pref: optimize feature model_load_balancing_enabled value fetch speed… (#4933)

Co-authored-by: Jyong <76649700+johnjyong@users.noreply.github.com>

.github/workflows/api-tests.yml

@@ -4,9 +4,17 @@ on:
   pull_request:
     branches:
       - main
+    paths:
+      - api/**
+      - docker/**
+
+concurrency:
+  group: api-tests-${{ github.head_ref || github.run_id }}
+  cancel-in-progress: true
 
 jobs:
   test:
+    name: API Tests
     runs-on: ubuntu-latest
     strategy:
       matrix:

.github/workflows/db-migration-test.yml

@@ -0,0 +1,53 @@
+name: DB Migration Test
+
+on:
+  pull_request:
+    branches:
+      - main
+    paths:
+      - api/migrations/**
+
+concurrency:
+  group: db-migration-test-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  db-migration-test:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version:
+          - "3.10"
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: 'pip'
+          cache-dependency-path: |
+            ./api/requirements.txt
+
+      - name: Install dependencies
+        run: pip install -r ./api/requirements.txt
+
+      - name: Set up Middleware
+        uses: hoverkraft-tech/compose-action@v2.0.0
+        with:
+          compose-file: |
+            docker/docker-compose.middleware.yaml
+          services: |
+            db
+
+      - name: Prepare configs
+        run: |
+          cd api
+          cp .env.example .env
+
+      - name: Run DB Migration
+        run: |
+          cd api
+          flask db upgrade

.github/workflows/style.yml

@@ -6,7 +6,7 @@ on:
       - main
 
 concurrency:
-  group: dep-${{ github.head_ref || github.run_id }}
+  group: style-${{ github.head_ref || github.run_id }}
   cancel-in-progress: true
 
 jobs:
@@ -18,54 +18,89 @@ jobs:
       - name: Checkout code
         uses: actions/checkout@v4
 
+      - name: Check changed files
+        id: changed-files
+        uses: tj-actions/changed-files@v44
+        with:
+          files: api/**
+
       - name: Set up Python
         uses: actions/setup-python@v5
+        if: steps.changed-files.outputs.any_changed == 'true'
         with:
           python-version: '3.10'
 
       - name: Python dependencies
+        if: steps.changed-files.outputs.any_changed == 'true'
         run: pip install ruff dotenv-linter
 
       - name: Ruff check
+        if: steps.changed-files.outputs.any_changed == 'true'
         run: ruff check ./api
 
       - name: Dotenv check
+        if: steps.changed-files.outputs.any_changed == 'true'
         run: dotenv-linter ./api/.env.example ./web/.env.example
 
       - name: Lint hints
         if: failure()
         run: echo "Please run 'dev/reformat' to fix the fixable linting errors."
 
-  test:
-    name: ESLint and SuperLinter
+  web-style:
+    name: Web Style
     runs-on: ubuntu-latest
-    needs: python-style
+    defaults:
+      run:
+        working-directory: ./web
 
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
+
+      - name: Check changed files
+        id: changed-files
+        uses: tj-actions/changed-files@v44
         with:
-          fetch-depth: 0
+          files: web/**
 
       - name: Setup NodeJS
         uses: actions/setup-node@v4
+        if: steps.changed-files.outputs.any_changed == 'true'
         with:
           node-version: 20
           cache: yarn
           cache-dependency-path: ./web/package.json
 
       - name: Web dependencies
-        run: |
-          cd ./web
-          yarn install --frozen-lockfile
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: yarn install --frozen-lockfile
 
       - name: Web style check
-        run: |
-          cd ./web
-          yarn run lint
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: yarn run lint
+
+
+  superlinter:
+    name: SuperLinter
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Check changed files
+        id: changed-files
+        uses: tj-actions/changed-files@v44
+        with:
+          files: |
+            **.sh
+            **.yaml
+            **.yml
+            Dockerfile
 
       - name: Super-linter
         uses: super-linter/super-linter/slim@v6
+        if: steps.changed-files.outputs.any_changed == 'true'
         env:
           BASH_SEVERITY: warning
           DEFAULT_BRANCH: main
@@ -76,4 +111,5 @@ jobs:
           VALIDATE_BASH_EXEC: true
           VALIDATE_GITHUB_ACTIONS: true
           VALIDATE_DOCKERFILE_HADOLINT: true
+          VALIDATE_XML: true
           VALIDATE_YAML: true

.github/workflows/tool-test-sdks.yaml

@@ -4,6 +4,13 @@ on:
   pull_request:
     branches:
       - main
+    paths:
+      - sdks/**
+
+concurrency:
+  group: sdk-tests-${{ github.head_ref || github.run_id }}
+  cancel-in-progress: true
+
 jobs:
   build:
     name: unit test for Node.js SDK

.gitignore

@@ -134,7 +134,8 @@ dmypy.json
 web/.vscode/settings.json
 
 # Intellij IDEA Files
-.idea/
+.idea/*
+!.idea/vcs.xml
 .ideaDataSources/
 
 api/.env

.idea/vcs.xml

@@ -0,0 +1,16 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="IssueNavigationConfiguration">
+    <option name="links">
+      <list>
+        <IssueNavigationLink>
+          <option name="issueRegexp" value="#(\d+)" />
+          <option name="linkRegexp" value="https://github.com/langgenius/dify/issues/$1" />
+        </IssueNavigationLink>
+      </list>
+    </option>
+  </component>
+  <component name="VcsDirectoryMappings">
+    <mapping directory="" vcs="Git" />
+  </component>
+</project>

CONTRIBUTING_JA.md

@@ -0,0 +1,160 @@
+Dify にコントリビュートしたいとお考えなのですね。それは素晴らしいことです。
+私たちは、LLM アプリケーションの構築と管理のための最も直感的なワークフローを設計するという壮大な野望を持っています。人数も資金も限られている新興企業として、コミュニティからの支援は本当に重要です。
+
+私たちは現状を鑑み、機敏かつ迅速に開発をする必要がありますが、同時にあなたのようなコントリビューターの方々に、可能な限りスムーズな貢献体験をしていただきたいと思っています。そのためにこのコントリビュートガイドを作成しました。
+コードベースやコントリビュータの方々と私たちがどのように仕事をしているのかに慣れていただき、楽しいパートにすぐに飛び込めるようにすることが目的です。
+
+このガイドは Dify そのものと同様に、継続的に改善されています。実際のプロジェクトに遅れをとることがあるかもしれませんが、ご理解をお願いします。
+
+ライセンスに関しては、私たちの短い[ライセンスおよびコントリビューター規約](./LICENSE)をお読みください。また、コミュニティは[行動規範](https://github.com/langgenius/.github/blob/main/CODE_OF_CONDUCT.md)を遵守しています。
+
+## 飛び込む前に
+
+[既存の Issue](https://github.com/langgenius/dify/issues?q=is:issue+is:closed) を探すか、[新しい Issue](https://github.com/langgenius/dify/issues/new/choose) を作成してください。私たちは Issue を 2 つのタイプに分類しています。
+
+### 機能リクエスト
+
+* 新しい機能要望を出す場合は、提案する機能が何を実現するものなのかを説明し、可能な限り多くの文脈を含めてください。[@perzeusss](https://github.com/perzeuss)は、あなたの要望を書き出すのに役立つ [Feature Request Copilot](https://udify.app/chat/MK2kVSnw1gakVwMX) を作ってくれました。気軽に試してみてください。
+
+* 既存の課題から 1 つ選びたい場合は、その下にコメントを書いてください。
+
+  関連する方向で作業しているチームメンバーが参加します。すべてが良好であれば、コーディングを開始する許可が与えられます。私たちが変更を提案した場合にあなたの作業が無駄になることがないよう、それまでこの機能の作業を控えていただくようお願いいたします。
+
+  提案された機能がどの分野に属するかによって、あなたは異なるチーム・メンバーと話をするかもしれません。以下は、各チームメンバーが現在取り組んでいる分野の概要です。
+
+| Member                                                                                  | Scope                                |
+| --------------------------------------------------------------------------------------- | ------------------------------------ |
+| [@yeuoly](https://github.com/Yeuoly)                                                    | エージェントアーキテクチャ           |
+| [@jyong](https://github.com/JohnJyong)                                                  | RAG パイプライン設計                 |
+| [@GarfieldDai](https://github.com/GarfieldDai)                                          | workflow orchestrations の構築       |
+| [@iamjoel](https://github.com/iamjoel) & [@zxhlyh](https://github.com/zxhlyh)           | フロントエンドを使いやすくする       |
+| [@guchenhe](https://github.com/guchenhe) & [@crazywoola](https://github.com/crazywoola) | 開発者体験、何でも相談できる窓口     |
+| [@takatost](https://github.com/takatost)                                                | 全体的な製品の方向性とアーキテクチャ |
+
+優先順位の付け方:
+
+| Feature Type                                                                                                          | Priority        |
+| --------------------------------------------------------------------------------------------------------------------- | --------------- |
+| チームメンバーによってラベル付けされた優先度の高い機能                                                                | High Priority   |
+| [community feedback board](https://github.com/langgenius/dify/discussions/categories/feedbacks)の人気の機能リクエスト | Medium Priority |
+| 非コア機能とマイナーな機能強化                                                                                        | Low Priority    |
+| 価値はあるが即効性はない                                                                                              | Future-Feature  |
+
+### その他 (バグレポート、パフォーマンスの最適化、誤字の修正など)
+
+* すぐにコーディングを始めてください
+
+優先順位の付け方:
+
+| Issue Type                                                                             | Priority        |
+| -------------------------------------------------------------------------------------- | --------------- |
+| コア機能のバグ（ログインできない、アプリケーションが動作しない、セキュリティの抜け穴） | Critical        |
+| 致命的でないバグ、パフォーマンス向上                                                   | Medium Priority |
+| 細かな修正（誤字脱字、機能はするが分かりにくい UI）                                  | Low Priority    |
+
+## インストール
+
+Dify を開発用にセットアップする手順は以下の通りです。
+
+### 1. このリポジトリをフォークする
+
+### 2. リポジトリをクローンする
+
+フォークしたリポジトリをターミナルからクローンします。
+
+```
+git clone git@github.com:<github_username>/dify.git
+```
+
+### 3. 依存関係の確認
+
+Dify を構築するには次の依存関係が必要です。それらがシステムにインストールされていることを確認してください。
+
+- [Docker](https://www.docker.com/)
+- [Docker Compose](https://docs.docker.com/compose/install/)
+- [Node.js v18.x (LTS)](http://nodejs.org)
+- [npm](https://www.npmjs.com/) version 8.x.x or [Yarn](https://yarnpkg.com/)
+- [Python](https://www.python.org/) version 3.10.x
+
+### 4. インストール
+
+Dify はバックエンドとフロントエンドから構成されています。
+まず`cd api/`でバックエンドのディレクトリに移動し、[Backend README](api/README.md)に従ってインストールします。
+次に別のターミナルで、`cd web/`でフロントエンドのディレクトリに移動し、[Frontend README](web/README.md)に従ってインストールしてください。
+
+よくある問題とトラブルシューティングの手順については、[installation FAQ](https://docs.dify.ai/getting-started/faq/install-faq) を確認してください。
+
+### 5. ブラウザで dify にアクセスする
+
+設定を確認するために、ブラウザで[http://localhost:3000](http://localhost:3000)(デフォルト、または自分で設定した URL とポート)にアクセスしてください。Dify が起動して実行中であることが確認できるはずです。
+
+## 開発中
+
+モデルプロバイダーを追加する場合は、[このガイド](https://github.com/langgenius/dify/blob/main/api/core/model_runtime/README.md)が役立ちます。
+
+Agent や Workflow にツールプロバイダーを追加する場合は、[このガイド](./api/core/tools/README.md)が役立ちます。
+
+Dify のバックエンドとフロントエンドの概要を簡単に説明します。
+
+### バックエンド
+
+Dify のバックエンドは[Flask](https://flask.palletsprojects.com/en/3.0.x/)を使って Python で書かれています。ORM には[SQLAlchemy](https://www.sqlalchemy.org/)を、タスクキューには[Celery](https://docs.celeryq.dev/en/stable/getting-started/introduction.html)を使っています。認証ロジックは Flask-login 経由で行われます。
+
+```
+[api/]
+├── constants             // コードベース全体で使用される定数設定
+├── controllers           // APIルート定義とリクエスト処理ロジック
+├── core                  // アプリケーションの中核的な管理、モデル統合、およびツール
+├── docker                // Dockerおよびコンテナ関連の設定
+├── events                // イベントのハンドリングと処理
+├── extensions            // 第三者のフレームワーク/プラットフォームとの拡張
+├── fields                // シリアライゼーション/マーシャリング用のフィールド定義
+├── libs                  // 再利用可能なライブラリとヘルパー
+├── migrations            // データベースマイグレーションスクリプト
+├── models                // データベースモデルとスキーマ定義
+├── services              // ビジネスロジックの定義
+├── storage               // 秘密鍵の保存
+├── tasks                 // 非同期タスクとバックグラウンドジョブの処理
+└── tests                 // テスト関連のファイル
+```
+
+### フロントエンド
+
+このウェブサイトは、Typescript の[Next.js](https://nextjs.org/)ボイラープレートでブートストラップされており、スタイリングには[Tailwind CSS](https://tailwindcss.com/)を使用しています。国際化には[React-i18next](https://react.i18next.com/)を使用しています。
+
+```
+[web/]
+├── app                   // レイアウト、ページ、コンポーネント
+│   ├── (commonLayout)    // アプリ全体で共通のレイアウト
+│   ├── (shareLayout)     // トークン特有のセッションで共有されるレイアウト
+│   ├── activate          // アクティベートページ
+│   ├── components        // ページやレイアウトで共有されるコンポーネント
+│   ├── install           // インストールページ
+│   ├── signin            // サインインページ
+│   └── styles            // グローバルに共有されるスタイル
+├── assets                // 静的アセット
+├── bin                   // ビルドステップで実行されるスクリプト
+├── config                // 調整可能な設定とオプション
+├── context               // アプリの異なる部分で使用される共有コンテキスト
+├── dictionaries          // 言語別の翻訳ファイル
+├── docker                // コンテナ設定
+├── hooks                 // 再利用可能なフック
+├── i18n                  // 国際化設定
+├── models                // データモデルとAPIレスポンスの形状を記述
+├── public                // ファビコンなどのメタアセット
+├── service               // APIアクションの形状を指定
+├── test
+├── types                 // 関数のパラメータと戻り値の記述
+└── utils                 // 共有ユーティリティ関数
+```
+
+## PR を投稿する
+
+いよいよ、私たちのリポジトリにプルリクエスト (PR) を提出する時が来ました。主要な機能については、まず `deploy/dev` ブランチにマージしてテストしてから `main` ブランチにマージします。
+マージ競合などの問題が発生した場合、またはプル リクエストを開く方法がわからない場合は、[GitHub's pull request tutorial](https://docs.github.com/en/pull-requests/collaborating-with-pull-requests) をチェックしてみてください。
+
+これで完了です！あなたの PR がマージされると、[README](https://github.com/langgenius/dify/blob/main/README.md) にコントリビューターとして紹介されます。
+
+## ヘルプを得る
+
+コントリビュート中に行き詰まったり、疑問が生じたりした場合は、GitHub の関連する issue から質問していただくか、[Discord](https://discord.gg/8Tpq4AcN9c)でチャットしてください。

README.md

@@ -35,6 +35,7 @@
   <a href="./README_ES.md"><img alt="README en Español" src="https://img.shields.io/badge/Español-d9d9d9"></a>
   <a href="./README_FR.md"><img alt="README en Français" src="https://img.shields.io/badge/Français-d9d9d9"></a>
   <a href="./README_KL.md"><img alt="README tlhIngan Hol" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
+  <a href="./README_KR.md"><img alt="README in Korean" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
 </p>
 
 

README_CN.md

@@ -35,6 +35,7 @@
   <a href="./README_ES.md"><img alt="上个月的提交次数" src="https://img.shields.io/badge/西班牙语-d9d9d9"></a>
   <a href="./README_KL.md"><img alt="上个月的提交次数" src="https://img.shields.io/badge/法语-d9d9d9"></a>
   <a href="./README_FR.md"><img alt="上个月的提交次数" src="https://img.shields.io/badge/克林贡语-d9d9d9"></a>
+  <a href="./README_KR.md"><img alt="上个月的提交次数" src="https://img.shields.io/badge/韓國語-d9d9d9"></a>
 </div>
 
 

README_ES.md

@@ -35,6 +35,7 @@
   <a href="./README_ES.md"><img alt="Actividad de Commits el último mes" src="https://img.shields.io/badge/Español-d9d9d9"></a>
   <a href="./README_KL.md"><img alt="Actividad de Commits el último mes" src="https://img.shields.io/badge/Français-d9d9d9"></a>
   <a href="./README_FR.md"><img alt="Actividad de Commits el último mes" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
+  <a href="./README_KR.md"><img alt="Actividad de Commits el último mes" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
 </p>
 
 #

README_FR.md

@@ -35,6 +35,7 @@
   <a href="./README_ES.md"><img alt="Commits le mois dernier" src="https://img.shields.io/badge/Español-d9d9d9"></a>
   <a href="./README_KL.md"><img alt="Commits le mois dernier" src="https://img.shields.io/badge/Français-d9d9d9"></a>
   <a href="./README_FR.md"><img alt="Commits le mois dernier" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
+  <a href="./README_KR.md"><img alt="Commits le mois dernier" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
 </p>
 
 #

README_JA.md

@@ -35,6 +35,7 @@
   <a href="./README_ES.md"><img alt="先月のコミット" src="https://img.shields.io/badge/Español-d9d9d9"></a>
   <a href="./README_KL.md"><img alt="先月のコミット" src="https://img.shields.io/badge/Français-d9d9d9"></a>
   <a href="./README_FR.md"><img alt="先月のコミット" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
+  <a href="./README_KR.md"><img alt="先月のコミット" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
 </p>
 
 #

README_KL.md

@@ -35,6 +35,7 @@
   <a href="./README_ES.md"><img alt="Commits last month" src="https://img.shields.io/badge/Español-d9d9d9"></a>
   <a href="./README_KL.md"><img alt="Commits last month" src="https://img.shields.io/badge/Français-d9d9d9"></a>
   <a href="./README_FR.md"><img alt="Commits last month" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
+  <a href="./README_KR.md"><img alt="Commits last month" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
 </p>
 
 #

README_KR.md

@@ -0,0 +1,243 @@
+![cover-v5-optimized](https://github.com/langgenius/dify/assets/13230914/f9e19af5-61ba-4119-b926-d10c4c06ebab)
+
+<p align="center">
+  <a href="https://cloud.dify.ai">Dify 클라우드</a> ·
+  <a href="https://docs.dify.ai/getting-started/install-self-hosted">셀프-호스팅</a> ·
+  <a href="https://docs.dify.ai">문서</a> ·
+  <a href="https://cal.com/guchenhe/60-min-meeting">기업 문의</a>
+</p>
+
+<p align="center">
+    <a href="https://dify.ai" target="_blank">
+        <img alt="Static Badge" src="https://img.shields.io/badge/Product-F04438"></a>
+    <a href="https://dify.ai/pricing" target="_blank">
+        <img alt="Static Badge" src="https://img.shields.io/badge/free-pricing?logo=free&color=%20%23155EEF&label=pricing&labelColor=%20%23528bff"></a>
+    <a href="https://discord.gg/FngNHpbcY7" target="_blank">
+        <img src="https://img.shields.io/discord/1082486657678311454?logo=discord&labelColor=%20%235462eb&logoColor=%20%23f5f5f5&color=%20%235462eb"
+            alt="chat on Discord"></a>
+    <a href="https://twitter.com/intent/follow?screen_name=dify_ai" target="_blank">
+        <img src="https://img.shields.io/twitter/follow/dify_ai?logo=X&color=%20%23f5f5f5"
+            alt="follow on Twitter"></a>
+    <a href="https://hub.docker.com/u/langgenius" target="_blank">
+        <img alt="Docker Pulls" src="https://img.shields.io/docker/pulls/langgenius/dify-web?labelColor=%20%23FDB062&color=%20%23f79009"></a>
+    <a href="https://github.com/langgenius/dify/graphs/commit-activity" target="_blank">
+        <img alt="Commits last month" src="https://img.shields.io/github/commit-activity/m/langgenius/dify?labelColor=%20%2332b583&color=%20%2312b76a"></a>
+    <a href="https://github.com/langgenius/dify/" target="_blank">
+        <img alt="Issues closed" src="https://img.shields.io/github/issues-search?query=repo%3Alanggenius%2Fdify%20is%3Aclosed&label=issues%20closed&labelColor=%20%237d89b0&color=%20%235d6b98"></a>
+    <a href="https://github.com/langgenius/dify/discussions/" target="_blank">
+        <img alt="Discussion posts" src="https://img.shields.io/github/discussions/langgenius/dify?labelColor=%20%239b8afb&color=%20%237a5af8"></a>
+</p>
+
+<p align="center">
+  <a href="./README.md"><img alt="README in English" src="https://img.shields.io/badge/English-d9d9d9"></a>
+  <a href="./README_CN.md"><img alt="简体中文版自述文件" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
+  <a href="./README_JA.md"><img alt="日本語のREADME" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
+  <a href="./README_ES.md"><img alt="README en Español" src="https://img.shields.io/badge/Español-d9d9d9"></a>
+  <a href="./README_FR.md"><img alt="README en Français" src="https://img.shields.io/badge/Français-d9d9d9"></a>
+  <a href="./README_KL.md"><img alt="README tlhIngan Hol" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
+  <a href="./README_KR.md"><img alt="한국어 README" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
+
+</p>
+
+
+   Dify는 오픈 소스 LLM 앱 개발 플랫폼입니다. 직관적인 인터페이스를 통해 AI 워크플로우, RAG 파이프라인, 에이전트 기능, 모델 관리, 관찰 기능 등을 결합하여 프로토타입에서 프로덕션까지 빠르게 전환할 수 있습니다. 주요 기능 목록은 다음과 같습니다:</br> </br>
+
+**1. 워크플로우**:
+    다음 기능들을 비롯한 다양한 기능을 활용하여 시각적 캔버스에서 강력한 AI 워크플로우를 구축하고 테스트하세요.
+
+
+  https://github.com/langgenius/dify/assets/13230914/356df23e-1604-483d-80a6-9517ece318aa
+
+
+
+**2. 포괄적인 모델 지원:**: 
+
+수십 개의 추론 제공업체와 자체 호스팅 솔루션에서 제공하는 수백 개의 독점 및 오픈 소스 LLM과 원활하게 통합되며, GPT, Mistral, Llama3 및 모든 OpenAI API 호환 모델을 포함합니다. 지원되는 모델 제공업체의 전체 목록은 [여기](https://docs.dify.ai/getting-started/readme/model-providers)에서 확인할 수 있습니다.
+![providers-v5](https://github.com/langgenius/dify/assets/13230914/5a17bdbe-097a-4100-8363-40255b70f6e3)
+
+
+**3. 통합 개발환경**:
+  프롬프트를 작성하고, 모델 성능을 비교하며, 텍스트-음성 변환과 같은 추가 기능을 채팅 기반 앱에 추가할 수 있는 직관적인 인터페이스를 제공합니다.
+
+**4. RAG 파이프라인**: 
+  문서 수집부터 검색까지 모든 것을 다루며, PDF, PPT 및 기타 일반적인 문서 형식에서 텍스트 추출을 위한 기본 지원이 포함되어 있는 광범위한 RAG 기능을 제공합니다.
+
+**5. 에이전트 기능**:
+  LLM 함수 호출 또는 ReAct를 기반으로 에이전트를 정의하고 에이전트에 대해 사전 구축된 도구나 사용자 정의 도구를 추가할 수 있습니다. Dify는 Google Search, DELL·E, Stable Diffusion, WolframAlpha 등 AI 에이전트를 위한 50개 이상의 내장 도구를 제공합니다.
+
+**6. LLMOps**:
+  시간 경과에 따른 애플리케이션 로그와 성능을 모니터링하고 분석합니다. 생산 데이터와 주석을 기반으로 프롬프트, 데이터세트, 모델을 지속적으로 개선할 수 있습니다.
+
+**7. Backend-as-a-Service**:
+  Dify의 모든 제품에는 해당 API가 함께 제공되므로 Dify를 자신의 비즈니스 로직에 쉽게 통합할 수 있습니다.
+
+## 기능 비교
+<table style="width: 100%;">
+  <tr>
+    <th align="center">기능</th>
+    <th align="center">Dify.AI</th>
+    <th align="center">LangChain</th>
+    <th align="center">Flowise</th>
+    <th align="center">OpenAI Assistants API</th>
+  </tr>
+  <tr>
+    <td align="center">프로그래밍 접근 방식</td>
+    <td align="center">API + 앱 중심</td>
+    <td align="center">Python 코드</td>
+    <td align="center">앱 중심</td>
+    <td align="center">API 중심</td>
+  </tr>
+  <tr>
+    <td align="center">지원되는 LLMs</td>
+    <td align="center">다양한 종류</td>
+    <td align="center">다양한 종류</td>
+    <td align="center">다양한 종류</td>
+    <td align="center">OpenAI 전용</td>
+  </tr>
+  <tr>
+    <td align="center">RAG 엔진</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+  </tr>
+  <tr>
+    <td align="center">에이전트</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+    <td align="center">✅</td>
+  </tr>
+  <tr>
+    <td align="center">워크플로우</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+  </tr>
+  <tr>
+    <td align="center">가시성</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+    <td align="center">❌</td>
+  </tr>
+  <tr>
+    <td align="center">기업용 기능 (SSO/접근 제어)</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+    <td align="center">❌</td>
+    <td align="center">❌</td>
+  </tr>
+  <tr>
+    <td align="center">로컬 배포</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+  </tr>
+</table>
+
+## Dify 사용하기
+
+- **클라우드 </br>**
+  우리는 누구나 설정이 필요 없이 사용해 볼 수 있도록 [Dify 클라우드](https://dify.ai) 서비스를 호스팅합니다. 이는 자체 배포 버전의 모든 기능을 제공하며, 샌드박스 플랜에서 무료로 200회의 GPT-4 호출을 포함합니다.
+
+- **셀프-호스팅 Dify 커뮤니티 에디션</br>**
+  환경에서 Dify를 빠르게 실행하려면 이  [스타터 가이드를](#quick-start) 참조하세요.
+  추가 참조 및 더 심층적인 지침은 [문서](https://docs.dify.ai)를 사용하세요.
+
+- **기업 / 조직을 위한 Dify</br>**
+  우리는 추가적인 기업 중심 기능을 제공합니다. 당사와 [미팅일정](https://cal.com/guchenhe/30min)을 잡거나  [이메일 보내기](mailto:business@dify.ai?subject=[GitHub]Business%20License%20Inquiry)를 통해 기업 요구 사항을 논의하십시오. </br>
+  > AWS를 사용하는 스타트업 및 중소기업의 경우 [AWS Marketplace에서 Dify Premium](https://aws.amazon.com/marketplace/pp/prodview-t22mebxzwjhu6)을 확인하고 한 번의 클릭으로 자체 AWS VPC에 배포하십시오. 맞춤형 로고와 브랜딩이 포함된 앱을 생성할 수 있는 옵션이 포함된 저렴한 AMI 제품입니다.
+
+
+
+## 앞서가기
+
+GitHub에서 Dify에 별표를 찍어 새로운 릴리스를 즉시 알림 받으세요.
+
+![star-us](https://github.com/langgenius/dify/assets/13230914/b823edc1-6388-4e25-ad45-2f6b187adbb4)
+
+
+
+## 빠른 시작
+>Dify를 설치하기 전에 컴퓨터가 다음과 같은 최소 시스템 요구 사항을 충족하는지 확인하세요 : 
+>- CPU >= 2 Core
+>- RAM >= 4GB
+
+</br>
+
+Dify 서버를 시작하는 가장 쉬운 방법은 [docker-compose.yml](docker/docker-compose.yaml) 파일을 실행하는 것입니다. 설치 명령을 실행하기 전에 [Docker](https://docs.docker.com/get-docker/) 및 [Docker Compose](https://docs.docker.com/compose/install/)가 머신에 설치되어 있는지 확인하세요.
+
+```bash
+cd docker
+docker compose up -d
+```
+
+실행 후 브라우저의 [http://localhost/install](http://localhost/install) 에서 Dify 대시보드에 액세스하고 초기화 프로세스를 시작할 수 있습니다.
+
+> Dify에 기여하거나 추가 개발을 하고 싶다면 소스 코드에서 [배포에 대한 가이드](https://docs.dify.ai/getting-started/install-self-hosted/local-source-code)를 참조하세요.
+
+## 다음 단계
+
+구성 커스터마이징이 필요한 경우, [docker-compose.yml](docker/docker-compose.yaml) 파일의 코멘트를 참조하여 환경 구성을 수동으로 설정하십시오. 변경 후 `docker-compose up -d` 를 다시 실행하십시오. 환경 변수의 전체 목록은 [여기](https://docs.dify.ai/getting-started/install-self-hosted/environments)에서 확인할 수 있습니다.
+
+
+고가용성 설정을 구성하려면 Dify를 Kubernetes에 배포할 수 있는 커뮤니티 제공 [Helm Charts](https://helm.sh/)가 있습니다.
+
+- [Helm Chart by @LeoQuote](https://github.com/douban/charts/tree/master/charts/dify)
+- [Helm Chart by @BorisPolonsky](https://github.com/BorisPolonsky/dify-helm)
+
+
+## 기여
+
+코드에 기여하고 싶은 분들은 [기여 가이드](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md)를 참조하세요.
+동시에 Dify를 소셜 미디어와 행사 및 컨퍼런스에 공유하여 지원하는 것을 고려해 주시기 바랍니다.
+
+
+> 우리는 Dify를 중국어나 영어 이외의 언어로 번역하는 데 도움을 줄 수 있는 기여자를 찾고 있습니다. 도움을 주고 싶으시다면 [i18n README](https://github.com/langgenius/dify/blob/main/web/i18n/README.md)에서 더 많은 정보를 확인하시고 [Discord 커뮤니티 서버](https://discord.gg/8Tpq4AcN9c)의 `global-users` 채널에 댓글을 남겨주세요. 
+
+**기여자**
+
+<a href="https://github.com/langgenius/dify/graphs/contributors">
+  <img src="https://contrib.rocks/image?repo=langgenius/dify" />
+</a>
+
+## 커뮤니티 & 연락처
+
+* [Github 토론](https://github.com/langgenius/dify/discussions). 피드백 공유 및 질문하기에 적합합니다.
+* [GitHub 이슈](https://github.com/langgenius/dify/issues). Dify.AI 사용 중 발견한 버그와 기능 제안에 적합합니다. [기여 가이드](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md)를 참조하세요. 
+* [이메일](mailto:support@dify.ai?subject=[GitHub]Questions%20About%20Dify). Dify.AI 사용에 대한 질문하기에 적합합니다.
+* [디스코드](https://discord.gg/FngNHpbcY7). 애플리케이션 공유 및 커뮤니티와 소통하기에 적합합니다.
+* [트위터](https://twitter.com/dify_ai). 애플리케이션 공유 및 커뮤니티와 소통하기에 적합합니다.
+
+또는 팀원과 직접 미팅을 예약하세요:
+
+<table>
+  <tr>
+    <th>연락처</th>
+    <th>목적</th>
+  </tr>
+  <tr>
+    <td><a href='https://cal.com/guchenhe/15min' target='_blank'><img class="schedule-button" src='https://github.com/langgenius/dify/assets/13230914/9ebcd111-1205-4d71-83d5-948d70b809f5' alt='Git-Hub-README-Button-3x' style="width: 180px; height: auto; object-fit: contain;"/></a></td>
+    <td>비즈니스 문의 및 제품 피드백</td>
+  </tr>
+  <tr>
+    <td><a href='https://cal.com/pinkbanana' target='_blank'><img class="schedule-button" src='https://github.com/langgenius/dify/assets/13230914/d1edd00a-d7e4-4513-be6c-e57038e143fd' alt='Git-Hub-README-Button-2x' style="width: 180px; height: auto; object-fit: contain;"/></a></td>
+    <td>기여, 이슈 및 기능 요청</td>
+  </tr>
+</table>
+
+## Star 히스토리
+
+[![Star History Chart](https://api.star-history.com/svg?repos=langgenius/dify&type=Date)](https://star-history.com/#langgenius/dify&Date)
+
+
+## 보안 공개
+
+개인정보 보호를 위해 보안 문제를 GitHub에 게시하지 마십시오. 대신 security@dify.ai로 질문을 보내주시면 더 자세한 답변을 드리겠습니다.
+
+## 라이선스
+
+이 저장소는 기본적으로 몇 가지 추가 제한 사항이 있는 Apache 2.0인 [Dify 오픈 소스 라이선스](LICENSE)에 따라 사용할 수 있습니다.

api/.env.example

@@ -17,6 +17,9 @@ APP_WEB_URL=http://127.0.0.1:3000
 # Files URL
 FILES_URL=http://127.0.0.1:5001
 
+# The time in seconds after the signature is rejected
+FILES_ACCESS_TIMEOUT=300
+
 # celery configuration
 CELERY_BROKER_URL=redis://:difyai123456@localhost:6379/1
 
@@ -124,10 +127,11 @@ RESEND_API_KEY=
 RESEND_API_URL=https://api.resend.com
 # smtp configuration
 SMTP_SERVER=smtp.gmail.com
-SMTP_PORT=587
+SMTP_PORT=465
 SMTP_USERNAME=123
 SMTP_PASSWORD=abc
-SMTP_USE_TLS=false
+SMTP_USE_TLS=true
+SMTP_OPPORTUNISTIC_TLS=false
 
 # Sentry configuration
 SENTRY_DSN=
@@ -144,6 +148,7 @@ NOTION_INTERNAL_SECRET=you-internal-secret
 
 ETL_TYPE=dify
 UNSTRUCTURED_API_URL=
+UNSTRUCTURED_API_KEY=
 
 SSRF_PROXY_HTTP_URL=
 SSRF_PROXY_HTTPS_URL=
@@ -175,3 +180,11 @@ HTTP_REQUEST_NODE_MAX_TEXT_SIZE=1048576 # 1MB
 
 # Log file path
 LOG_FILE=
+
+# Indexing configuration
+INDEXING_MAX_SEGMENTATION_TOKENS_LENGTH=1000
+
+# Workflow runtime configuration
+WORKFLOW_MAX_EXECUTION_STEPS=500
+WORKFLOW_MAX_EXECUTION_TIME=1200
+WORKFLOW_CALL_MAX_DEPTH=5

api/commands.py

@@ -448,9 +448,64 @@ def convert_to_agent_apps():
     click.echo(click.style('Congratulations! Converted {} agent apps.'.format(len(proceeded_app_ids)), fg='green'))
 
 
+@click.command('add-qdrant-doc-id-index', help='add qdrant doc_id index.')
+@click.option('--field', default='metadata.doc_id', prompt=False, help='index field , default is metadata.doc_id.')
+def add_qdrant_doc_id_index(field: str):
+    click.echo(click.style('Start add qdrant doc_id index.', fg='green'))
+    config = current_app.config
+    vector_type = config.get('VECTOR_STORE')
+    if vector_type != "qdrant":
+        click.echo(click.style('Sorry, only support qdrant vector store.', fg='red'))
+        return
+    create_count = 0
+
+    try:
+        bindings = db.session.query(DatasetCollectionBinding).all()
+        if not bindings:
+            click.echo(click.style('Sorry, no dataset collection bindings found.', fg='red'))
+            return
+        import qdrant_client
+        from qdrant_client.http.exceptions import UnexpectedResponse
+        from qdrant_client.http.models import PayloadSchemaType
+
+        from core.rag.datasource.vdb.qdrant.qdrant_vector import QdrantConfig
+        for binding in bindings:
+            qdrant_config = QdrantConfig(
+                endpoint=config.get('QDRANT_URL'),
+                api_key=config.get('QDRANT_API_KEY'),
+                root_path=current_app.root_path,
+                timeout=config.get('QDRANT_CLIENT_TIMEOUT'),
+                grpc_port=config.get('QDRANT_GRPC_PORT'),
+                prefer_grpc=config.get('QDRANT_GRPC_ENABLED')
+            )
+            try:
+                client = qdrant_client.QdrantClient(**qdrant_config.to_qdrant_params())
+                # create payload index
+                client.create_payload_index(binding.collection_name, field,
+                                            field_schema=PayloadSchemaType.KEYWORD)
+                create_count += 1
+            except UnexpectedResponse as e:
+                # Collection does not exist, so return
+                if e.status_code == 404:
+                    click.echo(click.style(f'Collection not found, collection_name:{binding.collection_name}.', fg='red'))
+                    continue
+                # Some other error occurred, so re-raise the exception
+                else:
+                    click.echo(click.style(f'Failed to create qdrant index, collection_name:{binding.collection_name}.', fg='red'))
+
+    except Exception as e:
+        click.echo(click.style('Failed to create qdrant client.', fg='red'))
+
+    click.echo(
+        click.style(f'Congratulations! Create {create_count} collection indexes.',
+                    fg='green'))
+
+
 def register_commands(app):
     app.cli.add_command(reset_password)
     app.cli.add_command(reset_email)
     app.cli.add_command(reset_encrypt_key_pair)
     app.cli.add_command(vdb_migrate)
     app.cli.add_command(convert_to_agent_apps)
+    app.cli.add_command(add_qdrant_doc_id_index)
+

api/config.py

@@ -23,14 +23,17 @@ DEFAULTS = {
     'SERVICE_API_URL': 'https://api.dify.ai',
     'APP_WEB_URL': 'https://udify.app',
     'FILES_URL': '',
+    'FILES_ACCESS_TIMEOUT': 300,
     'S3_ADDRESS_STYLE': 'auto',
     'STORAGE_TYPE': 'local',
     'STORAGE_LOCAL_PATH': 'storage',
     'CHECK_UPDATE_URL': 'https://updates.dify.ai',
     'DEPLOY_ENV': 'PRODUCTION',
+    'SQLALCHEMY_DATABASE_URI_SCHEME': 'postgresql',
     'SQLALCHEMY_POOL_SIZE': 30,
     'SQLALCHEMY_MAX_OVERFLOW': 10,
     'SQLALCHEMY_POOL_RECYCLE': 3600,
+    'SQLALCHEMY_POOL_PRE_PING': 'False',
     'SQLALCHEMY_ECHO': 'False',
     'SENTRY_TRACES_SAMPLE_RATE': 1.0,
     'SENTRY_PROFILES_SAMPLE_RATE': 1.0,
@@ -67,6 +70,7 @@ DEFAULTS = {
     'INVITE_EXPIRY_HOURS': 72,
     'BILLING_ENABLED': 'False',
     'CAN_REPLACE_LOGO': 'False',
+    'MODEL_LB_ENABLED': 'False',
     'ETL_TYPE': 'dify',
     'KEYWORD_STORE': 'jieba',
     'BATCH_UPLOAD_LIMIT': 20,
@@ -77,6 +81,10 @@ DEFAULTS = {
     'KEYWORD_DATA_SOURCE_TYPE': 'database',
     'INNER_API': 'False',
     'ENTERPRISE_ENABLED': 'False',
+    'INDEXING_MAX_SEGMENTATION_TOKENS_LENGTH': 1000,
+    'WORKFLOW_MAX_EXECUTION_STEPS': 500,
+    'WORKFLOW_MAX_EXECUTION_TIME': 1200,
+    'WORKFLOW_CALL_MAX_DEPTH': 5,
 }
 
 
@@ -107,7 +115,7 @@ class Config:
         # ------------------------
         # General Configurations.
         # ------------------------
-        self.CURRENT_VERSION = "0.6.8"
+        self.CURRENT_VERSION = "0.6.10"
         self.COMMIT_SHA = get_env('COMMIT_SHA')
         self.EDITION = get_env('EDITION')
         self.DEPLOY_ENV = get_env('DEPLOY_ENV')
@@ -116,6 +124,7 @@ class Config:
         self.LOG_FILE = get_env('LOG_FILE')
         self.LOG_FORMAT = get_env('LOG_FORMAT')
         self.LOG_DATEFORMAT = get_env('LOG_DATEFORMAT')
+        self.API_COMPRESSION_ENABLED = get_bool_env('API_COMPRESSION_ENABLED')
 
         # The backend URL prefix of the console API.
         # used to concatenate the login authorization callback or notion integration callback.
@@ -138,6 +147,10 @@ class Config:
         # Url is signed and has expiration time.
         self.FILES_URL = get_env('FILES_URL') if get_env('FILES_URL') else self.CONSOLE_API_URL
 
+        # File Access Time specifies a time interval in seconds for the file to be accessed.
+        # The default value is 300 seconds.
+        self.FILES_ACCESS_TIMEOUT = int(get_env('FILES_ACCESS_TIMEOUT'))
+
         # Your App secret key will be used for securely signing the session cookie
         # Make sure you are changing this key for your deployment with a strong key.
         # You can generate a strong key using `openssl rand -base64 42`.
@@ -165,14 +178,17 @@ class Config:
             key: get_env(key) for key in
             ['DB_USERNAME', 'DB_PASSWORD', 'DB_HOST', 'DB_PORT', 'DB_DATABASE', 'DB_CHARSET']
         }
+        self.SQLALCHEMY_DATABASE_URI_SCHEME = get_env('SQLALCHEMY_DATABASE_URI_SCHEME')
 
         db_extras = f"?client_encoding={db_credentials['DB_CHARSET']}" if db_credentials['DB_CHARSET'] else ""
 
-        self.SQLALCHEMY_DATABASE_URI = f"postgresql://{db_credentials['DB_USERNAME']}:{db_credentials['DB_PASSWORD']}@{db_credentials['DB_HOST']}:{db_credentials['DB_PORT']}/{db_credentials['DB_DATABASE']}{db_extras}"
+        self.SQLALCHEMY_DATABASE_URI = f"{self.SQLALCHEMY_DATABASE_URI_SCHEME}://{db_credentials['DB_USERNAME']}:{db_credentials['DB_PASSWORD']}@{db_credentials['DB_HOST']}:{db_credentials['DB_PORT']}/{db_credentials['DB_DATABASE']}{db_extras}"
         self.SQLALCHEMY_ENGINE_OPTIONS = {
             'pool_size': int(get_env('SQLALCHEMY_POOL_SIZE')),
             'max_overflow': int(get_env('SQLALCHEMY_MAX_OVERFLOW')),
-            'pool_recycle': int(get_env('SQLALCHEMY_POOL_RECYCLE'))
+            'pool_recycle': int(get_env('SQLALCHEMY_POOL_RECYCLE')),
+            'pool_pre_ping': get_bool_env('SQLALCHEMY_POOL_PRE_PING'),
+            'connect_args': {'options': '-c timezone=UTC'},
         }
 
         self.SQLALCHEMY_ECHO = get_bool_env('SQLALCHEMY_ECHO')
@@ -196,27 +212,41 @@ class Config:
             if self.CELERY_BACKEND == 'database' else self.CELERY_BROKER_URL
         self.BROKER_USE_SSL = self.CELERY_BROKER_URL.startswith('rediss://')
 
+        # ------------------------
+        # Code Execution Sandbox Configurations.
+        # ------------------------
+        self.CODE_EXECUTION_ENDPOINT = get_env('CODE_EXECUTION_ENDPOINT')
+        self.CODE_EXECUTION_API_KEY = get_env('CODE_EXECUTION_API_KEY')
+
         # ------------------------
         # File Storage Configurations.
         # ------------------------
         self.STORAGE_TYPE = get_env('STORAGE_TYPE')
         self.STORAGE_LOCAL_PATH = get_env('STORAGE_LOCAL_PATH')
+
+        # S3 Storage settings
         self.S3_ENDPOINT = get_env('S3_ENDPOINT')
         self.S3_BUCKET_NAME = get_env('S3_BUCKET_NAME')
         self.S3_ACCESS_KEY = get_env('S3_ACCESS_KEY')
         self.S3_SECRET_KEY = get_env('S3_SECRET_KEY')
         self.S3_REGION = get_env('S3_REGION')
         self.S3_ADDRESS_STYLE = get_env('S3_ADDRESS_STYLE')
+
+        # Azure Blob Storage settings
         self.AZURE_BLOB_ACCOUNT_NAME = get_env('AZURE_BLOB_ACCOUNT_NAME')
         self.AZURE_BLOB_ACCOUNT_KEY = get_env('AZURE_BLOB_ACCOUNT_KEY')
         self.AZURE_BLOB_CONTAINER_NAME = get_env('AZURE_BLOB_CONTAINER_NAME')
         self.AZURE_BLOB_ACCOUNT_URL = get_env('AZURE_BLOB_ACCOUNT_URL')
-        self.ALIYUN_OSS_BUCKET_NAME=get_env('ALIYUN_OSS_BUCKET_NAME')
-        self.ALIYUN_OSS_ACCESS_KEY=get_env('ALIYUN_OSS_ACCESS_KEY')
-        self.ALIYUN_OSS_SECRET_KEY=get_env('ALIYUN_OSS_SECRET_KEY')
-        self.ALIYUN_OSS_ENDPOINT=get_env('ALIYUN_OSS_ENDPOINT')
-        self.ALIYUN_OSS_REGION=get_env('ALIYUN_OSS_REGION')
-        self.ALIYUN_OSS_AUTH_VERSION=get_env('ALIYUN_OSS_AUTH_VERSION')
+
+        # Aliyun Storage settings
+        self.ALIYUN_OSS_BUCKET_NAME = get_env('ALIYUN_OSS_BUCKET_NAME')
+        self.ALIYUN_OSS_ACCESS_KEY = get_env('ALIYUN_OSS_ACCESS_KEY')
+        self.ALIYUN_OSS_SECRET_KEY = get_env('ALIYUN_OSS_SECRET_KEY')
+        self.ALIYUN_OSS_ENDPOINT = get_env('ALIYUN_OSS_ENDPOINT')
+        self.ALIYUN_OSS_REGION = get_env('ALIYUN_OSS_REGION')
+        self.ALIYUN_OSS_AUTH_VERSION = get_env('ALIYUN_OSS_AUTH_VERSION')
+
+        # Google Cloud Storage settings
         self.GOOGLE_STORAGE_BUCKET_NAME = get_env('GOOGLE_STORAGE_BUCKET_NAME')
         self.GOOGLE_STORAGE_SERVICE_ACCOUNT_JSON_BASE64 = get_env('GOOGLE_STORAGE_SERVICE_ACCOUNT_JSON_BASE64')
 
@@ -226,6 +256,7 @@ class Config:
         # ------------------------
         self.VECTOR_STORE = get_env('VECTOR_STORE')
         self.KEYWORD_STORE = get_env('KEYWORD_STORE')
+
         # qdrant settings
         self.QDRANT_URL = get_env('QDRANT_URL')
         self.QDRANT_API_KEY = get_env('QDRANT_API_KEY')
@@ -281,7 +312,8 @@ class Config:
         self.SMTP_USERNAME = get_env('SMTP_USERNAME')
         self.SMTP_PASSWORD = get_env('SMTP_PASSWORD')
         self.SMTP_USE_TLS = get_bool_env('SMTP_USE_TLS')
-        
+        self.SMTP_OPPORTUNISTIC_TLS = get_bool_env('SMTP_OPPORTUNISTIC_TLS')
+
         # ------------------------
         # Workspace Configurations.
         # ------------------------
@@ -308,6 +340,23 @@ class Config:
         self.UPLOAD_FILE_SIZE_LIMIT = int(get_env('UPLOAD_FILE_SIZE_LIMIT'))
         self.UPLOAD_FILE_BATCH_LIMIT = int(get_env('UPLOAD_FILE_BATCH_LIMIT'))
         self.UPLOAD_IMAGE_FILE_SIZE_LIMIT = int(get_env('UPLOAD_IMAGE_FILE_SIZE_LIMIT'))
+        self.BATCH_UPLOAD_LIMIT = get_env('BATCH_UPLOAD_LIMIT')
+
+        # RAG ETL Configurations.
+        self.ETL_TYPE = get_env('ETL_TYPE')
+        self.UNSTRUCTURED_API_URL = get_env('UNSTRUCTURED_API_URL')
+        self.UNSTRUCTURED_API_KEY = get_env('UNSTRUCTURED_API_KEY')
+        self.KEYWORD_DATA_SOURCE_TYPE = get_env('KEYWORD_DATA_SOURCE_TYPE')
+
+        # Indexing Configurations.
+        self.INDEXING_MAX_SEGMENTATION_TOKENS_LENGTH = get_env('INDEXING_MAX_SEGMENTATION_TOKENS_LENGTH')
+
+        # Tool Configurations.
+        self.TOOL_ICON_CACHE_MAX_AGE = get_env('TOOL_ICON_CACHE_MAX_AGE')
+
+        self.WORKFLOW_MAX_EXECUTION_STEPS = int(get_env('WORKFLOW_MAX_EXECUTION_STEPS'))
+        self.WORKFLOW_MAX_EXECUTION_TIME = int(get_env('WORKFLOW_MAX_EXECUTION_TIME'))
+        self.WORKFLOW_CALL_MAX_DEPTH = int(get_env('WORKFLOW_CALL_MAX_DEPTH'))
 
         # Moderation in app Configurations.
         self.OUTPUT_MODERATION_BUFFER_SIZE = int(get_env('OUTPUT_MODERATION_BUFFER_SIZE'))
@@ -359,18 +408,15 @@ class Config:
         self.HOSTED_FETCH_APP_TEMPLATES_MODE = get_env('HOSTED_FETCH_APP_TEMPLATES_MODE')
         self.HOSTED_FETCH_APP_TEMPLATES_REMOTE_DOMAIN = get_env('HOSTED_FETCH_APP_TEMPLATES_REMOTE_DOMAIN')
 
-        self.ETL_TYPE = get_env('ETL_TYPE')
-        self.UNSTRUCTURED_API_URL = get_env('UNSTRUCTURED_API_URL')
+        # Model Load Balancing Configurations.
+        self.MODEL_LB_ENABLED = get_bool_env('MODEL_LB_ENABLED')
+
+        # Platform Billing Configurations.
         self.BILLING_ENABLED = get_bool_env('BILLING_ENABLED')
-        self.CAN_REPLACE_LOGO = get_bool_env('CAN_REPLACE_LOGO')
 
-        self.BATCH_UPLOAD_LIMIT = get_env('BATCH_UPLOAD_LIMIT')
-
-        self.CODE_EXECUTION_ENDPOINT = get_env('CODE_EXECUTION_ENDPOINT')
-        self.CODE_EXECUTION_API_KEY = get_env('CODE_EXECUTION_API_KEY')
-
-        self.API_COMPRESSION_ENABLED = get_bool_env('API_COMPRESSION_ENABLED')
-        self.TOOL_ICON_CACHE_MAX_AGE = get_env('TOOL_ICON_CACHE_MAX_AGE')
-
-        self.KEYWORD_DATA_SOURCE_TYPE = get_env('KEYWORD_DATA_SOURCE_TYPE')
+        # ------------------------
+        # Enterprise feature Configurations.
+        # **Before using, please contact business@dify.ai by email to inquire about licensing matters.**
+        # ------------------------
         self.ENTERPRISE_ENABLED = get_bool_env('ENTERPRISE_ENABLED')
+        self.CAN_REPLACE_LOGO = get_bool_env('CAN_REPLACE_LOGO')

api/constants/languages.py

@@ -16,6 +16,7 @@ language_timezone_mapping = {
     'it-IT': 'Europe/Rome',
     'uk-UA': 'Europe/Kyiv',
     'vi-VN': 'Asia/Ho_Chi_Minh',
+    'ro-RO': 'Europe/Bucharest',
     'pl-PL': 'Europe/Warsaw',
 }
 

api/constants/recommended_apps.json

@@ -24,7 +24,8 @@
                     "description": "Welcome to your personalized Investment Analysis Copilot service, where we delve into the depths of stock analysis to provide you with comprehensive insights. \n",
                     "is_listed": true,
                     "position": 0,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -40,7 +41,8 @@
                     "description": "Code interpreter, clarifying the syntax and semantics of the code.",
                     "is_listed": true,
                     "position": 13,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -56,7 +58,8 @@
                     "description": "Hello, I am your creative partner in bringing ideas to vivid life! I can assist you in creating stunning designs by leveraging abilities of DALL\u00b7E 3.  ",
                     "is_listed": true,
                     "position": 4,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -72,7 +75,8 @@
                     "description": "Fully SEO Optimized Article including FAQs",
                     "is_listed": true,
                     "position": 1,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -88,7 +92,8 @@
                     "description": "Generate Flat Style Image",
                     "is_listed": true,
                     "position": 10,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -104,7 +109,8 @@
                     "description": "A multilingual translator that provides translation capabilities in multiple languages. Input the text you need to translate and select the target language.",
                     "is_listed": true,
                     "position": 10,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -120,7 +126,8 @@
                     "description": "I am a YouTube Channel Data Analysis Copilot, I am here to provide expert data analysis tailored to your needs. ",
                     "is_listed": true,
                     "position": 2,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -136,7 +143,8 @@
                     "description": "Meeting minutes generator",
                     "is_listed": true,
                     "position": 0,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -152,7 +160,8 @@
                     "description": "Tell me the main elements, I will generate a cyberpunk style image for you. ",
                     "is_listed": true,
                     "position": 10,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -168,7 +177,8 @@
                     "description": "Write SQL from natural language by pasting in your schema with the request.Please describe your query requirements in natural language and select the target database type.",
                     "is_listed": true,
                     "position": 13,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -184,7 +194,8 @@
                     "description": "Welcome to your personalized travel service with Consultant! \ud83c\udf0d\u2708\ufe0f Ready to embark on a journey filled with adventure and relaxation? Let's dive into creating your unforgettable travel experience. ",
                     "is_listed": true,
                     "position": 3,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -200,7 +211,8 @@
                     "description": "I can answer your questions related to strategic marketing.",
                     "is_listed": true,
                     "position": 10,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -216,7 +228,8 @@
                     "description": "A simulated front-end interviewer that tests the skill level of front-end development through questioning.",
                     "is_listed": true,
                     "position": 19,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -232,7 +245,8 @@
                     "description": "I'm here to hear about your feature request about Dify and help you flesh it out further. What's on your mind?",
                     "is_listed": true,
                     "position": 6,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 }
             ]
         },
@@ -261,7 +275,8 @@
                     "description": "\u4e00\u4e2a\u6a21\u62df\u7684\u524d\u7aef\u9762\u8bd5\u5b98\uff0c\u901a\u8fc7\u63d0\u95ee\u7684\u65b9\u5f0f\u5bf9\u524d\u7aef\u5f00\u53d1\u7684\u6280\u80fd\u6c34\u5e73\u8fdb\u884c\u68c0\u9a8c\u3002",
                     "is_listed": true,
                     "position": 20,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -277,7 +292,8 @@
                     "description": "\u8f93\u5165\u76f8\u5173\u5143\u7d20\uff0c\u4e3a\u4f60\u751f\u6210\u6241\u5e73\u63d2\u753b\u98ce\u683c\u7684\u5c01\u9762\u56fe\u7247",
                     "is_listed": true,
                     "position": 10,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -293,7 +309,8 @@
                     "description": "\u4e00\u4e2a\u591a\u8bed\u8a00\u7ffb\u8bd1\u5668\uff0c\u63d0\u4f9b\u591a\u79cd\u8bed\u8a00\u7ffb\u8bd1\u80fd\u529b\uff0c\u8f93\u5165\u4f60\u9700\u8981\u7ffb\u8bd1\u7684\u6587\u672c\uff0c\u9009\u62e9\u76ee\u6807\u8bed\u8a00\u5373\u53ef\u3002\u63d0\u793a\u8bcd\u6765\u81ea\u5b9d\u7389\u3002",
                     "is_listed": true,
                     "position": 10,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -309,7 +326,8 @@
                     "description": "\u6211\u5c06\u5e2e\u52a9\u4f60\u628a\u81ea\u7136\u8bed\u8a00\u8f6c\u5316\u6210\u6307\u5b9a\u7684\u6570\u636e\u5e93\u67e5\u8be2 SQL \u8bed\u53e5\uff0c\u8bf7\u5728\u4e0b\u65b9\u8f93\u5165\u4f60\u9700\u8981\u67e5\u8be2\u7684\u6761\u4ef6\uff0c\u5e76\u9009\u62e9\u76ee\u6807\u6570\u636e\u5e93\u7c7b\u578b\u3002",
                     "is_listed": true,
                     "position": 12,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -325,7 +343,8 @@
                     "description": "\u9610\u660e\u4ee3\u7801\u7684\u8bed\u6cd5\u548c\u8bed\u4e49\u3002",
                     "is_listed": true,
                     "position": 2,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -341,7 +360,8 @@
                     "description": "\u8f93\u5165\u76f8\u5173\u5143\u7d20\uff0c\u4e3a\u4f60\u751f\u6210\u8d5b\u535a\u670b\u514b\u98ce\u683c\u7684\u63d2\u753b",
                     "is_listed": true,
                     "position": 10,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -357,7 +377,8 @@
                     "description": "\u6211\u662f\u4e00\u540dSEO\u4e13\u5bb6\uff0c\u53ef\u4ee5\u6839\u636e\u60a8\u63d0\u4f9b\u7684\u6807\u9898\u3001\u5173\u952e\u8bcd\u3001\u76f8\u5173\u4fe1\u606f\u6765\u6279\u91cf\u751f\u6210SEO\u6587\u7ae0\u3002",
                     "is_listed": true,
                     "position": 10,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -373,7 +394,8 @@
                     "description": "\u5e2e\u4f60\u91cd\u65b0\u7ec4\u7ec7\u548c\u8f93\u51fa\u6df7\u4e71\u590d\u6742\u7684\u4f1a\u8bae\u7eaa\u8981\u3002",
                     "is_listed": true,
                     "position": 6,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -389,7 +411,8 @@
                     "description": "\u6b22\u8fce\u4f7f\u7528\u60a8\u7684\u4e2a\u6027\u5316\u7f8e\u80a1\u6295\u8d44\u5206\u6790\u52a9\u624b\uff0c\u5728\u8fd9\u91cc\u6211\u4eec\u6df1\u5165\u7684\u8fdb\u884c\u80a1\u7968\u5206\u6790\uff0c\u4e3a\u60a8\u63d0\u4f9b\u5168\u9762\u7684\u6d1e\u5bdf\u3002",
                     "is_listed": true,
                     "position": 0,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -405,7 +428,8 @@
                     "description": "\u60a8\u597d\uff0c\u6211\u662f\u60a8\u7684\u521b\u610f\u4f19\u4f34\uff0c\u5c06\u5e2e\u52a9\u60a8\u5c06\u60f3\u6cd5\u751f\u52a8\u5730\u5b9e\u73b0\uff01\u6211\u53ef\u4ee5\u534f\u52a9\u60a8\u5229\u7528DALL\u00b7E 3\u7684\u80fd\u529b\u521b\u9020\u51fa\u4ee4\u4eba\u60ca\u53f9\u7684\u8bbe\u8ba1\u3002",
                     "is_listed": true,
                     "position": 4,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -421,7 +445,8 @@
                     "description": "\u7ffb\u8bd1\u4e13\u5bb6\uff1a\u63d0\u4f9b\u4e2d\u82f1\u6587\u4e92\u8bd1",
                     "is_listed": true,
                     "position": 4,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -437,7 +462,8 @@
                     "description": "\u60a8\u7684\u79c1\u4eba\u5b66\u4e60\u5bfc\u5e08\uff0c\u5e2e\u60a8\u5236\u5b9a\u5b66\u4e60\u8ba1\u5212\u5e76\u8f85\u5bfc",
                     "is_listed": true,
                     "position": 26,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -453,7 +479,8 @@
                     "description": "\u5e2e\u4f60\u64b0\u5199\u8bba\u6587\u6587\u732e\u7efc\u8ff0",
                     "is_listed": true,
                     "position": 7,
-                    "privacy_policy": "https://dify.ai"
+                    "privacy_policy": "https://dify.ai",
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -469,7 +496,8 @@
                     "description": "\u4f60\u597d\uff0c\u544a\u8bc9\u6211\u60a8\u60f3\u5206\u6790\u7684 YouTube \u9891\u9053\uff0c\u6211\u5c06\u4e3a\u60a8\u6574\u7406\u4e00\u4efd\u5b8c\u6574\u7684\u6570\u636e\u5206\u6790\u62a5\u544a\u3002",
                     "is_listed": true,
                     "position": 0,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 },
                 {
                     "app": {
@@ -485,7 +513,8 @@
                     "description": "\u6b22\u8fce\u4f7f\u7528\u60a8\u7684\u4e2a\u6027\u5316\u65c5\u884c\u670d\u52a1\u987e\u95ee\uff01\ud83c\udf0d\u2708\ufe0f \u51c6\u5907\u597d\u8e0f\u4e0a\u4e00\u6bb5\u5145\u6ee1\u5192\u9669\u4e0e\u653e\u677e\u7684\u65c5\u7a0b\u4e86\u5417\uff1f\u8ba9\u6211\u4eec\u4e00\u8d77\u6df1\u5165\u6253\u9020\u60a8\u96be\u5fd8\u7684\u65c5\u884c\u4f53\u9a8c\u5427\u3002",
                     "is_listed": true,
                     "position": 0,
-                    "privacy_policy": null
+                    "privacy_policy": null,
+                    "custom_disclaimer": null
                 }
             ]
         },

api/controllers/console/__init__.py

@@ -37,9 +37,6 @@ from .billing import billing
 # Import datasets controllers
 from .datasets import data_source, datasets, datasets_document, datasets_segments, file, hit_testing
 
-# Import enterprise controllers
-from .enterprise import enterprise_sso
-
 # Import explore controllers
 from .explore import (
     audio,
@@ -57,4 +54,4 @@ from .explore import (
 from .tag import tags
 
 # Import workspace controllers
-from .workspace import account, members, model_providers, models, tool_providers, workspace
+from .workspace import account, load_balancing_config, members, model_providers, models, tool_providers, workspace

api/controllers/console/admin.py

@@ -48,6 +48,7 @@ class InsertExploreAppListApi(Resource):
         parser.add_argument('desc', type=str, location='json')
         parser.add_argument('copyright', type=str, location='json')
         parser.add_argument('privacy_policy', type=str, location='json')
+        parser.add_argument('custom_disclaimer', type=str, location='json')
         parser.add_argument('language', type=supported_language, required=True, nullable=False, location='json')
         parser.add_argument('category', type=str, required=True, nullable=False, location='json')
         parser.add_argument('position', type=int, required=True, nullable=False, location='json')
@@ -62,6 +63,7 @@ class InsertExploreAppListApi(Resource):
             desc = args['desc'] if args['desc'] else ''
             copy_right = args['copyright'] if args['copyright'] else ''
             privacy_policy = args['privacy_policy'] if args['privacy_policy'] else ''
+            custom_disclaimer = args['custom_disclaimer'] if args['custom_disclaimer'] else ''
         else:
             desc = site.description if site.description else \
                 args['desc'] if args['desc'] else ''
@@ -69,6 +71,8 @@ class InsertExploreAppListApi(Resource):
                 args['copyright'] if args['copyright'] else ''
             privacy_policy = site.privacy_policy if site.privacy_policy else \
                 args['privacy_policy'] if args['privacy_policy']  else ''
+            custom_disclaimer = site.custom_disclaimer if site.custom_disclaimer else \
+                args['custom_disclaimer'] if args['custom_disclaimer'] else ''
 
         recommended_app = RecommendedApp.query.filter(RecommendedApp.app_id == args['app_id']).first()
 
@@ -78,6 +82,7 @@ class InsertExploreAppListApi(Resource):
                 description=desc,
                 copyright=copy_right,
                 privacy_policy=privacy_policy,
+                custom_disclaimer=custom_disclaimer,
                 language=args['language'],
                 category=args['category'],
                 position=args['position']
@@ -93,6 +98,7 @@ class InsertExploreAppListApi(Resource):
             recommended_app.description = desc
             recommended_app.copyright = copy_right
             recommended_app.privacy_policy = privacy_policy
+            recommended_app.custom_disclaimer = custom_disclaimer
             recommended_app.language = args['language']
             recommended_app.category = args['category']
             recommended_app.position = args['position']

api/controllers/console/app/audio.py

@@ -85,7 +85,7 @@ class ChatMessageTextApi(Resource):
             response = AudioService.transcript_tts(
                 app_model=app_model,
                 text=request.form['text'],
-                voice=request.form.get('voice'),
+                voice=request.form['voice'],
                 streaming=False
             )
 

api/controllers/console/app/site.py

@@ -23,6 +23,7 @@ def parse_app_site_args():
     parser.add_argument('customize_domain', type=str, required=False, location='json')
     parser.add_argument('copyright', type=str, required=False, location='json')
     parser.add_argument('privacy_policy', type=str, required=False, location='json')
+    parser.add_argument('custom_disclaimer', type=str, required=False, location='json')
     parser.add_argument('customize_token_strategy', type=str, choices=['must', 'allow', 'not_allow'],
                         required=False,
                         location='json')
@@ -56,6 +57,7 @@ class AppSite(Resource):
             'customize_domain',
             'copyright',
             'privacy_policy',
+            'custom_disclaimer',
             'customize_token_strategy',
             'prompt_public'
         ]:

api/controllers/console/app/workflow.py

@@ -137,6 +137,71 @@ class AdvancedChatDraftWorkflowRunApi(Resource):
             logging.exception("internal server error.")
             raise InternalServerError()
 
+class AdvancedChatDraftRunIterationNodeApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @get_app_model(mode=[AppMode.ADVANCED_CHAT])
+    def post(self, app_model: App, node_id: str):
+        """
+        Run draft workflow iteration node
+        """
+        parser = reqparse.RequestParser()
+        parser.add_argument('inputs', type=dict, location='json')
+        args = parser.parse_args()
+
+        try:
+            response = AppGenerateService.generate_single_iteration(
+                app_model=app_model,
+                user=current_user,
+                node_id=node_id,
+                args=args,
+                streaming=True
+            )
+
+            return helper.compact_generate_response(response)
+        except services.errors.conversation.ConversationNotExistsError:
+            raise NotFound("Conversation Not Exists.")
+        except services.errors.conversation.ConversationCompletedError:
+            raise ConversationCompletedError()
+        except ValueError as e:
+            raise e
+        except Exception as e:
+            logging.exception("internal server error.")
+            raise InternalServerError()
+
+class WorkflowDraftRunIterationNodeApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @get_app_model(mode=[AppMode.WORKFLOW])
+    def post(self, app_model: App, node_id: str):
+        """
+        Run draft workflow iteration node
+        """
+        parser = reqparse.RequestParser()
+        parser.add_argument('inputs', type=dict, location='json')
+        args = parser.parse_args()
+
+        try:
+            response = AppGenerateService.generate_single_iteration(
+                app_model=app_model,
+                user=current_user,
+                node_id=node_id,
+                args=args,
+                streaming=True
+            )
+
+            return helper.compact_generate_response(response)
+        except services.errors.conversation.ConversationNotExistsError:
+            raise NotFound("Conversation Not Exists.")
+        except services.errors.conversation.ConversationCompletedError:
+            raise ConversationCompletedError()
+        except ValueError as e:
+            raise e
+        except Exception as e:
+            logging.exception("internal server error.")
+            raise InternalServerError()
 
 class DraftWorkflowRunApi(Resource):
     @setup_required
@@ -326,6 +391,8 @@ api.add_resource(AdvancedChatDraftWorkflowRunApi, '/apps/<uuid:app_id>/advanced-
 api.add_resource(DraftWorkflowRunApi, '/apps/<uuid:app_id>/workflows/draft/run')
 api.add_resource(WorkflowTaskStopApi, '/apps/<uuid:app_id>/workflow-runs/tasks/<string:task_id>/stop')
 api.add_resource(DraftWorkflowNodeRunApi, '/apps/<uuid:app_id>/workflows/draft/nodes/<string:node_id>/run')
+api.add_resource(AdvancedChatDraftRunIterationNodeApi, '/apps/<uuid:app_id>/advanced-chat/workflows/draft/iteration/nodes/<string:node_id>/run')
+api.add_resource(WorkflowDraftRunIterationNodeApi, '/apps/<uuid:app_id>/workflows/draft/iteration/nodes/<string:node_id>/run')
 api.add_resource(PublishedWorkflowApi, '/apps/<uuid:app_id>/workflows/publish')
 api.add_resource(DefaultBlockConfigsApi, '/apps/<uuid:app_id>/workflows/default-workflow-block-configs')
 api.add_resource(DefaultBlockConfigApi, '/apps/<uuid:app_id>/workflows/default-workflow-block-configs'

api/controllers/console/datasets/datasets_document.py

@@ -1,10 +1,12 @@
 import logging
+from argparse import ArgumentTypeError
 from datetime import datetime, timezone
 
 from flask import request
 from flask_login import current_user
 from flask_restful import Resource, fields, marshal, marshal_with, reqparse
 from sqlalchemy import asc, desc
+from transformers.hf_argparser import string_to_bool
 from werkzeug.exceptions import Forbidden, NotFound
 
 import services
@@ -141,7 +143,11 @@ class DatasetDocumentListApi(Resource):
         limit = request.args.get('limit', default=20, type=int)
         search = request.args.get('keyword', default=None, type=str)
         sort = request.args.get('sort', default='-created_at', type=str)
-        fetch = request.args.get('fetch', default=False, type=bool)
+        # "yes", "true", "t", "y", "1" convert to True, while others convert to False.
+        try:
+            fetch = string_to_bool(request.args.get('fetch', default='false'))
+        except (ArgumentTypeError, ValueError, Exception) as e:
+            fetch = False
         dataset = DatasetService.get_dataset(dataset_id)
         if not dataset:
             raise NotFound('Dataset not found.')
@@ -924,6 +930,28 @@ class DocumentRetryApi(DocumentResource):
         return {'result': 'success'}, 204
 
 
+class DocumentRenameApi(DocumentResource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @marshal_with(document_fields)
+    def post(self, dataset_id, document_id):
+        # The role of the current user in the ta table must be admin or owner
+        if not current_user.is_admin_or_owner:
+            raise Forbidden()
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('name', type=str, required=True, nullable=False, location='json')
+        args = parser.parse_args()
+
+        try:
+            document = DocumentService.rename_document(dataset_id, document_id, args['name'])
+        except services.errors.document.DocumentIndexingError:
+            raise DocumentIndexingError('Cannot delete document during indexing.')
+
+        return document
+
+
 api.add_resource(GetProcessRuleApi, '/datasets/process-rule')
 api.add_resource(DatasetDocumentListApi,
                  '/datasets/<uuid:dataset_id>/documents')
@@ -950,3 +978,5 @@ api.add_resource(DocumentStatusApi,
 api.add_resource(DocumentPauseApi, '/datasets/<uuid:dataset_id>/documents/<uuid:document_id>/processing/pause')
 api.add_resource(DocumentRecoverApi, '/datasets/<uuid:dataset_id>/documents/<uuid:document_id>/processing/resume')
 api.add_resource(DocumentRetryApi, '/datasets/<uuid:dataset_id>/retry')
+api.add_resource(DocumentRenameApi,
+                 '/datasets/<uuid:dataset_id>/documents/<uuid:document_id>/rename')

api/controllers/console/enterprise/enterprise_sso.py

@@ -1,59 +0,0 @@
-from flask import current_app, redirect
-from flask_restful import Resource, reqparse
-
-from controllers.console import api
-from controllers.console.setup import setup_required
-from services.enterprise.enterprise_sso_service import EnterpriseSSOService
-
-
-class EnterpriseSSOSamlLogin(Resource):
-
-    @setup_required
-    def get(self):
-        return EnterpriseSSOService.get_sso_saml_login()
-
-
-class EnterpriseSSOSamlAcs(Resource):
-
-    @setup_required
-    def post(self):
-        parser = reqparse.RequestParser()
-        parser.add_argument('SAMLResponse', type=str, required=True, location='form')
-        args = parser.parse_args()
-        saml_response = args['SAMLResponse']
-
-        try:
-            token = EnterpriseSSOService.post_sso_saml_acs(saml_response)
-            return redirect(f'{current_app.config.get("CONSOLE_WEB_URL")}/signin?console_token={token}')
-        except Exception as e:
-            return redirect(f'{current_app.config.get("CONSOLE_WEB_URL")}/signin?message={str(e)}')
-
-
-class EnterpriseSSOOidcLogin(Resource):
-
-    @setup_required
-    def get(self):
-        return EnterpriseSSOService.get_sso_oidc_login()
-
-
-class EnterpriseSSOOidcCallback(Resource):
-
-    @setup_required
-    def get(self):
-        parser = reqparse.RequestParser()
-        parser.add_argument('state', type=str, required=True, location='args')
-        parser.add_argument('code', type=str, required=True, location='args')
-        parser.add_argument('oidc-state', type=str, required=True, location='cookies')
-        args = parser.parse_args()
-
-        try:
-            token = EnterpriseSSOService.get_sso_oidc_callback(args)
-            return redirect(f'{current_app.config.get("CONSOLE_WEB_URL")}/signin?console_token={token}')
-        except Exception as e:
-            return redirect(f'{current_app.config.get("CONSOLE_WEB_URL")}/signin?message={str(e)}')
-
-
-api.add_resource(EnterpriseSSOSamlLogin, '/enterprise/sso/saml/login')
-api.add_resource(EnterpriseSSOSamlAcs, '/enterprise/sso/saml/acs')
-api.add_resource(EnterpriseSSOOidcLogin, '/enterprise/sso/oidc/login')
-api.add_resource(EnterpriseSSOOidcCallback, '/enterprise/sso/oidc/callback')

api/controllers/console/explore/audio.py

@@ -76,7 +76,7 @@ class ChatTextApi(InstalledAppResource):
             response = AudioService.transcript_tts(
                 app_model=app_model,
                 text=request.form['text'],
-                voice=request.form.get('voice'),
+                voice=request.form['voice'] if request.form.get('voice') else app_model.app_model_config.text_to_speech_dict.get('voice'),
                 streaming=False
             )
             return {'data': response.data.decode('latin1')}

api/controllers/console/explore/recommended_app.py

@@ -21,6 +21,7 @@ recommended_app_fields = {
     'description': fields.String(attribute='description'),
     'copyright': fields.String,
     'privacy_policy': fields.String,
+    'custom_disclaimer': fields.String,
     'category': fields.String,
     'position': fields.Integer,
     'is_listed': fields.Boolean

api/controllers/console/feature.py

@@ -1,24 +1,28 @@
 from flask_login import current_user
 from flask_restful import Resource
 
-from services.enterprise.enterprise_feature_service import EnterpriseFeatureService
+from libs.login import login_required
 from services.feature_service import FeatureService
 
 from . import api
-from .wraps import cloud_utm_record
+from .setup import setup_required
+from .wraps import account_initialization_required, cloud_utm_record
 
 
 class FeatureApi(Resource):
 
+    @setup_required
+    @login_required
+    @account_initialization_required
     @cloud_utm_record
     def get(self):
         return FeatureService.get_features(current_user.current_tenant_id).dict()
 
 
-class EnterpriseFeatureApi(Resource):
+class SystemFeatureApi(Resource):
     def get(self):
-        return EnterpriseFeatureService.get_enterprise_features().dict()
+        return FeatureService.get_system_features().dict()
 
 
 api.add_resource(FeatureApi, '/features')
-api.add_resource(EnterpriseFeatureApi, '/enterprise-features')
+api.add_resource(SystemFeatureApi, '/system-features')

api/controllers/console/version.py

@@ -17,13 +17,19 @@ class VersionApi(Resource):
         args = parser.parse_args()
         check_update_url = current_app.config['CHECK_UPDATE_URL']
 
-        if not check_update_url:
-            return {
-                'version': '0.0.0',
-                'release_date': '',
-                'release_notes': '',
-                'can_auto_update': False
+        result = {
+            'version': current_app.config['CURRENT_VERSION'],
+            'release_date': '',
+            'release_notes': '',
+            'can_auto_update': False,
+            'features': {
+                'can_replace_logo': current_app.config['CAN_REPLACE_LOGO'],
+                'model_load_balancing_enabled': current_app.config['MODEL_LB_ENABLED']
             }
+        }
+
+        if not check_update_url:
+            return result
 
         try:
             response = requests.get(check_update_url, {
@@ -31,20 +37,15 @@ class VersionApi(Resource):
             })
         except Exception as error:
             logging.warning("Check update version error: {}.".format(str(error)))
-            return {
-                'version': args.get('current_version'),
-                'release_date': '',
-                'release_notes': '',
-                'can_auto_update': False
-            }
+            result['version'] = args.get('current_version')
+            return result
 
         content = json.loads(response.content)
-        return {
-            'version': content['version'],
-            'release_date': content['releaseDate'],
-            'release_notes': content['releaseNotes'],
-            'can_auto_update': content['canAutoUpdate']
-        }
+        result['version'] = content['version']
+        result['release_date'] = content['releaseDate']
+        result['release_notes'] = content['releaseNotes']
+        result['can_auto_update'] = content['canAutoUpdate']
+        return result
 
 
 api.add_resource(VersionApi, '/version')

api/controllers/console/workspace/load_balancing_config.py

@@ -0,0 +1,106 @@
+from flask_restful import Resource, reqparse
+from werkzeug.exceptions import Forbidden
+
+from controllers.console import api
+from controllers.console.setup import setup_required
+from controllers.console.wraps import account_initialization_required
+from core.model_runtime.entities.model_entities import ModelType
+from core.model_runtime.errors.validate import CredentialsValidateFailedError
+from libs.login import current_user, login_required
+from models.account import TenantAccountRole
+from services.model_load_balancing_service import ModelLoadBalancingService
+
+
+class LoadBalancingCredentialsValidateApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def post(self, provider: str):
+        if not TenantAccountRole.is_privileged_role(current_user.current_tenant.current_role):
+            raise Forbidden()
+
+        tenant_id = current_user.current_tenant_id
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('model', type=str, required=True, nullable=False, location='json')
+        parser.add_argument('model_type', type=str, required=True, nullable=False,
+                            choices=[mt.value for mt in ModelType], location='json')
+        parser.add_argument('credentials', type=dict, required=True, nullable=False, location='json')
+        args = parser.parse_args()
+
+        # validate model load balancing credentials
+        model_load_balancing_service = ModelLoadBalancingService()
+
+        result = True
+        error = None
+
+        try:
+            model_load_balancing_service.validate_load_balancing_credentials(
+                tenant_id=tenant_id,
+                provider=provider,
+                model=args['model'],
+                model_type=args['model_type'],
+                credentials=args['credentials']
+            )
+        except CredentialsValidateFailedError as ex:
+            result = False
+            error = str(ex)
+
+        response = {'result': 'success' if result else 'error'}
+
+        if not result:
+            response['error'] = error
+
+        return response
+
+
+class LoadBalancingConfigCredentialsValidateApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def post(self, provider: str, config_id: str):
+        if not TenantAccountRole.is_privileged_role(current_user.current_tenant.current_role):
+            raise Forbidden()
+
+        tenant_id = current_user.current_tenant_id
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('model', type=str, required=True, nullable=False, location='json')
+        parser.add_argument('model_type', type=str, required=True, nullable=False,
+                            choices=[mt.value for mt in ModelType], location='json')
+        parser.add_argument('credentials', type=dict, required=True, nullable=False, location='json')
+        args = parser.parse_args()
+
+        # validate model load balancing config credentials
+        model_load_balancing_service = ModelLoadBalancingService()
+
+        result = True
+        error = None
+
+        try:
+            model_load_balancing_service.validate_load_balancing_credentials(
+                tenant_id=tenant_id,
+                provider=provider,
+                model=args['model'],
+                model_type=args['model_type'],
+                credentials=args['credentials'],
+                config_id=config_id,
+            )
+        except CredentialsValidateFailedError as ex:
+            result = False
+            error = str(ex)
+
+        response = {'result': 'success' if result else 'error'}
+
+        if not result:
+            response['error'] = error
+
+        return response
+
+
+# Load Balancing Config
+api.add_resource(LoadBalancingCredentialsValidateApi,
+                 '/workspaces/current/model-providers/<string:provider>/models/load-balancing-configs/credentials-validate')
+
+api.add_resource(LoadBalancingConfigCredentialsValidateApi,
+                 '/workspaces/current/model-providers/<string:provider>/models/load-balancing-configs/<string:config_id>/credentials-validate')

api/controllers/console/workspace/models.py

@@ -12,6 +12,7 @@ from core.model_runtime.errors.validate import CredentialsValidateFailedError
 from core.model_runtime.utils.encoders import jsonable_encoder
 from libs.login import login_required
 from models.account import TenantAccountRole
+from services.model_load_balancing_service import ModelLoadBalancingService
 from services.model_provider_service import ModelProviderService
 
 
@@ -104,21 +105,56 @@ class ModelProviderModelApi(Resource):
         parser.add_argument('model', type=str, required=True, nullable=False, location='json')
         parser.add_argument('model_type', type=str, required=True, nullable=False,
                             choices=[mt.value for mt in ModelType], location='json')
-        parser.add_argument('credentials', type=dict, required=True, nullable=False, location='json')
+        parser.add_argument('credentials', type=dict, required=False, nullable=True, location='json')
+        parser.add_argument('load_balancing', type=dict, required=False, nullable=True, location='json')
+        parser.add_argument('config_from', type=str, required=False, nullable=True, location='json')
         args = parser.parse_args()
 
-        model_provider_service = ModelProviderService()
+        model_load_balancing_service = ModelLoadBalancingService()
 
-        try:
-            model_provider_service.save_model_credentials(
+        if ('load_balancing' in args and args['load_balancing'] and
+                'enabled' in args['load_balancing'] and args['load_balancing']['enabled']):
+            if 'configs' not in args['load_balancing']:
+                raise ValueError('invalid load balancing configs')
+
+            # save load balancing configs
+            model_load_balancing_service.update_load_balancing_configs(
                 tenant_id=tenant_id,
                 provider=provider,
                 model=args['model'],
                 model_type=args['model_type'],
-                credentials=args['credentials']
+                configs=args['load_balancing']['configs']
             )
-        except CredentialsValidateFailedError as ex:
-            raise ValueError(str(ex))
+
+            # enable load balancing
+            model_load_balancing_service.enable_model_load_balancing(
+                tenant_id=tenant_id,
+                provider=provider,
+                model=args['model'],
+                model_type=args['model_type']
+            )
+        else:
+            # disable load balancing
+            model_load_balancing_service.disable_model_load_balancing(
+                tenant_id=tenant_id,
+                provider=provider,
+                model=args['model'],
+                model_type=args['model_type']
+            )
+
+            if args.get('config_from', '') != 'predefined-model':
+                model_provider_service = ModelProviderService()
+
+                try:
+                    model_provider_service.save_model_credentials(
+                        tenant_id=tenant_id,
+                        provider=provider,
+                        model=args['model'],
+                        model_type=args['model_type'],
+                        credentials=args['credentials']
+                    )
+                except CredentialsValidateFailedError as ex:
+                    raise ValueError(str(ex))
 
         return {'result': 'success'}, 200
 
@@ -170,11 +206,73 @@ class ModelProviderModelCredentialApi(Resource):
             model=args['model']
         )
 
+        model_load_balancing_service = ModelLoadBalancingService()
+        is_load_balancing_enabled, load_balancing_configs = model_load_balancing_service.get_load_balancing_configs(
+            tenant_id=tenant_id,
+            provider=provider,
+            model=args['model'],
+            model_type=args['model_type']
+        )
+
         return {
-            "credentials": credentials
+            "credentials": credentials,
+            "load_balancing": {
+                "enabled": is_load_balancing_enabled,
+                "configs": load_balancing_configs
+            }
         }
 
 
+class ModelProviderModelEnableApi(Resource):
+
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def patch(self, provider: str):
+        tenant_id = current_user.current_tenant_id
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('model', type=str, required=True, nullable=False, location='json')
+        parser.add_argument('model_type', type=str, required=True, nullable=False,
+                            choices=[mt.value for mt in ModelType], location='json')
+        args = parser.parse_args()
+
+        model_provider_service = ModelProviderService()
+        model_provider_service.enable_model(
+            tenant_id=tenant_id,
+            provider=provider,
+            model=args['model'],
+            model_type=args['model_type']
+        )
+
+        return {'result': 'success'}
+
+
+class ModelProviderModelDisableApi(Resource):
+
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def patch(self, provider: str):
+        tenant_id = current_user.current_tenant_id
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('model', type=str, required=True, nullable=False, location='json')
+        parser.add_argument('model_type', type=str, required=True, nullable=False,
+                            choices=[mt.value for mt in ModelType], location='json')
+        args = parser.parse_args()
+
+        model_provider_service = ModelProviderService()
+        model_provider_service.disable_model(
+            tenant_id=tenant_id,
+            provider=provider,
+            model=args['model'],
+            model_type=args['model_type']
+        )
+
+        return {'result': 'success'}
+
+
 class ModelProviderModelValidateApi(Resource):
 
     @setup_required
@@ -259,6 +357,10 @@ class ModelProviderAvailableModelApi(Resource):
 
 
 api.add_resource(ModelProviderModelApi, '/workspaces/current/model-providers/<string:provider>/models')
+api.add_resource(ModelProviderModelEnableApi, '/workspaces/current/model-providers/<string:provider>/models/enable',
+                 endpoint='model-provider-model-enable')
+api.add_resource(ModelProviderModelDisableApi, '/workspaces/current/model-providers/<string:provider>/models/disable',
+                 endpoint='model-provider-model-disable')
 api.add_resource(ModelProviderModelCredentialApi,
                  '/workspaces/current/model-providers/<string:provider>/models/credentials')
 api.add_resource(ModelProviderModelValidateApi,

api/controllers/console/workspace/tool_providers.py

@@ -9,8 +9,13 @@ from controllers.console import api
 from controllers.console.setup import setup_required
 from controllers.console.wraps import account_initialization_required
 from core.model_runtime.utils.encoders import jsonable_encoder
+from libs.helper import alphanumeric, uuid_value
 from libs.login import login_required
-from services.tools_manage_service import ToolManageService
+from services.tools.api_tools_manage_service import ApiToolManageService
+from services.tools.builtin_tools_manage_service import BuiltinToolManageService
+from services.tools.tool_labels_service import ToolLabelsService
+from services.tools.tools_manage_service import ToolCommonService
+from services.tools.workflow_tools_manage_service import WorkflowToolManageService
 
 
 class ToolProviderListApi(Resource):
@@ -21,7 +26,11 @@ class ToolProviderListApi(Resource):
         user_id = current_user.id
         tenant_id = current_user.current_tenant_id
 
-        return ToolManageService.list_tool_providers(user_id, tenant_id)
+        req = reqparse.RequestParser()
+        req.add_argument('type', type=str, choices=['builtin', 'model', 'api', 'workflow'], required=False, nullable=True, location='args')
+        args = req.parse_args()
+
+        return ToolCommonService.list_tool_providers(user_id, tenant_id, args.get('type', None))
 
 class ToolBuiltinProviderListToolsApi(Resource):
     @setup_required
@@ -31,7 +40,7 @@ class ToolBuiltinProviderListToolsApi(Resource):
         user_id = current_user.id
         tenant_id = current_user.current_tenant_id
 
-        return jsonable_encoder(ToolManageService.list_builtin_tool_provider_tools(
+        return jsonable_encoder(BuiltinToolManageService.list_builtin_tool_provider_tools(
             user_id,
             tenant_id,
             provider,
@@ -48,7 +57,7 @@ class ToolBuiltinProviderDeleteApi(Resource):
         user_id = current_user.id
         tenant_id = current_user.current_tenant_id
 
-        return ToolManageService.delete_builtin_tool_provider(
+        return BuiltinToolManageService.delete_builtin_tool_provider(
             user_id,
             tenant_id,
             provider,
@@ -70,7 +79,7 @@ class ToolBuiltinProviderUpdateApi(Resource):
 
         args = parser.parse_args()
 
-        return ToolManageService.update_builtin_tool_provider(
+        return BuiltinToolManageService.update_builtin_tool_provider(
             user_id,
             tenant_id,
             provider,
@@ -85,7 +94,7 @@ class ToolBuiltinProviderGetCredentialsApi(Resource):
         user_id = current_user.id
         tenant_id = current_user.current_tenant_id
 
-        return ToolManageService.get_builtin_tool_provider_credentials(
+        return BuiltinToolManageService.get_builtin_tool_provider_credentials(
             user_id,
             tenant_id,
             provider,
@@ -94,35 +103,10 @@ class ToolBuiltinProviderGetCredentialsApi(Resource):
 class ToolBuiltinProviderIconApi(Resource):
     @setup_required
     def get(self, provider):
-        icon_bytes, mimetype = ToolManageService.get_builtin_tool_provider_icon(provider)
+        icon_bytes, mimetype = BuiltinToolManageService.get_builtin_tool_provider_icon(provider)
         icon_cache_max_age = int(current_app.config.get('TOOL_ICON_CACHE_MAX_AGE'))
         return send_file(io.BytesIO(icon_bytes), mimetype=mimetype, max_age=icon_cache_max_age)
 
-class ToolModelProviderIconApi(Resource):
-    @setup_required
-    def get(self, provider):
-        icon_bytes, mimetype = ToolManageService.get_model_tool_provider_icon(provider)
-        return send_file(io.BytesIO(icon_bytes), mimetype=mimetype)
-    
-class ToolModelProviderListToolsApi(Resource):
-    @setup_required
-    @login_required
-    @account_initialization_required
-    def get(self):
-        user_id = current_user.id
-        tenant_id = current_user.current_tenant_id
-
-        parser = reqparse.RequestParser()
-        parser.add_argument('provider', type=str, required=True, nullable=False, location='args')
-
-        args = parser.parse_args()
-
-        return jsonable_encoder(ToolManageService.list_model_tool_provider_tools(
-            user_id,
-            tenant_id,
-            args['provider'],
-        ))
-
 class ToolApiProviderAddApi(Resource):
     @setup_required
     @login_required
@@ -141,10 +125,12 @@ class ToolApiProviderAddApi(Resource):
         parser.add_argument('provider', type=str, required=True, nullable=False, location='json')
         parser.add_argument('icon', type=dict, required=True, nullable=False, location='json')
         parser.add_argument('privacy_policy', type=str, required=False, nullable=True, location='json')
+        parser.add_argument('labels', type=list[str], required=False, nullable=True, location='json', default=[])
+        parser.add_argument('custom_disclaimer', type=str, required=False, nullable=True, location='json')
 
         args = parser.parse_args()
 
-        return ToolManageService.create_api_tool_provider(
+        return ApiToolManageService.create_api_tool_provider(
             user_id,
             tenant_id,
             args['provider'],
@@ -153,6 +139,8 @@ class ToolApiProviderAddApi(Resource):
             args['schema_type'],
             args['schema'],
             args.get('privacy_policy', ''),
+            args.get('custom_disclaimer', ''),
+            args.get('labels', []),
         )
 
 class ToolApiProviderGetRemoteSchemaApi(Resource):
@@ -166,7 +154,7 @@ class ToolApiProviderGetRemoteSchemaApi(Resource):
 
         args = parser.parse_args()
 
-        return ToolManageService.get_api_tool_provider_remote_schema(
+        return ApiToolManageService.get_api_tool_provider_remote_schema(
             current_user.id,
             current_user.current_tenant_id,
             args['url'],
@@ -186,7 +174,7 @@ class ToolApiProviderListToolsApi(Resource):
 
         args = parser.parse_args()
 
-        return jsonable_encoder(ToolManageService.list_api_tool_provider_tools(
+        return jsonable_encoder(ApiToolManageService.list_api_tool_provider_tools(
             user_id,
             tenant_id,
             args['provider'],
@@ -211,10 +199,12 @@ class ToolApiProviderUpdateApi(Resource):
         parser.add_argument('original_provider', type=str, required=True, nullable=False, location='json')
         parser.add_argument('icon', type=dict, required=True, nullable=False, location='json')
         parser.add_argument('privacy_policy', type=str, required=True, nullable=True, location='json')
+        parser.add_argument('labels', type=list[str], required=False, nullable=True, location='json')
+        parser.add_argument('custom_disclaimer', type=str, required=True, nullable=True, location='json')
 
         args = parser.parse_args()
 
-        return ToolManageService.update_api_tool_provider(
+        return ApiToolManageService.update_api_tool_provider(
             user_id,
             tenant_id,
             args['provider'],
@@ -224,6 +214,8 @@ class ToolApiProviderUpdateApi(Resource):
             args['schema_type'],
             args['schema'],
             args['privacy_policy'],
+            args['custom_disclaimer'],
+            args.get('labels', []),
         )
 
 class ToolApiProviderDeleteApi(Resource):
@@ -243,7 +235,7 @@ class ToolApiProviderDeleteApi(Resource):
 
         args = parser.parse_args()
 
-        return ToolManageService.delete_api_tool_provider(
+        return ApiToolManageService.delete_api_tool_provider(
             user_id,
             tenant_id,
             args['provider'],
@@ -263,7 +255,7 @@ class ToolApiProviderGetApi(Resource):
 
         args = parser.parse_args()
 
-        return ToolManageService.get_api_tool_provider(
+        return ApiToolManageService.get_api_tool_provider(
             user_id,
             tenant_id,
             args['provider'],
@@ -274,7 +266,7 @@ class ToolBuiltinProviderCredentialsSchemaApi(Resource):
     @login_required
     @account_initialization_required
     def get(self, provider):
-        return ToolManageService.list_builtin_provider_credentials_schema(provider)
+        return BuiltinToolManageService.list_builtin_provider_credentials_schema(provider)
 
 class ToolApiProviderSchemaApi(Resource):
     @setup_required
@@ -287,7 +279,7 @@ class ToolApiProviderSchemaApi(Resource):
 
         args = parser.parse_args()
 
-        return ToolManageService.parser_api_schema(
+        return ApiToolManageService.parser_api_schema(
             schema=args['schema'],
         )
 
@@ -307,7 +299,7 @@ class ToolApiProviderPreviousTestApi(Resource):
 
         args = parser.parse_args()
 
-        return ToolManageService.test_api_tool_preview(
+        return ApiToolManageService.test_api_tool_preview(
             current_user.current_tenant_id,
             args['provider_name'] if args['provider_name'] else '',
             args['tool_name'],
@@ -317,6 +309,153 @@ class ToolApiProviderPreviousTestApi(Resource):
             args['schema'],
         )
 
+class ToolWorkflowProviderCreateApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def post(self):
+        if not current_user.is_admin_or_owner:
+            raise Forbidden()
+        
+        user_id = current_user.id
+        tenant_id = current_user.current_tenant_id
+
+        reqparser = reqparse.RequestParser()
+        reqparser.add_argument('workflow_app_id', type=uuid_value, required=True, nullable=False, location='json')
+        reqparser.add_argument('name', type=alphanumeric, required=True, nullable=False, location='json')
+        reqparser.add_argument('label', type=str, required=True, nullable=False, location='json')
+        reqparser.add_argument('description', type=str, required=True, nullable=False, location='json')
+        reqparser.add_argument('icon', type=dict, required=True, nullable=False, location='json')
+        reqparser.add_argument('parameters', type=list[dict], required=True, nullable=False, location='json')
+        reqparser.add_argument('privacy_policy', type=str, required=False, nullable=True, location='json', default='')
+        reqparser.add_argument('labels', type=list[str], required=False, nullable=True, location='json')
+
+        args = reqparser.parse_args()
+
+        return WorkflowToolManageService.create_workflow_tool(
+            user_id,
+            tenant_id,
+            args['workflow_app_id'],
+            args['name'],
+            args['label'],
+            args['icon'],
+            args['description'],
+            args['parameters'],
+            args['privacy_policy'],
+            args.get('labels', []),
+        )
+
+class ToolWorkflowProviderUpdateApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def post(self):
+        if not current_user.is_admin_or_owner:
+            raise Forbidden()
+        
+        user_id = current_user.id
+        tenant_id = current_user.current_tenant_id
+
+        reqparser = reqparse.RequestParser()
+        reqparser.add_argument('workflow_tool_id', type=uuid_value, required=True, nullable=False, location='json')
+        reqparser.add_argument('name', type=alphanumeric, required=True, nullable=False, location='json')
+        reqparser.add_argument('label', type=str, required=True, nullable=False, location='json')
+        reqparser.add_argument('description', type=str, required=True, nullable=False, location='json')
+        reqparser.add_argument('icon', type=dict, required=True, nullable=False, location='json')
+        reqparser.add_argument('parameters', type=list[dict], required=True, nullable=False, location='json')
+        reqparser.add_argument('privacy_policy', type=str, required=False, nullable=True, location='json', default='')
+        reqparser.add_argument('labels', type=list[str], required=False, nullable=True, location='json')
+        
+        args = reqparser.parse_args()
+
+        if not args['workflow_tool_id']:
+            raise ValueError('incorrect workflow_tool_id')
+        
+        return WorkflowToolManageService.update_workflow_tool(
+            user_id,
+            tenant_id,
+            args['workflow_tool_id'],
+            args['name'],
+            args['label'],
+            args['icon'],
+            args['description'],
+            args['parameters'],
+            args['privacy_policy'],
+            args.get('labels', []),
+        )
+
+class ToolWorkflowProviderDeleteApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def post(self):
+        if not current_user.is_admin_or_owner:
+            raise Forbidden()
+        
+        user_id = current_user.id
+        tenant_id = current_user.current_tenant_id
+
+        reqparser = reqparse.RequestParser()
+        reqparser.add_argument('workflow_tool_id', type=uuid_value, required=True, nullable=False, location='json')
+
+        args = reqparser.parse_args()
+
+        return WorkflowToolManageService.delete_workflow_tool(
+            user_id,
+            tenant_id,
+            args['workflow_tool_id'],
+        )
+        
+class ToolWorkflowProviderGetApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def get(self):
+        user_id = current_user.id
+        tenant_id = current_user.current_tenant_id
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('workflow_tool_id', type=uuid_value, required=False, nullable=True, location='args')
+        parser.add_argument('workflow_app_id', type=uuid_value, required=False, nullable=True, location='args')
+
+        args = parser.parse_args()
+
+        if args.get('workflow_tool_id'):
+            tool = WorkflowToolManageService.get_workflow_tool_by_tool_id(
+                user_id,
+                tenant_id,
+                args['workflow_tool_id'],
+            )
+        elif args.get('workflow_app_id'):
+            tool = WorkflowToolManageService.get_workflow_tool_by_app_id(
+                user_id,
+                tenant_id,
+                args['workflow_app_id'],
+            )
+        else:
+            raise ValueError('incorrect workflow_tool_id or workflow_app_id')
+
+        return jsonable_encoder(tool)
+    
+class ToolWorkflowProviderListToolApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def get(self):
+        user_id = current_user.id
+        tenant_id = current_user.current_tenant_id
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('workflow_tool_id', type=uuid_value, required=True, nullable=False, location='args')
+
+        args = parser.parse_args()
+
+        return jsonable_encoder(WorkflowToolManageService.list_single_workflow_tools(
+            user_id,
+            tenant_id,
+            args['workflow_tool_id'],
+        ))
+
 class ToolBuiltinListApi(Resource):
     @setup_required
     @login_required
@@ -325,7 +464,7 @@ class ToolBuiltinListApi(Resource):
         user_id = current_user.id
         tenant_id = current_user.current_tenant_id
 
-        return jsonable_encoder([provider.to_dict() for provider in ToolManageService.list_builtin_tools(
+        return jsonable_encoder([provider.to_dict() for provider in BuiltinToolManageService.list_builtin_tools(
             user_id,
             tenant_id,
         )])
@@ -338,20 +477,43 @@ class ToolApiListApi(Resource):
         user_id = current_user.id
         tenant_id = current_user.current_tenant_id
 
-        return jsonable_encoder([provider.to_dict() for provider in ToolManageService.list_api_tools(
+        return jsonable_encoder([provider.to_dict() for provider in ApiToolManageService.list_api_tools(
             user_id,
             tenant_id,
         )])
+    
+class ToolWorkflowListApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def get(self):
+        user_id = current_user.id
+        tenant_id = current_user.current_tenant_id
 
+        return jsonable_encoder([provider.to_dict() for provider in WorkflowToolManageService.list_tenant_workflow_tools(
+            user_id,
+            tenant_id,
+        )])
+    
+class ToolLabelsApi(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def get(self):
+        return jsonable_encoder(ToolLabelsService.list_tool_labels())
+
+# tool provider
 api.add_resource(ToolProviderListApi, '/workspaces/current/tool-providers')
+
+# builtin tool provider
 api.add_resource(ToolBuiltinProviderListToolsApi, '/workspaces/current/tool-provider/builtin/<provider>/tools')
 api.add_resource(ToolBuiltinProviderDeleteApi, '/workspaces/current/tool-provider/builtin/<provider>/delete')
 api.add_resource(ToolBuiltinProviderUpdateApi, '/workspaces/current/tool-provider/builtin/<provider>/update')
 api.add_resource(ToolBuiltinProviderGetCredentialsApi, '/workspaces/current/tool-provider/builtin/<provider>/credentials')
 api.add_resource(ToolBuiltinProviderCredentialsSchemaApi, '/workspaces/current/tool-provider/builtin/<provider>/credentials_schema')
 api.add_resource(ToolBuiltinProviderIconApi, '/workspaces/current/tool-provider/builtin/<provider>/icon')
-api.add_resource(ToolModelProviderIconApi, '/workspaces/current/tool-provider/model/<provider>/icon')
-api.add_resource(ToolModelProviderListToolsApi, '/workspaces/current/tool-provider/model/tools')
+
+# api tool provider
 api.add_resource(ToolApiProviderAddApi, '/workspaces/current/tool-provider/api/add')
 api.add_resource(ToolApiProviderGetRemoteSchemaApi, '/workspaces/current/tool-provider/api/remote')
 api.add_resource(ToolApiProviderListToolsApi, '/workspaces/current/tool-provider/api/tools')
@@ -361,5 +523,15 @@ api.add_resource(ToolApiProviderGetApi, '/workspaces/current/tool-provider/api/g
 api.add_resource(ToolApiProviderSchemaApi, '/workspaces/current/tool-provider/api/schema')
 api.add_resource(ToolApiProviderPreviousTestApi, '/workspaces/current/tool-provider/api/test/pre')
 
+# workflow tool provider
+api.add_resource(ToolWorkflowProviderCreateApi, '/workspaces/current/tool-provider/workflow/create')
+api.add_resource(ToolWorkflowProviderUpdateApi, '/workspaces/current/tool-provider/workflow/update')
+api.add_resource(ToolWorkflowProviderDeleteApi, '/workspaces/current/tool-provider/workflow/delete')
+api.add_resource(ToolWorkflowProviderGetApi, '/workspaces/current/tool-provider/workflow/get')
+api.add_resource(ToolWorkflowProviderListToolApi, '/workspaces/current/tool-provider/workflow/tools')
+
 api.add_resource(ToolBuiltinListApi, '/workspaces/current/tools/builtin')
-api.add_resource(ToolApiListApi, '/workspaces/current/tools/api')
+api.add_resource(ToolApiListApi, '/workspaces/current/tools/api')
+api.add_resource(ToolWorkflowListApi, '/workspaces/current/tools/workflow')
+
+api.add_resource(ToolLabelsApi, '/workspaces/current/tool-labels')

api/controllers/console/workspace/workspace.py

@@ -161,13 +161,13 @@ class CustomConfigWorkspaceApi(Resource):
         parser.add_argument('replace_webapp_logo', type=str,  location='json')
         args = parser.parse_args()
 
+        tenant = db.session.query(Tenant).filter(Tenant.id == current_user.current_tenant_id).one_or_404()
+
         custom_config_dict = {
             'remove_webapp_brand': args['remove_webapp_brand'],
-            'replace_webapp_logo': args['replace_webapp_logo'],
+            'replace_webapp_logo': args['replace_webapp_logo'] if args['replace_webapp_logo'] is not None else tenant.custom_config_dict.get('replace_webapp_logo') ,
         }
 
-        tenant = db.session.query(Tenant).filter(Tenant.id == current_user.current_tenant_id).one_or_404()
-
         tenant.custom_config_dict = custom_config_dict
         db.session.commit()
 

api/controllers/service_api/app/message.py

@@ -97,7 +97,7 @@ class MessageListApi(Resource):
 
 
 class MessageFeedbackApi(Resource):
-    @validate_app_token(fetch_user_arg=FetchUserArg(fetch_from=WhereisUserArg.JSON))
+    @validate_app_token(fetch_user_arg=FetchUserArg(fetch_from=WhereisUserArg.JSON, required=True))
     def post(self, app_model: App, end_user: EndUser, message_id):
         message_id = str(message_id)
 
@@ -114,7 +114,7 @@ class MessageFeedbackApi(Resource):
 
 
 class MessageSuggestedApi(Resource):
-    @validate_app_token(fetch_user_arg=FetchUserArg(fetch_from=WhereisUserArg.QUERY))
+    @validate_app_token(fetch_user_arg=FetchUserArg(fetch_from=WhereisUserArg.QUERY, required=True))
     def get(self, app_model: App, end_user: EndUser, message_id):
         message_id = str(message_id)
         app_mode = AppMode.value_of(app_model.mode)

api/controllers/service_api/wraps.py

@@ -8,7 +8,7 @@ from flask import current_app, request
 from flask_login import user_logged_in
 from flask_restful import Resource
 from pydantic import BaseModel
-from werkzeug.exceptions import Forbidden, NotFound, Unauthorized
+from werkzeug.exceptions import Forbidden, Unauthorized
 
 from extensions.ext_database import db
 from libs.login import _get_user
@@ -39,17 +39,17 @@ def validate_app_token(view: Optional[Callable] = None, *, fetch_user_arg: Optio
 
             app_model = db.session.query(App).filter(App.id == api_token.app_id).first()
             if not app_model:
-                raise NotFound()
+                raise Forbidden("The app no longer exists.")
 
             if app_model.status != 'normal':
-                raise NotFound()
+                raise Forbidden("The app's status is abnormal.")
 
             if not app_model.enable_api:
-                raise NotFound()
+                raise Forbidden("The app's API service has been disabled.")
 
             tenant = db.session.query(Tenant).filter(Tenant.id == app_model.tenant_id).first()
             if tenant.status == TenantStatus.ARCHIVE:
-                raise NotFound()
+                raise Forbidden("The workspace's status is archived.")
 
             kwargs['app_model'] = app_model
 

api/controllers/web/__init__.py

@@ -6,4 +6,4 @@ bp = Blueprint('web', __name__, url_prefix='/api')
 api = ExternalApi(bp)
 
 
-from . import app, audio, completion, conversation, file, message, passport, saved_message, site, workflow
+from . import app, audio, completion, conversation, feature, file, message, passport, saved_message, site, workflow

api/controllers/web/app.py

@@ -1,14 +1,10 @@
-import json
-
 from flask import current_app
 from flask_restful import fields, marshal_with
 
 from controllers.web import api
 from controllers.web.error import AppUnavailableError
 from controllers.web.wraps import WebApiResource
-from extensions.ext_database import db
-from models.model import App, AppMode, AppModelConfig
-from models.tools import ApiToolProvider
+from models.model import App, AppMode
 from services.app_service import AppService
 
 

api/controllers/web/audio.py

@@ -74,7 +74,7 @@ class TextApi(WebApiResource):
                 app_model=app_model,
                 text=request.form['text'],
                 end_user=end_user.external_user_id,
-                voice=request.form.get('voice'),
+                voice=request.form['voice'] if request.form.get('voice') else app_model.app_model_config.text_to_speech_dict.get('voice'),
                 streaming=False
             )
 

api/controllers/web/error.py

@@ -115,3 +115,9 @@ class UnsupportedFileTypeError(BaseHTTPException):
     error_code = 'unsupported_file_type'
     description = "File type not allowed."
     code = 415
+
+
+class WebSSOAuthRequiredError(BaseHTTPException):
+    error_code = 'web_sso_auth_required'
+    description = "Web SSO authentication required."
+    code = 401

api/controllers/web/feature.py

@@ -0,0 +1,12 @@
+from flask_restful import Resource
+
+from controllers.web import api
+from services.feature_service import FeatureService
+
+
+class SystemFeatureApi(Resource):
+    def get(self):
+        return FeatureService.get_system_features().dict()
+
+
+api.add_resource(SystemFeatureApi, '/system-features')

api/controllers/web/passport.py

@@ -5,14 +5,21 @@ from flask_restful import Resource
 from werkzeug.exceptions import NotFound, Unauthorized
 
 from controllers.web import api
+from controllers.web.error import WebSSOAuthRequiredError
 from extensions.ext_database import db
 from libs.passport import PassportService
 from models.model import App, EndUser, Site
+from services.feature_service import FeatureService
 
 
 class PassportResource(Resource):
     """Base resource for passport."""
     def get(self):
+
+        system_features = FeatureService.get_system_features()
+        if system_features.sso_enforced_for_web:
+            raise WebSSOAuthRequiredError()
+
         app_code = request.headers.get('X-App-Code')
         if app_code is None:
             raise Unauthorized('X-App-Code header is missing.')
@@ -28,7 +35,7 @@ class PassportResource(Resource):
         app_model = db.session.query(App).filter(App.id == site.app_id).first()
         if not app_model or app_model.status != 'normal' or not app_model.enable_site:
             raise NotFound()
-        
+
         end_user = EndUser(
             tenant_id=app_model.tenant_id,
             app_id=app_model.id,
@@ -36,6 +43,7 @@ class PassportResource(Resource):
             is_anonymous=True,
             session_id=generate_session_id(),
         )
+
         db.session.add(end_user)
         db.session.commit()
 
@@ -53,8 +61,10 @@ class PassportResource(Resource):
             'access_token': tk,
         }
 
+
 api.add_resource(PassportResource, '/passport')
 
+
 def generate_session_id():
     """
     Generate a unique session ID.

api/controllers/web/site.py

@@ -31,6 +31,7 @@ class AppSiteApi(WebApiResource):
         'description': fields.String,
         'copyright': fields.String,
         'privacy_policy': fields.String,
+        'custom_disclaimer': fields.String,
         'default_language': fields.String,
         'prompt_public': fields.Boolean
     }

api/controllers/web/wraps.py

@@ -2,11 +2,13 @@ from functools import wraps
 
 from flask import request
 from flask_restful import Resource
-from werkzeug.exceptions import NotFound, Unauthorized
+from werkzeug.exceptions import BadRequest, NotFound, Unauthorized
 
+from controllers.web.error import WebSSOAuthRequiredError
 from extensions.ext_database import db
 from libs.passport import PassportService
 from models.model import App, EndUser, Site
+from services.feature_service import FeatureService
 
 
 def validate_jwt_token(view=None):
@@ -21,34 +23,60 @@ def validate_jwt_token(view=None):
         return decorator(view)
     return decorator
 
+
 def decode_jwt_token():
-    auth_header = request.headers.get('Authorization')
-    if auth_header is None:
-        raise Unauthorized('Authorization header is missing.')
+    system_features = FeatureService.get_system_features()
 
-    if ' ' not in auth_header:
-        raise Unauthorized('Invalid Authorization header format. Expected \'Bearer <api-key>\' format.')
-    
-    auth_scheme, tk = auth_header.split(None, 1)
-    auth_scheme = auth_scheme.lower()
+    try:
+        auth_header = request.headers.get('Authorization')
+        if auth_header is None:
+            raise Unauthorized('Authorization header is missing.')
 
-    if auth_scheme != 'bearer':
-        raise Unauthorized('Invalid Authorization header format. Expected \'Bearer <api-key>\' format.')
-    decoded = PassportService().verify(tk)
-    app_code = decoded.get('app_code')
-    app_model = db.session.query(App).filter(App.id == decoded['app_id']).first()
-    site = db.session.query(Site).filter(Site.code == app_code).first()
-    if not app_model:
-        raise NotFound()
-    if not app_code or not site:
-        raise Unauthorized('Site URL is no longer valid.')
-    if app_model.enable_site is False:
-        raise Unauthorized('Site is disabled.')
-    end_user = db.session.query(EndUser).filter(EndUser.id == decoded['end_user_id']).first()
-    if not end_user:
-        raise NotFound()
+        if ' ' not in auth_header:
+            raise Unauthorized('Invalid Authorization header format. Expected \'Bearer <api-key>\' format.')
+
+        auth_scheme, tk = auth_header.split(None, 1)
+        auth_scheme = auth_scheme.lower()
+
+        if auth_scheme != 'bearer':
+            raise Unauthorized('Invalid Authorization header format. Expected \'Bearer <api-key>\' format.')
+        decoded = PassportService().verify(tk)
+        app_code = decoded.get('app_code')
+        app_model = db.session.query(App).filter(App.id == decoded['app_id']).first()
+        site = db.session.query(Site).filter(Site.code == app_code).first()
+        if not app_model:
+            raise NotFound()
+        if not app_code or not site:
+            raise BadRequest('Site URL is no longer valid.')
+        if app_model.enable_site is False:
+            raise BadRequest('Site is disabled.')
+        end_user = db.session.query(EndUser).filter(EndUser.id == decoded['end_user_id']).first()
+        if not end_user:
+            raise NotFound()
+
+        _validate_web_sso_token(decoded, system_features)
+
+        return app_model, end_user
+    except Unauthorized as e:
+        if system_features.sso_enforced_for_web:
+            raise WebSSOAuthRequiredError()
+
+        raise Unauthorized(e.description)
+
+
+def _validate_web_sso_token(decoded, system_features):
+    # Check if SSO is enforced for web, and if the token source is not SSO, raise an error and redirect to SSO login
+    if system_features.sso_enforced_for_web:
+        source = decoded.get('token_source')
+        if not source or source != 'sso':
+            raise WebSSOAuthRequiredError()
+
+    # Check if SSO is not enforced for web, and if the token source is SSO, raise an error and redirect to normal passport login
+    if not system_features.sso_enforced_for_web:
+        source = decoded.get('token_source')
+        if source and source == 'sso':
+            raise Unauthorized('sso token expired.')
 
-    return app_model, end_user
 
 class WebApiResource(Resource):
     method_decorators = [validate_jwt_token]

api/core/agent/base_agent_runner.py

@@ -39,6 +39,7 @@ from core.tools.entities.tool_entities import (
 from core.tools.tool.dataset_retriever_tool import DatasetRetrieverTool
 from core.tools.tool.tool import Tool
 from core.tools.tool_manager import ToolManager
+from core.tools.utils.tool_parameter_converter import ToolParameterConverter
 from extensions.ext_database import db
 from models.model import Conversation, Message, MessageAgentThought
 from models.tools import ToolConversationVariables
@@ -128,6 +129,8 @@ class BaseAgentRunner(AppRunner):
             self.files = application_generate_entity.files
         else:
             self.files = []
+        self.query = None
+        self._current_thoughts: list[PromptMessage] = []
 
     def _repack_app_generate_entity(self, app_generate_entity: AgentChatAppGenerateEntity) \
             -> AgentChatAppGenerateEntity:
@@ -165,6 +168,7 @@ class BaseAgentRunner(AppRunner):
             tenant_id=self.tenant_id,
             app_id=self.app_config.app_id,
             agent_tool=tool,
+            invoke_from=self.application_generate_entity.invoke_from
         )
         tool_entity.load_variables(self.variables_pool)
 
@@ -183,21 +187,11 @@ class BaseAgentRunner(AppRunner):
             if parameter.form != ToolParameter.ToolParameterForm.LLM:
                 continue
 
-            parameter_type = 'string'
+            parameter_type = ToolParameterConverter.get_parameter_type(parameter.type)
             enum = []
-            if parameter.type == ToolParameter.ToolParameterType.STRING:
-                parameter_type = 'string'
-            elif parameter.type == ToolParameter.ToolParameterType.BOOLEAN:
-                parameter_type = 'boolean'
-            elif parameter.type == ToolParameter.ToolParameterType.NUMBER:
-                parameter_type = 'number'
-            elif parameter.type == ToolParameter.ToolParameterType.SELECT:
-                for option in parameter.options:
-                    enum.append(option.value)
-                parameter_type = 'string'
-            else:
-                raise ValueError(f"parameter type {parameter.type} is not supported")
-            
+            if parameter.type == ToolParameter.ToolParameterType.SELECT:
+                enum = [option.value for option in parameter.options]
+
             message_tool.parameters['properties'][parameter.name] = {
                 "type": parameter_type,
                 "description": parameter.llm_description or '',
@@ -278,20 +272,10 @@ class BaseAgentRunner(AppRunner):
             if parameter.form != ToolParameter.ToolParameterForm.LLM:
                 continue
 
-            parameter_type = 'string'
+            parameter_type = ToolParameterConverter.get_parameter_type(parameter.type)
             enum = []
-            if parameter.type == ToolParameter.ToolParameterType.STRING:
-                parameter_type = 'string'
-            elif parameter.type == ToolParameter.ToolParameterType.BOOLEAN:
-                parameter_type = 'boolean'
-            elif parameter.type == ToolParameter.ToolParameterType.NUMBER:
-                parameter_type = 'number'
-            elif parameter.type == ToolParameter.ToolParameterType.SELECT:
-                for option in parameter.options:
-                    enum.append(option.value)
-                parameter_type = 'string'
-            else:
-                raise ValueError(f"parameter type {parameter.type} is not supported")
+            if parameter.type == ToolParameter.ToolParameterType.SELECT:
+                enum = [option.value for option in parameter.options]
         
             prompt_tool.parameters['properties'][parameter.name] = {
                 "type": parameter_type,
@@ -463,7 +447,7 @@ class BaseAgentRunner(AppRunner):
         for message in messages:
             if message.id == self.message.id:
                 continue
-            
+
             result.append(self.organize_agent_user_prompt(message))
             agent_thoughts: list[MessageAgentThought] = message.agent_thoughts
             if agent_thoughts:
@@ -544,3 +528,4 @@ class BaseAgentRunner(AppRunner):
                 return UserPromptMessage(content=prompt_message_contents)
         else:
             return UserPromptMessage(content=message.query)
+         

api/core/agent/cot_agent_runner.py

@@ -15,6 +15,7 @@ from core.model_runtime.entities.message_entities import (
     ToolPromptMessage,
     UserPromptMessage,
 )
+from core.prompt.agent_history_prompt_transform import AgentHistoryPromptTransform
 from core.tools.entities.tool_entities import ToolInvokeMeta
 from core.tools.tool.tool import Tool
 from core.tools.tool_engine import ToolEngine
@@ -121,7 +122,7 @@ class CotAgentRunner(BaseAgentRunner, ABC):
                 raise ValueError("failed to invoke llm")
             
             usage_dict = {}
-            react_chunks = CotAgentOutputParser.handle_react_stream_output(chunks)
+            react_chunks = CotAgentOutputParser.handle_react_stream_output(chunks, usage_dict)
             scratchpad = AgentScratchpadUnit(
                 agent_response='',
                 thought='',
@@ -189,7 +190,7 @@ class CotAgentRunner(BaseAgentRunner, ABC):
 
             if not scratchpad.action:
                 # failed to extract action, return final answer directly
-                final_answer = scratchpad.agent_response or ''
+                final_answer = ''
             else:
                 if scratchpad.action.action_name.lower() == "final answer":
                     # action is final answer, return final answer directly
@@ -373,7 +374,7 @@ class CotAgentRunner(BaseAgentRunner, ABC):
 
         return message
 
-    def _organize_historic_prompt_messages(self) -> list[PromptMessage]:
+    def _organize_historic_prompt_messages(self, current_session_messages: list[PromptMessage] = None) -> list[PromptMessage]:
         """
             organize historic prompt messages
         """
@@ -381,6 +382,13 @@ class CotAgentRunner(BaseAgentRunner, ABC):
         scratchpad: list[AgentScratchpadUnit] = []
         current_scratchpad: AgentScratchpadUnit = None
 
+        self.history_prompt_messages = AgentHistoryPromptTransform(
+            model_config=self.model_config,
+            prompt_messages=current_session_messages or [],
+            history_messages=self.history_prompt_messages,
+            memory=self.memory
+        ).get_prompt()
+
         for message in self.history_prompt_messages:
             if isinstance(message, AssistantPromptMessage):
                 current_scratchpad = AgentScratchpadUnit(

api/core/agent/cot_chat_agent_runner.py

@@ -32,9 +32,6 @@ class CotChatAgentRunner(CotAgentRunner):
         # organize system prompt
         system_message = self._organize_system_prompt()
 
-        # organize historic prompt messages
-        historic_messages = self._historic_prompt_messages
-
         # organize current assistant messages
         agent_scratchpad = self._agent_scratchpad
         if not agent_scratchpad:
@@ -57,6 +54,13 @@ class CotChatAgentRunner(CotAgentRunner):
         query_messages = UserPromptMessage(content=self._query)
 
         if assistant_messages:
+            # organize historic prompt messages
+            historic_messages = self._organize_historic_prompt_messages([
+                system_message,
+                query_messages,
+                *assistant_messages,
+                UserPromptMessage(content='continue')
+            ])            
             messages = [
                 system_message,
                 *historic_messages,
@@ -65,6 +69,8 @@ class CotChatAgentRunner(CotAgentRunner):
                 UserPromptMessage(content='continue')
             ]
         else:
+            # organize historic prompt messages
+            historic_messages = self._organize_historic_prompt_messages([system_message, query_messages])
             messages = [system_message, *historic_messages, query_messages]
 
         # join all messages

api/core/agent/cot_completion_agent_runner.py

@@ -19,11 +19,11 @@ class CotCompletionAgentRunner(CotAgentRunner):
         
         return system_prompt
 
-    def _organize_historic_prompt(self) -> str:
+    def _organize_historic_prompt(self, current_session_messages: list[PromptMessage] = None) -> str:
         """
         Organize historic prompt
         """
-        historic_prompt_messages = self._historic_prompt_messages
+        historic_prompt_messages = self._organize_historic_prompt_messages(current_session_messages)
         historic_prompt = ""
 
         for message in historic_prompt_messages:

api/core/agent/entities.py

@@ -8,7 +8,7 @@ class AgentToolEntity(BaseModel):
     """
     Agent Tool Entity.
     """
-    provider_type: Literal["builtin", "api"]
+    provider_type: Literal["builtin", "api", "workflow"]
     provider_id: str
     tool_name: str
     tool_parameters: dict[str, Any] = {}

api/core/agent/fc_agent_runner.py

@@ -17,6 +17,7 @@ from core.model_runtime.entities.message_entities import (
     ToolPromptMessage,
     UserPromptMessage,
 )
+from core.prompt.agent_history_prompt_transform import AgentHistoryPromptTransform
 from core.tools.entities.tool_entities import ToolInvokeMeta
 from core.tools.tool_engine import ToolEngine
 from models.model import Message
@@ -24,21 +25,18 @@ from models.model import Message
 logger = logging.getLogger(__name__)
 
 class FunctionCallAgentRunner(BaseAgentRunner):
+
     def run(self, 
             message: Message, query: str, **kwargs: Any
     ) -> Generator[LLMResultChunk, None, None]:
         """
         Run FunctionCall agent application
         """
+        self.query = query
         app_generate_entity = self.application_generate_entity
 
         app_config = self.app_config
 
-        prompt_template = app_config.prompt_template.simple_prompt_template or ''
-        prompt_messages = self.history_prompt_messages
-        prompt_messages = self._init_system_message(prompt_template, prompt_messages)
-        prompt_messages = self._organize_user_query(query, prompt_messages)
-
         # convert tools into ModelRuntime Tool format
         tool_instances, prompt_messages_tools = self._init_prompt_tools()
 
@@ -81,6 +79,7 @@ class FunctionCallAgentRunner(BaseAgentRunner):
             )
 
             # recalc llm max tokens
+            prompt_messages = self._organize_prompt_messages()
             self.recalc_llm_max_tokens(self.model_config, prompt_messages)
             # invoke model
             chunks: Union[Generator[LLMResultChunk, None, None], LLMResult] = model_instance.invoke_llm(
@@ -203,7 +202,7 @@ class FunctionCallAgentRunner(BaseAgentRunner):
             else:
                 assistant_message.content = response
             
-            prompt_messages.append(assistant_message)
+            self._current_thoughts.append(assistant_message)
 
             # save thought
             self.save_agent_thought(
@@ -265,12 +264,14 @@ class FunctionCallAgentRunner(BaseAgentRunner):
                     }
                 
                 tool_responses.append(tool_response)
-                prompt_messages = self._organize_assistant_message(
-                    tool_call_id=tool_call_id,
-                    tool_call_name=tool_call_name,
-                    tool_response=tool_response['tool_response'],
-                    prompt_messages=prompt_messages,
-                )
+                if tool_response['tool_response'] is not None:
+                    self._current_thoughts.append(
+                        ToolPromptMessage(
+                            content=tool_response['tool_response'],
+                            tool_call_id=tool_call_id,
+                            name=tool_call_name,
+                        )
+                    ) 
 
             if len(tool_responses) > 0:
                 # save agent thought
@@ -300,8 +301,6 @@ class FunctionCallAgentRunner(BaseAgentRunner):
 
             iteration_step += 1
 
-            prompt_messages = self._clear_user_prompt_image_messages(prompt_messages)
-
         self.update_db_variables(self.variables_pool, self.db_variables_pool)
         # publish end event
         self.queue_manager.publish(QueueMessageEndEvent(llm_result=LLMResult(
@@ -393,24 +392,6 @@ class FunctionCallAgentRunner(BaseAgentRunner):
 
         return prompt_messages
     
-    def _organize_assistant_message(self, tool_call_id: str = None, tool_call_name: str = None, tool_response: str = None, 
-                                    prompt_messages: list[PromptMessage] = None) -> list[PromptMessage]:
-        """
-        Organize assistant message
-        """
-        prompt_messages = deepcopy(prompt_messages)
-
-        if tool_response is not None:
-            prompt_messages.append(
-                ToolPromptMessage(
-                    content=tool_response,
-                    tool_call_id=tool_call_id,
-                    name=tool_call_name,
-                )
-            )
-
-        return prompt_messages
-    
     def _clear_user_prompt_image_messages(self, prompt_messages: list[PromptMessage]) -> list[PromptMessage]:
         """
         As for now, gpt supports both fc and vision at the first iteration.
@@ -428,4 +409,26 @@ class FunctionCallAgentRunner(BaseAgentRunner):
                         for content in prompt_message.content 
                     ])
 
-        return prompt_messages
+        return prompt_messages
+
+    def _organize_prompt_messages(self):
+        prompt_template = self.app_config.prompt_template.simple_prompt_template or ''
+        self.history_prompt_messages = self._init_system_message(prompt_template, self.history_prompt_messages)
+        query_prompt_messages = self._organize_user_query(self.query, [])
+
+        self.history_prompt_messages = AgentHistoryPromptTransform(
+            model_config=self.model_config,
+            prompt_messages=[*query_prompt_messages, *self._current_thoughts],
+            history_messages=self.history_prompt_messages,
+            memory=self.memory
+        ).get_prompt()
+
+        prompt_messages = [
+            *self.history_prompt_messages,
+            *query_prompt_messages,
+            *self._current_thoughts
+        ]
+        if len(self._current_thoughts) != 0:
+            # clear messages after the first iteration
+            prompt_messages = self._clear_user_prompt_image_messages(prompt_messages)
+        return prompt_messages

api/core/agent/output_parser/cot_output_parser.py

@@ -9,7 +9,7 @@ from core.model_runtime.entities.llm_entities import LLMResultChunk
 
 class CotAgentOutputParser:
     @classmethod
-    def handle_react_stream_output(cls, llm_response: Generator[LLMResultChunk, None, None]) -> \
+    def handle_react_stream_output(cls, llm_response: Generator[LLMResultChunk, None, None], usage_dict: dict) -> \
         Generator[Union[str, AgentScratchpadUnit.Action], None, None]:
         def parse_action(json_str):
             try:
@@ -58,6 +58,8 @@ class CotAgentOutputParser:
         thought_idx = 0
 
         for response in llm_response:
+            if response.delta.usage:
+                usage_dict['usage'] = response.delta.usage
             response = response.delta.message.content
             if not isinstance(response, str):
                 continue

api/core/app/app_config/common/sensitive_word_avoidance/manager.py

@@ -11,7 +11,7 @@ class SensitiveWordAvoidanceConfigManager:
         if not sensitive_word_avoidance_dict:
             return None
 
-        if 'enabled' in sensitive_word_avoidance_dict and sensitive_word_avoidance_dict['enabled']:
+        if sensitive_word_avoidance_dict.get('enabled'):
             return SensitiveWordAvoidanceEntity(
                 type=sensitive_word_avoidance_dict.get('type'),
                 config=sensitive_word_avoidance_dict.get('config'),

api/core/app/app_config/easy_ui_based_app/agent/manager.py

@@ -1,7 +1,7 @@
 from typing import Optional
 
 from core.agent.entities import AgentEntity, AgentPromptEntity, AgentToolEntity
-from core.tools.prompt.template import REACT_PROMPT_TEMPLATES
+from core.agent.prompt.template import REACT_PROMPT_TEMPLATES
 
 
 class AgentConfigManager:

api/core/app/app_config/entities.py

@@ -239,4 +239,4 @@ class WorkflowUIBasedAppConfig(AppConfig):
     """
     Workflow UI Based App Config Entity.
     """
-    workflow_id: str
+    workflow_id: str

api/core/app/app_config/features/file_upload/manager.py

@@ -14,7 +14,7 @@ class FileUploadConfigManager:
         """
         file_upload_dict = config.get('file_upload')
         if file_upload_dict:
-            if 'image' in file_upload_dict and file_upload_dict['image']:
+            if file_upload_dict.get('image'):
                 if 'enabled' in file_upload_dict['image'] and file_upload_dict['image']['enabled']:
                     image_config = {
                         'number_limits': file_upload_dict['image']['number_limits'],

api/core/app/app_config/features/more_like_this/manager.py

@@ -9,7 +9,7 @@ class MoreLikeThisConfigManager:
         more_like_this = False
         more_like_this_dict = config.get('more_like_this')
         if more_like_this_dict:
-            if 'enabled' in more_like_this_dict and more_like_this_dict['enabled']:
+            if more_like_this_dict.get('enabled'):
                 more_like_this = True
 
         return more_like_this

api/core/app/app_config/features/retrieval_resource/manager.py

@@ -4,7 +4,7 @@ class RetrievalResourceConfigManager:
         show_retrieve_source = False
         retriever_resource_dict = config.get('retriever_resource')
         if retriever_resource_dict:
-            if 'enabled' in retriever_resource_dict and retriever_resource_dict['enabled']:
+            if retriever_resource_dict.get('enabled'):
                 show_retrieve_source = True
 
         return show_retrieve_source

api/core/app/app_config/features/speech_to_text/manager.py

@@ -9,7 +9,7 @@ class SpeechToTextConfigManager:
         speech_to_text = False
         speech_to_text_dict = config.get('speech_to_text')
         if speech_to_text_dict:
-            if 'enabled' in speech_to_text_dict and speech_to_text_dict['enabled']:
+            if speech_to_text_dict.get('enabled'):
                 speech_to_text = True
 
         return speech_to_text

api/core/app/app_config/features/suggested_questions_after_answer/manager.py

@@ -9,7 +9,7 @@ class SuggestedQuestionsAfterAnswerConfigManager:
         suggested_questions_after_answer = False
         suggested_questions_after_answer_dict = config.get('suggested_questions_after_answer')
         if suggested_questions_after_answer_dict:
-            if 'enabled' in suggested_questions_after_answer_dict and suggested_questions_after_answer_dict['enabled']:
+            if suggested_questions_after_answer_dict.get('enabled'):
                 suggested_questions_after_answer = True
 
         return suggested_questions_after_answer

api/core/app/app_config/features/text_to_speech/manager.py

@@ -12,7 +12,7 @@ class TextToSpeechConfigManager:
         text_to_speech = False
         text_to_speech_dict = config.get('text_to_speech')
         if text_to_speech_dict:
-            if 'enabled' in text_to_speech_dict and text_to_speech_dict['enabled']:
+            if text_to_speech_dict.get('enabled'):
                 text_to_speech = TextToSpeechEntity(
                     enabled=text_to_speech_dict.get('enabled'),
                     voice=text_to_speech_dict.get('voice'),

api/core/app/apps/advanced_chat/app_generator.py

@@ -66,7 +66,7 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
             conversation = self._get_conversation_by_user(app_model, args.get('conversation_id'), user)
 
         # parse files
-        files = args['files'] if 'files' in args and args['files'] else []
+        files = args['files'] if args.get('files') else []
         message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
         file_extra_config = FileUploadConfigManager.convert(workflow.features_dict, is_vision=False)
         if file_extra_config:
@@ -98,6 +98,90 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
             extras=extras
         )
 
+        return self._generate(
+            app_model=app_model,
+            workflow=workflow,
+            user=user,
+            invoke_from=invoke_from,
+            application_generate_entity=application_generate_entity,
+            conversation=conversation,
+            stream=stream
+        )
+    
+    def single_iteration_generate(self, app_model: App,
+                                  workflow: Workflow,
+                                  node_id: str,
+                                  user: Account,
+                                  args: dict,
+                                  stream: bool = True) \
+            -> Union[dict, Generator[dict, None, None]]:
+        """
+        Generate App response.
+
+        :param app_model: App
+        :param workflow: Workflow
+        :param user: account or end user
+        :param args: request args
+        :param invoke_from: invoke from source
+        :param stream: is stream
+        """
+        if not node_id:
+            raise ValueError('node_id is required')
+        
+        if args.get('inputs') is None:
+            raise ValueError('inputs is required')
+        
+        extras = {
+            "auto_generate_conversation_name": False
+        }
+
+        # get conversation
+        conversation = None
+        if args.get('conversation_id'):
+            conversation = self._get_conversation_by_user(app_model, args.get('conversation_id'), user)
+
+        # convert to app config
+        app_config = AdvancedChatAppConfigManager.get_app_config(
+            app_model=app_model,
+            workflow=workflow
+        )
+
+        # init application generate entity
+        application_generate_entity = AdvancedChatAppGenerateEntity(
+            task_id=str(uuid.uuid4()),
+            app_config=app_config,
+            conversation_id=conversation.id if conversation else None,
+            inputs={},
+            query='',
+            files=[],
+            user_id=user.id,
+            stream=stream,
+            invoke_from=InvokeFrom.DEBUGGER,
+            extras=extras,
+            single_iteration_run=AdvancedChatAppGenerateEntity.SingleIterationRunEntity(
+                node_id=node_id,
+                inputs=args['inputs']
+            )
+        )
+
+        return self._generate(
+            app_model=app_model,
+            workflow=workflow,
+            user=user,
+            invoke_from=InvokeFrom.DEBUGGER,
+            application_generate_entity=application_generate_entity,
+            conversation=conversation,
+            stream=stream
+        )
+
+    def _generate(self, app_model: App,
+                 workflow: Workflow,
+                 user: Union[Account, EndUser],
+                 invoke_from: InvokeFrom,
+                 application_generate_entity: AdvancedChatAppGenerateEntity,
+                 conversation: Conversation = None,
+                 stream: bool = True) \
+            -> Union[dict, Generator[dict, None, None]]:
         is_first_conversation = False
         if not conversation:
             is_first_conversation = True
@@ -167,18 +251,30 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
         """
         with flask_app.app_context():
             try:
-                # get conversation and message
-                conversation = self._get_conversation(conversation_id)
-                message = self._get_message(message_id)
-
-                # chatbot app
                 runner = AdvancedChatAppRunner()
-                runner.run(
-                    application_generate_entity=application_generate_entity,
-                    queue_manager=queue_manager,
-                    conversation=conversation,
-                    message=message
-                )
+                if application_generate_entity.single_iteration_run:
+                    single_iteration_run = application_generate_entity.single_iteration_run
+                    runner.single_iteration_run(
+                        app_id=application_generate_entity.app_config.app_id,
+                        workflow_id=application_generate_entity.app_config.workflow_id,
+                        queue_manager=queue_manager,
+                        inputs=single_iteration_run.inputs,
+                        node_id=single_iteration_run.node_id,
+                        user_id=application_generate_entity.user_id
+                    )
+                else:
+                    # get conversation and message
+                    conversation = self._get_conversation(conversation_id)
+                    message = self._get_message(message_id)
+
+                    # chatbot app
+                    runner = AdvancedChatAppRunner()
+                    runner.run(
+                        application_generate_entity=application_generate_entity,
+                        queue_manager=queue_manager,
+                        conversation=conversation,
+                        message=message
+                    )
             except GenerateTaskStoppedException:
                 pass
             except InvokeAuthorizationError:

api/core/app/apps/advanced_chat/app_runner.py

@@ -102,6 +102,7 @@ class AdvancedChatAppRunner(AppRunner):
             user_from=UserFrom.ACCOUNT
             if application_generate_entity.invoke_from in [InvokeFrom.EXPLORE, InvokeFrom.DEBUGGER]
             else UserFrom.END_USER,
+            invoke_from=application_generate_entity.invoke_from,
             user_inputs=inputs,
             system_inputs={
                 SystemVariable.QUERY: query,
@@ -109,6 +110,35 @@ class AdvancedChatAppRunner(AppRunner):
                 SystemVariable.CONVERSATION_ID: conversation.id,
                 SystemVariable.USER_ID: user_id
             },
+            callbacks=workflow_callbacks,
+            call_depth=application_generate_entity.call_depth
+        )
+
+    def single_iteration_run(self, app_id: str, workflow_id: str,
+                             queue_manager: AppQueueManager,
+                             inputs: dict, node_id: str, user_id: str) -> None:
+        """
+        Single iteration run
+        """
+        app_record: App = db.session.query(App).filter(App.id == app_id).first()
+        if not app_record:
+            raise ValueError("App not found")
+        
+        workflow = self.get_workflow(app_model=app_record, workflow_id=workflow_id)
+        if not workflow:
+            raise ValueError("Workflow not initialized")
+        
+        workflow_callbacks = [WorkflowEventTriggerCallback(
+            queue_manager=queue_manager,
+            workflow=workflow
+        )]
+
+        workflow_engine_manager = WorkflowEngineManager()
+        workflow_engine_manager.single_step_run_iteration_workflow_node(
+            workflow=workflow,
+            node_id=node_id,
+            user_id=user_id,
+            user_inputs=inputs,
             callbacks=workflow_callbacks
         )
 

api/core/app/apps/advanced_chat/generate_task_pipeline.py

@@ -12,6 +12,9 @@ from core.app.entities.queue_entities import (
     QueueAdvancedChatMessageEndEvent,
     QueueAnnotationReplyEvent,
     QueueErrorEvent,
+    QueueIterationCompletedEvent,
+    QueueIterationNextEvent,
+    QueueIterationStartEvent,
     QueueMessageReplaceEvent,
     QueueNodeFailedEvent,
     QueueNodeStartedEvent,
@@ -64,6 +67,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
     _workflow: Workflow
     _user: Union[Account, EndUser]
     _workflow_system_variables: dict[SystemVariable, Any]
+    _iteration_nested_relations: dict[str, list[str]]
 
     def __init__(self, application_generate_entity: AdvancedChatAppGenerateEntity,
                  workflow: Workflow,
@@ -104,6 +108,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
         )
 
         self._stream_generate_routes = self._get_stream_generate_routes()
+        self._iteration_nested_relations = self._get_iteration_nested_relations(self._workflow.graph_dict)
         self._conversation_name_generate_thread = None
 
     def process(self) -> Union[ChatbotAppBlockingResponse, Generator[ChatbotAppStreamResponse, None, None]]:
@@ -204,6 +209,8 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                 # search stream_generate_routes if node id is answer start at node
                 if not self._task_state.current_stream_generate_state and event.node_id in self._stream_generate_routes:
                     self._task_state.current_stream_generate_state = self._stream_generate_routes[event.node_id]
+                    # reset current route position to 0
+                    self._task_state.current_stream_generate_state.current_route_position = 0
 
                     # generate stream outputs when node started
                     yield from self._generate_stream_outputs_when_node_started()
@@ -225,6 +232,22 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                     task_id=self._application_generate_entity.task_id,
                     workflow_node_execution=workflow_node_execution
                 )
+
+                if isinstance(event, QueueNodeFailedEvent):
+                    yield from self._handle_iteration_exception(
+                        task_id=self._application_generate_entity.task_id,
+                        error=f'Child node failed: {event.error}'
+                    )
+            elif isinstance(event, QueueIterationStartEvent | QueueIterationNextEvent | QueueIterationCompletedEvent):
+                if isinstance(event, QueueIterationNextEvent):
+                    # clear ran node execution infos of current iteration
+                    iteration_relations = self._iteration_nested_relations.get(event.node_id)
+                    if iteration_relations:
+                        for node_id in iteration_relations:
+                            self._task_state.ran_node_execution_infos.pop(node_id, None)
+
+                yield self._handle_iteration_to_stream_response(self._application_generate_entity.task_id, event)
+                self._handle_iteration_operation(event)
             elif isinstance(event, QueueStopEvent | QueueWorkflowSucceededEvent | QueueWorkflowFailedEvent):
                 workflow_run = self._handle_workflow_finished(event)
                 if workflow_run:
@@ -263,10 +286,6 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                 self._handle_retriever_resources(event)
             elif isinstance(event, QueueAnnotationReplyEvent):
                 self._handle_annotation_reply(event)
-            # elif isinstance(event, QueueMessageFileEvent):
-            #     response = self._message_file_to_stream_response(event)
-            #     if response:
-            #         yield response
             elif isinstance(event, QueueTextChunkEvent):
                 delta_text = event.text
                 if delta_text is None:
@@ -342,7 +361,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
             id=self._message.id,
             **extras
         )
-
+    
     def _get_stream_generate_routes(self) -> dict[str, ChatflowStreamGenerateRoute]:
         """
         Get stream generate routes.
@@ -372,7 +391,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                 )
 
         return stream_generate_routes
-
+    
     def _get_answer_start_at_node_ids(self, graph: dict, target_node_id: str) \
             -> list[str]:
         """
@@ -401,14 +420,23 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                 continue
 
             node_type = source_node.get('data', {}).get('type')
+            node_iteration_id = source_node.get('data', {}).get('iteration_id')
+            iteration_start_node_id = None
+            if node_iteration_id:
+                iteration_node = next((node for node in nodes if node.get('id') == node_iteration_id), None)
+                iteration_start_node_id = iteration_node.get('data', {}).get('start_node_id')
+
             if node_type in [
                 NodeType.ANSWER.value,
                 NodeType.IF_ELSE.value,
-                NodeType.QUESTION_CLASSIFIER.value
+                NodeType.QUESTION_CLASSIFIER.value,
+                NodeType.ITERATION.value,
+                NodeType.LOOP.value
             ]:
                 start_node_id = target_node_id
                 start_node_ids.append(start_node_id)
-            elif node_type == NodeType.START.value:
+            elif node_type == NodeType.START.value or \
+                node_iteration_id is not None and iteration_start_node_id == source_node.get('id'):
                 start_node_id = source_node_id
                 start_node_ids.append(start_node_id)
             else:
@@ -417,7 +445,27 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                     start_node_ids.extend(sub_start_node_ids)
 
         return start_node_ids
+    
+    def _get_iteration_nested_relations(self, graph: dict) -> dict[str, list[str]]:
+        """
+        Get iteration nested relations.
+        :param graph: graph
+        :return:
+        """
+        nodes = graph.get('nodes')
 
+        iteration_ids = [node.get('id') for node in nodes 
+                         if node.get('data', {}).get('type') in [
+                             NodeType.ITERATION.value,
+                             NodeType.LOOP.value,
+                        ]]
+
+        return {
+            iteration_id: [
+                node.get('id') for node in nodes if node.get('data', {}).get('iteration_id') == iteration_id
+            ] for iteration_id in iteration_ids
+        }
+    
     def _generate_stream_outputs_when_node_started(self) -> Generator:
         """
         Generate stream outputs.
@@ -425,7 +473,8 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
         """
         if self._task_state.current_stream_generate_state:
             route_chunks = self._task_state.current_stream_generate_state.generate_route[
-                           self._task_state.current_stream_generate_state.current_route_position:]
+                self._task_state.current_stream_generate_state.current_route_position:
+            ]
 
             for route_chunk in route_chunks:
                 if route_chunk.type == 'text':
@@ -458,7 +507,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
 
         route_chunks = self._task_state.current_stream_generate_state.generate_route[
                        self._task_state.current_stream_generate_state.current_route_position:]
-
+        
         for route_chunk in route_chunks:
             if route_chunk.type == 'text':
                 route_chunk = cast(TextGenerateRouteChunk, route_chunk)

api/core/app/apps/advanced_chat/workflow_event_trigger_callback.py

@@ -1,8 +1,11 @@
-from typing import Optional
+from typing import Any, Optional
 
 from core.app.apps.base_app_queue_manager import AppQueueManager, PublishFrom
 from core.app.entities.queue_entities import (
     AppQueueEvent,
+    QueueIterationCompletedEvent,
+    QueueIterationNextEvent,
+    QueueIterationStartEvent,
     QueueNodeFailedEvent,
     QueueNodeStartedEvent,
     QueueNodeSucceededEvent,
@@ -130,6 +133,66 @@ class WorkflowEventTriggerCallback(BaseWorkflowCallback):
             ), PublishFrom.APPLICATION_MANAGER
         )
 
+    def on_workflow_iteration_started(self, 
+                                      node_id: str,
+                                      node_type: NodeType,
+                                      node_run_index: int = 1,
+                                      node_data: Optional[BaseNodeData] = None,
+                                      inputs: dict = None,
+                                      predecessor_node_id: Optional[str] = None,
+                                      metadata: Optional[dict] = None) -> None:
+        """
+        Publish iteration started
+        """
+        self._queue_manager.publish(
+            QueueIterationStartEvent(
+                node_id=node_id,
+                node_type=node_type,
+                node_run_index=node_run_index,
+                node_data=node_data,
+                inputs=inputs,
+                predecessor_node_id=predecessor_node_id,
+                metadata=metadata
+            ),
+            PublishFrom.APPLICATION_MANAGER
+        )
+
+    def on_workflow_iteration_next(self, node_id: str, 
+                                   node_type: NodeType,
+                                   index: int, 
+                                   node_run_index: int,
+                                   output: Optional[Any]) -> None:
+        """
+        Publish iteration next
+        """
+        self._queue_manager._publish(
+            QueueIterationNextEvent(
+                node_id=node_id,
+                node_type=node_type,
+                index=index,
+                node_run_index=node_run_index,
+                output=output
+            ),
+            PublishFrom.APPLICATION_MANAGER
+        )
+
+    def on_workflow_iteration_completed(self, node_id: str, 
+                                        node_type: NodeType,
+                                        node_run_index: int,
+                                        outputs: dict) -> None:
+        """
+        Publish iteration completed
+        """
+        self._queue_manager._publish(
+            QueueIterationCompletedEvent(
+                node_id=node_id,
+                node_type=node_type,
+                node_run_index=node_run_index,
+                outputs=outputs
+            ),
+            PublishFrom.APPLICATION_MANAGER
+        )
+
     def on_event(self, event: AppQueueEvent) -> None:
         """
         Publish event

api/core/app/apps/agent_chat/app_generator.py

@@ -83,7 +83,7 @@ class AgentChatAppGenerator(MessageBasedAppGenerator):
             )
 
         # parse files
-        files = args['files'] if 'files' in args and args['files'] else []
+        files = args['files'] if args.get('files') else []
         message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
         file_extra_config = FileUploadConfigManager.convert(override_model_config_dict or app_model_config.to_dict())
         if file_extra_config:
@@ -115,7 +115,8 @@ class AgentChatAppGenerator(MessageBasedAppGenerator):
             user_id=user.id,
             stream=stream,
             invoke_from=invoke_from,
-            extras=extras
+            extras=extras,
+            call_depth=0
         )
 
         # init generate records

api/core/app/apps/base_app_runner.py

@@ -1,6 +1,6 @@
 import time
 from collections.abc import Generator
-from typing import Optional, Union, cast
+from typing import Optional, Union
 
 from core.app.app_config.entities import ExternalDataVariableEntity, PromptTemplateEntity
 from core.app.apps.base_app_queue_manager import AppQueueManager, PublishFrom
@@ -16,11 +16,11 @@ from core.app.features.hosting_moderation.hosting_moderation import HostingModer
 from core.external_data_tool.external_data_fetch import ExternalDataFetch
 from core.file.file_obj import FileVar
 from core.memory.token_buffer_memory import TokenBufferMemory
+from core.model_manager import ModelInstance
 from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk, LLMResultChunkDelta, LLMUsage
 from core.model_runtime.entities.message_entities import AssistantPromptMessage, PromptMessage
 from core.model_runtime.entities.model_entities import ModelPropertyKey
 from core.model_runtime.errors.invoke import InvokeBadRequestError
-from core.model_runtime.model_providers.__base.large_language_model import LargeLanguageModel
 from core.moderation.input_moderation import InputModeration
 from core.prompt.advanced_prompt_transform import AdvancedPromptTransform
 from core.prompt.entities.advanced_prompt_entities import ChatModelMessage, CompletionModelPromptTemplate, MemoryConfig
@@ -45,8 +45,11 @@ class AppRunner:
         :param query: query
         :return:
         """
-        model_type_instance = model_config.provider_model_bundle.model_type_instance
-        model_type_instance = cast(LargeLanguageModel, model_type_instance)
+        # Invoke model
+        model_instance = ModelInstance(
+            provider_model_bundle=model_config.provider_model_bundle,
+            model=model_config.model
+        )
 
         model_context_tokens = model_config.model_schema.model_properties.get(ModelPropertyKey.CONTEXT_SIZE)
 
@@ -73,9 +76,7 @@ class AppRunner:
             query=query
         )
 
-        prompt_tokens = model_type_instance.get_num_tokens(
-            model_config.model,
-            model_config.credentials,
+        prompt_tokens = model_instance.get_llm_num_tokens(
             prompt_messages
         )
 
@@ -89,8 +90,10 @@ class AppRunner:
     def recalc_llm_max_tokens(self, model_config: ModelConfigWithCredentialsEntity,
                               prompt_messages: list[PromptMessage]):
         # recalc max_tokens if sum(prompt_token +  max_tokens) over model token limit
-        model_type_instance = model_config.provider_model_bundle.model_type_instance
-        model_type_instance = cast(LargeLanguageModel, model_type_instance)
+        model_instance = ModelInstance(
+            provider_model_bundle=model_config.provider_model_bundle,
+            model=model_config.model
+        )
 
         model_context_tokens = model_config.model_schema.model_properties.get(ModelPropertyKey.CONTEXT_SIZE)
 
@@ -107,9 +110,7 @@ class AppRunner:
         if max_tokens is None:
             max_tokens = 0
 
-        prompt_tokens = model_type_instance.get_num_tokens(
-            model_config.model,
-            model_config.credentials,
+        prompt_tokens = model_instance.get_llm_num_tokens(
             prompt_messages
         )
 

api/core/app/apps/chat/app_generator.py

@@ -80,7 +80,7 @@ class ChatAppGenerator(MessageBasedAppGenerator):
             )
 
         # parse files
-        files = args['files'] if 'files' in args and args['files'] else []
+        files = args['files'] if args.get('files') else []
         message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
         file_extra_config = FileUploadConfigManager.convert(override_model_config_dict or app_model_config.to_dict())
         if file_extra_config:

api/core/app/apps/completion/app_generator.py

@@ -75,7 +75,7 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
             )
 
         # parse files
-        files = args['files'] if 'files' in args and args['files'] else []
+        files = args['files'] if args.get('files') else []
         message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
         file_extra_config = FileUploadConfigManager.convert(override_model_config_dict or app_model_config.to_dict())
         if file_extra_config:

api/core/app/apps/workflow/app_generator.py

@@ -34,7 +34,8 @@ class WorkflowAppGenerator(BaseAppGenerator):
                  user: Union[Account, EndUser],
                  args: dict,
                  invoke_from: InvokeFrom,
-                 stream: bool = True) \
+                 stream: bool = True,
+                 call_depth: int = 0) \
             -> Union[dict, Generator[dict, None, None]]:
         """
         Generate App response.
@@ -49,7 +50,7 @@ class WorkflowAppGenerator(BaseAppGenerator):
         inputs = args['inputs']
 
         # parse files
-        files = args['files'] if 'files' in args and args['files'] else []
+        files = args['files'] if args.get('files') else []
         message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
         file_extra_config = FileUploadConfigManager.convert(workflow.features_dict, is_vision=False)
         if file_extra_config:
@@ -75,9 +76,38 @@ class WorkflowAppGenerator(BaseAppGenerator):
             files=file_objs,
             user_id=user.id,
             stream=stream,
-            invoke_from=invoke_from
+            invoke_from=invoke_from,
+            call_depth=call_depth
         )
 
+        return self._generate(
+            app_model=app_model,
+            workflow=workflow,
+            user=user,
+            application_generate_entity=application_generate_entity,
+            invoke_from=invoke_from,
+            stream=stream,
+            call_depth=call_depth
+        )
+
+    def _generate(self, app_model: App,
+                 workflow: Workflow,
+                 user: Union[Account, EndUser],
+                 application_generate_entity: WorkflowAppGenerateEntity,
+                 invoke_from: InvokeFrom,
+                 stream: bool = True,
+                 call_depth: int = 0) \
+            -> Union[dict, Generator[dict, None, None]]:
+        """
+        Generate App response.
+
+        :param app_model: App
+        :param workflow: Workflow
+        :param user: account or end user
+        :param application_generate_entity: application generate entity
+        :param invoke_from: invoke from source
+        :param stream: is stream
+        """
         # init queue manager
         queue_manager = WorkflowAppQueueManager(
             task_id=application_generate_entity.task_id,
@@ -109,6 +139,64 @@ class WorkflowAppGenerator(BaseAppGenerator):
             invoke_from=invoke_from
         )
 
+    def single_iteration_generate(self, app_model: App,
+                                  workflow: Workflow,
+                                  node_id: str,
+                                  user: Account,
+                                  args: dict,
+                                  stream: bool = True) \
+            -> Union[dict, Generator[dict, None, None]]:
+        """
+        Generate App response.
+
+        :param app_model: App
+        :param workflow: Workflow
+        :param user: account or end user
+        :param args: request args
+        :param invoke_from: invoke from source
+        :param stream: is stream
+        """
+        if not node_id:
+            raise ValueError('node_id is required')
+        
+        if args.get('inputs') is None:
+            raise ValueError('inputs is required')
+        
+        extras = {
+            "auto_generate_conversation_name": False
+        }
+
+        # convert to app config
+        app_config = WorkflowAppConfigManager.get_app_config(
+            app_model=app_model,
+            workflow=workflow
+        )
+
+        # init application generate entity
+        application_generate_entity = WorkflowAppGenerateEntity(
+            task_id=str(uuid.uuid4()),
+            app_config=app_config,
+            inputs={},
+            files=[],
+            user_id=user.id,
+            stream=stream,
+            invoke_from=InvokeFrom.DEBUGGER,
+            extras=extras,
+            single_iteration_run=WorkflowAppGenerateEntity.SingleIterationRunEntity(
+                node_id=node_id,
+                inputs=args['inputs']
+            )
+        )
+
+        return self._generate(
+            app_model=app_model,
+            workflow=workflow,
+            user=user,
+            invoke_from=InvokeFrom.DEBUGGER,
+            application_generate_entity=application_generate_entity,
+            stream=stream
+        )
+
     def _generate_worker(self, flask_app: Flask,
                          application_generate_entity: WorkflowAppGenerateEntity,
                          queue_manager: AppQueueManager) -> None:
@@ -123,10 +211,21 @@ class WorkflowAppGenerator(BaseAppGenerator):
             try:
                 # workflow app
                 runner = WorkflowAppRunner()
-                runner.run(
-                    application_generate_entity=application_generate_entity,
-                    queue_manager=queue_manager
-                )
+                if application_generate_entity.single_iteration_run:
+                    single_iteration_run = application_generate_entity.single_iteration_run
+                    runner.single_iteration_run(
+                        app_id=application_generate_entity.app_config.app_id,
+                        workflow_id=application_generate_entity.app_config.workflow_id,
+                        queue_manager=queue_manager,
+                        inputs=single_iteration_run.inputs,
+                        node_id=single_iteration_run.node_id,
+                        user_id=application_generate_entity.user_id
+                    )
+                else:
+                    runner.run(
+                        application_generate_entity=application_generate_entity,
+                        queue_manager=queue_manager
+                    )
             except GenerateTaskStoppedException:
                 pass
             except InvokeAuthorizationError:

api/core/app/apps/workflow/app_runner.py

@@ -73,11 +73,44 @@ class WorkflowAppRunner:
             user_from=UserFrom.ACCOUNT
             if application_generate_entity.invoke_from in [InvokeFrom.EXPLORE, InvokeFrom.DEBUGGER]
             else UserFrom.END_USER,
+            invoke_from=application_generate_entity.invoke_from,
             user_inputs=inputs,
             system_inputs={
                 SystemVariable.FILES: files,
                 SystemVariable.USER_ID: user_id
             },
+            callbacks=workflow_callbacks,
+            call_depth=application_generate_entity.call_depth
+        )
+
+    def single_iteration_run(self, app_id: str, workflow_id: str,
+                             queue_manager: AppQueueManager,
+                             inputs: dict, node_id: str, user_id: str) -> None:
+        """
+        Single iteration run
+        """
+        app_record: App = db.session.query(App).filter(App.id == app_id).first()
+        if not app_record:
+            raise ValueError("App not found")
+        
+        if not app_record.workflow_id:
+            raise ValueError("Workflow not initialized")
+
+        workflow = self.get_workflow(app_model=app_record, workflow_id=workflow_id)
+        if not workflow:
+            raise ValueError("Workflow not initialized")
+        
+        workflow_callbacks = [WorkflowEventTriggerCallback(
+            queue_manager=queue_manager,
+            workflow=workflow
+        )]
+
+        workflow_engine_manager = WorkflowEngineManager()
+        workflow_engine_manager.single_step_run_iteration_workflow_node(
+            workflow=workflow,
+            node_id=node_id,
+            user_id=user_id,
+            user_inputs=inputs,
             callbacks=workflow_callbacks
         )
 

api/core/app/apps/workflow/generate_task_pipeline.py

@@ -9,6 +9,9 @@ from core.app.entities.app_invoke_entities import (
 )
 from core.app.entities.queue_entities import (
     QueueErrorEvent,
+    QueueIterationCompletedEvent,
+    QueueIterationNextEvent,
+    QueueIterationStartEvent,
     QueueMessageReplaceEvent,
     QueueNodeFailedEvent,
     QueueNodeStartedEvent,
@@ -58,6 +61,7 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
     _task_state: WorkflowTaskState
     _application_generate_entity: WorkflowAppGenerateEntity
     _workflow_system_variables: dict[SystemVariable, Any]
+    _iteration_nested_relations: dict[str, list[str]]
 
     def __init__(self, application_generate_entity: WorkflowAppGenerateEntity,
                  workflow: Workflow,
@@ -85,8 +89,11 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
             SystemVariable.USER_ID: user_id
         }
 
-        self._task_state = WorkflowTaskState()
+        self._task_state = WorkflowTaskState(
+            iteration_nested_node_ids=[]
+        )
         self._stream_generate_nodes = self._get_stream_generate_nodes()
+        self._iteration_nested_relations = self._get_iteration_nested_relations(self._workflow.graph_dict)
 
     def process(self) -> Union[WorkflowAppBlockingResponse, Generator[WorkflowAppStreamResponse, None, None]]:
         """
@@ -191,6 +198,22 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
                     task_id=self._application_generate_entity.task_id,
                     workflow_node_execution=workflow_node_execution
                 )
+
+                if isinstance(event, QueueNodeFailedEvent):
+                    yield from self._handle_iteration_exception(
+                        task_id=self._application_generate_entity.task_id,
+                        error=f'Child node failed: {event.error}'
+                    )
+            elif isinstance(event, QueueIterationStartEvent | QueueIterationNextEvent | QueueIterationCompletedEvent):
+                if isinstance(event, QueueIterationNextEvent):
+                    # clear ran node execution infos of current iteration
+                    iteration_relations = self._iteration_nested_relations.get(event.node_id)
+                    if iteration_relations:
+                        for node_id in iteration_relations:
+                            self._task_state.ran_node_execution_infos.pop(node_id, None)
+
+                yield self._handle_iteration_to_stream_response(self._application_generate_entity.task_id, event)
+                self._handle_iteration_operation(event)
             elif isinstance(event, QueueStopEvent | QueueWorkflowSucceededEvent | QueueWorkflowFailedEvent):
                 workflow_run = self._handle_workflow_finished(event)
 
@@ -331,13 +354,20 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
                 continue
 
             node_type = source_node.get('data', {}).get('type')
+            node_iteration_id = source_node.get('data', {}).get('iteration_id')
+            iteration_start_node_id = None
+            if node_iteration_id:
+                iteration_node = next((node for node in nodes if node.get('id') == node_iteration_id), None)
+                iteration_start_node_id = iteration_node.get('data', {}).get('start_node_id')
+
             if node_type in [
                 NodeType.IF_ELSE.value,
                 NodeType.QUESTION_CLASSIFIER.value
             ]:
                 start_node_id = target_node_id
                 start_node_ids.append(start_node_id)
-            elif node_type == NodeType.START.value:
+            elif node_type == NodeType.START.value or \
+                node_iteration_id is not None and iteration_start_node_id == source_node.get('id'):
                 start_node_id = source_node_id
                 start_node_ids.append(start_node_id)
             else:
@@ -411,3 +441,24 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
             return False
 
         return True
+
+    def _get_iteration_nested_relations(self, graph: dict) -> dict[str, list[str]]:
+        """
+        Get iteration nested relations.
+        :param graph: graph
+        :return:
+        """
+        nodes = graph.get('nodes')
+
+        iteration_ids = [node.get('id') for node in nodes 
+                         if node.get('data', {}).get('type') in [
+                             NodeType.ITERATION.value,
+                             NodeType.LOOP.value,
+                        ]]
+
+        return {
+            iteration_id: [
+                node.get('id') for node in nodes if node.get('data', {}).get('iteration_id') == iteration_id
+            ] for iteration_id in iteration_ids
+        }
+    

api/core/app/apps/workflow/workflow_event_trigger_callback.py

@@ -1,8 +1,11 @@
-from typing import Optional
+from typing import Any, Optional
 
 from core.app.apps.base_app_queue_manager import AppQueueManager, PublishFrom
 from core.app.entities.queue_entities import (
     AppQueueEvent,
+    QueueIterationCompletedEvent,
+    QueueIterationNextEvent,
+    QueueIterationStartEvent,
     QueueNodeFailedEvent,
     QueueNodeStartedEvent,
     QueueNodeSucceededEvent,
@@ -130,6 +133,66 @@ class WorkflowEventTriggerCallback(BaseWorkflowCallback):
             ), PublishFrom.APPLICATION_MANAGER
         )
 
+    def on_workflow_iteration_started(self, 
+                                      node_id: str,
+                                      node_type: NodeType,
+                                      node_run_index: int = 1,
+                                      node_data: Optional[BaseNodeData] = None,
+                                      inputs: dict = None,
+                                      predecessor_node_id: Optional[str] = None,
+                                      metadata: Optional[dict] = None) -> None:
+        """
+        Publish iteration started
+        """
+        self._queue_manager.publish(
+            QueueIterationStartEvent(
+                node_id=node_id,
+                node_type=node_type,
+                node_run_index=node_run_index,
+                node_data=node_data,
+                inputs=inputs,
+                predecessor_node_id=predecessor_node_id,
+                metadata=metadata
+            ),
+            PublishFrom.APPLICATION_MANAGER
+        )
+
+    def on_workflow_iteration_next(self, node_id: str, 
+                                   node_type: NodeType,
+                                   index: int, 
+                                   node_run_index: int,
+                                   output: Optional[Any]) -> None:
+        """
+        Publish iteration next
+        """
+        self._queue_manager.publish(
+            QueueIterationNextEvent(
+                node_id=node_id,
+                node_type=node_type,
+                index=index,
+                node_run_index=node_run_index,
+                output=output
+            ),
+            PublishFrom.APPLICATION_MANAGER
+        )
+
+    def on_workflow_iteration_completed(self, node_id: str, 
+                                        node_type: NodeType,
+                                        node_run_index: int,
+                                        outputs: dict) -> None:
+        """
+        Publish iteration completed
+        """
+        self._queue_manager.publish(
+            QueueIterationCompletedEvent(
+                node_id=node_id,
+                node_type=node_type,
+                node_run_index=node_run_index,
+                outputs=outputs
+            ),
+            PublishFrom.APPLICATION_MANAGER
+        )
+        
     def on_event(self, event: AppQueueEvent) -> None:
         """
         Publish event

api/core/app/apps/workflow_logging_callback.py

@@ -102,6 +102,39 @@ class WorkflowLoggingCallback(BaseWorkflowCallback):
 
         self.print_text(text, color="pink", end="")
 
+    def on_workflow_iteration_started(self, 
+                                      node_id: str,
+                                      node_type: NodeType,
+                                      node_run_index: int = 1,
+                                      node_data: Optional[BaseNodeData] = None,
+                                      inputs: dict = None,
+                                      predecessor_node_id: Optional[str] = None,
+                                      metadata: Optional[dict] = None) -> None:
+        """
+        Publish iteration started
+        """
+        self.print_text("\n[on_workflow_iteration_started]", color='blue')
+        self.print_text(f"Node ID: {node_id}", color='blue')
+
+    def on_workflow_iteration_next(self, node_id: str, 
+                                   node_type: NodeType,
+                                   index: int, 
+                                   node_run_index: int,
+                                   output: Optional[dict]) -> None:
+        """
+        Publish iteration next
+        """
+        self.print_text("\n[on_workflow_iteration_next]", color='blue')
+
+    def on_workflow_iteration_completed(self, node_id: str, 
+                                        node_type: NodeType,
+                                        node_run_index: int,
+                                        outputs: dict) -> None:
+        """
+        Publish iteration completed
+        """
+        self.print_text("\n[on_workflow_iteration_completed]", color='blue')
+
     def on_event(self, event: AppQueueEvent) -> None:
         """
         Publish event

api/core/app/entities/app_invoke_entities.py

@@ -80,6 +80,9 @@ class AppGenerateEntity(BaseModel):
     stream: bool
     invoke_from: InvokeFrom
 
+    # invoke call depth
+    call_depth: int = 0
+
     # extra parameters, like: auto_generate_conversation_name
     extras: dict[str, Any] = {}
 
@@ -126,6 +129,14 @@ class AdvancedChatAppGenerateEntity(AppGenerateEntity):
     conversation_id: Optional[str] = None
     query: Optional[str] = None
 
+    class SingleIterationRunEntity(BaseModel):
+        """
+        Single Iteration Run Entity.
+        """
+        node_id: str
+        inputs: dict
+
+    single_iteration_run: Optional[SingleIterationRunEntity] = None
 
 class WorkflowAppGenerateEntity(AppGenerateEntity):
     """
@@ -133,3 +144,12 @@ class WorkflowAppGenerateEntity(AppGenerateEntity):
     """
     # app config
     app_config: WorkflowUIBasedAppConfig
+
+    class SingleIterationRunEntity(BaseModel):
+        """
+        Single Iteration Run Entity.
+        """
+        node_id: str
+        inputs: dict
+
+    single_iteration_run: Optional[SingleIterationRunEntity] = None

api/core/app/entities/queue_entities.py

@@ -1,7 +1,7 @@
 from enum import Enum
 from typing import Any, Optional
 
-from pydantic import BaseModel
+from pydantic import BaseModel, validator
 
 from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk
 from core.workflow.entities.base_node_data_entities import BaseNodeData
@@ -21,6 +21,9 @@ class QueueEvent(Enum):
     WORKFLOW_STARTED = "workflow_started"
     WORKFLOW_SUCCEEDED = "workflow_succeeded"
     WORKFLOW_FAILED = "workflow_failed"
+    ITERATION_START = "iteration_start"
+    ITERATION_NEXT = "iteration_next"
+    ITERATION_COMPLETED = "iteration_completed"
     NODE_STARTED = "node_started"
     NODE_SUCCEEDED = "node_succeeded"
     NODE_FAILED = "node_failed"
@@ -47,6 +50,55 @@ class QueueLLMChunkEvent(AppQueueEvent):
     event = QueueEvent.LLM_CHUNK
     chunk: LLMResultChunk
 
+class QueueIterationStartEvent(AppQueueEvent):
+    """
+    QueueIterationStartEvent entity
+    """
+    event = QueueEvent.ITERATION_START
+    node_id: str
+    node_type: NodeType
+    node_data: BaseNodeData
+
+    node_run_index: int
+    inputs: dict = None
+    predecessor_node_id: Optional[str] = None
+    metadata: Optional[dict] = None
+
+class QueueIterationNextEvent(AppQueueEvent):
+    """
+    QueueIterationNextEvent entity
+    """
+    event = QueueEvent.ITERATION_NEXT
+
+    index: int
+    node_id: str
+    node_type: NodeType
+
+    node_run_index: int
+    output: Optional[Any] # output for the current iteration
+
+    @validator('output', pre=True, always=True)
+    def set_output(cls, v):
+        """
+        Set output
+        """
+        if v is None:
+            return None
+        if isinstance(v, int | float | str | bool | dict | list):
+            return v
+        raise ValueError('output must be a valid type')
+
+class QueueIterationCompletedEvent(AppQueueEvent):
+    """
+    QueueIterationCompletedEvent entity
+    """
+    event = QueueEvent.ITERATION_COMPLETED
+
+    node_id: str
+    node_type: NodeType
+    
+    node_run_index: int
+    outputs: dict
 
 class QueueTextChunkEvent(AppQueueEvent):
     """

api/core/app/entities/task_entities.py

@@ -1,12 +1,14 @@
 from enum import Enum
-from typing import Optional
+from typing import Any, Optional
 
 from pydantic import BaseModel
 
 from core.model_runtime.entities.llm_entities import LLMResult, LLMUsage
 from core.model_runtime.utils.encoders import jsonable_encoder
+from core.workflow.entities.base_node_data_entities import BaseNodeData
 from core.workflow.entities.node_entities import NodeType
 from core.workflow.nodes.answer.entities import GenerateRouteChunk
+from models.workflow import WorkflowNodeExecutionStatus
 
 
 class WorkflowStreamGenerateNodes(BaseModel):
@@ -65,6 +67,7 @@ class WorkflowTaskState(TaskState):
 
     current_stream_generate_state: Optional[WorkflowStreamGenerateNodes] = None
 
+    iteration_nested_node_ids: list[str] = None
 
 class AdvancedChatTaskState(WorkflowTaskState):
     """
@@ -91,6 +94,9 @@ class StreamEvent(Enum):
     WORKFLOW_FINISHED = "workflow_finished"
     NODE_STARTED = "node_started"
     NODE_FINISHED = "node_finished"
+    ITERATION_STARTED = "iteration_started"
+    ITERATION_NEXT = "iteration_next"
+    ITERATION_COMPLETED = "iteration_completed"
     TEXT_CHUNK = "text_chunk"
     TEXT_REPLACE = "text_replace"
 
@@ -319,6 +325,74 @@ class NodeFinishStreamResponse(StreamResponse):
             }
         }
 
+class IterationNodeStartStreamResponse(StreamResponse):
+    """
+    NodeStartStreamResponse entity
+    """
+    class Data(BaseModel):
+        """
+        Data entity
+        """
+        id: str
+        node_id: str
+        node_type: str
+        title: str
+        created_at: int
+        extras: dict = {}
+        metadata: dict = {}
+        inputs: dict = {}
+
+    event: StreamEvent = StreamEvent.ITERATION_STARTED
+    workflow_run_id: str
+    data: Data
+
+class IterationNodeNextStreamResponse(StreamResponse):
+    """
+    NodeStartStreamResponse entity
+    """
+    class Data(BaseModel):
+        """
+        Data entity
+        """
+        id: str
+        node_id: str
+        node_type: str
+        title: str
+        index: int
+        created_at: int
+        pre_iteration_output: Optional[Any]
+        extras: dict = {}
+
+    event: StreamEvent = StreamEvent.ITERATION_NEXT
+    workflow_run_id: str
+    data: Data
+
+class IterationNodeCompletedStreamResponse(StreamResponse):
+    """
+    NodeStartStreamResponse entity
+    """
+    class Data(BaseModel):
+        """
+        Data entity
+        """
+        id: str
+        node_id: str
+        node_type: str
+        title: str
+        outputs: Optional[dict]
+        created_at: int
+        extras: dict = None
+        inputs: dict = None
+        status: WorkflowNodeExecutionStatus
+        error: Optional[str]
+        elapsed_time: float
+        total_tokens: int
+        finished_at: int
+        steps: int
+
+    event: StreamEvent = StreamEvent.ITERATION_COMPLETED
+    workflow_run_id: str
+    data: Data
 
 class TextChunkStreamResponse(StreamResponse):
     """
@@ -454,3 +528,23 @@ class WorkflowAppBlockingResponse(AppBlockingResponse):
 
     workflow_run_id: str
     data: Data
+
+class WorkflowIterationState(BaseModel):
+    """
+    WorkflowIterationState entity
+    """
+    class Data(BaseModel):
+        """
+        Data entity
+        """
+        parent_iteration_id: Optional[str] = None
+        iteration_id: str
+        current_index: int
+        iteration_steps_boundary: list[int] = None
+        node_execution_id: str
+        started_at: float
+        inputs: dict = None
+        total_tokens: int = 0
+        node_data: BaseNodeData
+
+    current_iterations: dict[str, Data] = None

api/core/app/task_pipeline/easy_ui_based_generate_task_pipeline.py

@@ -37,6 +37,7 @@ from core.app.entities.task_entities import (
 )
 from core.app.task_pipeline.based_generate_task_pipeline import BasedGenerateTaskPipeline
 from core.app.task_pipeline.message_cycle_manage import MessageCycleManage
+from core.model_manager import ModelInstance
 from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk, LLMResultChunkDelta, LLMUsage
 from core.model_runtime.entities.message_entities import (
     AssistantPromptMessage,
@@ -317,29 +318,30 @@ class EasyUIBasedGenerateTaskPipeline(BasedGenerateTaskPipeline, MessageCycleMan
         """
         model_config = self._model_config
         model = model_config.model
-        model_type_instance = model_config.provider_model_bundle.model_type_instance
-        model_type_instance = cast(LargeLanguageModel, model_type_instance)
+
+        model_instance = ModelInstance(
+            provider_model_bundle=model_config.provider_model_bundle,
+            model=model_config.model
+        )
 
         # calculate num tokens
         prompt_tokens = 0
         if event.stopped_by != QueueStopEvent.StopBy.ANNOTATION_REPLY:
-            prompt_tokens = model_type_instance.get_num_tokens(
-                model,
-                model_config.credentials,
+            prompt_tokens = model_instance.get_llm_num_tokens(
                 self._task_state.llm_result.prompt_messages
             )
 
         completion_tokens = 0
         if event.stopped_by == QueueStopEvent.StopBy.USER_MANUAL:
-            completion_tokens = model_type_instance.get_num_tokens(
-                model,
-                model_config.credentials,
+            completion_tokens = model_instance.get_llm_num_tokens(
                 [self._task_state.llm_result.message]
             )
 
         credentials = model_config.credentials
 
         # transform usage
+        model_type_instance = model_config.provider_model_bundle.model_type_instance
+        model_type_instance = cast(LargeLanguageModel, model_type_instance)
         self._task_state.llm_result.usage = model_type_instance._calc_response_usage(
             model,
             credentials,

api/core/app/task_pipeline/workflow_cycle_manage.py

@@ -1,9 +1,9 @@
 import json
 import time
 from datetime import datetime, timezone
-from typing import Any, Optional, Union, cast
+from typing import Optional, Union, cast
 
-from core.app.entities.app_invoke_entities import AdvancedChatAppGenerateEntity, InvokeFrom, WorkflowAppGenerateEntity
+from core.app.entities.app_invoke_entities import InvokeFrom
 from core.app.entities.queue_entities import (
     QueueNodeFailedEvent,
     QueueNodeStartedEvent,
@@ -13,18 +13,17 @@ from core.app.entities.queue_entities import (
     QueueWorkflowSucceededEvent,
 )
 from core.app.entities.task_entities import (
-    AdvancedChatTaskState,
     NodeExecutionInfo,
     NodeFinishStreamResponse,
     NodeStartStreamResponse,
     WorkflowFinishStreamResponse,
     WorkflowStartStreamResponse,
-    WorkflowTaskState,
 )
+from core.app.task_pipeline.workflow_iteration_cycle_manage import WorkflowIterationCycleManage
 from core.file.file_obj import FileVar
 from core.model_runtime.utils.encoders import jsonable_encoder
 from core.tools.tool_manager import ToolManager
-from core.workflow.entities.node_entities import NodeRunMetadataKey, NodeType, SystemVariable
+from core.workflow.entities.node_entities import NodeRunMetadataKey, NodeType
 from core.workflow.nodes.tool.entities import ToolNodeData
 from core.workflow.workflow_engine_manager import WorkflowEngineManager
 from extensions.ext_database import db
@@ -42,13 +41,7 @@ from models.workflow import (
 )
 
 
-class WorkflowCycleManage:
-    _application_generate_entity: Union[AdvancedChatAppGenerateEntity, WorkflowAppGenerateEntity]
-    _workflow: Workflow
-    _user: Union[Account, EndUser]
-    _task_state: Union[AdvancedChatTaskState, WorkflowTaskState]
-    _workflow_system_variables: dict[SystemVariable, Any]
-
+class WorkflowCycleManage(WorkflowIterationCycleManage):
     def _init_workflow_run(self, workflow: Workflow,
                            triggered_from: WorkflowRunTriggeredFrom,
                            user: Union[Account, EndUser],
@@ -237,6 +230,7 @@ class WorkflowCycleManage:
                                         inputs: Optional[dict] = None,
                                         process_data: Optional[dict] = None,
                                         outputs: Optional[dict] = None,
+                                        execution_metadata: Optional[dict] = None
                                         ) -> WorkflowNodeExecution:
         """
         Workflow node execution failed
@@ -255,6 +249,8 @@ class WorkflowCycleManage:
         workflow_node_execution.inputs = json.dumps(inputs) if inputs else None
         workflow_node_execution.process_data = json.dumps(process_data) if process_data else None
         workflow_node_execution.outputs = json.dumps(outputs) if outputs else None
+        workflow_node_execution.execution_metadata = json.dumps(jsonable_encoder(execution_metadata)) \
+            if execution_metadata else None
 
         db.session.commit()
         db.session.refresh(workflow_node_execution)
@@ -444,6 +440,23 @@ class WorkflowCycleManage:
         current_node_execution = self._task_state.ran_node_execution_infos[event.node_id]
         workflow_node_execution = db.session.query(WorkflowNodeExecution).filter(
             WorkflowNodeExecution.id == current_node_execution.workflow_node_execution_id).first()
+        
+        execution_metadata = event.execution_metadata if isinstance(event, QueueNodeSucceededEvent) else None
+        
+        if self._iteration_state and self._iteration_state.current_iterations:
+            if not execution_metadata:
+                execution_metadata = {}
+            current_iteration_data = None
+            for iteration_node_id in self._iteration_state.current_iterations:
+                data = self._iteration_state.current_iterations[iteration_node_id]
+                if data.parent_iteration_id == None:
+                    current_iteration_data = data
+                    break
+
+            if current_iteration_data:
+                execution_metadata[NodeRunMetadataKey.ITERATION_ID] = current_iteration_data.iteration_id
+                execution_metadata[NodeRunMetadataKey.ITERATION_INDEX] = current_iteration_data.current_index
+
         if isinstance(event, QueueNodeSucceededEvent):
             workflow_node_execution = self._workflow_node_execution_success(
                 workflow_node_execution=workflow_node_execution,
@@ -451,12 +464,18 @@ class WorkflowCycleManage:
                 inputs=event.inputs,
                 process_data=event.process_data,
                 outputs=event.outputs,
-                execution_metadata=event.execution_metadata
+                execution_metadata=execution_metadata
             )
 
-            if event.execution_metadata and event.execution_metadata.get(NodeRunMetadataKey.TOTAL_TOKENS):
+            if execution_metadata and execution_metadata.get(NodeRunMetadataKey.TOTAL_TOKENS):
                 self._task_state.total_tokens += (
-                    int(event.execution_metadata.get(NodeRunMetadataKey.TOTAL_TOKENS)))
+                    int(execution_metadata.get(NodeRunMetadataKey.TOTAL_TOKENS)))
+                
+                if self._iteration_state:
+                    for iteration_node_id in self._iteration_state.current_iterations:
+                        data = self._iteration_state.current_iterations[iteration_node_id]
+                        if execution_metadata.get(NodeRunMetadataKey.TOTAL_TOKENS):
+                            data.total_tokens += int(execution_metadata.get(NodeRunMetadataKey.TOTAL_TOKENS))
 
             if workflow_node_execution.node_type == NodeType.LLM.value:
                 outputs = workflow_node_execution.outputs_dict
@@ -469,7 +488,8 @@ class WorkflowCycleManage:
                 error=event.error,
                 inputs=event.inputs,
                 process_data=event.process_data,
-                outputs=event.outputs
+                outputs=event.outputs,
+                execution_metadata=execution_metadata
             )
 
         db.session.close()

api/core/app/task_pipeline/workflow_cycle_state_manager.py

@@ -0,0 +1,16 @@
+from typing import Any, Union
+
+from core.app.entities.app_invoke_entities import AdvancedChatAppGenerateEntity, WorkflowAppGenerateEntity
+from core.app.entities.task_entities import AdvancedChatTaskState, WorkflowTaskState
+from core.workflow.entities.node_entities import SystemVariable
+from models.account import Account
+from models.model import EndUser
+from models.workflow import Workflow
+
+
+class WorkflowCycleStateManager:
+    _application_generate_entity: Union[AdvancedChatAppGenerateEntity, WorkflowAppGenerateEntity]
+    _workflow: Workflow
+    _user: Union[Account, EndUser]
+    _task_state: Union[AdvancedChatTaskState, WorkflowTaskState]
+    _workflow_system_variables: dict[SystemVariable, Any]

api/core/app/task_pipeline/workflow_iteration_cycle_manage.py

@@ -0,0 +1,281 @@
+import json
+import time
+from collections.abc import Generator
+from typing import Optional, Union
+
+from core.app.entities.queue_entities import (
+    QueueIterationCompletedEvent,
+    QueueIterationNextEvent,
+    QueueIterationStartEvent,
+)
+from core.app.entities.task_entities import (
+    IterationNodeCompletedStreamResponse,
+    IterationNodeNextStreamResponse,
+    IterationNodeStartStreamResponse,
+    NodeExecutionInfo,
+    WorkflowIterationState,
+)
+from core.app.task_pipeline.workflow_cycle_state_manager import WorkflowCycleStateManager
+from core.workflow.entities.node_entities import NodeType
+from extensions.ext_database import db
+from models.workflow import (
+    WorkflowNodeExecution,
+    WorkflowNodeExecutionStatus,
+    WorkflowNodeExecutionTriggeredFrom,
+    WorkflowRun,
+)
+
+
+class WorkflowIterationCycleManage(WorkflowCycleStateManager):
+    _iteration_state: WorkflowIterationState = None
+
+    def _init_iteration_state(self) -> WorkflowIterationState:
+        if not self._iteration_state:
+            self._iteration_state = WorkflowIterationState(
+                current_iterations={}
+            )
+
+    def _handle_iteration_to_stream_response(self, task_id: str, event: QueueIterationStartEvent | QueueIterationNextEvent | QueueIterationCompletedEvent) \
+    -> Union[IterationNodeStartStreamResponse, IterationNodeNextStreamResponse, IterationNodeCompletedStreamResponse]:
+        """
+        Handle iteration to stream response
+        :param task_id: task id
+        :param event: iteration event
+        :return:
+        """
+        if isinstance(event, QueueIterationStartEvent):
+            return IterationNodeStartStreamResponse(
+                task_id=task_id,
+                workflow_run_id=self._task_state.workflow_run_id,
+                data=IterationNodeStartStreamResponse.Data(
+                    id=event.node_id,
+                    node_id=event.node_id,
+                    node_type=event.node_type.value,
+                    title=event.node_data.title,
+                    created_at=int(time.time()),
+                    extras={},
+                    inputs=event.inputs,
+                    metadata=event.metadata
+                )
+            )
+        elif isinstance(event, QueueIterationNextEvent):
+            current_iteration = self._iteration_state.current_iterations[event.node_id]
+
+            return IterationNodeNextStreamResponse(
+                task_id=task_id,
+                workflow_run_id=self._task_state.workflow_run_id,
+                data=IterationNodeNextStreamResponse.Data(
+                    id=event.node_id,
+                    node_id=event.node_id,
+                    node_type=event.node_type.value,
+                    title=current_iteration.node_data.title,
+                    index=event.index,
+                    pre_iteration_output=event.output,
+                    created_at=int(time.time()),
+                    extras={}
+                )
+            )
+        elif isinstance(event, QueueIterationCompletedEvent):
+            current_iteration = self._iteration_state.current_iterations[event.node_id]
+
+            return IterationNodeCompletedStreamResponse(
+                task_id=task_id,
+                workflow_run_id=self._task_state.workflow_run_id,
+                data=IterationNodeCompletedStreamResponse.Data(
+                    id=event.node_id,
+                    node_id=event.node_id,
+                    node_type=event.node_type.value,
+                    title=current_iteration.node_data.title,
+                    outputs=event.outputs,
+                    created_at=int(time.time()),
+                    extras={},
+                    inputs=current_iteration.inputs,
+                    status=WorkflowNodeExecutionStatus.SUCCEEDED,
+                    error=None,
+                    elapsed_time=time.perf_counter() - current_iteration.started_at,
+                    total_tokens=current_iteration.total_tokens,
+                    finished_at=int(time.time()),
+                    steps=current_iteration.current_index
+                )
+            )
+        
+    def _init_iteration_execution_from_workflow_run(self, 
+        workflow_run: WorkflowRun,
+        node_id: str,
+        node_type: NodeType,
+        node_title: str,
+        node_run_index: int = 1,
+        inputs: Optional[dict] = None,
+        predecessor_node_id: Optional[str] = None
+    ) -> WorkflowNodeExecution:
+        workflow_node_execution = WorkflowNodeExecution(
+            tenant_id=workflow_run.tenant_id,
+            app_id=workflow_run.app_id,
+            workflow_id=workflow_run.workflow_id,
+            triggered_from=WorkflowNodeExecutionTriggeredFrom.WORKFLOW_RUN.value,
+            workflow_run_id=workflow_run.id,
+            predecessor_node_id=predecessor_node_id,
+            index=node_run_index,
+            node_id=node_id,
+            node_type=node_type.value,
+            inputs=json.dumps(inputs) if inputs else None,
+            title=node_title,
+            status=WorkflowNodeExecutionStatus.RUNNING.value,
+            created_by_role=workflow_run.created_by_role,
+            created_by=workflow_run.created_by,
+            execution_metadata=json.dumps({
+                'started_run_index': node_run_index + 1,
+                'current_index': 0,
+                'steps_boundary': [],
+            })
+        )
+
+        db.session.add(workflow_node_execution)
+        db.session.commit()
+        db.session.refresh(workflow_node_execution)
+        db.session.close()
+
+        return workflow_node_execution
+    
+    def _handle_iteration_operation(self, event: QueueIterationStartEvent | QueueIterationNextEvent | QueueIterationCompletedEvent) -> WorkflowNodeExecution:
+        if isinstance(event, QueueIterationStartEvent):
+            return self._handle_iteration_started(event)
+        elif isinstance(event, QueueIterationNextEvent):
+            return self._handle_iteration_next(event)
+        elif isinstance(event, QueueIterationCompletedEvent):
+            return self._handle_iteration_completed(event)
+    
+    def _handle_iteration_started(self, event: QueueIterationStartEvent) -> WorkflowNodeExecution:
+        self._init_iteration_state()
+
+        workflow_run = db.session.query(WorkflowRun).filter(WorkflowRun.id == self._task_state.workflow_run_id).first()
+        workflow_node_execution = self._init_iteration_execution_from_workflow_run(
+            workflow_run=workflow_run,
+            node_id=event.node_id,
+            node_type=NodeType.ITERATION,
+            node_title=event.node_data.title,
+            node_run_index=event.node_run_index,
+            inputs=event.inputs,
+            predecessor_node_id=event.predecessor_node_id
+        )
+
+        latest_node_execution_info = NodeExecutionInfo(
+            workflow_node_execution_id=workflow_node_execution.id,
+            node_type=NodeType.ITERATION,
+            start_at=time.perf_counter()
+        )
+
+        self._task_state.ran_node_execution_infos[event.node_id] = latest_node_execution_info
+        self._task_state.latest_node_execution_info = latest_node_execution_info
+
+        self._iteration_state.current_iterations[event.node_id] = WorkflowIterationState.Data(
+            parent_iteration_id=None,
+            iteration_id=event.node_id,
+            current_index=0,
+            iteration_steps_boundary=[],
+            node_execution_id=workflow_node_execution.id,
+            started_at=time.perf_counter(),
+            inputs=event.inputs,
+            total_tokens=0,
+            node_data=event.node_data
+        )
+
+        db.session.close()
+
+        return workflow_node_execution
+    
+    def _handle_iteration_next(self, event: QueueIterationNextEvent) -> WorkflowNodeExecution:
+        if event.node_id not in self._iteration_state.current_iterations:
+            return
+        current_iteration = self._iteration_state.current_iterations[event.node_id]
+        current_iteration.current_index = event.index
+        current_iteration.iteration_steps_boundary.append(event.node_run_index)
+        workflow_node_execution: WorkflowNodeExecution = db.session.query(WorkflowNodeExecution).filter(
+            WorkflowNodeExecution.id == current_iteration.node_execution_id
+        ).first()
+
+        original_node_execution_metadata = workflow_node_execution.execution_metadata_dict
+        if original_node_execution_metadata:
+            original_node_execution_metadata['current_index'] = event.index
+            original_node_execution_metadata['steps_boundary'] = current_iteration.iteration_steps_boundary
+            original_node_execution_metadata['total_tokens'] = current_iteration.total_tokens
+            workflow_node_execution.execution_metadata = json.dumps(original_node_execution_metadata)
+
+            db.session.commit()
+
+        db.session.close()
+
+    def _handle_iteration_completed(self, event: QueueIterationCompletedEvent) -> WorkflowNodeExecution:
+        if event.node_id not in self._iteration_state.current_iterations:
+            return
+        
+        current_iteration = self._iteration_state.current_iterations[event.node_id]
+        workflow_node_execution: WorkflowNodeExecution = db.session.query(WorkflowNodeExecution).filter(
+            WorkflowNodeExecution.id == current_iteration.node_execution_id
+        ).first()
+
+        workflow_node_execution.status = WorkflowNodeExecutionStatus.SUCCEEDED.value
+        workflow_node_execution.outputs = json.dumps(event.outputs) if event.outputs else None
+        workflow_node_execution.elapsed_time = time.perf_counter() - current_iteration.started_at
+        
+        original_node_execution_metadata = workflow_node_execution.execution_metadata_dict
+        if original_node_execution_metadata:
+            original_node_execution_metadata['steps_boundary'] = current_iteration.iteration_steps_boundary
+            original_node_execution_metadata['total_tokens'] = current_iteration.total_tokens
+            workflow_node_execution.execution_metadata = json.dumps(original_node_execution_metadata)
+
+        db.session.commit()
+
+        # remove current iteration
+        self._iteration_state.current_iterations.pop(event.node_id, None)
+
+        # set latest node execution info
+        latest_node_execution_info = NodeExecutionInfo(
+            workflow_node_execution_id=workflow_node_execution.id,
+            node_type=NodeType.ITERATION,
+            start_at=time.perf_counter()
+        )
+
+        self._task_state.latest_node_execution_info = latest_node_execution_info
+        
+        db.session.close()
+
+    def _handle_iteration_exception(self, task_id: str, error: str) -> Generator[IterationNodeCompletedStreamResponse, None, None]:
+        """
+        Handle iteration exception
+        """
+        if not self._iteration_state or not self._iteration_state.current_iterations:
+            return
+        
+        for node_id, current_iteration in self._iteration_state.current_iterations.items():
+            workflow_node_execution: WorkflowNodeExecution = db.session.query(WorkflowNodeExecution).filter(
+                WorkflowNodeExecution.id == current_iteration.node_execution_id
+            ).first()
+
+            workflow_node_execution.status = WorkflowNodeExecutionStatus.FAILED.value
+            workflow_node_execution.error = error
+            workflow_node_execution.elapsed_time = time.perf_counter() - current_iteration.started_at
+
+            db.session.commit()
+            db.session.close()
+
+            yield IterationNodeCompletedStreamResponse(
+                task_id=task_id,
+                workflow_run_id=self._task_state.workflow_run_id,
+                data=IterationNodeCompletedStreamResponse.Data(
+                    id=node_id,
+                    node_id=node_id,
+                    node_type=NodeType.ITERATION.value,
+                    title=current_iteration.node_data.title,
+                    outputs={},
+                    created_at=int(time.time()),
+                    extras={},
+                    inputs=current_iteration.inputs,
+                    status=WorkflowNodeExecutionStatus.FAILED,
+                    error=error,
+                    elapsed_time=time.perf_counter() - current_iteration.started_at,
+                    total_tokens=current_iteration.total_tokens,
+                    finished_at=int(time.time()),
+                    steps=current_iteration.current_index
+                )
+            )

api/core/entities/model_entities.py

@@ -16,6 +16,7 @@ class ModelStatus(Enum):
     NO_CONFIGURE = "no-configure"
     QUOTA_EXCEEDED = "quota-exceeded"
     NO_PERMISSION = "no-permission"
+    DISABLED = "disabled"
 
 
 class SimpleModelProviderEntity(BaseModel):
@@ -43,12 +44,19 @@ class SimpleModelProviderEntity(BaseModel):
         )
 
 
-class ModelWithProviderEntity(ProviderModel):
+class ProviderModelWithStatusEntity(ProviderModel):
+    """
+    Model class for model response.
+    """
+    status: ModelStatus
+    load_balancing_enabled: bool = False
+
+
+class ModelWithProviderEntity(ProviderModelWithStatusEntity):
     """
     Model with provider entity.
     """
     provider: SimpleModelProviderEntity
-    status: ModelStatus
 
 
 class DefaultModelProviderEntity(BaseModel):

api/core/entities/provider_configuration.py

@@ -1,6 +1,7 @@
 import datetime
 import json
 import logging
+from collections import defaultdict
 from collections.abc import Iterator
 from json import JSONDecodeError
 from typing import Optional
@@ -8,7 +9,12 @@ from typing import Optional
 from pydantic import BaseModel
 
 from core.entities.model_entities import ModelStatus, ModelWithProviderEntity, SimpleModelProviderEntity
-from core.entities.provider_entities import CustomConfiguration, SystemConfiguration, SystemConfigurationStatus
+from core.entities.provider_entities import (
+    CustomConfiguration,
+    ModelSettings,
+    SystemConfiguration,
+    SystemConfigurationStatus,
+)
 from core.helper import encrypter
 from core.helper.model_provider_cache import ProviderCredentialsCache, ProviderCredentialsCacheType
 from core.model_runtime.entities.model_entities import FetchFrom, ModelType
@@ -22,7 +28,14 @@ from core.model_runtime.model_providers import model_provider_factory
 from core.model_runtime.model_providers.__base.ai_model import AIModel
 from core.model_runtime.model_providers.__base.model_provider import ModelProvider
 from extensions.ext_database import db
-from models.provider import Provider, ProviderModel, ProviderType, TenantPreferredModelProvider
+from models.provider import (
+    LoadBalancingModelConfig,
+    Provider,
+    ProviderModel,
+    ProviderModelSetting,
+    ProviderType,
+    TenantPreferredModelProvider,
+)
 
 logger = logging.getLogger(__name__)
 
@@ -39,6 +52,7 @@ class ProviderConfiguration(BaseModel):
     using_provider_type: ProviderType
     system_configuration: SystemConfiguration
     custom_configuration: CustomConfiguration
+    model_settings: list[ModelSettings]
 
     def __init__(self, **data):
         super().__init__(**data)
@@ -62,6 +76,14 @@ class ProviderConfiguration(BaseModel):
         :param model: model name
         :return:
         """
+        if self.model_settings:
+            # check if model is disabled by admin
+            for model_setting in self.model_settings:
+                if (model_setting.model_type == model_type
+                        and model_setting.model == model):
+                    if not model_setting.enabled:
+                        raise ValueError(f'Model {model} is disabled.')
+
         if self.using_provider_type == ProviderType.SYSTEM:
             restrict_models = []
             for quota_configuration in self.system_configuration.quota_configurations:
@@ -80,15 +102,17 @@ class ProviderConfiguration(BaseModel):
 
             return copy_credentials
         else:
+            credentials = None
             if self.custom_configuration.models:
                 for model_configuration in self.custom_configuration.models:
                     if model_configuration.model_type == model_type and model_configuration.model == model:
-                        return model_configuration.credentials
+                        credentials = model_configuration.credentials
+                        break
 
             if self.custom_configuration.provider:
-                return self.custom_configuration.provider.credentials
-            else:
-                return None
+                credentials = self.custom_configuration.provider.credentials
+
+            return credentials
 
     def get_system_configuration_status(self) -> SystemConfigurationStatus:
         """
@@ -130,7 +154,7 @@ class ProviderConfiguration(BaseModel):
             return credentials
 
         # Obfuscate credentials
-        return self._obfuscated_credentials(
+        return self.obfuscated_credentials(
             credentials=credentials,
             credential_form_schemas=self.provider.provider_credential_schema.credential_form_schemas
             if self.provider.provider_credential_schema else []
@@ -151,7 +175,7 @@ class ProviderConfiguration(BaseModel):
         ).first()
 
         # Get provider credential secret variables
-        provider_credential_secret_variables = self._extract_secret_variables(
+        provider_credential_secret_variables = self.extract_secret_variables(
             self.provider.provider_credential_schema.credential_form_schemas
             if self.provider.provider_credential_schema else []
         )
@@ -274,7 +298,7 @@ class ProviderConfiguration(BaseModel):
                     return credentials
 
                 # Obfuscate credentials
-                return self._obfuscated_credentials(
+                return self.obfuscated_credentials(
                     credentials=credentials,
                     credential_form_schemas=self.provider.model_credential_schema.credential_form_schemas
                     if self.provider.model_credential_schema else []
@@ -302,7 +326,7 @@ class ProviderConfiguration(BaseModel):
         ).first()
 
         # Get provider credential secret variables
-        provider_credential_secret_variables = self._extract_secret_variables(
+        provider_credential_secret_variables = self.extract_secret_variables(
             self.provider.model_credential_schema.credential_form_schemas
             if self.provider.model_credential_schema else []
         )
@@ -402,6 +426,160 @@ class ProviderConfiguration(BaseModel):
 
             provider_model_credentials_cache.delete()
 
+    def enable_model(self, model_type: ModelType, model: str) -> ProviderModelSetting:
+        """
+        Enable model.
+        :param model_type: model type
+        :param model: model name
+        :return:
+        """
+        model_setting = db.session.query(ProviderModelSetting) \
+            .filter(
+            ProviderModelSetting.tenant_id == self.tenant_id,
+            ProviderModelSetting.provider_name == self.provider.provider,
+            ProviderModelSetting.model_type == model_type.to_origin_model_type(),
+            ProviderModelSetting.model_name == model
+        ).first()
+
+        if model_setting:
+            model_setting.enabled = True
+            model_setting.updated_at = datetime.datetime.now(datetime.timezone.utc).replace(tzinfo=None)
+            db.session.commit()
+        else:
+            model_setting = ProviderModelSetting(
+                tenant_id=self.tenant_id,
+                provider_name=self.provider.provider,
+                model_type=model_type.to_origin_model_type(),
+                model_name=model,
+                enabled=True
+            )
+            db.session.add(model_setting)
+            db.session.commit()
+
+        return model_setting
+
+    def disable_model(self, model_type: ModelType, model: str) -> ProviderModelSetting:
+        """
+        Disable model.
+        :param model_type: model type
+        :param model: model name
+        :return:
+        """
+        model_setting = db.session.query(ProviderModelSetting) \
+            .filter(
+            ProviderModelSetting.tenant_id == self.tenant_id,
+            ProviderModelSetting.provider_name == self.provider.provider,
+            ProviderModelSetting.model_type == model_type.to_origin_model_type(),
+            ProviderModelSetting.model_name == model
+        ).first()
+
+        if model_setting:
+            model_setting.enabled = False
+            model_setting.updated_at = datetime.datetime.now(datetime.timezone.utc).replace(tzinfo=None)
+            db.session.commit()
+        else:
+            model_setting = ProviderModelSetting(
+                tenant_id=self.tenant_id,
+                provider_name=self.provider.provider,
+                model_type=model_type.to_origin_model_type(),
+                model_name=model,
+                enabled=False
+            )
+            db.session.add(model_setting)
+            db.session.commit()
+
+        return model_setting
+
+    def get_provider_model_setting(self, model_type: ModelType, model: str) -> Optional[ProviderModelSetting]:
+        """
+        Get provider model setting.
+        :param model_type: model type
+        :param model: model name
+        :return:
+        """
+        return db.session.query(ProviderModelSetting) \
+            .filter(
+            ProviderModelSetting.tenant_id == self.tenant_id,
+            ProviderModelSetting.provider_name == self.provider.provider,
+            ProviderModelSetting.model_type == model_type.to_origin_model_type(),
+            ProviderModelSetting.model_name == model
+        ).first()
+
+    def enable_model_load_balancing(self, model_type: ModelType, model: str) -> ProviderModelSetting:
+        """
+        Enable model load balancing.
+        :param model_type: model type
+        :param model: model name
+        :return:
+        """
+        load_balancing_config_count = db.session.query(LoadBalancingModelConfig) \
+            .filter(
+            LoadBalancingModelConfig.tenant_id == self.tenant_id,
+            LoadBalancingModelConfig.provider_name == self.provider.provider,
+            LoadBalancingModelConfig.model_type == model_type.to_origin_model_type(),
+            LoadBalancingModelConfig.model_name == model
+        ).count()
+
+        if load_balancing_config_count <= 1:
+            raise ValueError('Model load balancing configuration must be more than 1.')
+
+        model_setting = db.session.query(ProviderModelSetting) \
+            .filter(
+            ProviderModelSetting.tenant_id == self.tenant_id,
+            ProviderModelSetting.provider_name == self.provider.provider,
+            ProviderModelSetting.model_type == model_type.to_origin_model_type(),
+            ProviderModelSetting.model_name == model
+        ).first()
+
+        if model_setting:
+            model_setting.load_balancing_enabled = True
+            model_setting.updated_at = datetime.datetime.now(datetime.timezone.utc).replace(tzinfo=None)
+            db.session.commit()
+        else:
+            model_setting = ProviderModelSetting(
+                tenant_id=self.tenant_id,
+                provider_name=self.provider.provider,
+                model_type=model_type.to_origin_model_type(),
+                model_name=model,
+                load_balancing_enabled=True
+            )
+            db.session.add(model_setting)
+            db.session.commit()
+
+        return model_setting
+
+    def disable_model_load_balancing(self, model_type: ModelType, model: str) -> ProviderModelSetting:
+        """
+        Disable model load balancing.
+        :param model_type: model type
+        :param model: model name
+        :return:
+        """
+        model_setting = db.session.query(ProviderModelSetting) \
+            .filter(
+            ProviderModelSetting.tenant_id == self.tenant_id,
+            ProviderModelSetting.provider_name == self.provider.provider,
+            ProviderModelSetting.model_type == model_type.to_origin_model_type(),
+            ProviderModelSetting.model_name == model
+        ).first()
+
+        if model_setting:
+            model_setting.load_balancing_enabled = False
+            model_setting.updated_at = datetime.datetime.now(datetime.timezone.utc).replace(tzinfo=None)
+            db.session.commit()
+        else:
+            model_setting = ProviderModelSetting(
+                tenant_id=self.tenant_id,
+                provider_name=self.provider.provider,
+                model_type=model_type.to_origin_model_type(),
+                model_name=model,
+                load_balancing_enabled=False
+            )
+            db.session.add(model_setting)
+            db.session.commit()
+
+        return model_setting
+
     def get_provider_instance(self) -> ModelProvider:
         """
         Get provider instance.
@@ -453,7 +631,7 @@ class ProviderConfiguration(BaseModel):
 
         db.session.commit()
 
-    def _extract_secret_variables(self, credential_form_schemas: list[CredentialFormSchema]) -> list[str]:
+    def extract_secret_variables(self, credential_form_schemas: list[CredentialFormSchema]) -> list[str]:
         """
         Extract secret input form variables.
 
@@ -467,7 +645,7 @@ class ProviderConfiguration(BaseModel):
 
         return secret_input_form_variables
 
-    def _obfuscated_credentials(self, credentials: dict, credential_form_schemas: list[CredentialFormSchema]) -> dict:
+    def obfuscated_credentials(self, credentials: dict, credential_form_schemas: list[CredentialFormSchema]) -> dict:
         """
         Obfuscated credentials.
 
@@ -476,7 +654,7 @@ class ProviderConfiguration(BaseModel):
         :return:
         """
         # Get provider credential secret variables
-        credential_secret_variables = self._extract_secret_variables(
+        credential_secret_variables = self.extract_secret_variables(
             credential_form_schemas
         )
 
@@ -522,15 +700,22 @@ class ProviderConfiguration(BaseModel):
         else:
             model_types = provider_instance.get_provider_schema().supported_model_types
 
+        # Group model settings by model type and model
+        model_setting_map = defaultdict(dict)
+        for model_setting in self.model_settings:
+            model_setting_map[model_setting.model_type][model_setting.model] = model_setting
+
         if self.using_provider_type == ProviderType.SYSTEM:
             provider_models = self._get_system_provider_models(
                 model_types=model_types,
-                provider_instance=provider_instance
+                provider_instance=provider_instance,
+                model_setting_map=model_setting_map
             )
         else:
             provider_models = self._get_custom_provider_models(
                 model_types=model_types,
-                provider_instance=provider_instance
+                provider_instance=provider_instance,
+                model_setting_map=model_setting_map
             )
 
         if only_active:
@@ -541,18 +726,27 @@ class ProviderConfiguration(BaseModel):
 
     def _get_system_provider_models(self,
                                     model_types: list[ModelType],
-                                    provider_instance: ModelProvider) -> list[ModelWithProviderEntity]:
+                                    provider_instance: ModelProvider,
+                                    model_setting_map: dict[ModelType, dict[str, ModelSettings]]) \
+            -> list[ModelWithProviderEntity]:
         """
         Get system provider models.
 
         :param model_types: model types
         :param provider_instance: provider instance
+        :param model_setting_map: model setting map
         :return:
         """
         provider_models = []
         for model_type in model_types:
-            provider_models.extend(
-                [
+            for m in provider_instance.models(model_type):
+                status = ModelStatus.ACTIVE
+                if m.model_type in model_setting_map and m.model in model_setting_map[m.model_type]:
+                    model_setting = model_setting_map[m.model_type][m.model]
+                    if model_setting.enabled is False:
+                        status = ModelStatus.DISABLED
+
+                provider_models.append(
                     ModelWithProviderEntity(
                         model=m.model,
                         label=m.label,
@@ -562,11 +756,9 @@ class ProviderConfiguration(BaseModel):
                         model_properties=m.model_properties,
                         deprecated=m.deprecated,
                         provider=SimpleModelProviderEntity(self.provider),
-                        status=ModelStatus.ACTIVE
+                        status=status
                     )
-                    for m in provider_instance.models(model_type)
-                ]
-            )
+                )
 
         if self.provider.provider not in original_provider_configurate_methods:
             original_provider_configurate_methods[self.provider.provider] = []
@@ -586,7 +778,8 @@ class ProviderConfiguration(BaseModel):
                 break
 
             if should_use_custom_model:
-                if original_provider_configurate_methods[self.provider.provider] == [ConfigurateMethod.CUSTOMIZABLE_MODEL]:
+                if original_provider_configurate_methods[self.provider.provider] == [
+                    ConfigurateMethod.CUSTOMIZABLE_MODEL]:
                     # only customizable model
                     for restrict_model in restrict_models:
                         copy_credentials = self.system_configuration.credentials.copy()
@@ -611,6 +804,13 @@ class ProviderConfiguration(BaseModel):
                         if custom_model_schema.model_type not in model_types:
                             continue
 
+                        status = ModelStatus.ACTIVE
+                        if (custom_model_schema.model_type in model_setting_map
+                                and custom_model_schema.model in model_setting_map[custom_model_schema.model_type]):
+                            model_setting = model_setting_map[custom_model_schema.model_type][custom_model_schema.model]
+                            if model_setting.enabled is False:
+                                status = ModelStatus.DISABLED
+
                         provider_models.append(
                             ModelWithProviderEntity(
                                 model=custom_model_schema.model,
@@ -621,7 +821,7 @@ class ProviderConfiguration(BaseModel):
                                 model_properties=custom_model_schema.model_properties,
                                 deprecated=custom_model_schema.deprecated,
                                 provider=SimpleModelProviderEntity(self.provider),
-                                status=ModelStatus.ACTIVE
+                                status=status
                             )
                         )
 
@@ -632,16 +832,20 @@ class ProviderConfiguration(BaseModel):
                     m.status = ModelStatus.NO_PERMISSION
                 elif not quota_configuration.is_valid:
                     m.status = ModelStatus.QUOTA_EXCEEDED
+
         return provider_models
 
     def _get_custom_provider_models(self,
                                     model_types: list[ModelType],
-                                    provider_instance: ModelProvider) -> list[ModelWithProviderEntity]:
+                                    provider_instance: ModelProvider,
+                                    model_setting_map: dict[ModelType, dict[str, ModelSettings]]) \
+            -> list[ModelWithProviderEntity]:
         """
         Get custom provider models.
 
         :param model_types: model types
         :param provider_instance: provider instance
+        :param model_setting_map: model setting map
         :return:
         """
         provider_models = []
@@ -656,6 +860,16 @@ class ProviderConfiguration(BaseModel):
 
             models = provider_instance.models(model_type)
             for m in models:
+                status = ModelStatus.ACTIVE if credentials else ModelStatus.NO_CONFIGURE
+                load_balancing_enabled = False
+                if m.model_type in model_setting_map and m.model in model_setting_map[m.model_type]:
+                    model_setting = model_setting_map[m.model_type][m.model]
+                    if model_setting.enabled is False:
+                        status = ModelStatus.DISABLED
+
+                    if len(model_setting.load_balancing_configs) > 1:
+                        load_balancing_enabled = True
+
                 provider_models.append(
                     ModelWithProviderEntity(
                         model=m.model,
@@ -666,7 +880,8 @@ class ProviderConfiguration(BaseModel):
                         model_properties=m.model_properties,
                         deprecated=m.deprecated,
                         provider=SimpleModelProviderEntity(self.provider),
-                        status=ModelStatus.ACTIVE if credentials else ModelStatus.NO_CONFIGURE
+                        status=status,
+                        load_balancing_enabled=load_balancing_enabled
                     )
                 )
 
@@ -690,6 +905,17 @@ class ProviderConfiguration(BaseModel):
             if not custom_model_schema:
                 continue
 
+            status = ModelStatus.ACTIVE
+            load_balancing_enabled = False
+            if (custom_model_schema.model_type in model_setting_map
+                    and custom_model_schema.model in model_setting_map[custom_model_schema.model_type]):
+                model_setting = model_setting_map[custom_model_schema.model_type][custom_model_schema.model]
+                if model_setting.enabled is False:
+                    status = ModelStatus.DISABLED
+
+                if len(model_setting.load_balancing_configs) > 1:
+                    load_balancing_enabled = True
+
             provider_models.append(
                 ModelWithProviderEntity(
                     model=custom_model_schema.model,
@@ -700,7 +926,8 @@ class ProviderConfiguration(BaseModel):
                     model_properties=custom_model_schema.model_properties,
                     deprecated=custom_model_schema.deprecated,
                     provider=SimpleModelProviderEntity(self.provider),
-                    status=ModelStatus.ACTIVE
+                    status=status,
+                    load_balancing_enabled=load_balancing_enabled
                 )
             )
 

api/core/entities/provider_entities.py

@@ -72,3 +72,22 @@ class CustomConfiguration(BaseModel):
     """
     provider: Optional[CustomProviderConfiguration] = None
     models: list[CustomModelConfiguration] = []
+
+
+class ModelLoadBalancingConfiguration(BaseModel):
+    """
+    Class for model load balancing configuration.
+    """
+    id: str
+    name: str
+    credentials: dict
+
+
+class ModelSettings(BaseModel):
+    """
+    Model class for model settings.
+    """
+    model: str
+    model_type: ModelType
+    enabled: bool = True
+    load_balancing_configs: list[ModelLoadBalancingConfiguration] = []

api/core/extension/extensible.py

@@ -7,7 +7,7 @@ from typing import Any, Optional
 
 from pydantic import BaseModel
 
-from core.utils.position_helper import sort_to_dict_by_position_map
+from core.helper.position_helper import sort_to_dict_by_position_map
 
 
 class ExtensionModule(enum.Enum):

api/core/file/message_file_parser.py

@@ -42,6 +42,8 @@ class MessageFileParser:
                     raise ValueError('Invalid file url')
             if file.get('transfer_method') == FileTransferMethod.LOCAL_FILE.value and not file.get('upload_file_id'):
                 raise ValueError('Missing file upload_file_id')
+            if file.get('transform_method') == FileTransferMethod.TOOL_FILE.value and not file.get('tool_file_id'):
+                raise ValueError('Missing file tool_file_id')
 
         # transform files to file objs
         type_file_objs = self._to_file_objs(files, file_extra_config)
@@ -149,12 +151,21 @@ class MessageFileParser:
         """
         if isinstance(file, dict):
             transfer_method = FileTransferMethod.value_of(file.get('transfer_method'))
+            if transfer_method != FileTransferMethod.TOOL_FILE:
+                return FileVar(
+                    tenant_id=self.tenant_id,
+                    type=FileType.value_of(file.get('type')),
+                    transfer_method=transfer_method,
+                    url=file.get('url') if transfer_method == FileTransferMethod.REMOTE_URL else None,
+                    related_id=file.get('upload_file_id') if transfer_method == FileTransferMethod.LOCAL_FILE else None,
+                    extra_config=file_extra_config
+                )
             return FileVar(
                 tenant_id=self.tenant_id,
                 type=FileType.value_of(file.get('type')),
                 transfer_method=transfer_method,
-                url=file.get('url') if transfer_method == FileTransferMethod.REMOTE_URL else None,
-                related_id=file.get('upload_file_id') if transfer_method == FileTransferMethod.LOCAL_FILE else None,
+                url=None,
+                related_id=file.get('tool_file_id'),
                 extra_config=file_extra_config
             )
         else:

api/core/file/upload_file_parser.py

@@ -77,4 +77,4 @@ class UploadFileParser:
             return False
 
         current_time = int(time.time())
-        return current_time - int(timestamp) <= 300  # expired after 5 minutes
+        return current_time - int(timestamp) <= current_app.config.get('FILES_ACCESS_TIMEOUT')

api/core/helper/code_executor/code_executor.py

@@ -10,9 +10,10 @@ from yarl import URL
 
 from config import get_env
 from core.helper.code_executor.entities import CodeDependency
-from core.helper.code_executor.javascript_transformer import NodeJsTemplateTransformer
-from core.helper.code_executor.jinja2_transformer import Jinja2TemplateTransformer
-from core.helper.code_executor.python_transformer import PYTHON_STANDARD_PACKAGES, PythonTemplateTransformer
+from core.helper.code_executor.javascript.javascript_transformer import NodeJsTemplateTransformer
+from core.helper.code_executor.jinja2.jinja2_transformer import Jinja2TemplateTransformer
+from core.helper.code_executor.python3.python3_transformer import Python3TemplateTransformer
+from core.helper.code_executor.template_transformer import TemplateTransformer
 
 logger = logging.getLogger(__name__)
 
@@ -34,6 +35,7 @@ class CodeExecutionResponse(BaseModel):
     message: str
     data: Data
 
+
 class CodeLanguage(str, Enum):
     PYTHON3 = 'python3'
     JINJA2 = 'jinja2'
@@ -44,8 +46,8 @@ class CodeExecutor:
     dependencies_cache = {}
     dependencies_cache_lock = Lock()
 
-    code_template_transformers = {
-        CodeLanguage.PYTHON3: PythonTemplateTransformer,
+    code_template_transformers: dict[CodeLanguage, type[TemplateTransformer]] = {
+        CodeLanguage.PYTHON3: Python3TemplateTransformer,
         CodeLanguage.JINJA2: Jinja2TemplateTransformer,
         CodeLanguage.JAVASCRIPT: NodeJsTemplateTransformer,
     }
@@ -56,6 +58,10 @@ class CodeExecutor:
         CodeLanguage.PYTHON3: CodeLanguage.PYTHON3,
     }
 
+    supported_dependencies_languages: set[CodeLanguage] = {
+        CodeLanguage.PYTHON3
+    }
+
     @classmethod
     def execute_code(cls, 
                      language: Literal['python3', 'javascript', 'jinja2'], 
@@ -93,7 +99,9 @@ class CodeExecutor:
         except CodeExecutionException as e:
             raise e
         except Exception as e:
-            raise CodeExecutionException('Failed to execute code, this is likely a network issue, please check if the sandbox service is running')
+            raise CodeExecutionException('Failed to execute code, which is likely a network issue,'
+                                         ' please check if the sandbox service is running.'
+                                         f' ( Error: {str(e)} )')
         
         try:
             response = response.json()
@@ -133,7 +141,10 @@ class CodeExecutor:
         return template_transformer.transform_response(response)
     
     @classmethod
-    def list_dependencies(cls, language: Literal['python3']) -> list[CodeDependency]:
+    def list_dependencies(cls, language: str) -> list[CodeDependency]:
+        if language not in cls.supported_dependencies_languages:
+            return []
+
         with cls.dependencies_cache_lock:
             if language in cls.dependencies_cache:
                 # check expiration
@@ -178,7 +189,8 @@ class CodeExecutor:
             response = response.json()
             dependencies = response.get('data', {}).get('dependencies', [])
             return [
-                CodeDependency(**dependency) for dependency in dependencies if dependency.get('name') not in PYTHON_STANDARD_PACKAGES
+                CodeDependency(**dependency) for dependency in dependencies
+                if dependency.get('name') not in Python3TemplateTransformer.get_standard_packages()
             ]
         except Exception as e:
             logger.exception(f'Failed to list dependencies: {e}')

api/core/helper/code_executor/code_node_provider.py

@@ -0,0 +1,55 @@
+from abc import abstractmethod
+
+from pydantic import BaseModel
+
+from core.helper.code_executor.code_executor import CodeExecutor
+
+
+class CodeNodeProvider(BaseModel):
+    @staticmethod
+    @abstractmethod
+    def get_language() -> str:
+        pass
+
+    @classmethod
+    def is_accept_language(cls, language: str) -> bool:
+        return language == cls.get_language()
+
+    @classmethod
+    @abstractmethod
+    def get_default_code(cls) -> str:
+        """
+        get default code in specific programming language for the code node
+        """
+        pass
+
+    @classmethod
+    def get_default_available_packages(cls) -> list[dict]:
+        return [p.dict() for p in CodeExecutor.list_dependencies(cls.get_language())]
+
+    @classmethod
+    def get_default_config(cls) -> dict:
+        return {
+            "type": "code",
+            "config": {
+                "variables": [
+                    {
+                        "variable": "arg1",
+                        "value_selector": []
+                    },
+                    {
+                        "variable": "arg2",
+                        "value_selector": []
+                    }
+                ],
+                "code_language": cls.get_language(),
+                "code": cls.get_default_code(),
+                "outputs": {
+                    "result": {
+                        "type": "string",
+                        "children": None
+                    }
+                }
+            },
+            "available_dependencies": cls.get_default_available_packages(),
+        }

api/core/helper/code_executor/entities.py

@@ -3,4 +3,4 @@ from pydantic import BaseModel
 
 class CodeDependency(BaseModel):
     name: str
-    version: str
+    version: str

api/core/helper/code_executor/javascript/javascript_code_provider.py

@@ -0,0 +1,21 @@
+from textwrap import dedent
+
+from core.helper.code_executor.code_executor import CodeLanguage
+from core.helper.code_executor.code_node_provider import CodeNodeProvider
+
+
+class JavascriptCodeProvider(CodeNodeProvider):
+    @staticmethod
+    def get_language() -> str:
+        return CodeLanguage.JAVASCRIPT
+
+    @classmethod
+    def get_default_code(cls) -> str:
+        return dedent(
+            """
+            function main({arg1, arg2}) {
+                return {
+                    result: arg1 + arg2
+                }
+            }
+            """)

api/core/helper/code_executor/javascript/javascript_transformer.py

@@ -0,0 +1,25 @@
+from textwrap import dedent
+
+from core.helper.code_executor.template_transformer import TemplateTransformer
+
+
+class NodeJsTemplateTransformer(TemplateTransformer):
+    @classmethod
+    def get_runner_script(cls) -> str:
+        runner_script = dedent(
+            f"""
+            // declare main function
+            {cls._code_placeholder}
+            
+            // decode and prepare input object
+            var inputs_obj = JSON.parse(Buffer.from('{cls._inputs_placeholder}', 'base64').toString('utf-8'))
+            
+            // execute main function
+            var output_obj = main(inputs_obj)
+            
+            // convert output to json and print
+            var output_json = JSON.stringify(output_obj)
+            var result = `<<RESULT>>${{output_json}}<<RESULT>>`
+            console.log(result)
+            """)
+        return runner_script

api/core/helper/code_executor/javascript_transformer.py

@@ -1,57 +0,0 @@
-import json
-import re
-from typing import Optional
-
-from core.helper.code_executor.entities import CodeDependency
-from core.helper.code_executor.template_transformer import TemplateTransformer
-
-NODEJS_RUNNER = """// declare main function here
-{{code}}
-
-// execute main function, and return the result
-// inputs is a dict, unstructured inputs
-output = main({{inputs}})
-
-// convert output to json and print
-output = JSON.stringify(output)
-
-result = `<<RESULT>>${output}<<RESULT>>`
-
-console.log(result)
-"""
-
-NODEJS_PRELOAD = """"""
-
-class NodeJsTemplateTransformer(TemplateTransformer):
-    @classmethod
-    def transform_caller(cls, code: str, inputs: dict, 
-                         dependencies: Optional[list[CodeDependency]] = None) -> tuple[str, str, list[CodeDependency]]:
-        """
-        Transform code to python runner
-        :param code: code
-        :param inputs: inputs
-        :return:
-        """
-
-        # transform inputs to json string
-        inputs_str = json.dumps(inputs, indent=4, ensure_ascii=False)
-
-        # replace code and inputs
-        runner = NODEJS_RUNNER.replace('{{code}}', code)
-        runner = runner.replace('{{inputs}}', inputs_str)
-
-        return runner, NODEJS_PRELOAD, []
-
-    @classmethod
-    def transform_response(cls, response: str) -> dict:
-        """
-        Transform response to dict
-        :param response: response
-        :return:
-        """
-        # extract result
-        result = re.search(r'<<RESULT>>(.*)<<RESULT>>', response, re.DOTALL)
-        if not result:
-            raise ValueError('Failed to parse result')
-        result = result.group(1)
-        return json.loads(result)

api/core/helper/code_executor/jinja2/jinja2_formatter.py

@@ -1,4 +1,4 @@
-from core.helper.code_executor.code_executor import CodeExecutor
+from core.helper.code_executor.code_executor import CodeExecutor, CodeLanguage
 
 
 class Jinja2Formatter:
@@ -11,7 +11,7 @@ class Jinja2Formatter:
         :return:
         """
         result = CodeExecutor.execute_workflow_code_template(
-            language='jinja2', code=template, inputs=inputs
+            language=CodeLanguage.JINJA2, code=template, inputs=inputs
         )
 
-        return result['result']
+        return result['result']

api/core/helper/code_executor/jinja2/jinja2_transformer.py

@@ -0,0 +1,64 @@
+from textwrap import dedent
+
+from core.helper.code_executor.python3.python3_transformer import Python3TemplateTransformer
+from core.helper.code_executor.template_transformer import TemplateTransformer
+
+
+class Jinja2TemplateTransformer(TemplateTransformer):
+    @classmethod
+    def get_standard_packages(cls) -> set[str]:
+        return {'jinja2'} | Python3TemplateTransformer.get_standard_packages()
+
+    @classmethod
+    def transform_response(cls, response: str) -> dict:
+        """
+        Transform response to dict
+        :param response: response
+        :return:
+        """
+        return {
+            'result': cls.extract_result_str_from_response(response)
+        }
+
+    @classmethod
+    def get_runner_script(cls) -> str:
+        runner_script = dedent(f"""
+            # declare main function
+            def main(**inputs):
+                import jinja2
+                template = jinja2.Template('''{cls._code_placeholder}''')
+                return template.render(**inputs)
+                
+            import json
+            from base64 import b64decode
+            
+            # decode and prepare input dict
+            inputs_obj = json.loads(b64decode('{cls._inputs_placeholder}').decode('utf-8'))
+            
+            # execute main function
+            output = main(**inputs_obj)
+            
+            # convert output and print
+            result = f'''<<RESULT>>{{output}}<<RESULT>>'''
+            print(result)
+            
+            """)
+        return runner_script
+
+    @classmethod
+    def get_preload_script(cls) -> str:
+        preload_script = dedent("""
+            import jinja2
+            from base64 import b64decode
+            
+            def _jinja2_preload_():
+                # prepare jinja2 environment, load template and render before to avoid sandbox issue
+                template = jinja2.Template('{{s}}')
+                template.render(s='a')
+            
+            if __name__ == '__main__':
+                _jinja2_preload_()
+            
+            """)
+
+        return preload_script

api/core/helper/code_executor/jinja2_transformer.py

@@ -1,108 +0,0 @@
-import json
-import re
-from base64 import b64encode
-from typing import Optional
-
-from core.helper.code_executor.entities import CodeDependency
-from core.helper.code_executor.python_transformer import PYTHON_STANDARD_PACKAGES
-from core.helper.code_executor.template_transformer import TemplateTransformer
-
-PYTHON_RUNNER = """
-import jinja2
-from json import loads
-from base64 import b64decode
-
-template = jinja2.Template('''{{code}}''')
-
-def main(**inputs):
-    return template.render(**inputs)
-
-# execute main function, and return the result
-inputs = b64decode('{{inputs}}').decode('utf-8')
-output = main(**loads(inputs))
-
-result = f'''<<RESULT>>{output}<<RESULT>>'''
-
-print(result)
-
-"""
-
-JINJA2_PRELOAD_TEMPLATE = """{% set fruits = ['Apple'] %}
-{{ 'a' }}
-{% for fruit in fruits %}
-    <li>{{ fruit }}</li>
-{% endfor %}
-{% if fruits|length > 1 %}
-1
-{% endif %}
-{% for i in range(5) %}
-    {% if i == 3 %}{{ i }}{% else %}{% endif %}
-{% endfor %}
-    {% for i in range(3) %}
-        {{ i + 1 }}
-    {% endfor %}
-{% macro say_hello() %}a{{ 'b' }}{% endmacro %}
-{{ s }}{{ say_hello() }}"""
-
-JINJA2_PRELOAD = f"""
-import jinja2
-from base64 import b64decode
-
-def _jinja2_preload_():
-    # prepare jinja2 environment, load template and render before to avoid sandbox issue
-    template = jinja2.Template('''{JINJA2_PRELOAD_TEMPLATE}''')
-    template.render(s='a')
-
-if __name__ == '__main__':
-    _jinja2_preload_()
-
-"""
-
-
-class Jinja2TemplateTransformer(TemplateTransformer):
-    @classmethod
-    def transform_caller(cls, code: str, inputs: dict, 
-                         dependencies: Optional[list[CodeDependency]] = None) -> tuple[str, str, list[CodeDependency]]:
-        """
-        Transform code to python runner
-        :param code: code
-        :param inputs: inputs
-        :return:
-        """
-
-        inputs_str = b64encode(json.dumps(inputs, ensure_ascii=False).encode()).decode('utf-8')
-
-        # transform jinja2 template to python code
-        runner = PYTHON_RUNNER.replace('{{code}}', code)
-        runner = runner.replace('{{inputs}}', inputs_str)
-
-        if not dependencies:
-            dependencies = []
-
-        # add native packages and jinja2
-        for package in PYTHON_STANDARD_PACKAGES.union(['jinja2']):
-            dependencies.append(CodeDependency(name=package, version=''))
-
-        # deduplicate
-        dependencies = list({
-            dep.name: dep for dep in dependencies if dep.name
-        }.values())
-
-        return runner, JINJA2_PRELOAD, dependencies
-
-    @classmethod
-    def transform_response(cls, response: str) -> dict:
-        """
-        Transform response to dict
-        :param response: response
-        :return:
-        """
-        # extract result
-        result = re.search(r'<<RESULT>>(.*)<<RESULT>>', response, re.DOTALL)
-        if not result:
-            raise ValueError('Failed to parse result')
-        result = result.group(1)
-
-        return {
-            'result': result
-        }