: "text",
+              name: "uri",
+              message:
+                "Please enter the connection string (URI) for the database.",
+              initial: "mysql+pymysql://user:pass@localhost:3306/mydb",
+              validate: (value: string) => {
+                if (!value) {
+                  return "Please provide a valid connection string";
+                } else if (
+                  !(
+                    value.startsWith("mysql+pymysql://") ||
+                    value.startsWith("postgresql+psycopg://")
+                  )
+                ) {
+                  return "The connection string must start with 'mysql+pymysql://' for MySQL or 'postgresql+psycopg://' for PostgreSQL";
+                }
+                return true;
+              },
+            },
+            // Only ask for a query, user can provide more complex queries in the config file later
+            {
+              type: (prev) => (prev ? "text" : null),
+              name: "queries",
+              message: "Please enter the SQL query to fetch data:",
+              initial: "SELECT * FROM mytable",
+            },
+          ];
+          program.dataSources.push({
+            type: "db",
+            config: await prompts(dbPrompts, questionHandlers),
+          });
+          break;
+        }
+      }
+    }
+  }
+
+  const isUsingLlamaCloud = program.vectorDb === "llamacloud";
+
+  // Asking for LlamaParse if user selected file data source
+  if (isUsingLlamaCloud) {
+    // default to use LlamaParse if using LlamaCloud
+    program.useLlamaParse = true;
+  } else {
+    // Reflex template doesn't support LlamaParse right now (cannot use asyncio loop in Reflex)
+    if (program.useLlamaParse === undefined && program.template !== "reflex") {
+      // if already set useLlamaParse, don't ask again
+      if (program.dataSources.some((ds) => ds.type === "file")) {
+        const { useLlamaParse } = await prompts(
+          {
+            type: "toggle",
+            name: "useLlamaParse",
+            message:
+              "Would you like to use LlamaParse (improved parser for RAG - requires API key)?",
+            initial: false,
+            active: "Yes",
+            inactive: "No",
+          },
+          questionHandlers,
+        );
+        program.useLlamaParse = useLlamaParse;
+      }
+    }
+  }
+
+  // Ask for LlamaCloud API key when using a LlamaCloud index or LlamaParse
+  if (isUsingLlamaCloud || program.useLlamaParse) {
+    if (!program.llamaCloudKey && !isCI) {
+      // if already set, don't ask again
+      // Ask for LlamaCloud API key
+      const { llamaCloudKey } = await prompts(
+        {
+          type: "text",
+          name: "llamaCloudKey",
+          message:
+            "Please provide your LlamaCloud API key (leave blank to skip):",
+        },
+        questionHandlers,
+      );
+      program.llamaCloudKey = llamaCloudKey || process.env.LLAMA_CLOUD_API_KEY;
+    }
+  }
+
+  if (
+    !program.tools &&
+    (program.template === "streaming" || program.template === "multiagent")
+  ) {
+    const options = supportedTools.filter((t) =>
+      t.supportedFrameworks?.includes(program.framework),
+    );
+    const toolChoices = options.map((tool) => ({
+      title: `${tool.display}${toolRequiresConfig(tool) ? " (needs configuration)" : ""}`,
+      value: tool.name,
+    }));
+    const { toolsName } = await prompts({
+      type: "multiselect",
+      name: "toolsName",
+      message:
+        "Would you like to build an agent using tools? If so, select the tools here, otherwise just press enter",
+      choices: toolChoices,
+    });
+    const tools = toolsName?.map((tool: string) =>
+      supportedTools.find((t) => t.name === tool),
+    );
+    program.tools = tools;
+  }
+
+  if (!program.postInstallAction) {
+    program.postInstallAction = await askPostInstallAction(program);
+  }
+};

questions/simple.ts

@@ -0,0 +1,166 @@
+import prompts from "prompts";
+import { EXAMPLE_10K_SEC_FILES, EXAMPLE_FILE } from "../helpers/datasources";
+import { askModelConfig } from "../helpers/providers";
+import { getTools } from "../helpers/tools";
+import { ModelConfig, TemplateFramework } from "../helpers/types";
+import { PureQuestionArgs, QuestionResults } from "./types";
+import { askPostInstallAction, questionHandlers } from "./utils";
+
+type AppType = "agentic_rag" | "financial_report" | "deep_research";
+
+type SimpleAnswers = {
+  appType: AppType;
+  language: TemplateFramework;
+  useLlamaCloud: boolean;
+  llamaCloudKey?: string;
+};
+
+export const askSimpleQuestions = async (
+  args: PureQuestionArgs,
+): Promise<QuestionResults> => {
+  const { appType } = await prompts(
+    {
+      type: "select",
+      name: "appType",
+      message: "What use case do you want to build?",
+      choices: [
+        {
+          title: "Agentic RAG",
+          value: "agentic_rag",
+          description:
+            "Chatbot that answers questions based on provided documents.",
+        },
+        {
+          title: "Financial Report",
+          value: "financial_report",
+          description:
+            "Agent that analyzes data and generates visualizations by using a code interpreter.",
+        },
+        {
+          title: "Deep Research",
+          value: "deep_research",
+          description:
+            "Researches and analyzes provided documents from multiple perspectives, generating a comprehensive report with citations to support key findings and insights.",
+        },
+      ],
+    },
+    questionHandlers,
+  );
+
+  let language: TemplateFramework = "fastapi";
+  let llamaCloudKey = args.llamaCloudKey;
+
+  let useLlamaCloud = false;
+
+  if (appType !== "extractor" && appType !== "contract_review") {
+    const { language: newLanguage } = await prompts(
+      {
+        type: "select",
+        name: "language",
+        message: "What language do you want to use?",
+        choices: [
+          { title: "Python (FastAPI)", value: "fastapi" },
+          { title: "Typescript (NextJS)", value: "nextjs" },
+        ],
+      },
+      questionHandlers,
+    );
+    language = newLanguage;
+  }
+
+  const { useLlamaCloud: newUseLlamaCloud } = await prompts(
+    {
+      type: "toggle",
+      name: "useLlamaCloud",
+      message: "Do you want to use LlamaCloud services?",
+      initial: false,
+      active: "Yes",
+      inactive: "No",
+      hint: "see https://www.llamaindex.ai/enterprise for more info",
+    },
+    questionHandlers,
+  );
+  useLlamaCloud = newUseLlamaCloud;
+
+  if (useLlamaCloud && !llamaCloudKey) {
+    // Ask for LlamaCloud API key, if not set
+    const { llamaCloudKey: newLlamaCloudKey } = await prompts(
+      {
+        type: "text",
+        name: "llamaCloudKey",
+        message:
+          "Please provide your LlamaCloud API key (leave blank to skip):",
+      },
+      questionHandlers,
+    );
+    llamaCloudKey = newLlamaCloudKey || process.env.LLAMA_CLOUD_API_KEY;
+  }
+
+  const results = await convertAnswers(args, {
+    appType,
+    language,
+    useLlamaCloud,
+    llamaCloudKey,
+  });
+
+  results.postInstallAction = await askPostInstallAction(results);
+  return results;
+};
+
+const convertAnswers = async (
+  args: PureQuestionArgs,
+  answers: SimpleAnswers,
+): Promise<QuestionResults> => {
+  const MODEL_GPT4o: ModelConfig = {
+    provider: "openai",
+    apiKey: args.openAiKey,
+    model: "gpt-4o",
+    embeddingModel: "text-embedding-3-large",
+    dimensions: 1536,
+    isConfigured(): boolean {
+      return !!args.openAiKey;
+    },
+  };
+  const lookup: Record<
+    AppType,
+    Pick<QuestionResults, "template" | "tools" | "dataSources" | "useCase"> & {
+      modelConfig?: ModelConfig;
+    }
+  > = {
+    agentic_rag: {
+      template: "llamaindexserver",
+      dataSources: [EXAMPLE_FILE],
+    },
+    financial_report: {
+      template: "llamaindexserver",
+      dataSources: EXAMPLE_10K_SEC_FILES,
+      tools: getTools(["interpreter", "document_generator"]),
+      modelConfig: MODEL_GPT4o,
+    },
+    deep_research: {
+      template: "llamaindexserver",
+      dataSources: EXAMPLE_10K_SEC_FILES,
+      tools: [],
+      modelConfig: MODEL_GPT4o,
+    },
+  };
+
+  const results = lookup[answers.appType];
+  return {
+    framework: answers.language,
+    useCase: answers.appType,
+    ui: "shadcn",
+    llamaCloudKey: answers.llamaCloudKey,
+    useLlamaParse: answers.useLlamaCloud,
+    vectorDb: answers.useLlamaCloud ? "llamacloud" : "none",
+    ...results,
+    modelConfig:
+      results.modelConfig ??
+      (await askModelConfig({
+        openAiKey: args.openAiKey,
+        askModels: args.askModels ?? false,
+        framework: answers.language,
+      })),
+    frontend: true,
+  };
+};

questions/stores.ts

@@ -0,0 +1,36 @@
+import fs from "fs";
+import path from "path";
+import { TemplateFramework } from "../helpers";
+import { templatesDir } from "../helpers/dir";
+
+export const getVectorDbChoices = (framework: TemplateFramework) => {
+  const choices = [
+    {
+      title: "No, just store the data in the file system",
+      value: "none",
+    },
+    { title: "MongoDB", value: "mongo" },
+    { title: "PostgreSQL", value: "pg" },
+    { title: "Pinecone", value: "pinecone" },
+    { title: "Milvus", value: "milvus" },
+    { title: "Astra", value: "astra" },
+    { title: "Qdrant", value: "qdrant" },
+    { title: "ChromaDB", value: "chroma" },
+    { title: "Weaviate", value: "weaviate" },
+    { title: "LlamaCloud (use Managed Index)", value: "llamacloud" },
+  ];
+
+  const vectordbLang = framework === "fastapi" ? "python" : "typescript";
+  const compPath = path.join(templatesDir, "components");
+  const vectordbPath = path.join(compPath, "vectordbs", vectordbLang);
+
+  const availableChoices = fs
+    .readdirSync(vectordbPath)
+    .filter((file) => fs.statSync(path.join(vectordbPath, file)).isDirectory());
+
+  const displayedChoices = choices.filter((choice) =>
+    availableChoices.includes(choice.value),
+  );
+
+  return displayedChoices;
+};

questions/types.ts

@@ -0,0 +1,15 @@
+import { InstallAppArgs } from "../create-app";
+
+export type QuestionResults = Omit<
+  InstallAppArgs,
+  "appPath" | "packageManager"
+>;
+
+export type PureQuestionArgs = {
+  askModels?: boolean;
+  pro?: boolean;
+  openAiKey?: string;
+  llamaCloudKey?: string;
+};
+
+export type QuestionArgs = QuestionResults & PureQuestionArgs;

questions/utils.ts

@@ -0,0 +1,178 @@
+import { execSync } from "child_process";
+import fs from "fs";
+import path from "path";
+import { red } from "picocolors";
+import prompts from "prompts";
+import { TemplateDataSourceType, TemplatePostInstallAction } from "../helpers";
+import { toolsRequireConfig } from "../helpers/tools";
+import { QuestionResults } from "./types";
+
+export const supportedContextFileTypes = [
+  ".pdf",
+  ".doc",
+  ".docx",
+  ".xls",
+  ".xlsx",
+  ".csv",
+];
+
+const MACOS_FILE_SELECTION_SCRIPT = `
+osascript -l JavaScript -e '
+  a = Application.currentApplication();
+  a.includeStandardAdditions = true;
+  a.chooseFile({ withPrompt: "Please select files to process:", multipleSelectionsAllowed: true }).map(file => file.toString())
+'`;
+
+const MACOS_FOLDER_SELECTION_SCRIPT = `
+osascript -l JavaScript -e '
+  a = Application.currentApplication();
+  a.includeStandardAdditions = true;
+  a.chooseFolder({ withPrompt: "Please select folders to process:", multipleSelectionsAllowed: true }).map(folder => folder.toString())
+'`;
+
+const WINDOWS_FILE_SELECTION_SCRIPT = `
+Add-Type -AssemblyName System.Windows.Forms
+$openFileDialog = New-Object System.Windows.Forms.OpenFileDialog
+$openFileDialog.InitialDirectory = [Environment]::GetFolderPath('Desktop')
+$openFileDialog.Multiselect = $true
+$result = $openFileDialog.ShowDialog()
+if ($result -eq 'OK') {
+  $openFileDialog.FileNames
+}
+`;
+
+const WINDOWS_FOLDER_SELECTION_SCRIPT = `
+Add-Type -AssemblyName System.windows.forms
+$folderBrowser = New-Object System.Windows.Forms.FolderBrowserDialog
+$dialogResult = $folderBrowser.ShowDialog()
+if ($dialogResult -eq [System.Windows.Forms.DialogResult]::OK)
+{
+    $folderBrowser.SelectedPath
+}
+`;
+
+export const selectLocalContextData = async (type: TemplateDataSourceType) => {
+  try {
+    let selectedPath: string = "";
+    let execScript: string;
+    let execOpts: any = {};
+    switch (process.platform) {
+      case "win32": // Windows
+        execScript =
+          type === "file"
+            ? WINDOWS_FILE_SELECTION_SCRIPT
+            : WINDOWS_FOLDER_SELECTION_SCRIPT;
+        execOpts = { shell: "powershell.exe" };
+        break;
+      case "darwin": // MacOS
+        execScript =
+          type === "file"
+            ? MACOS_FILE_SELECTION_SCRIPT
+            : MACOS_FOLDER_SELECTION_SCRIPT;
+        break;
+      default: // Unsupported OS
+        console.log(red("Unsupported OS error!"));
+        process.exit(1);
+    }
+    selectedPath = execSync(execScript, execOpts).toString().trim();
+    const paths =
+      process.platform === "win32"
+        ? selectedPath.split("\r\n")
+        : selectedPath.split(", ");
+
+    for (const p of paths) {
+      if (
+        fs.statSync(p).isFile() &&
+        !supportedContextFileTypes.includes(path.extname(p))
+      ) {
+        console.log(
+          red(
+            `Please select a supported file type: ${supportedContextFileTypes}`,
+          ),
+        );
+        process.exit(1);
+      }
+    }
+    return paths;
+  } catch (error) {
+    console.log(
+      red(
+        "Got an error when trying to select local context data! Please try again or select another data source option.",
+      ),
+    );
+    process.exit(1);
+  }
+};
+
+export const onPromptState = (state: any) => {
+  if (state.aborted) {
+    // If we don't re-enable the terminal cursor before exiting
+    // the program, the cursor will remain hidden
+    process.stdout.write("\x1B[?25h");
+    process.stdout.write("\n");
+    process.exit(1);
+  }
+};
+
+export const toChoice = (value: string) => {
+  return { title: value, value };
+};
+
+export const questionHandlers = {
+  onCancel: () => {
+    console.error("Exiting.");
+    process.exit(1);
+  },
+};
+
+// Ask for next action after installation
+export async function askPostInstallAction(
+  args: QuestionResults,
+): Promise<TemplatePostInstallAction> {
+  const actionChoices = [
+    {
+      title: "Just generate code (~1 sec)",
+      value: "none",
+    },
+    {
+      title: "Start in VSCode (~1 sec)",
+      value: "VSCode",
+    },
+    {
+      title: "Generate code and install dependencies (~2 min)",
+      value: "dependencies",
+    },
+  ];
+
+  const modelConfigured = !args.llamapack && args.modelConfig.isConfigured();
+  // If using LlamaParse, require LlamaCloud API key
+  const llamaCloudKeyConfigured = args.useLlamaParse
+    ? args.llamaCloudKey || process.env["LLAMA_CLOUD_API_KEY"]
+    : true;
+  const hasVectorDb = args.vectorDb && args.vectorDb !== "none";
+  // Can run the app if all tools do not require configuration
+  if (
+    !hasVectorDb &&
+    modelConfigured &&
+    llamaCloudKeyConfigured &&
+    !toolsRequireConfig(args.tools)
+  ) {
+    actionChoices.push({
+      title: "Generate code, install dependencies, and run the app (~2 min)",
+      value: "runApp",
+    });
+  }
+
+  const { action } = await prompts(
+    {
+      type: "select",
+      name: "action",
+      message: "How would you like to proceed?",
+      choices: actionChoices,
+      initial: 1,
+    },
+    questionHandlers,
+  );
+
+  return action;
+}

templates/README-fullstack.md

@@ -1,18 +0,0 @@
-This is a [LlamaIndex](https://www.llamaindex.ai/) project bootstrapped with [`create-llama`](https://github.com/run-llama/LlamaIndexTS/tree/main/packages/create-llama).
-
-## Getting Started
-
-First, startup the backend as described in the [backend README](./backend/README.md).
-
-Second, run the development server of the frontend as described in the [frontend README](./frontend/README.md).
-
-Open [http://localhost:3000](http://localhost:3000) with your browser to see the result.
-
-## Learn More
-
-To learn more about LlamaIndex, take a look at the following resources:
-
-- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex (Python features).
-- [LlamaIndexTS Documentation](https://ts.llamaindex.ai) - learn about LlamaIndex (Typescript features).
-
-You can check out [the LlamaIndexTS GitHub repository](https://github.com/run-llama/LlamaIndexTS) - your feedback and contributions are welcome!

templates/components/agents/python/blog/README-template.md

@@ -0,0 +1,68 @@
+## Overview
+
+This example is using three agents to generate a blog post:
+
+- a researcher that retrieves content via a RAG pipeline,
+- a writer that specializes in writing blog posts and
+- a reviewer that is reviewing the blog post.
+
+There are three different methods how the agents can interact to reach their goal:
+
+1. [Choreography](./app/agents/choreography.py) - the agents decide themselves to delegate a task to another agent
+1. [Orchestrator](./app/agents/orchestrator.py) - a central orchestrator decides which agent should execute a task
+1. [Explicit Workflow](./app/agents/workflow.py) - a pre-defined workflow specific for the task is used to execute the tasks
+
+## Getting Started
+
+First, setup the environment with poetry:
+
+> **_Note:_** This step is not needed if you are using the dev-container.
+
+```shell
+poetry install
+```
+
+Then check the parameters that have been pre-configured in the `.env` file in this directory. (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:
+
+```shell
+poetry run generate
+```
+
+Third, run the development server:
+
+```shell
+poetry run dev
+```
+
+Per default, the example is using the explicit workflow. You can change the example by setting the `EXAMPLE_TYPE` environment variable to `choreography` or `orchestrator`.
+The example provides one streaming API endpoint `/api/chat`.
+You can test the endpoint with the following curl request:
+
+```
+curl --location 'localhost:8000/api/chat' \
+--header 'Content-Type: application/json' \
+--data '{ "messages": [{ "role": "user", "content": "Write a blog post about physical standards for letters" }] }'
+```
+
+You can start editing the API by modifying `app/api/routers/chat.py` or `app/examples/workflow.py`. The API auto-updates as you save the files.
+
+Open [http://localhost:8000](http://localhost:8000) with your browser to start the app.
+
+To start the app optimized for **production**, run:
+
+```
+poetry run prod
+```
+
+## Deployments
+
+For production deployments, check the [DEPLOY.md](DEPLOY.md) file.
+
+## Learn More
+
+To learn more about LlamaIndex, take a look at the following resources:
+
+- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex.
+- [Workflows Introduction](https://docs.llamaindex.ai/en/stable/understanding/workflows/) - learn about LlamaIndex workflows.
+  You can check out [the LlamaIndex GitHub repository](https://github.com/run-llama/llama_index) - your feedback and contributions are welcome!

templates/components/agents/python/blog/app/agents/choreography.py

@@ -0,0 +1,34 @@
+from textwrap import dedent
+from typing import List, Optional
+
+from app.agents.publisher import create_publisher
+from app.agents.researcher import create_researcher
+from app.workflows.multi import AgentCallingAgent
+from app.workflows.single import FunctionCallingAgent
+from llama_index.core.chat_engine.types import ChatMessage
+
+
+def create_choreography(chat_history: Optional[List[ChatMessage]] = None, **kwargs):
+    researcher = create_researcher(chat_history, **kwargs)
+    publisher = create_publisher(chat_history)
+    reviewer = FunctionCallingAgent(
+        name="reviewer",
+        description="expert in reviewing blog posts, needs a written post to review",
+        system_prompt="You are an expert in reviewing blog posts. You are given a task to review a blog post. Review the post for logical inconsistencies, ask critical questions, and provide suggestions for improvement. Furthermore, proofread the post for grammar and spelling errors. If the post is good, you can say 'The post is good.'",
+        chat_history=chat_history,
+    )
+    return AgentCallingAgent(
+        name="writer",
+        agents=[researcher, reviewer, publisher],
+        description="expert in writing blog posts, needs researched information and images to write a blog post",
+        system_prompt=dedent(
+            """
+            You are an expert in writing blog posts. You are given a task to write a blog post. Before starting to write the post, consult the researcher agent to get the information you need. Don't make up any information yourself.
+            After creating a draft for the post, send it to the reviewer agent to receive feedback and make sure to incorporate the feedback from the reviewer.
+            You can consult the reviewer and researcher a maximum of two times. Your output should contain only the blog post.
+            Finally, always request the publisher to create a document (PDF, HTML) and publish the blog post.
+        """
+        ),
+        # TODO: add chat_history support to AgentCallingAgent
+        # chat_history=chat_history,
+    )

templates/components/agents/python/blog/app/agents/orchestrator.py

@@ -0,0 +1,44 @@
+from textwrap import dedent
+from typing import List, Optional
+
+from app.agents.publisher import create_publisher
+from app.agents.researcher import create_researcher
+from app.workflows.multi import AgentOrchestrator
+from app.workflows.single import FunctionCallingAgent
+from llama_index.core.chat_engine.types import ChatMessage
+
+
+def create_orchestrator(chat_history: Optional[List[ChatMessage]] = None, **kwargs):
+    researcher = create_researcher(chat_history, **kwargs)
+    writer = FunctionCallingAgent(
+        name="writer",
+        description="expert in writing blog posts, need information and images to write a post",
+        system_prompt=dedent(
+            """
+            You are an expert in writing blog posts.
+            You are given a task to write a blog post. Do not make up any information yourself.
+            If you don't have the necessary information to write a blog post, reply "I need information about the topic to write the blog post".
+            If you need to use images, reply "I need images about the topic to write the blog post". Do not use any dummy images made up by you.
+            If you have all the information needed, write the blog post.
+        """
+        ),
+        chat_history=chat_history,
+    )
+    reviewer = FunctionCallingAgent(
+        name="reviewer",
+        description="expert in reviewing blog posts, needs a written blog post to review",
+        system_prompt=dedent(
+            """
+            You are an expert in reviewing blog posts. You are given a task to review a blog post. Review the post and fix any issues found yourself. You must output a final blog post.
+            A post must include at least one valid image. If not, reply "I need images about the topic to write the blog post". An image URL starting with "example" or "your website" is not valid.
+            Especially check for logical inconsistencies and proofread the post for grammar and spelling errors.
+        """
+        ),
+        chat_history=chat_history,
+    )
+    publisher = create_publisher(chat_history)
+    return AgentOrchestrator(
+        agents=[writer, reviewer, researcher, publisher],
+        refine_plan=False,
+        chat_history=chat_history,
+    )

templates/components/agents/python/blog/app/agents/publisher.py

@@ -0,0 +1,35 @@
+from textwrap import dedent
+from typing import List, Tuple
+
+from app.engine.tools import ToolFactory
+from app.workflows.single import FunctionCallingAgent
+from llama_index.core.chat_engine.types import ChatMessage
+from llama_index.core.tools import FunctionTool
+
+
+def get_publisher_tools() -> Tuple[List[FunctionTool], str, str]:
+    tools = []
+    # Get configured tools from the tools.yaml file
+    configured_tools = ToolFactory.from_env(map_result=True)
+    if "generate_document" in configured_tools.keys():
+        tools.append(configured_tools["generate_document"])
+        prompt_instructions = dedent("""
+            Normally, reply the blog post content to the user directly. 
+            But if user requested to generate a file, use the generate_document tool to generate the file and reply the link to the file.
+        """)
+        description = "Expert in publishing the blog post, able to publish the blog post in PDF or HTML format."
+    else:
+        prompt_instructions = "You don't have a tool to generate document. Please reply the content directly."
+        description = "Expert in publishing the blog post"
+    return tools, prompt_instructions, description
+
+
+def create_publisher(chat_history: List[ChatMessage]):
+    tools, prompt_instructions, description = get_publisher_tools()
+    return FunctionCallingAgent(
+        name="publisher",
+        tools=tools,
+        description=description,
+        system_prompt=prompt_instructions,
+        chat_history=chat_history,
+    )

templates/components/agents/python/blog/app/agents/researcher.py

@@ -0,0 +1,71 @@
+from textwrap import dedent
+from typing import List
+
+from app.engine.index import IndexConfig, get_index
+from app.engine.tools import ToolFactory
+from app.workflows.single import FunctionCallingAgent
+from llama_index.core.chat_engine.types import ChatMessage
+from app.engine.tools.query_engine import get_query_engine_tool
+
+
+def _get_research_tools(**kwargs):
+    """
+    Researcher take responsibility for retrieving information.
+    Try init wikipedia or duckduckgo tool if available.
+    """
+    tools = []
+    # Create query engine tool
+    index_config = IndexConfig(**kwargs)
+    index = get_index(index_config)
+    if index is not None:
+        query_engine_tool = get_query_engine_tool(index=index)
+        if query_engine_tool is not None:
+            tools.append(query_engine_tool)
+
+    # Create duckduckgo tool
+    researcher_tool_names = [
+        "duckduckgo_search",
+        "duckduckgo_image_search",
+        "wikipedia.WikipediaToolSpec",
+    ]
+    configured_tools = ToolFactory.from_env(map_result=True)
+    for tool_name, tool in configured_tools.items():
+        if tool_name in researcher_tool_names:
+            tools.append(tool)
+    return tools
+
+
+def create_researcher(chat_history: List[ChatMessage], **kwargs):
+    """
+    Researcher is an agent that take responsibility for using tools to complete a given task.
+    """
+    tools = _get_research_tools(**kwargs)
+    return FunctionCallingAgent(
+        name="researcher",
+        tools=tools,
+        description="expert in retrieving any unknown content or searching for images from the internet",
+        system_prompt=dedent(
+            """
+            You are a researcher agent. You are given a research task.
+            
+            If the conversation already includes the information and there is no new request for additional information from the user, you should return the appropriate content to the writer.
+            Otherwise, you must use tools to retrieve information or images needed for the task.
+
+            It's normal for the task to include some ambiguity. You must always think carefully about the context of the user's request to understand what are the main content needs to be retrieved.
+            Example:
+                Request: "Create a blog post about the history of the internet, write in English and publish in PDF format."
+                ->Though: The main content is "history of the internet", while "write in English and publish in PDF format" is a requirement for other agents.
+                Your task: Look for information in English about the history of the Internet.
+                This is not your task: Create a blog post or look for how to create a PDF.
+
+                Next request: "Publish the blog post in HTML format."
+                ->Though: User just asking for a format change, the previous content is still valid.
+                Your task: Return the previous content of the post to the writer. No need to do any research.
+                This is not your task: Look for how to create an HTML file.
+
+            If you use the tools but don't find any related information, please return "I didn't find any new information for {the topic}." along with the content you found. Don't try to make up information yourself.
+            If the request doesn't need any new information because it was in the conversation history, please return "The task doesn't need any new information. Please reuse the existing content in the conversation history."
+        """
+        ),
+        chat_history=chat_history,
+    )

templates/components/agents/python/blog/app/agents/workflow.py

@@ -0,0 +1,267 @@
+from textwrap import dedent
+from typing import AsyncGenerator, List, Optional
+
+from app.agents.publisher import create_publisher
+from app.agents.researcher import create_researcher
+from app.workflows.single import AgentRunEvent, AgentRunResult, FunctionCallingAgent
+from llama_index.core.chat_engine.types import ChatMessage
+from llama_index.core.prompts import PromptTemplate
+from llama_index.core.settings import Settings
+from llama_index.core.workflow import (
+    Context,
+    Event,
+    StartEvent,
+    StopEvent,
+    Workflow,
+    step,
+)
+
+
+def create_workflow(chat_history: Optional[List[ChatMessage]] = None, **kwargs):
+    researcher = create_researcher(
+        chat_history=chat_history,
+        **kwargs,
+    )
+    publisher = create_publisher(
+        chat_history=chat_history,
+    )
+    writer = FunctionCallingAgent(
+        name="writer",
+        description="expert in writing blog posts, need information and images to write a post.",
+        system_prompt=dedent(
+            """
+            You are an expert in writing blog posts.
+            You are given the task of writing a blog post based on research content provided by the researcher agent. Do not invent any information yourself. 
+            It's important to read the entire conversation history to write the blog post accurately.
+            If you receive a review from the reviewer, update the post according to the feedback and return the new post content.
+            If the content is not valid (e.g., broken link, broken image, etc.), do not use it.
+            It's normal for the task to include some ambiguity, so you must define the user's initial request to write the post correctly.
+            If you update the post based on the reviewer's feedback, first explain what changes you made to the post, then provide the new post content. Do not include the reviewer's comments.
+            Example:
+                Task: "Here is the information I found about the history of the internet: 
+                Create a blog post about the history of the internet, write in English, and publish in PDF format."
+                -> Your task: Use the research content {...} to write a blog post in English.
+                -> This is not your task: Create a PDF
+                Please note that a localhost link is acceptable, but dummy links like "example.com" or "your-website.com" are not valid.
+        """
+        ),
+        chat_history=chat_history,
+    )
+    reviewer = FunctionCallingAgent(
+        name="reviewer",
+        description="expert in reviewing blog posts, needs a written blog post to review.",
+        system_prompt=dedent(
+            """
+            You are an expert in reviewing blog posts.
+            You are given a task to review a blog post. As a reviewer, it's important that your review aligns with the user's request. Please focus on the user's request when reviewing the post.
+            Review the post for logical inconsistencies, ask critical questions, and provide suggestions for improvement.
+            Furthermore, proofread the post for grammar and spelling errors.
+            Only if the post is good enough for publishing should you return 'The post is good.' In all other cases, return your review.
+            It's normal for the task to include some ambiguity, so you must define the user's initial request to review the post correctly.
+            Please note that a localhost link is acceptable, but dummy links like "example.com" or "your-website.com" are not valid.
+            Example:
+                Task: "Create a blog post about the history of the internet, write in English and publish in PDF format."
+                -> Your task: Review whether the main content of the post is about the history of the internet and if it is written in English.
+                -> This is not your task: Create blog post, create PDF, write in English.
+        """
+        ),
+        chat_history=chat_history,
+    )
+    workflow = BlogPostWorkflow(
+        timeout=360, chat_history=chat_history
+    )  # Pass chat_history here
+    workflow.add_workflows(
+        researcher=researcher,
+        writer=writer,
+        reviewer=reviewer,
+        publisher=publisher,
+    )
+    return workflow
+
+
+class ResearchEvent(Event):
+    input: str
+
+
+class WriteEvent(Event):
+    input: str
+    is_good: bool = False
+
+
+class ReviewEvent(Event):
+    input: str
+
+
+class PublishEvent(Event):
+    input: str
+
+
+class BlogPostWorkflow(Workflow):
+    def __init__(
+        self, timeout: int = 360, chat_history: Optional[List[ChatMessage]] = None
+    ):
+        super().__init__(timeout=timeout)
+        self.chat_history = chat_history or []
+
+    @step()
+    async def start(self, ctx: Context, ev: StartEvent) -> ResearchEvent | PublishEvent:
+        # set streaming
+        ctx.data["streaming"] = getattr(ev, "streaming", False)
+        # start the workflow with researching about a topic
+        ctx.data["task"] = ev.input
+        ctx.data["user_input"] = ev.input
+
+        # Decision-making process
+        decision = await self._decide_workflow(ev.input, self.chat_history)
+
+        if decision != "publish":
+            return ResearchEvent(input=f"Research for this task: {ev.input}")
+        else:
+            chat_history_str = "\n".join(
+                [f"{msg.role}: {msg.content}" for msg in self.chat_history]
+            )
+            return PublishEvent(
+                input=f"Please publish content based on the chat history\n{chat_history_str}\n\n and task: {ev.input}"
+            )
+
+    async def _decide_workflow(
+        self, input: str, chat_history: List[ChatMessage]
+    ) -> str:
+        prompt_template = PromptTemplate(
+            dedent(
+                """
+                You are an expert in decision-making, helping people write and publish blog posts.
+                If the user is asking for a file or to publish content, respond with 'publish'.
+                If the user requests to write or update a blog post, respond with 'not_publish'.
+
+                Here is the chat history:
+                {chat_history}
+
+                The current user request is:
+                {input}
+
+                Given the chat history and the new user request, decide whether to publish based on existing information.
+                Decision (respond with either 'not_publish' or 'publish'):
+            """
+            )
+        )
+
+        chat_history_str = "\n".join(
+            [f"{msg.role}: {msg.content}" for msg in chat_history]
+        )
+        prompt = prompt_template.format(chat_history=chat_history_str, input=input)
+
+        output = await Settings.llm.acomplete(prompt)
+        decision = output.text.strip().lower()
+
+        return "publish" if decision == "publish" else "research"
+
+    @step()
+    async def research(
+        self, ctx: Context, ev: ResearchEvent, researcher: FunctionCallingAgent
+    ) -> WriteEvent:
+        result: AgentRunResult = await self.run_agent(ctx, researcher, ev.input)
+        content = result.response.message.content
+        return WriteEvent(
+            input=f"Write a blog post given this task: {ctx.data['task']} using this research content: {content}"
+        )
+
+    @step()
+    async def write(
+        self, ctx: Context, ev: WriteEvent, writer: FunctionCallingAgent
+    ) -> ReviewEvent | StopEvent:
+        MAX_ATTEMPTS = 2
+        ctx.data["attempts"] = ctx.data.get("attempts", 0) + 1
+        too_many_attempts = ctx.data["attempts"] > MAX_ATTEMPTS
+        if too_many_attempts:
+            ctx.write_event_to_stream(
+                AgentRunEvent(
+                    name=writer.name,
+                    msg=f"Too many attempts ({MAX_ATTEMPTS}) to write the blog post. Proceeding with the current version.",
+                )
+            )
+        if ev.is_good or too_many_attempts:
+            # too many attempts or the blog post is good - stream final response if requested
+            result = await self.run_agent(
+                ctx,
+                writer,
+                f"Based on the reviewer's feedback, refine the post and return only the final version of the post. Here's the current version: {ev.input}",
+                streaming=ctx.data["streaming"],
+            )
+            return StopEvent(result=result)
+        result: AgentRunResult = await self.run_agent(ctx, writer, ev.input)
+        ctx.data["result"] = result
+        return ReviewEvent(input=result.response.message.content)
+
+    @step()
+    async def review(
+        self, ctx: Context, ev: ReviewEvent, reviewer: FunctionCallingAgent
+    ) -> WriteEvent:
+        result: AgentRunResult = await self.run_agent(ctx, reviewer, ev.input)
+        review = result.response.message.content
+        old_content = ctx.data["result"].response.message.content
+        post_is_good = "post is good" in review.lower()
+        ctx.write_event_to_stream(
+            AgentRunEvent(
+                name=reviewer.name,
+                msg=f"The post is {'not ' if not post_is_good else ''}good enough for publishing. Sending back to the writer{' for publication.' if post_is_good else '.'}",
+            )
+        )
+        if post_is_good:
+            return WriteEvent(
+                input=f"You're blog post is ready for publication. Please respond with just the blog post. Blog post: ```{old_content}```",
+                is_good=True,
+            )
+        else:
+            return WriteEvent(
+                input=dedent(
+                    f"""
+                    Improve the writing of a given blog post by using a given review.
+                    Blog post:
+                    ```
+                    {old_content}
+                    ``` 
+
+                    Review: 
+                    ```
+                    {review}
+                    ```
+                    """
+                ),
+            )
+
+    @step()
+    async def publish(
+        self,
+        ctx: Context,
+        ev: PublishEvent,
+        publisher: FunctionCallingAgent,
+    ) -> StopEvent:
+        try:
+            result: AgentRunResult = await self.run_agent(
+                ctx, publisher, ev.input, streaming=ctx.data["streaming"]
+            )
+            return StopEvent(result=result)
+        except Exception as e:
+            ctx.write_event_to_stream(
+                AgentRunEvent(
+                    name=publisher.name,
+                    msg=f"Error publishing: {e}",
+                )
+            )
+            return StopEvent(result=None)
+
+    async def run_agent(
+        self,
+        ctx: Context,
+        agent: FunctionCallingAgent,
+        input: str,
+        streaming: bool = False,
+    ) -> AgentRunResult | AsyncGenerator:
+        handler = agent.run(input=input, streaming=streaming)
+        # bubble all events while running the executor to the planner
+        async for event in handler.stream_events():
+            # Don't write the StopEvent from sub task to the stream
+            if type(event) is not StopEvent:
+                ctx.write_event_to_stream(event)
+        return await handler